research and development - gov.ukrandd.defra.gov.uk/document.aspx?document=wa0802_1… · web...

MINISTRY OF AGRICULTURE, FISHERIES AND FOOD CSG 15Research and Development

Final Project Report(Not to be used for LINK projects)

Two hard copies of this form should be returned to:Research Policy and International Division, Final Reports UnitMAFF, Area 6/011A Page Street, London SW1P 4PQ

An electronic version should be e-mailed to [email protected]

Project title Atmospheric emissions of particulates from agriculture: a scoping study

MAFF project code WA 0802

Contractor organisation and location

I C Consultants47 Princes GateExhibition Road, London SW7 2QA

Total MAFF project costs £ 30,600

Project start date 01/01/00 Project end date 30/06/00

Executive summary (maximum 2 sides A4)

This project was carried out jointly by I C Consultants (main contractor) and Silsoe Research Institute (sub-contractor). The objectives of the project were:

i) To undertake a literature review of potential sources of particulates from agricultural activities.ii) To prepare a preliminary quantitative inventory of UK agricultural particulate emissions.iii) To assess the toxicity and allergenicity of agricultural particulates.iv) To compare the role of rural as opposed to urban emissions in population exposure.v) To identify potential abatement techniques.vi) To identify gaps in knowledge and understanding, and priorities for further research.

The findings of the project were as follows:

Literature review of potential sources of particulates from agricultural activities

In the agricultural sector the range of types of particulate emission is broader than from other sectors. Emissions from the agriculture sector include not only mineral particulates (e.g. soils, lime) but also organic particulates (e.g. grain dust, feather dust fungal spores, pollen grains, bacteria and viruses). Most of the literature identified reports the concentrations of the airborne particulates measured in various agricultural scenarios, rather than emission rates of particulates therefrom: this is because traditionally the motive for the research was worker health rather than public health. Moreover, a wide range of different forms of

CSG 15 (Rev. 12/99) 1

Projecttitle

Atmospheric emissions of particulates from agriculture: a scoping study

MAFFproject code

WA 0802

concentration of particulate has been measured e.g. concentrations of inhalable, respirable and PM10 dust. As realisation grows that the finer fractions of particulates (e.g. PM2.5) are more correlated with human health risk, more recent studies report concentrations of the finer fractions, but older studies mostly only report concentrations of "total" (generally equivalent to inhalable) dust.

For comparisons between the results of different studies to be meaningful (and equally, for inventories to be meaningful), it is necessary always to compare the same forms of particulate concentration. However there is only very incomplete knowledge of conversion factors between the different forms e.g. for converting between inhalable and PM10 concentrations. The sector of agriculture for which the most literature was available was housed livestock, and of the little information overall on directly-measured emission rates, most was available for this same sector. With other sectors of agriculture it was possible in some circumstances to estimate emission rates from concentration values, but the estimation is only crude.

Preliminary quantitative inventory of UK agricultural particulate emissions

The results of this inventory are shown in the table below. All values are expressed in PM10, but it is stressed that many of these represent conversions from originally-measured values expressed in other forms (usually, concentrations of inhalable dust). It is also stressed that the emissions reported are gross not net. In other words, no attempt has yet been made (because of lack of data!) to allow for re-deposition of emitted dust. It may well be that much of the emitted dust is redeposited in the short range e.g. within the farm on which it has been generated. As the table below shows, housed livestock contribute by far the largest proportion of total annual UK PM10 emissions. Of the total of 11,445 t from housed livestock, 8,669 t come from housed poultry, 2,033 t from housed pigs and 743 t from housed cattle.

Source PM10 annual emission, tEstimated lower bound Average Estimated upper bound

Housed livestock 6,900 11,500 16,100Arable farming 800 2,400 4,000Crop storage 10 20 30Energy used on farms 2,300 2,900 3,500Unpaved roads on farms 20 40 60Total 10,000 16,900 23,700

Toxicity and allergenicity of agricultural particulates

Based upon both direct and indirect evidence, our conclusions are:

(i) occupational exposure to agricultural bioaerosols causes respiratory disease in a significant proportion of agricultural workers;

(ii) agricultural bioaerosols should be considered as a potential hazard to public health just as with other types of PM10 particulates;

(iii) with the exception of pollinosis, there is conflicting evidence whether the rural population is at a special risk from agricultural bioaerosols, over and above that arising from their contribution to PM10

particulates; and(iv) further research and analysis of existing information is needed to resolve the question of the potential

health hazards of agricultural bioaerosols to the general public.

The role of rural as opposed to urban emissions in population exposure

CSG 15 (1/00) 2

Projecttitle


MAFFproject code

WA 0802

Illustrative calculations have been undertaken to show how dilution reduces the contribution of agricultural emissions with distance from the exposed population. It is concluded that this significantly reduces the importance of agricultural emissions relative to urban emissions.

Potential abatement techniques

A range of "processing" techniques are available for abating airborne dusts e.g. electrostatic precipitators, wet scrubbers. However, most of these are considered too expensive for widespread adoption by UK agriculture, especially in the case of animal houses. No inexpensive sure-fire method of abatement suitable for agriculture was identified, although promising research has recently been reported on the use of oil spraying inside animal houses to reduce dust emissions from them. Modification of the animals' diet e.g. going from dry feeding to liquid feeding, in the case of pigs, also has promise as a means of abating dust from animal houses.

Gaps in knowledge and understanding, and priorities for further research

CSG 15 (1/00) 3

Projecttitle


MAFFproject code

WA 0802

Scientific report (maximum 20 sides A4)TABLE OF CONTENTS

1. Introduction................................................................................................................. 52. Review of published literature on PM10 emissions from agricultural sources

2.1 Introduction 52.2 Livestock buildings 62.3. Arable farming 82.4. Post-harvest 102.5. Combustion 102.6. Unpaved roads on farms 102.7. Energy use 112.8. Conclusions 11

3. Preliminary inventory for quantifiable sources 114. The potential health hazards of agricultural bioaerosols

4.1 Introduction 134.2 Direct evidence for the health hazards of agricultural bioaerosols 134.3 Indirect evidence for the health hazards of agricultural bioaerosols 164.4 Preliminary assessment of the potential hazards of agricultural bioaerosols to human health

following occupational and non-occupational exposures 175. Aspects of dispersion affecting risk 186. Abatement techniques

6.1 Introduction 196.2 Aarhus symposium, May 1999 196.3 Brainstorm on abatement approaches 20

7. Gaps in knowledge and needs for future research7.1 Characterisation of rural particulates and comparison with urban particulate composition 217.2 Research to quantify major sources identified in the emissions inventory 217.3 Research on abatement measures 227.4 Research on toxic and allergenic components and agricultural bioaerosols 227.5 Secondary aerosol formation 23

8. Summary and conclusions 239. Acknowledgements 23

Annex A: Glossary of technical terms and abbreviations 24Annex B: Detailed review of potential sources 27Annex C: Sources of error in measured values of particulate concentrations in air 76Annex D: Spreadsheets for preliminary inventory 78Annex E: Feasible options for mitigating PM10 emissions from agriculture (i) livestock 79Annex E: Feasible options for mitigating PM10 emissions from agriculture (ii) arable 80Annex E: Feasible options for mitigating PM10 emissions from agriculture (iii) other sources 81Annex F: Other relevant studies in progress 82Annex G: Full list of abatement ideas (before any assessment for feasibility) from brainstorm held at

SRI on 8 May 00 84Annex H: Full reference list 87

CSG 15 (1/00) 4

1. Introduction

Concerns about correlations between exposure to fine particulate and impacts on public health have led to measures to regulate atmospheric concentrations of fine particulate. At present the mechanism leading to human health risks is not well understood, and, as a precautionary measure, limits have been placed on PM10

(i.e. fine particles of aerodynamic diameter below 10 m) as the inhalable fraction, with no differentiation as to chemical speciation or origin. Thus in keeping with European limits, the UK Air Quality Strategy has set a limit of 50 g.m-3 over 24 hours not to be exceeded more than 35 times per year, and an annual average limit of 40 g.m-3. The highest concentrations occur in urban areas, where local authorities are in the process of defining Air Quality Management Areas (AQMAs) where these standards may be exceeded in 2004, and drawing up plans to avoid such exceedance. Clarification of the particle characteristics which are potentially harmful may put future emphasis on smaller (e.g. 2.5 m) or ultrafine particles, or on particular components.

Atmospheric particulates are a mixture of primary emissions and of secondary particles formed in the atmosphere (e.g. SO4

", NO3', NH4

+ and secondary organics). There are large uncertainties about sources of primary particles, including those from the agricultural sector. This project has therefore reviewed the information available on agricultural emissions, including biological components, and has quantified UK emissions where possible. Factors affecting the importance of agricultural emissions relative to other PM10

emissions have also been indicated, including toxicity and allergenicity, and some potential abatement measures identified. Finally gaps in knowledge have been addressed and needs for future research indicated.

2. Review of published literature on PM10 emissions from agricultural sources

2.1. Introduction

A large number of studies, in different countries, have been identified, which report on some aspect of particulate emissions from agricultural activities. These are described in greater detail in Annex B. This review included information from journals, from the USEPA, and from communication with people working in the fields of animal and poultry husbandry and health, general farming practice, plant pathology, aerobiology, and compost production. This cannot include all information available on all aspects of this subject, as it crosses a number of disciplines and detailed study was beyond the scope of this preliminary study, but this review is believed to give a reasonable indication of the current state of knowledge.

Quantitative estimates of emissions from livestock buildings in Europe (Takai et al, 1998), and USEPA emission factors (deduced from large numbers of trials in the USA) for soil preparation, cereal harvesting, and unpaved roads, are given. (Publication of USEPA emission factors for other activities is expected later in 2000.) However, the majority of the other studies measured occupational exposure, and so are not necessarily representative of actual emissions. The various studies used different types of samplers, and results from different studies should be compared with caution. All measurements are subject to some degree of error, and even small changes in sampler position can cause a large difference to the results obtained, as can the environmental conditions such as humidity, windspeed, and temperature.

A further complication is, this study aims to quantify particulate emissions in terms of PM10 (for a definition refer to the glossary, Annex A). However, the majority of the studies did not measure amounts of PM10 itself. All studies measured mass of total suspended dust, and some measured respirable dust (defined in the glossary, Annex A), which for most studies was taken as only the particles smaller than around 5 m diameter. An estimate for PM10 and other size fractions can be made from this information (PM10 is between total and respirable dust), by using an appropriate size distribution factor, but these factors are not well defined, and are different for different dusts.2.2. Livestock buildings

A recent study by Takai et al, 1998, measured concentrations of inhalable and respirable particulate in the air of cattle, pig and poultry buildings. Ventilation rates were also estimated for each building at each measurement, using a CO2 balance method. This information was used to estimate emissions of particulates. Each building was measured over 24 hours usually both in summer and winter, and was considered representative of current

Projecttitle


MAFFproject code WA 0802

practice, in each of the UK, Netherlands, Germany and Denmark. The results of these measurements are considered a good indication of current emissions from these buildings. However, measurements were not taken during operations such as moving animals or birds, or cleaning the buildings. These operations are relatively infrequent, but can generate high dust concentrations, albeit temporarily. The estimated emission from each type of building is tabulated below.

TABLE 1 Emission rates of particulates in kg per year per 500 kg liveweight of animal housed, from buildings in England (adapted from Takai et al, 1998)Pig buildings 20/20* Poultry buildings 12/12 Cattle buildings 16/4Inhalable dust

Respirable dust

Inhalable dust

Respirable dust

Inhalable dust

Respirable dust

5.5 0.82 27.5 3.3 0.85 0.27* number of buildings surveyed in winter/summer.

TNO made estimates of emissions from pig and poultry buildings in the Netherlands (Berdowski et al, 1997), which are somewhat higher than those reported by Takai et al (1998). The TNO estimates were made before the data from Takai et al study (1998) became available, and the TNO authors state these are initial, first order estimates, to be taken as an indication of the scale of emissions, and that subsequent research suggests that actual emissions may be somewhat lower. A number of other studies have also looked at dust levels in animal buildings, and details are given of a selection of reports in the UK, the USA, Finland, and Sweden. These studies were all aiming to estimate exposure to workers, and so were not necessarily representative of average dust concentrations in buildings, as the animals are generally more active when farmworkers are in the buildings, and this increases dust concentrations. No study other than that of Takai et al measured particulate levels at night, which may partly explain why Takai et al. found consistently lower average concentrations of particulate. However, many factors influence the particulate concentrations within buildings, including: type of feed and feeding system, type of manure system, type and method of applying bedding (where used), temperature, humidity, activity of animals, and of course, ventilation rate. Wide variation between individual buildings has been found (e.g. Crook et al, 1991) so a large number of buildings would need to be measured in order to be confident that the sample is indeed representative of current practice in the UK. Results from a selection of reports are tabulated below.

6

Projecttitle



TABLE 2 Concentrations of particulate in air measured in PIG buildingsReference Country and comments Mean dust concentrations

(mg per m3 of air)Total suspended dust

Respirable dust

Takai et al, 1998 England (Southern) 1.87 0.24England, Germany Netherlands, Denmark

2.19 0.23

Crook et al, 1991 Scotland (Northern)summer minimum mean 3.5 winter maximum mean 11.5

Roelofs, 1995, cited in Berdowski et al, 1997

Netherlands (2.5-2.6 PM10)

Preller et al, 1993, cited in Berdowski et al, 1997

Netherlands 5.4 - 6.4

Donham et al, 1989 Sweden area sampling 4.6 0.33personal sampling 6.4 0.34

Donham et al, 1986 USA area sampling 7.6 2.5personal sampling 6.25 0.53

TABLE 3 Concentrations of particulate in air, measured in POULTRY buildingsReference Country

and commentsMean dust concentrations (mg per m3 of air)Total suspended dust

Respirable dust

Takai et al, 1998 England (Southern) 3.31 0.51England, NL, Germany, Denmark

3.60 0.45

Whyte, 2000 England Overall mean exposure, barns 9.8Overall mean exposure, cages 4.2Routine tasks 5 – 10Cleaning, sweeping 35 – 65

Evers, 1995, cited in Berdowski et al, 1996

Netherlands 5 – 20

Jones et al, 1984 USA 7-day old broilers 1 – 5 0.05 - 0.330-day old broilers 7 – 11 0.4 - 0.6

7

Projecttitle



TABLE 4 Concentrations of particulate in air measured in CATTLE buildingsReference Country and

commentsMean dust concentrations (mg per m3 of air)Total suspended dust

Respirable dust

Takai et al, 1998

England (Southern) 0.22 0.15England, NL, Germany, Denmark

0.38 0.07

Louhelainen et al, 1987 Finland air sampling 1Personal sampling 5.6

Nieuwenhuijsen et al, 1998

USA, California feeding 26.1 (2.9 PM10) 0.59Milking 0.8 (0.1 PM10 ) 0.06Manure removal 3.7 (0.5 PM10 ) 0.15

Olenchock, et al, 1990 USA (worst case conditions during bedding chopping)

8 to 41 1.6 to 2.5

It is considered that Takai et al’s results should be seen as a conservative estimate of the likely emissions from UK livestock buildings.

The nature of dust from pig buildings in the USA has been thoroughly studied in two independent studies (Donham et al, 1986, and Heber et al, 1988). These found that the bulk of the dust was feed, but the finer material (particles smaller than around 10 m) was mostly dried manure. Skin flakes, hair and other material were found in much smaller amounts. Bedding was evidently not used in these pig units. Dust in poultry buildings is expected to be similar, with added contributions from litter (for meat birds) and feathers; however, further study of this would be beneficial. Particulate matter from cow and sheep buildings is rather different. Where good quality feed and bedding are used, particulate concentrations are low, and the majority of particles are fragments made from chopping and spreading the bedding. In vitro tests for different feed and bedding materials for horses found that hay is inherently dustier than alternatives such as silage, and straw is dustier than other bedding materials (Vandeput et al, 1997). If the bedding or hay is stored in a damp condition, very large numbers of moulds develop over time, and these come to far outnumber all other particulate material. This is particularly serious, because these moulds include species such as Aspergillus fumigatus, which are known to cause respiratory problems to farmworkers and animals.

2.3. Arable farming

No published information, relevant to UK conditions, has been found, from which a good estimate of particulate emissions can be made. However, the USEPA has calculated emission factors for cereal harvesting and soil preparation. The estimate for soil preparation requires the input of data concerning soil moisture and silt content; general average values for dry soil have been used, but for more reliable estimates UK data should be used.

8

Projecttitle



TABLE 5 Concentrations of particulates measured in air near agricultural operationsReference Country Mean dust concentration,

mg per m3 of airComments

Total suspen-ded dust

Particles less than 10 m diameter

Respirable particles (under 4 m diameter)

Soil preparationClausnitzer and Singer, 1996

USA, California

0.5 – 10 Depending on operation. Sampler <1 m above ground


USA, California

10-150 0.5 -9 0.1-1 Personal exposure with no protective cab, using cascade impactor for total and PM10, and cyclone for respirable dust


USA, California

5 0.3 As above, using IOM sampler for inhalable dust

Norén, 1985 Sweden 100, 150

Mean dust concs at 2 points on outside of tractor

Batel, 1979 Germany 40 Mean concentration at tractor driver level. (Max conc was 630)

Cereal harvestingClausnitzer and Singer, 1996

USA, California

1 Sampler 2.4 m above ground

Batel, 1979 Germany 20 Max conc 80. Sampler at tractor driver level

Darke et al, 1976

England 2-10 Approximate mass of spores collected behind combine

Several other studies have measured concentrations of particulates (or in one case, spores) near to operations. Studies performed in England, Sweden, Germany and Finland, are considered to be representative of UK conditions: however results from California and the Canadian prairies are considered less relevant, because of significant climatic differences. A wide range of dust concentrations is reported, with several authors reporting values over 100 mg m-3, and mean concentrations well over 10 mg m-3 for some operations. Unfortunately, most of the studies are rather old, as they were aiming to assess exposure to tractor drivers, who are nowadays enclosed in protective cabs. These reports do not all give complete details of procedure, numbers of measurements, or statistical treatments of results, so can only be interpreted as an indication of particulate concentration, rather than absolute values. Furthermore, great variations in measurements were found with even small differences in sampling location and weather. Thus, it is considered that emissions from arable farming could be significant. In order to perform a meaningful assessment, a large number of measurements, as part of a carefully prepared study, would be needed.

Emissions of biological particles, mainly fungal spores and pollen grains, can outnumber all other particles at certain times of year, particularly around harvest time, though their contribution to annual emissions is probably not high. There is not sufficient information to make a reliable quantitative estimate.

Wind erosion of soil is another source of particulate emissions which is probably very low on the national scale, but may be significant at certain times of year in local areas with soils prone to this type of erosion. Fine sandy and peaty soils are the types reported to suffer most from wind erosion: however, it is not known if PM10

emissions from these soils are significant, as the bulk of the particles are much larger. Research on this matter in the UK has been primarily concerned with reducing crop losses, rather than limiting emissions. This topic has been widely studied in the USA, but this data is not considered relevant to conditions in the UK.

9

Projecttitle



The production of compost, whether for mushroom farming, or to turn raw organic wastes into more useful material, also has the potential to emit particulates, especially large numbers of spores of the microbes which are present in the composting process. The material remains moist at all times, and as long as it is undisturbed, emits only low levels. However, during turning, which is regularly done to aerate the material, a lot of steam and spores are released. Some studies on this subject concluded that this is certainly a potential occupational hazard to workers (who can become sensitised and develop allergic symptoms, though this is a fairly uncommon occurrence), but risks to the general public are not significant.

Various applications are made to arable land, to improve the quality or fertility of the soil, or to dispose of organic wastes such as manure or sewage sludge. The relatively small amount of slurry spread these days by gun spraying is obviously a potential source of emissions, though a study of the spread of pathogens through air was fairly inconclusive. Poultry litter can be very powdery if dry, and applying this material to land could cause significant emissions to air. Lime is routinely applied, often in fine powder form, and is considered to be the application with the greatest potential for making significant emissions. Fertiliser application could also emit particulates, but most modern fertilisers are pelleted, so the potential for dust emissions is thought to be low. Pesticide application also has the potential to emit particles of either liquid or powder, particularly when applied to orchards, where the substance is entrained in air and blown onto the trees. Most other crops are sprayed by boom sprayers, which direct the substance down onto the crop and the chance for escape is small. The Code of Good Agricultural Practice for the Protection of Water (MAFF, 1998) specifies that pesticides ought not to be applied near watercourses in windy weather, to avoid contamination by the expected stray spray. Other applications are not expected to contribute significantly to airborne particulate emissions.

2.4. Post-harvest

Cereal drying is commonly carried out on the farm, and is reported to be a high emitter of particulates, though this is not covered by any regulations, and no reliable data have been collected. (Effective cereal drying is also necessary to limit the growth of moulds during storage.)

Transporting grain, milling, and producing animal feed are all processes which can potentially emit large quantities of particulates. The USEPA have developed emission factors for these, though it is not known how relevant they are to UK practices. In the UK, production of food and animal feed is covered by IPC regulations under The Environment Protection Act 1990, and emissions from plants are regulated by Local Authority Environmental Health Enforcers. Some examples of emissions from some plants have been collected but UK data nationwide have not been collated.

2.5. Combustion

Many farms have small incinerators used to dispose of items such as dead birds. Incinerators treating less than 50 kg/hour are not covered by regulations, but will nevertheless emit particulates. Data are not available. The banning of burning cereal stubble has contributed to reduced emissions from arable farming, but some stubbles (of oilseeds and flax, which are difficult to plough into the ground) can still be burnt in the field. These will cause emissions of particulates which will be small on the national scale, but may be significant locally.

2.6. Unpaved roads on farms

No UK data have been found for estimating of emissions from unpaved roads on farms, but this is certainly a potential source of emissions. The USEPA have calculated a formula for emission factor, and other studies in the USA found this was a reasonable approximation to measured values. Applying this formula to roads in a dry condition suggests that emissions from this source could be significant. Further study would be needed to check if this is definitely the case.

10

Projecttitle



2.7. Energy use

Data are available to estimate PM10 emissions from energy and fuel use in agriculture, in 1998. The largest on-farm emissions come from burning straw as a fuel. Diesel and oil, used for vehicles, heaters and dryers, account for the next largest emissions of PM10. Smaller emissions are generated by coal and gas. Use of electricity was estimated to produce significant amounts of PM10, though this was produced at power stations, rather than at the farm.

2.8. Conclusions

Most existing studies measured total suspended particulate or respirable dust, and extrapolations to estimate PM10 can only be approximate. Also, estimates from all activities are as measured close to source, and a large proportion of this particulate is likely to be deposited nearby. Researchers at the University of California at Davis are starting to work to produce a factor to allow for this, and thus estimate net emissions from the farm. Further research on this topic is needed.

Livestock buildings appear to be the largest source of emissions of particulate matter from agriculture. Some reasonable estimations of these emissions are available, though further work would be valuable to improve on these estimates, and deduce more about the nature of these dusts in UK buildings.

Certain operations in arable farming, namely soil preparation and harvesting, can also emit large quantities of particulate matter. There is insufficient good data available to make realistic estimates of emissions, though a number of studies have measured dust concentrations, and report wide variations in results. Some emissions will also occur from applying material, in particular lime, to land, but no quantitative information is available. Other applications, such as manure, pesticide and fertiliser, may make a small contribution of particulate matter. At certain times of year, biological particles, such as pollen grains and fungal spores, can make significant contributions to particulate levels. Wind erosion of soils may be a significant source of particulates in some areas, but no suitable data have been identified.

Post harvest operations such as cereal drying, transport and handling, can also be high emitters of dust: some US emission factors are available, but not UK data. Preparation of animal feed is covered by IPC regulations. Emissions from unpaved roads can be predicted from a USEPA formula, which suggests that they may be significant, but verification in UK conditions would be needed. Energy and fuel use for agriculture has been well quantified, and emissions of PM10 are reasonably well known.

3. Preliminary inventory for quantifiable sources

A preliminary inventory of particulate emissions from UK agriculture was drawn up, as a first step in assessing the relative magnitudes of the various sources, and hence in deciding on optimum abatement strategies.

It is stressed that the inventory described below is preliminary, as, except for housed livestock, the data on which it is based are very incomplete. There may , moreover, be other agricultural emission sources which are as yet unrecognised. There was also the problem that different authors have reported measurements on particulates in a wide range of different forms e.g. total dust, inhalable dust etc. To construct a meaningful inventory, it is of course necessary to "compare apples with apples and oranges with oranges", and a decision was therefore made to base the inventory on PM10 values, as a compromise between relevance to human health risks and availability of data. (PM2.5 values are now believed to correlate better with human health risks than do PM10 values, but very few data on PM2.5 were found.)

11

Projecttitle



Where results were reported as PM10 itself, these were of course used for the inventory directly, but where it was necessary to fall back on particulates reported in other forms, correction factors were used to convert these other forms to estimated PM10. It is accepted that this is a crude procedure, but no better procedure was available. In the interests of transparency, the sources of all such correction factors, where used, have been stated.

Table 6 presents a summary of the resulting PM10 inventory, which was constructed using an Excel spreadsheet.

TABLE 6 First estimate of total annual dust (PM10) emission from UK agriculture

Source PM10 annual emission, tEstimated lower bound Average Estimated upper bound

Housed livestock 6,900 11,500 16,100Arable farming 800 2,400 4,000Crop storage 10 20 30Energy used on farms 2,300 2,900 3,500Unpaved roads on farms 20 40 60Total 10,000 16,900 23,700

The Excel file dust emission inventory.xls contains all the calculations used to produce the average values shown in Table 6, and is supplied with the electronic version of this report. In hard copies of this report, the spreadsheet outputs can be found in Annex D.

As Table 6 shows, housed livestock contribute by far the largest proportion of total annual UK PM 10 emissions. Of the total average of 11,500 t from housed livestock, 8,700 t come from housed poultry, 2,000 t from housed pigs and 800 t from housed cattle. The second largest term is from energy used on farms: here the largest contributor is the combustion of straw as a fuel in small boilers.

In the following paragraphs, comments will be made on the individual spreadsheet outputs (shown in hard copies of this report as Tables D.2 to D.8, in Annex D), and on our estimates of the upper and lower bounds on each component of the total emissions.

Thanks to the work of Louhelainen et al, 1987b, a good value can be deduced, for animal house dust, for the ratio not only of PM10 to total dust (0.45) but also of PM2.5 to total dust (0.08). Therefore, in Tables D.2, D.3 and D.4 only, estimated emissions of PM2.5, as well as of PM10, are shown. Compared with some of the other sources of emission the housed livestock emissions are fairly well characterized (Takai et al, 1998): an uncertainty band of I 40% has thus been used. It is fortunate that this, the largest item in the inventory, has the smallest uncertainty bands.

For arable farming (Table D.5), no data reporting emission rates directly were found, but many data were available on dust concentrations arising near different arable operations. A simple model as described below was therefore used to estimate the former from the latter. The reported dust concentrations were assumed to prevail across the cross-section of a "plume" downwind of the dust-generating operation. A "plume" height of 3 m was assumed, and the "plume" was assumed to cover the whole UK field area for the operation concerned. For hay-making, four passes by a machine annually were assumed, but, for all other operations, only one pass annually was assumed. All the dust in each of these imaginary boxes was assumed to be emitted off the farm, with no re-deposition. For this source of emission, the plume height was believed to be the major source of uncertainty. Upper and lower bounds were therefore estimated by assuming extremes of plume height of 5 m and 1 m respectively.

12

Projecttitle



In the case of crop storage (Table D.6) data reporting emission rates directly were found only for grain handling in the USA (Shaw et al, 1997). As with arable field operations, many data were found on concentrations, around other storage facilities but, unlike with arable field operations, no sensible estimate could be made of air flow rates around these facilities, so no attempt was made in this case to estimate emission rates from the concentration data. The uncertainty band on the grain handling term was estimated at 50%.

Energy use on UK farms (Table D.7) includes off-road emissions from tractors and other vehicles, from heaters, dryers and other on-farm energy generation. The largest contribution comes from combustion of straw in small boilers. It has been assumed that all such boilers are in agriculture e.g for glasshouse heating, but this needs investigation. The statistics on straw burning boilers has not been revised since 1994, and the present numbers may be smaller.

Finally, for unpaved roads on UK farms (Table D.8) a simple model was used to estimate what total length of unpaved road might be found per unit of area of UK farmland, since the emission factors used, from the USEPA, are expressed as weights of dust per vehicle-mile travelled. The reported total annual emission of 40 t PM10 assumes an average road surface moisture content of 10 %, which may be low by UK standards. If a value of 50 % is used in the USEPA formula instead, the total annual emission falls to 25 t PM10.

4. The potential health hazards of agricultural bioaerosols4.1. Introduction

A bioaerosol is an aerosol comprising particles of biological origin and/or activity with aerodynamic diameters between ca. 0.5 and 100 μm (Cox and Wathes, 1995). Agricultural bioaerosols can be distinguished from other particulate sources because their composition may include a substantial mass or number of micro-organisms, pollens or other biological materials. Thus the potential health hazards of agricultural bioaerosols include specific diseases, such as pollinosis and zoonoses, as well as the general hazard that may arise from exposure to non-biological particles and, in this latter respect, agricultural bioaerosols should be given the same attention as any other major source of PM10 particulates. Indeed, concern over the health hazards of PM10 has arisen in the UK, US and elsewhere even though the pathophysiological mechanisms are not understood and the harmful components of PM10 have not been identified (Williams, 1999). There are now sufficient epidemiological studies which demonstrate an association, if not a direct causal relationship, between particulate air pollution and ill health in urban populations (see Committee on the Medical Effects of Air Pollutants (COMEAP), 1998). COMEAP estimates that "PM10 contributes to the advancement of around 8,100 deaths and 10,500 hospital admissions for respiratory disease in the urban population of Great Britain annually": clearly, other less severe effects may arise from PM10. The lower and upper bounds were therefore taken as 20 and 40 t PM10, respectively.

This brief review summarises some of the direct and indirect evidence that agricultural bioaerosols may pose a health hazard to the population at large. In some cases, the evidence is based upon a significant body of scientific literature on topics in which there is now a common consensus, but in other areas further research may be needed or existing information may need to be collated (as indicated in italics at the end of each section).

4.2. Direct evidence for the health hazards of agricultural bioaerosols

Zoonoses

Zoonoses form a special class of communicable disease which may infect both animals and man and are of particular importance in public health. The aerial route is the natural pathway for transmission of some microbial pathogens, and inhalation has been suggested as an inoculation route for the following zoonoses (see Table 7) which may involve agricultural sources (HSE, 1993).

13

Projecttitle



TABLE 7 Recent incidence in the UK of some zoonoses which may involve agricultural sources

Disease Organism Approximate human incidence in UK, cases per year

Anthrax Bacillus anthracis 0.1Bovine tuberculosis Myrobacterium bovis Not knownBrucellosis Brucella abortus 7 in 1984Newcastle disease Newcastle disease virus RareOvine chlamydiosis Chlamydia psittaci 532 in 1988Psittacosis Chlamydia psittaciQ fever Coxiella burnetti 185 in 1989

Single organisms are typically smaller than ≈ 2-3 μm. There is no information on the size of the aerosols of these zoonoses though the fact that some may infect the respiratory tract (e.g. in anthrax and psittacosis) implies the involvement of respirable particles. The incidence of these zoonoses is very low and mainly involves stockmen, abattoir workers and veterinarians. This suggests that a hazard to the general public from zoonotic bioaerosols exists but its risk is low and this is supported, in part, by the minimal hazard of bioaerosols during land spreading of agricultural wastes (Boutin et al, 1988).

The current incidence of agricultural zoonoses in the UK

Occupational respiratory diseases in agricultural workers

Agricultural workers are routinely exposed to high concentrations of bioaerosols and toxic gases and information on occupational respiratory diseases can help determine the likely health hazards of agricultural bioaerosols to the general public.

There is now a substantial body of evidence that acute and chronic exposures to bioaerosols cause occupational respiratory diseases in pig and poultry stockmen (e.g. Whyte, 1993). Fungi and bacteria in agricultural dust are of particular concern because of the high concentrations, varying between 103 and 1010 cfu m-3 (Dutkiewicz, 1997). The specific diseases include bronchitis, hypersensitivity pneumonitis, toxin fever and asthma with associated respiratory symptoms such as a decline in lung function, eye irritation and headache. The overall prevalence depends on the disease or symptom; e.g. 15 - 25 % for rhinitis, sneezing and coughing (Choudat et al, 1994), 10 % for asthma (Donham, 1987) and 33 % for at least one chronic respiratory symptom in pig farmers (Preller et al, 1995); and 38 % for hypersensitivity pneumonitis (Stahuljak-Beritic et al, 1977) and 17 % for a 5 % decline in FEV1 in poultry workers (Donham et al, 2000). In pig farmers, the decline in lung function is progressive with continual exposure (Senthilselvan et al, 1997). The aetiology of occupational respiratory diseases is not fully understood but dust, ammonia and, especially, endotoxin have all been implicated to the extent that tentative recommendations for exposure limits have been made (Table 8) on the basis of lung function (Donham et al, 1995; 2000). These limits are far more stringent than those specified in the UK (HSE, 1997) and may reflect the synergy between bioaerosols and toxic gases in pig and poultry houses. Equally importantly, measured concentrations of bioaerosols and ammonia are of this magnitude (Wathes et al, 1998).

14

Projecttitle



TABLE 8 Recommended daily exposure limit (Donham et al 1995; 2000)

Total dust, mg m-3

Respirable dust,mg m-3

EndotoxinEU m-3

Ammonia,ppm

Poultry workers

2.4 0.16 614 12

Pig workers

2.5 0.23 100 7

Large scale composting to produce agricultural products can also generate substantial quantities of bioaerosols of which Aspergillus fumigatus, endotoxins and β-1,3 glucans (from the cell walls of fungi and mycotoxins) are of particular concern. A major review by Millner et al (1994) concluded that "where worker health was studied for periods of up to ten years on a composting site, no significant adverse health impacts were found". However, they qualified this conclusion because of the limitations of the designs of the underlying studies.

Other agricultural work in which there is chronic exposure to bioaerosols includes harvesting and storage of hay and grain (which leads to farmers lung), processing of mushroom compost and wood bark chips, and glasshouse production of tomatoes, lettuce and other crops (Crook and Olenchook, 1995). An association between chronic respiratory symptoms and the use of chemical disinfectants has also been demonstrated in pig farmers (Preller et al, 1995).

Overall, the existence of occupational respiratory disease in agricultural workers associated with chronic exposure to agricultural bioaerosols provides evidence that such bioaerosols could create a hazard to the general public, especially atopic, immuno-compromised or other highly susceptible individuals.

Current incidence of occupational respiratory diseases in UK agricultural workers and its association with bioaerosol exposure.

Health of neighbours of an intensive pig farm

There has been only one study of the health of neighbours of an intensive livestock farm, which is a potent source of bioaerosols and other aerial pollutants (Thu et al, 1997). In this study, 18 neighbours living within a 2 mile radius of a large (4000 sow) farm were matched with controls, who also lived in the countryside. An interview and questionnaire showed a significantly higher prevalence of two clusters of physical symptoms (sputum, cough, breath shortness, chest tightness and wheezing; nausea, dizziness, weakness and fainting; P<0.05) in the farm's neighbours. These symptoms are similar to those found in pig stockmen. This preliminary study provides some tentative evidence, but needs to be enlarged into an epidemiological study with better controls and evaluation of health, and quantification of the bioaerosol exposure.

An epidemiological study of health in neighbours of major sources of agricultural bioaerosols.

Pollinosis

Pollinosis (hay fever) is mainly caused by approximately 100 pollen-producing plant species of which the important agricultural sources include grasses and oil seed rape, though the role of the latter is controversial (McSharry, 1992). Although intact pollen grains are usually larger than 10 μm and should therefore not be respirable, the symptoms of asthma in people who are allergic to pollen indicate antigen activity in the respirable size fraction (Rantio-Lehtimäki, 1995). However, the most frequent symptoms are allergic rhinitis or conjunctivitis, characterised by sneezing, watering eyes, nasal obstruction, itchy eyes and nose and also

15

Projecttitle



coughing, all of which can be attributed to the large particles. Within Europe, the prevalence of pollinosis is rising and estimates range from 4.9 % in Spain to 16.5 % in the UK (D'Amato et al, 1998; Strachan, 1995). The reasons for the rise are not fully understood but one explanation is the increased exposure to air pollutants, though this is inconsistent with the lack of difference in prevalence between urban and rural dwellers in some, but not all, studies (Strachan, 1995). The European Aeroallergen Network provides weekly reports on pollen counts and more local reports are available in the UK.

The relative contributions of pollen from agricultural and non-agricultural sources The emission rate and concentration of agricultural pollens throughout the pollen season

4.3. Indirect evidence for the health hazards of agricultural bioaerosols

Health of urban and rural inhabitants

A priori, children brought up in a rural environment should be exposed to more bioaerosols than those in urban areas. A recent cross-sectional study of 10,667 Finnish students showed that when compared with urban living, a childhood farm environment reduced the risk of allergic rhinitis and/or conjunctivitis, and asthma in association with episodic wheezing but not atopic dermatitis (Kilpelinen et al, 2000). However, the benefit of a rural environment did not extend to rural children who did not live on a farm, implying that exposure to 'farm protective factors' was essential. Kilpelinen et al suggested that the protection arose from exposure to immune modulating agents, such as mycobacteria and actinomycetes. It would be almost impossible to quantify the life-time exposure to such agents, though some information could be gained from antibody titres to specific antigens.

Involvement of bioaerosols in human disease

It is common knowledge that bioaerosols from non-agricultural sources are involved in many infectious and allergic diseases of medical importance. The topic has been reviewed regularly (e.g. various chapters in Cox and Wathes, 1995) and notable examples include influenza, Legionnaire's disease and various allergies (e.g. to house dust). Exposure is ubiquitous and can occur in the office, factory, hospital or home.

Airborne diseases of livestock

Many pathogens of livestock are known to be transmitted aerially and include the species of at least 20 bacteria, 17 viruses, 8 fungi, 1 rickettsia and 1 protozoa, which can cause respiratory and other diseases (Wathes, 1995). For some hardy spores, e.g. A. fumigatus, the aerial route is the natural pathway for transmission though it can be used by more fragile micro-organisms when the shedding rate is rapid and the infectious or allergic dose is low. Airborne pathogens can travel long distances, > 20 km, as illustrated by foot and mouth disease virus, which has crossed the English Channel on several occasions to infect livestock in Southern England (Donaldson, 1983). Indeed, a computational model has been devised (Gloster et al, 1981) to forecast the likely spread of the virus and utilises experimental data on shedding rates, aerosol survival and infective doses for the major host species; meteorological and topographical information are also included.

Odour and other nuisances from agricultural processes

Voluntary returns from local authorities in England and Wales on the incidence of odour nuisance give an underestimate of public concern. In 1997/8, the number of complaints per million population was 240 and 452 for agricultural and industrial processes, respectively, and involved 2980 and 5924 sources, respectively. (Personal communication, K. Willis, Chartered Institute of Environmental Health). The greatest source of complaints was livestock, especially pig, farms. Odorants are not only transported in the gas phase but can also

16

Projecttitle



be adsorbed on airborne particles, which will normally be inhalable. This evidence implies that agricultural bioaerosols have the potential to be transported aerially to people living or passing near to farms.

UK and European Legislation on air pollution from farming

The existence of UK and European legislation on air pollution from farming provides evidence of concern for the adverse consequences of agricultural emissions. While the primary rationale may be connected with climate change or odour nuisance, the benefits extend to the potential hazards of bioaerosols either implicitly or explicitly. Relevant legislation and codes include:

Town and Country Planning General Development Order (1988)Town and Country Planning (Assessment of Environmental Effects) Regulations (1988)Environmental Protection Act (1990)EU Directive on Integrated Pollution Prevention and Control (1996)Code of Good Agricultural Practice for the Protection of Air (MAFF, 1998)

Interests of the Health and Safety Executive

A further indication of the potential hazards of agricultural bioaerosols is given by the HSE's current research (HSE, 1999), which includes:

Exposure response in occupational asthma – aims to define the relative risk of developing occupational asthma after exposure to various dusts and vapours.

Low toxicity dusts workshops – examined the reliability of occupational exposure levels with particular attention to dose-response data.

Information on current concerns of Agricultural Inspectors in the HSE with respect to agricultural bioaerosols.

4.4. Preliminary assessment of the potential hazards of agricultural bioaerosols to human health following occupational and non-occupational exposures

By comparison with other sources of particulates, agricultural bioaerosols comprise a higher proportion of material of biological origin, e.g. animal dander, feed grains and pollens. COMEAP's (1998) concern over PM10 was mainly based on studies of urban air pollutants, which would include many mineral and other inorganic dusts. A priori, there does not seem to be a case to exclude agricultural bioaerosols from the general attention given to other sources of PM10 particulates.

Nevertheless, any assessment of the potential hazards of agricultural bioaerosols must take account of the available scientific evidence. In the case of agricultural workers, it is now well established that chronic exposure to bioaerosols does cause ill health in a substantial proportion of the workforce. The hazards are well recognised and various technical means are in place to limit exposure.

However, the hazards of agricultural bioaerosols to the general public are less clear. Rural inhabitants living adjacent to major sources of bioaerosols such as large pig, poultry and mushroom farms or fields of pollen-producing plants at flowering, are most at risk. The only direct evidence for a significant hazard is pollinosis and the preliminary findings of Thu et al (1997) on the respiratory symptoms in people living near to a pig farm. On the other hand, the superior health of Finnish children raised on farms versus those brought up in urban environments (Kilpelinen et al, 2000) appears to discount the general hazards from agricultural bioaerosols, though there is a caveat that the pattern of Finnish agriculture is quite different from that in the

17

Projecttitle



UK, especially the lack of intensive livestock production on a large scale. With regard to other risk assessments of major bioaerosol sources, Millner et al (1994) concluded that "composting facilities do not pose any unique endangerment to health and welfare of the general public ...... though there are sub-populations that may be at increased risk from exposure to composting bioaerosols". They argued for epidemiological studies to identify hazards and dose responses to bioaerosols.

Before definitive conclusions can be drawn about the health hazards of agricultural bioaerosols to the rural population (as well as to casual visitors to the countryside), we need to know:

The physical, chemical and microbiological properties of agricultural bioaerosols and how they compare with those of urban bioaerosols.

The likely exposure to agricultural bioaerosols in the rural population and their health effects. The size of the rural population at risk.

The conclusions of this preliminary assessment are therefore that:

(i) occupational exposure to agricultural bioaerosols causes respiratory disease in a significant proportion of agricultural workers;

(ii) agricultural bioaerosols should be considered as a potential hazard to public health just as with other types of PM10 particulates;

(iii) with the exception of pollinosis, there is conflicting evidence whether the rural population is at a special risk from agricultural bioaerosols, over and above that arising from their contribution to PM10

particulates; and(iv) further research and analysis of existing information is needed to resolve the question of the potential

health hazards of agricultural bioaerosols to the general public.

5. Aspects of dispersion affecting risk

The concern to reduce particulate concentrations arises from concerns about effects on human health. The setting of standards is thus a precautionary approach aimed at reducing population exposure and associated risk. Since the agricultural emissions arise in rural areas remote from centres of population, this reduces their contribution to population exposure as compared with that from emissions from traffic and urban sources. The latter are also exacerbated by restricted dispersion in busy streets confined by buildings ("street canyons"): these provide hot-spots of high concentrations, as is clearly illustrated in assessments by local authorities under the UK Air Quality Strategy.

As an illustration of the effect of distance on concentration, the dispersion model ADMS has been used to show how an agricultural source would affect airborne particulate concentration in the surrounding area. Annual average concentrations have been calculated across a 30 km by 30 km grid, assuming a uniform source of emissions across a 1 km by 1 km square (100 hectares) at the centre. The total emission rate is 1 tonne per year, released at a low level over typical agricultural land (roughness length 20 cm). In the first row, Table 9 shows the annual average concentration within the emitting area, in g.m-3 . The second row gives the range of corresponding concentrations at the centres of the ring of eight 1 km grid squares immediately surrounding this (i.e. the minimum and maximum), together with the total over the eight squares. The latter is a comparative measure of the contribution to population exposure from a 1 km wide ring of agricultural land surrounding a 1 by 1 km populated area. The next row in the table corresponds to the next ring of 1 kilometre grid squares surrounding the squares in the previous row, and its cumulative contribution with the inner ring; and so on to the next row. It is clear that concentrations fall off very rapidly with distance from the emitting source. The contribution from the emitting square to population exposure in that local square is three times as large as from an equivalent density of emissions in a surrounding belt 1 km wide. Integrating over an area of 30 km by

18

Projecttitle



30 km surrounding the central 1 km square, a uniform source density of 1 tonne per km2 outside this central square would contribute 0.28 g.m-3 in the central square, as compared with 0.18 g.m-3 in this illustration from a single tonne emitted within the central square.

TABLE 9 Comparison of contributions over different distancesRegion Range of concn. due to

source in centre squareConcn. in centre square due to ring of emission

Cumulative conc. in centre from rings

Central 1x1km grid 0.18 - -8 neighbouring grids in 3x3 km

0.0025 to 0.017 0.063 0.063

Next ring in 5x5km 0.0005 to 0.0035 0.029 0.092Next ring in 7x7km 0.0002 to 0.0018 0.019 0.111Next ring in 9x9km 0.0001 to 0.001 0.014 0.125Next ring to 11 km max 0.007 0.011 0.136Out to 30 km 0.279

In the above illustration, no allowance has been made for deposition of large particles by gravitational settling. The settling velocity of a 10 micron diameter particle is in the range 0.3 to 1 cm.s-1, depending on density; this velocity varies as the square of the particle diameter. Over longer distances than those considered, the larger particles contributing to PM10 would settle out gradually, but the PM2.5 would largely persist until removed in rain.

One consideration which may add to the importance of agricultural emissions is their episodic nature. The air quality standards are concerned with the number of days on which concentrations exceed 50 g.m-3, and agriculture may yield peak emissions concentrated within certain dry periods of the year - for example at harvest time. This needs to be considered together with the toxic and allergenic nature of some of these emissions, as discussed in Section 4 of this report.

6. Abatement techniques

6.1. Introduction

"Processing" techniques for the abatement of dust and fume are discussed in detail in a booklet published by the Institution of Chemical Engineers (Muir, ed., 1992) and in a TNO report (Visschedijk et al, 1997). We consider many of the abatement techniques described in the booklet and report (e.g. electrostatic precipitators, high energy wet washers) to be unsuitable for agricultural applications, for which the air flows to be treated are generally large and the money available to purchase capital equipment is generally small. While these conditions prevail particularly for animal housing, nevertheless certain of the "processing" techniques for dust abatement are used elsewhere in agriculture, particularly on the off-gases from grain dryers. For example, Carier Ltd., Braintree, Essex, offer ballistic-type dust removal devices in which the dust particles carried on off-gases from a grain dryer are forced by the geometry of the device to follow a trajectory which separates at least the larger particles from the main air flow. Ballistic type dust separators are favoured over other types for this application because they can treat large air flow rates and cause only a low pressure drop, thus increasing blower energy costs no more than necessary. The penalty is that dust removal efficiency is lower than with some other, more energy-intensive, types of separator.

19

Projecttitle



6.2. Aarhus symposium, May 1999

It was expected, when a preliminary inventory was drawn up, that housed animals would prove to be a large source of agricultural particulate emissions. This has indeed been the case: see Section 3. As it happened, the control of dusts in animal production had been the subject of a specialised symposium held in Aarhus, Denmark, in May 1999, and for which printed proceedings are available. (Danish Institute of Agricultural Science, 1999).

As one of the keynote speakers at the Aarhus symposium noted (Zhang, 1999), 12 of the 40 papers at the symposium related to the use of oil spraying inside animal houses (mostly pig houses, but not excluding poultry houses) as a means of abating dust. Zhang concluded that "while spreading oil is effective in reducing dust, many unknowns need to be investigated. Needed research areas include alternative liquids that are easy to clean yet effective in dust reduction, economical and reliable automatic sprinkling systems and the long-term effect of the oil application."

Zhang (1999) also discussed the potential for using modified ventilation to create relatively dust-free areas in parts of an animal house (e.g. those parts where stockmen spend most of their time). This approach, however, is unlikely to reduce the overall dust emissions leaving an animal house.

Bottcher et al (1999) reported a study of dust abatement using solid windbreak walls (positioned a few metres outside the exhausts from fan-ventilated pig houses with wall-mounted fans), or using wet pad scrubbers (installed to form a full partition inside a pig house, 1.2 m upstream of wall-mounted exhaust fans). Although the solid windbreak walls captured modest amounts of dust, their major benefit was seen as deflecting dust-laden exhaust air upwards. This encouraged dispersion of the dust but did not in itself reduce the emission rate greatly. The wet pad scrubbers did not create high additional pressure drop for the existing fans to cope with. They achieved a 65 % reduction of the concentration of total dust at a low ventilation rate, but only a 16 % reduction at a high ventilation rate.

Parbst (1998), meanwhile, reported pilot scale studies on other low pressure drop means to remove dust from air streams, concluding that using a labyrinthine parallel plate arrangement was promising, as it achieved 41 % reduction of the concentration of total dust with an additional pressure drop of only 0.3 Pa at a superficial air velocity of 8 cm s-1.

6.3. Brainstorm on abatement approaches

A brainstorm session was held at SRI on 8 May 2000, with the aim of teasing out either new angles on existing approaches to abatement, or perhaps novel approaches. Both SRI and IC were represented at the brainstorm, at which some participants were experienced in the topic and some, deliberately, were inexperienced. Full lists of the ideas generated, before any assessment for feasibility, are shown in Annex G. A feasibility assessment was subsequently carried out on the ideas which were felt to be the most promising ones: tables summarising the assessment results are given in Annex E.

The general rule that prevention is better than cure applies, no less than elsewhere, to the abatement of dust from animal houses (the major source of particulates from UK agriculture: see Section 3). Given that much (but not all) of the dust in animal houses arises from feed, we recommend that modifying the feed properties should be pursued as a reliable means of reducing dust emissions from animal houses. There is currently interest in encouraging a change from dry feeding to wet feeding of UK fattening pigs (pers. comm., P. Gill, Meat and Livestock Commission; 2000). The main incentive is economic, but there should also be a considerable benefit in reduced dust emissions.

20

Projecttitle



One novel approach which appears to be worthy of research (even though it is curative rather than preventive) is the use of strategically placed vegetation e.g. tree belts around animal houses as a means of capturing dusts. This is essentially a passive form of capture, so should be reliable. Tree belts or similar can be designed so as to offer any desired level of porosity to air flow and thus should capture dusts more effectively than does a solid windbreak wall. The use of tree belts for dust capture is already practised by certain industries in China (pers. comm., J. J. Colls, 2000).Turning to arable operations, it will not be easy to recommend effective abatement options before better data for UK conditions are available. For example, many of the available emission factors have been researched in California with its very different climate.

7. Gaps in knowledge and needs for future research

Sections 2 and 3 together revealed a wide range of potential agricultural sources of particulate material, and a paucity of information to quantify them. Although agricultural emissions are more remote from population centres, reducing their importance relative to urban emissions, certain biological components have been identified as potentially toxic or allergenic (Section 4).

The following are recommendations for research:-

7.1. Characterisation of rural particulates and comparison with urban particulate composition

Monitoring at 3 to 4 sites selected to represent different types of farming (pig, poultry, arable) would be a valuable extension of the current UK monitoring network, and would greatly help to characterise the daily variations as well as annual average concentrations in rural areas as compared with urban areas. This would help to distinguish the contribution of different agricultural activities, including those generating occasional peak emissions. Such monitoring could be conducted on experimental farms where daily activities are continuously registered, and relevant data such as weather and soil conditions can be collected. Liaison with DETR is important to ensure compatibility with the TEOMs used to make measurements at existing UK monitoring sites.

Detailed consideration should be given to selecting and developing appropriate measurement techniques for sampling of agricultural particles, since traditional methods for PM10 such as TEOMS may dehydrate or destroy critical components. Rural aerosols may have quite different characteristics from urban aerosols in this respect. Detailed examination to characterise different components, especially aerobiological components (see below), is very time consuming and needs critical selection procedures and automation as far as possible .

7.2. Research to quantify major sources identified in emissions inventory

Livestock farming

Pig and poultry buildings have been identified as significant sources of PM10, and there have already been useful studies of these in the CAMAR project (Takai et al, 1998) which need further analysis to draw out different husbandry practices. However the measurements indicate a wide variability of concentrations, and were not specifically designed to define emissions of PM10 and PM2.5. Further work is required on pig and poultry units, representative of UK conditions and trends in agriculture, to understand the variability of particle emissions depending on feed, bedding, ventilation etc. Beef and dairy cattle appear less important. Characterisation of particle composition as well as size will also help with source apportionment to indicate the origin of particulates, and hence indicate effective abatement techniques. The overall cycle of management needs to be considered including cleaning out, and manure storage and spreading of livestock wastes and poultry manures.

21

Projecttitle



Arable farming

Estimates in Section 3 indicate that some cultivation and harvesting processes may lead to significant emissions, and may be concentrated in short periods of the year contributing to episodes. But there are large uncertainties. Measurements available are not sufficient to quantify emissions, and field experiments are required designed to assess emissions from different cultivation and harvesting processes with different soil types and conditions. These include soil tilling, fertiliser application and liming, spraying, harvesting and ploughing. There is also resuspended dust from farm roads. Some initial studies undertaken with UK soils with high silt content in dry conditions would help to give initial upper estimates. Some post-harvest processes such as grain drying are known to be dusty but there is little quantitative information available. It is recommended that some measurements are made for a large UK grain drying installation to estimate the emissions.

Other farming activities

Energy use and combustion by agriculture can also generate significant emissions, but there is limited information about incineration on farms. Some other specific farming activities have been identified which may generate high particulate emissions - for example compost production for mushroom farming. Some targeted spot measurements would help to indicate whether such processes can be significant for local communities.

7.3. Research on abatement measures

Section 6 has identified some potential abatement techniques, and further information is given in Annex E. Many of these techniques may be impractical, but others need further investigation - particularly those that might lead to significant reductions. Only by farm-scale experiments to test the effectiveness of these measures can they be recommended for adoption as proven and available techniques. Since emissions from livestock housing and arable farming are potentially the largest, abatement measures applicable to these should preferably be executed first. e.g. converting from dry feeding to liquid feeding of pigs.

7.4. Research on toxic and allergenic components and agricultural bioaerosols

The APEG report (DETR, 1999) suggests that up to 10% of PM10 may be of biological origin; however there are few measurements in the UK to support this or to identify the origins. This is particularly important, because as indicated in Section 4, some components of the bioaerosols generated by agriculture may be toxic or allergenic, and hence particularly related to health effects This includes endotoxins, bacteria, fungal spores and storage mites. The needs for further work are summarised below:-

7.4.1. Studies to characterise and quantify key components of agricultural bioaerosols

Measurements to establish the physical, chemical and microbiological properties of agricultural aerosols and how they compare with those of urban bioaerosols; the relative contributions of toxic or allergenic components from agricultural and non-agricultural sources.Field studies and controlled experiments to quantify emissions and concentrations of these potentially harmful components.

7.4.2. Assessment of exposure to toxic and allergenic components and associated risk

Assessment of the likely exposure patterns of both rural and urban populations.Studies of the current incidence of agricultural zoonoses, and of occupational respiratory deseases in UK agricultural workers and their association with bioaerosols: epidemiological studies of health in areas close to major sources of agricultural bioaerosols. Information on current concerns of Agricultural Inspectors in the HSE with respect to agricultural bioaerosols. Deductions concerning populations at risk.

22

Projecttitle



7.5. Secondary aerosol formation

This project has been concerned with primary particulate emissions from agriculture. However the overall particulate exposure also includes secondary particulates, i.e. these formed in the atmosphere. In this context the contribution of agriculture to secondary organic aerosol formation also need to be investigated, and a small scoping study on this topic is also recommended.

8. Summary and conclusions

This report indicates that there is a great paucity of quantitative information on many aspects of agriculture with regard to particulate emissions. Most of what is available is concerned with concentrations rather than emissions, and does not relate directly to PM10. However initial estimates have been made of UK emissions where possible. Animal housing is estimated to be the largest source, and although more is known about these emissions than about those from other agricultural sources, further research is nevertheless required. There is also considerable uncertainty about emissions from arable farming, which may also be large, and about emissions from on-farm combustion. An initial review has been undertaken to identify potential measures to reduce emissions, but these are tentative and unexplored.

Research needs include a mixture of monitoring, field campaigns, and controlled experiments in different situations. Emphasis has been placed on improved understanding of the origins and characterisation of emissions, which will also help to indicate what abatement measures are likely to be effective. Such measures can then be tested and assessed.

Although atmospheric dilution will reduce the contribution from agricultural emissions to population exposure compared with urban emissions, attention has been drawn to particular components of agricultural bioaerosols which may be toxic or allergenic. These include endotoxins, bacteria, fungal spores and storage mites. Research is urgently needed to investigate exposure to these components and the associated risks.

9. Acknowledgements

The authors are grateful for assistance and information from many people. Besides their IC and SRI colleagues, particular thanks are due to Andrew Fergusson and Ian Lean of Wye College, Fiona Nicholson and Fiona Short of ADAS Gleadthorpe, Bernard Fisher, Peter Burt and Poonam Sharma of NRI/University of Greenwich at Chatham, Patrick Gaffney of the University of California at Davis, and Jens Seedorf from the Hannover Veterinary High School.

The authors are also grateful for advice and information from: Mark Nieuwenhuijsen of Imperial College, Michael Jeger, Joe Lopez Real and Howard Lee of Wye College, David Jones of ADAS Rosemaund, Nigel Hardwick of MAFF Central Science Laboratories, Phil Metcalf of ADAS Wolverhampton, Nick Bradshaw of ADAS Cardiff, Sally Runham of ADAS Arthur Rickwood, Colin Roberts of the Animal Health Trust, Tim Murrels and Justin Goodwin of AEA Technology, Jean Emberlin of UK Pollen Network, Eric Caulton of Scottish Centre for Pollen Studies, Jim Taylor of the Bracken Advisory Commission, and the staff of Ecobeds, Marksway Horsehage and Mollichaff, and Dobson and Horrell (manufacturers of bedding and feed for horses).

23

Projecttitle



Annex A: Glossary of technical terms and abbreviations

A.1. Definitions and properties of particulate matter

Aerodynamic diameter The aerodynamic diameter of a particle is the diameter of a spherical particle (of the same density) which has the same sedimentation rate.

Inhalable dust (formerly termed inspirable dust) The fraction of particulate matter which can be taken into the human airway. This includes all size fractions of particles suspended in air.

Inorganic material Mineral material, derived from geological sources.

Mass median aerodynamic diameter (MMAD) Term used to characterise the distribution of sizes of particulate in cloud. MMAD is the size where half the mass is contained in particles of aerodynamic diameter smaller than the stated aerodynamic diameter, and half the mass in larger particles.

Mass Median Diameter (MMD) As for MMAD except the diameter considered is the actual diameter rather than the aerodynamic diameter.

Organic material Material derived from, or consisting of, living organisms, either alive or dead.

PM Particulate matter suspended in air.

PM10 Particulate matter smaller than 10 m aerodynamic diameter, or more strictly, particles which pass through a size selective inlet with a 50 % efficiency cut-off at 10 m diameter, with higher efficiency for smaller particles, and lower efficiency for larger ones.

PM2.5 Particulate matter smaller than 2.5 m aerodynamic diameter, or more strictly, particles which pass through a size selective inlet with a 50 % efficiency cut-off at 2.5 m diameter, with higher efficiency for smaller particles, and lower efficiency for larger ones.

Respirable dust Particles which can penetrate to the unciliated regions of the deep lung. This is currently taken as particles smaller than 3.5 - 4 m diameter (more precisely, with a 50 % efficiency at 3.5 or 4 m diameter particles, and a well defined curve ). Older articles considered 5 m as the 50 % cut-off point.

Thoracic fraction of dust The fraction of particulate which can enter the thorax. This is approximately PM10.

Total suspended particulate (TSP) A term describing the gravimetrically determined mass loading of airborne particles, most commonly associated with use of the US high volume air sampler in which particles are collected on a filter for weighing. (This is approximately the same as inhalable dust.)

A.2. Types of device used for measuring particulate concentrations in air

Burkhard trap, Hirst trap Standard apparatus for measuring levels of spores and pollen grains in air. Generally operated continuously, particles being collected on adhesive tape and observed microscopically.

Cascade impactor Sampler commonly used in the USA and Europe for measuring particulates of all, or a selection of, size fractions. Also known as a Anderson sampler.

Cyclone sampler Generally used to measure respirable dust (see above). A cyclone is first used to remove larger particles, so that only smaller particles are captured by a fibrous filter downstream.

24

Projecttitle



IOM sampler Institute of Occupational Medicine approved sampler for measuring inhalable dust concentrations.

Personal sampler Any sampler attached to a person, usually near the breathing zone, to estimate personal exposure.

TEOM Tapered element oscillating microbalance. A device for continuously measuring concentrations of total dust. Several versions are available commercially.

A.3. Terms concerning biological particles

Allergen Substance which causes allergy.

Colony forming unit (cfu) Growth of a single bacterium or fungal spore on a growth medium, e.g. agar, leads to a colony of microorganisms, which can then be observed and counted.

Endotoxin A substance found in cell wall of many Gram-negative bacteria, originating largely from faecal material. Endotoxins are powerful lung irritants, even in small doses.

Pathogen Organism which causes animal or human disease.

Spore Microbes in a state adapted for dispersal in the air.

A.4. Types of livestock units

Broilers Chickens raised for meat (normally in a building with a solid floor on which litter, e.g. wood shavings or straw, has been spread). Cages Laying hens are normally kept in halting cages.

Farrowing unit A pig house designed to hold sows from immediately before the birth of their piglets to several weeks after.

Finishing unit A pig house designed to hold pigs being fattened to their final weight prior to slaughter.

Laying hens Hens kept for egg-laying.

Percheries A system for housing laying hens which allows the hens to move around inside the building and offers them facilities for perching.

Silage Forage (normally grass or maize) fermented and stored anaerobically.

Weaners Young pigs which are no longer suckling their mother.

A.5. Abbreviations (where not already included in one of the sections above)

ADAS A consultancy involved in agricultural research at many sites in England and Wales.

25

Projecttitle



ADMS Atmospheric Dispersion Modelling System. Used to model the dispersion of pollutants in air, using pollution monitoring data and allowing for local topography.

AQMA Air Quality Management Area.

HSE Health and Safety Executive..

IC Imperial College.

IPC Integrated Pollution and Control.

SRI Silsoe Research Institute.

TNO Toegepast Natuurwetenschappelijk Onderzoek (the Netherlands Organisation for Applied Scientific Research)

USEPA United States Environmental Protection Agency (part of the US government).

26

Projecttitle



Annex B: Detailed review of potential sources

TABLE OF CONTENTS OF ANNEX B

B.1. Introduction 29

B.2. Livestock

B.2.1. Livestock buildingsB.2.1.1. Direct measurements of particulate emissions in England and 29

other European countriesB.2.1.2. Measurements of concentrations of particulates in air within buildings 32B.2.1.3. Studies referring to dust levels measured during hay making 43B.2.1.4. Harmful substances contained within dust in some livestock buildings 44B.2.1.5. Studies which give information about the character and sources of particulates in

animal buildings 44B.2.1.6. Materials which contribute to dust in animal buildings 45B.2.1.7. Influence of environmental factors on particulate levels 47

B.2.2. Outdoor livestock 47

B.3. Arable Farming

B.3.1. Emission factors 50B.3.1.1. Land preparation 50B.3.1.2. Cereal harvesting 50

B.3.2. Measurements of concentrations of particulate matter in air during arable farming activitiesB.3.2.1. Californian measurements 51B.3.2.2. Batel, 1979 54B.3.2.3. Norén, 1985 55B.3.2.4. Louhelainen et al, 1987a 55B.3.2.5. Burg et al, 1982 55B.3.2.6. Darke et al, 1976 56B.3.2.7. Atiemo et al, 1978 56B.3.2.8. Size distribution and characterisation of dust emitted from arable operations 57B.3.2.9. Conclusion 58

B.3.3. Spores and pollen grains relevant to agricultureB.3.3.1. Fungal spores 59B.3.3.2. Emissions of spores from compost, mulches, and other decaying organic matter 61B.3.3.3. Bracken 62B.3.3.4. Pollen 62B.3.3.5. Methods of counting and identifying spores and pollen grains 63

B.3.4. Wind erosion of UK soils 63B.3.5. Applications to agricultural land

B.3.5.1. Manure 64B.3.5.2. Sewage sludge 65B.3.5.3. Fertiliser 65B.3.5.4. Lime 65B.3.5.5. Pesticides 66B.3.5.6. Other wastes 67

27

Projecttitle



B.4. Post harvest and miscellaneous processesB.4.1. Grain storage and transport, and processing into animal feed 67B.4.2. Cereal drying 69B.4.3. Potato grading 69B.4.4. Incinerators, fires 69B.4.5. Burning of stubble 69

B.5. Unpaved roads om farms 70

B.6. Energy use 72

B.7. ConclusionsB.7.1. Livestock 72B.7.2. Arable 73B.7.3. Post harvest and miscellaneous processes 74B.7.4. Unpaved roads 74B.7.5. Energy 75

28

Projecttitle



B.1. Introduction

There have been a number of studies on particulate concentrations generated by agricultural activities. A small number of these aimed to measure emission rates of particulate material. However, the majority of the studies are concerned with effects on the health of farmworkers, and so monitor concentrations of particulate to which workers are exposed. Thus, they can only be an approximate guide to average concentrations within a building or near a process, as the studies never aimed to find this information. Thus, these studies would not give such a reliable measurement of ambient dust levels in air, as would studies which are specifically designed to obtain this information. Furthermore, many of the occupational exposure studies collect information on dust levels together with epidemiological data, and attempt to deduce correlations between exposure to dust, and bronchial problems. Many studies also measure concentrations of certain substances within the dusts, which are believed to be particularly harmful, namely bacteria, certain fungal spores, and endotoxin. (The effects these on health is discussed in greater detail in Section 4 of the main report.) Thus, for a number of these studies, determining levels of particulate in the air was not the main aim of the study.

Another important point to note is that different types of samplers were used in the studies listed. Thus, direct comparisons of results should be made with caution, as not all samplers measure exactly what they are stated to measure. All the samples referred to in these studies used a gravimetric method, where air was sucked through a filter or other capturing device, and later weighed. Some degree of error is inherent in all operations, and the main sources of error are detailed in Annex C. It is also important to state that studies using numerous measurements, would normally give more reliable estimates than studies using few measurements, and that longer sampling times (and also higher volumes of air sampling) also improve the quality of results.

It is important to regard these values given as a guide to particulate concentrations, not as exact measurements, especially in the case of arable operations, where measurements are extremely variable for a number of reasons. Another point to note is that all emissions quoted are gross, i.e. as measured close to source. These do not allow for the proportion of material, which would deposit on the nearby fields, so never be emitted from the farm itself.

B.2. Livestock

B.2.1. Livestock buildings

Emissions of gases and particulate matter from livestock buildings have been described in numerous papers, as a matter of considerable concern. The largest numbers of papers refer to occupational hazards of farmworkers, including respiratory problems which are believed to be triggered by exposure to dust and other substances in livestock buildings. High concentrations of air pollutants can also affect the health and viability of the livestock, thus having a commercial implication. A more recent concern is the potential effect on the environment beyond the farm, both for ecological reasons (such as damage to wildlife, or greenhouse gases) and nuisance or the effect on populations.

B.2.1.1. Direct measurements of particulate emissions in England and other European countries

(i) Study by Takai et al, 1998

Livestock buildings are the only sources of agricultural emissions from which rates have been directly measured in the UK. A large collaborative survey, involving institutes in Denmark, The Netherlands, and Germany, as well as in England, measured both concentrations in, and emissions from, buildings housing pigs, poultry and cattle (Takai et al, 1998, Wathes et al, 1998). Measurements were made of "inhalable dust" (total suspended particulate), and "respirable dust" (below approx. 5 m diameter), as well as of gaseous and other pollutants.

29

Projecttitle



Measurements were made over 24 hr periods, for several days, at usually 4 different buildings of each type in each country. In accordance with husbandry practice, measurements were made at each building once in summer and once in winter.

Samplers were located in a vertical cross section at 7 locations to represent the location of the animals, the stockmen and the exhaust to the ventilation system. Inhalable (total) dust was measured with Institute of Occupational Medicine dust samplers (sampling at 2 litres / min, collecting dust on filters and weighing); respirable dust was collected using cyclone dust samplers with the 50 % cutoff at 5 m particle diameter.

The emission rates were calculated by multiplying concentrations of particulate and the ventilation rate. The concentrations used were those at ventilation exhaust level (in the case of mechanically ventilated buildings) or a mean building concentration level (naturally ventilated buildings). The ventilation rate was calculated using a balance of heat and CO2, produced by the animals. This allowed calculation of rates in naturally, as well as force-ventilated buildings. It was less accurate than using fan wheel anemometers, but more practicable, and agreement with more thoroughly tested buildings was within around 20 %. To test this approach, a more detailed study was done on 4 buildings, measuring ventilation rate by CO2 mass balance and also fan-wheel anemometers, showed good agreement, though this was still considered to be the main source of uncertainty, particularly in naturally ventilated buildings. (Further details are given in Phillips et al, 1998.)

These calculations are therefore considered a reasonable estimate of the emissions from livestock buildings.

The results showed significant variation in emission rates, between countries, between housing types, and in some cases, between season, based on emissions per animal, or the more common unit, emissions per 500 kg liveweight of animal. In general terms, cows and pigs were found to cause similar emissions per animal. Per livestock unit (500 kg of liveweight), cows gave the lowest emissions, pigs much higher and poultry emitted the greatest amount of particulate. Variations were attributed to differences in housing design, which can considerably affect the dust production.

TABLE B.1. Overall Mean Inhalable and Respirable Dust Emission Rates from cattle, pig and poultry buildings in England, Denmark, Germany and The Netherlands (Adapted from Takai, et al, 1998)

Cattle buildings Pig buildings Poultry buildingsInhalable dust

Respirable dust

Inhalable dust

Respirable dust

Inhalable dust

Respirable dust

Emission in kg/y per animal or bird

0.96 0.17 0.97 0.12 0.11 0.02

Emission in kg/y per 500 kg weight of animal or bird

1.27 0.21 6.68 0.75 27.7 4.42

TABLE B.2. Overall Mean Inhalable and Respirable Dust Emission Rates from cattle, pig and poultry buildings in England alone (Adapted from Takai, et al, 1998)

Cattle buildings Pig buildings Poultry buildingsInhalable dust

Respirable dust

Inhalable dust Respirable dust

Inhalable dust

Respirable dust

Emission in kg/y per 500 kg weight of animal or bird

0.85 0.27 5.55 0.82 27.5 3.27

30

Projecttitle



Separate results are also listed, for different categories of pig (sows, weaners and fatteners), cattle (dairy, beef and calves), and poultry (layers and broilers), and types of housing (for cattle: litter, cubicles and slats; for pigs: litter and slats, and for hens: perchery (barn), cage and litter). It was not possible to generalise about whether bedding increased or decreased dust concentrations, as the effect varied between type of animal and from country to country.

The English measurements were taken during 1993-1995, in both summer and winter, in SE England. They are considered to be reasonably representative of current practice in the UK. However, it must be noted that:

Pig buildingsMost of the pigs studied were fed on pelleted feed. This is considerably less dusty than meal feed (ground but not pelleted), which is used on some farms. Such farms would be expected to have considerably higher particulate emissions. (The nature and composition of dust from animal buildings is discussed in Section B.2.1.5.)

Poultry buildingsThe poultry buildings studied were relatively old, and it is possible (as the poultry industry have suggested) that these measurements slightly over-estimate present-day emissions.

Occasional operationsThe approach of measuring over a 24 hour cycle gives a measurement which allows for diurnal variations in dust levels. However, some operations which are done infrequently, (such as moving animals or birds, or cleaning out units) can generate large amounts of particulate matter (e.g. Whyte, 2000), but such operations were not measured in this study. This would suggest that these measurements could underestimate total annual emissions of particulate matter.

(ii) Study of emissions from buildings in The Netherlands, by TNO (Berdowski et al, 1997)

Estimations of emissions from pig and poultry buildings in the Netherlands were made by TNO. This is an earlier study, made before the data from Takai et al’s study became available, and the authors state that this is a first order estimate of emissions, and should be considered as a first approximation. They also state that more recent research in the Netherlands suggest that the above study might be overestimating emissions.

Pig buildingsFor pig buildings, a ventilation rate of 100 m3 per hour per animal was assumed. The authors state that this is relatively high, and on colder days a lower rate might be used. The concentrations of total dust were taken as 5.4 - 6.4 mg.m-3, and of PM10 as 2.5 - 2.6 mg.m-3. This yielded an estimated PM10 emission factor of 2.2 kg/year.animal, or 0.016 kg per year per kg of liveweight. The standard deviation was 1.3 kg /year.animal.

Poultry buildingsIn poultry buildings, a ventilation rate of 1 m3 per hour per kg liveweight housed was assumed. Concentrations of inhalable dust of 5 - 20 mg.m-3 were used, of which 70 % was assumed to be particles smaller than 10 m diameter. Thus, a mean PM10 concentration of 10 mg.m-3 is assumed in the calculation. This yielded an emission factor of 0.086 kg per year per kg of liveweight, with a standard deviation of 0.05 kg/year.kg of liveweight.

PM2.5 and PM0.1 fractionsThe authors state that very few data are available on very fine particulate material from these types of emission sources, and that emissions are reported to be in the range of 1 - 10 m. A PM2.5 fraction of 0.3 for pig buildings, and 0.5 for poultry buildings was assumed. This yields the following emission estimates:

Pig buildings: 0.75 kg PM2.5 per year per animal (standard deviation of 0.6 kg/y.animal)

31

Projecttitle



Poultry buildings: 0.043 kg PM2.5 per year per kg lw (standard deviation of 0.032 kg/y.kg lw)

TABLE B.3. Comparison of the above estimations of emission rates of particulates from pig and poultry buildings in kg per year per 500 kg liveweightReference Country Pig Poultry

Inhalable dust

PM10 Respirable dust

Inhalable dust

PM10 Respirable dust

Takai et al, 19981

England 5.5 0.82 27.5 3.3Netherlands 5.9 0.65 31.9 6.3

Berdowski et al,19972

Netherlands 203 8 614 43

1. adapted from Takai et al, 19982. adapted from Berdowski et al, 1997 (TNO Report)3. taking the respirable dust fraction to be 40 % of the total dust4. taking the respirable dust fraction to be 70 % of the total dust

The TNO report notes significantly higher emission rates than the study by Takai et al. For pig buildings, concentrations of total dust in the air are a little higher than in Takai et al’s study, though well within the range of other studies. However, PM10 estimations appear to be rather higher than other studies would suggest. However, the variability between surveys is wide.

The ventilation rate used for the TNO calculation is 100m3 per hour per animal. If one takes a mean pig mass of 50 kg, this is equivalent to a rate of 1000 m3 per hour per 500 kg of animal. In Takai et al’s study, the ventilation rate was calculated for each building for each measurement, and ventilation rates were much lower. Average rates were calculated separately for sows, weaners and fatteners, in winter and in summer, but assuming equal numbers of all of these types of pigs, would yield a mean ventilation rate in the Netherlands of 200 m3 per hour per 500 kg of pig. This is only one fifth of the rate assumed by Berdowski et al. The average ventilation rate in English pig buildings, using the same assumptions, was 340 m3 per hour per 500 kg of pig. These differences alone would account for the difference.

For poultry buildings, emission rates again are much higher in the TNO calculation than in Takai et al’s study. The concentrations used in TNO's calculation (5 - 20 mg.m-3) are a wide range also reported in other studies: 5 - 10 mg.m-3 is confirmed by other studies, while 20 mg.m-3 would probably correspond to operations such as moving birds. Takai et al found slightly lower mean concentrations, of around 3 mg.m-3. The two calculations used similar ventilation rates, though, again, in Takai et al’s study, this was calculated individually for every building, and TNO used a single representative rate.

Discussion of concentrations from different studies is detailed in the following section.

The proportion of fine material assumed by TNO for both pig and poultry buildings appears to be rather high compared to results of other studies. This is described in more detail below.

B.2.1.2. Measurements of concentrations of particulates in air within buildings

A number of studies, in several countries, have included measurements of levels of particulate matter in air within buildings. The most recent and relevant of these studies is the one described above, as these were in England and other countries in northern Europe; they were recent, and most importantly, measurements were conducted over a 24 hour period and sampling locations to give a representative mean building concentration.

32

Projecttitle



Other studies generally sampled for much shorter durations (though in some cases, on more occasions). They were generally concerned with assessing the concentrations to which farm workers are exposed, which is not necessarily representative of the building concentration. All but one study used a gravimetric method of measuring mass of dust, though different types of samplers were used. All studies measured mass of inhalable dust i.e. all size fractions suspended in air: some studies also measured respirable dust (particles below 5 m diameter); PM10 and 2.5 were not measured. These can be estimated from total particulate using an estimated size distribution factor; this can obviously entail some error as there are differences in the size distributions of particulates found in different studies.

Some studies also measured concentrations of other pollutants, particularly ammonia, bacteria, fungal spores, and endotoxin. Many also detail medical information concerning the workers. There is a further wealth of literature on occupational exposure to dust and other harmful substances in farm buildings, particularly pig and poultry buildings, but many of these articles do not include measurements of conditions within the buildings.

Details of individual studies are given below.

Pig buildings

(i) Study by Takai et al, 1998

Mean concentrations of total suspended particulate were found to be around 2 mg.m-3 in each country; the overall mean across the whole study was 2.19 mg.m-3, and the English buildings averaged a slightly lower concentration of 1.87 mg.m-3. Respirable dust concentrations were also similar across the 4 countries, around 0.2 mg.m-3, with an overall average of 0.23 mg.m-3, and the average in the English buildings of 0.24 mg.m-3.

Types of buildings sampled were: sows on litter, sows on mesh or slats (slurry system), weaner pigs on mesh or slats (slurry system), fattening pigs on litter, and fattening pigs on slats (slurry system). The number of samples at each location was small (one in summer, and one in winter) but the long sampling duration and fairly large sample size (130 buildings altogether, of which 20 were English) suggest that these results are fairly representative. They also include different types of pig units (sows on litter and slats, weaner pigs on slats, and fattening pigs on litter and slats), which occur on commercial farms in the UK.

(ii) Crook et al, 1991

Measurements on 20 units on 12 farms in northern Scotland found mean concentrations across all units ranged from 3.5 mg.m-3 in August, to 11.5 mg.m-3 in November. In this study, measurements were made every 4 weeks from July to December. Each measurement was for the duration of a working shift, at a point above the central walkway 1.5 m above the floor. Air was sampled at a rate of 2 l/min. All the units were for growers and finishers, weighing 40 - 90 kg.

Mean concentrations over the interval July – December measured in individual buildings ranged from 1.7 to 21.0 mg.m-3, though the majority of building means (83 %) were below 10 mg.m-3. Individual measurements ranged from below 1 mg.m-3 to over 40 mg.m-3. An increase in concentration from summer to winter is normal, as the ventilation rate is decreased in order to conserve heat. It was noted that most of the buildings used milled feed, which is a lot dustier than the pellets; and that of the 4 buildings with the lowest dust concentrations, 3 were using pelleted feed. The type of feed system used also affected the dust levels. The mean dust level among buildings with a restricted feed system (where feed is distributed every few hours) was 10.0 mg.m-3, twice that for buildings using an unrestricted feed system (in which the feed system is enclosed) (4.6 mg.m-3). This difference was found to be significant at the 5 % level).

33

Projecttitle



This is considered to be another comprehensive study of particulate concentrations within buildings. The concentrations found are higher than those of the first study. The differences are probably because of 1) older buildings 2) dustier feed 3) only sampling in the daytime, when the pigs are more active than at night.

(iii) Netherlands surveys cited in TNO report by Berdowski et al, 1997

Berdowski et al (1997) cite two main studies which quoted concentrations of dust in pig buildings. One (Preller et al, 1993) found total dust levels in south Netherlands stocks were in the range 5.4 - 6.4 mg.m-3. These were based on measurements concerned with occupational exposure of workers. Another study (Roelofs, 1995) measured PM10 concentrations in buildings for meat pigs ("growers") and for piglet breeding ("weaners" or "farrowing units"), and found these to be 2.5 mg.m-3 and 2.6 mg.m-3 respectively. From these, the authors deduced that around 40 % of the dust in pig buildings is PM10.

(iv) Heber et al, 1988

This study of buildings in Kansas, USA, found mean particulate levels of 8.3 mg.m-3, with individual measurements ranging from 0.4 to 38.2 mg.m-3. There was a modest difference between levels in naturally and mechanically ventilated buildings (8.8 and 6.9 mg.m-3 respectively).

Measurements were taken monthly between July 1985 and March 1986. Sampling was for 60 minutes, using 2 high volume samplers located towards the centre of the building.

(v) Louhalainen et al, 1987a and b

Two separate studies are reported, both of which include levels of particulates in pig buildings in Finland in the early 1980's. Mean concentrations of between 5 and 13 mg.m-3 were found.

Buildings with sows and fatteners were measured separately, by different samplers, both at the breathing zone of workers, and at stationary sites chosen to represent exposure of workers. Sampling was in accordance with the Finnish standard for assessing occupational exposure to dust. There was some variation between these two measurements, though apart from in the earlier study, these differences were not significant. Measurements were of relatively short duration, 19 - 90 minutes, on a fairly small number of units (between 4 and 11). Samples were conducted on between 1 and 25 occasions, with more measurements being taken at larger units.

In view of the fairly small sample size, it is considered that this should be taken only as an indication of levels of particulates which can be found in pig buildings.

(vi) Donham et al, 1989

A study of 28 Swedish pig buildings yielded average dust concentrations of 4.6 mg.m-3 (measured by area sampling, and 6.4 mg.m-3 by personal sampling. (Both methods using personal sampling pumps of 1.9 l/min.) Respirable dust was also measured by cyclones, and found to be 0.33 - 0.34 mg.m-3 (using area/personal sampling) as determined by cyclone samplers.

These authors also refer to other studies, documenting concentrations in a similar range.

34

Projecttitle



(vii) Donham et al, 1986a

A study of dusts collected from 21 pig buildings in Iowa was conducted during the winter months December to February. Total particulates were collected in cassettes, drawing air at 1.7 l/min; respirable dust was measured in cyclone preselectors. Six pairs of total mass and respirable samplers, were collected side-by-side in a grid fashion to cover the entire working area of each building.

Mean mass concentrations were found to be 7.6 mg.m-3 (using cascade impactors) and 6.25 mg.m-3 (using cassette samplers), which is fairly good agreement. Both devices found farrowing units were the least dusty (3 - 4 mg.m-3), but differed as to whether nursery/grower or finishing units were the dustiest.

The respirable fraction, as measured by the cyclone, was 0.53 mg.m-3, with considerably higher concentrations being found in finishing units than in the other building types. Cascade impactors measured much higher concentrations of respirable dust (2.5 mg.m-3 overall, similar in all types of units) which the authors consider to be erroneously high (Donham et al, 1986b). Further studies were made to characterise the size distribution and composition of the dust: these are described later.

To sum up, a wide range of concentrations in pig buildings has been reported, but most mean values for total dust lie between 2 and 12 mg.m-3.

The study documented by Takai et al gives concentrations lower, though not usually outside the range, of dust levels reported elsewhere. This study involved 2 samples each of 24 hours, taken at 20 buildings in England. The other study in the UK, by Crook, 1991, found higher levels of particulate, ranging from a minimum of 3.5 mg.m-3 in summer, to a maximum of 11.5 mg.m-3 in winter. Reasons for the differences could include

(a) Diurnal variations

The study of Crook et al (1991) (as with all other studies except that of Takai et al, 1998) only made measurements during the daytime. Pigs are diurnal animals, so more active during the day. Animal movements disperse settled dust and generate surges in concentrations of particulates (as measured by Perkins et al, 1997), so dust levels will be higher during the day than at night.

(b) Age of buildings

The Crook et al study was conducted a few years prior to the Takai study, so older buildings may have been partly responsible for differences. In Crook's study, a considerable range of concentrations was found between buildings (from 1.7 to 21.0 mg.m-3), though the worst building had levels approaching twice those in the second dustiest unit.

(c) Feeding system

In Crook et al's study, almost half the buildings used restricted batch feeding, where the feed is distributed several times a day; most of the others had unrestricted systems where feed is delivered in closed conveyors to hoppers and feeders. The lower dust levels found in buildings with unrestricted systems was found to be significant.

Takai's study included buildings with both feeding systems.

35

Projecttitle



(d) Feed type

Most of the feed given to pigs in Takai's study was pelleted, whereas in Crook's study was unpelleted meal feed. The meal feed is inherently dustier, and this is probably a major cause of difference between results.

(e) Location/climate

Buildings in the north of Scotland will experience, on average, a cooler climate than those in the south of England, so ventilation rates, particularly in winter would be expected to be lower, to conserve heat, so concentrations of airborne dust would be higher.

The other countries measured in the Takai et al study had small differences in concentration between each other. Other studies in the USA, Sweden and Finland, found mean concentration of between 4.6 and 13 mg.m-3, though individual measurements varied much more widely (0.4 to 50 mg.m-3). Donham's measurements of respirable dust in Swedish buildings are fairly close to levels found in English and other northern European buildings (Donham's mean = 0.34; Takai et al for England: mean = 0.24 mg.m-3). These measurements were of shorter duration, on older buildings (1980s) and in other countries where differences in climate and husbandry practice could affect results. Their closeness with the results should be taken as confirmation of the validity of the former measurements. Another factor of note is the differences in results from the same buildings, at the same time, with different samplers (Donham et al, 1989, and Louhelainen et al, 1987b).

36

Projecttitle



TABLE B.4. Comparison of different surveys of dust concentrations in pig buildings

Reference Country Mean dust concentrations (mg.m-3)

Sample position

Sampling time / season

Other comments

Total dust

Respir-able dust

Takai et al, 1998

England (S) 1.87 0.24 Air sampling, 7 locations in 20 buildings

24h mean, summer + winter

Included farrowing units, weaners, finishers, on bedding and slats.Most fed pelleted feedEngland, NL,

Germany, Denmark

2.19 0.23 As above, in 130 buildings

As above

Crook et al, 1991

Scotland (N) 3.5 * Over central walkway, in 20 buildings

Monthly measurementsJuly – December*(min overall summer mean)**max overall winter mean

Finishing units. 11.5 **

Roelofs, 1995 Netherlands 2.5-2.6 PM10

2.5 for meat pigs, 2.6 for piglet breeding

Preller et al 1993

Netherlands 5.4-6.4 Occupational exposure

Heber et al, 1988

USA, Kansas 8.3 2 high vol. samplers, building centre

Monthly, July-March

Louhelainen et al, 1987a, 1987b

Finland 5 – 13 Short measurement duration

Donham et al, 1989

Sweden 4.6 0.33 Area sampling 28 buildings6.4 0.34 Personal

samplingDonham et al, 1986

USA, Iowa 7.6 2.5 6 cascade impactors in air

December-February21 buildings

Included farrowing, nursery and finishing unitsFarrowers least dusty, with highest % respirable6.25 0.53 6 personal

samplers

Poultry buildings

(i) Takai et al, 1998

Mean concentrations of inhalable dust in English buildings were found to be 3.31 mg.m-3. This compares with mean levels in The Netherlands, Denmark and Germany of 4.58, 4.52, and 2.22 mg.m-3, and an overall mean concentration of 3.60 mg.m-3. Mean respirable dust levels were 0.51 mg.m-3 for English buildings, near to the overall mean of all countries of 0.45 mg.m-3. Highest mean respirable dust levels were found in Denmark (0.64 mg.m-3) and the lowest in Germany (0.19 mg.m-3).

These were based on a total of measurements from 12 buildings in England, and a total of 81 buildings in all the 4 countries. Buildings measured included laying units, broilers and percheries, though "free range" poultry were not included. Sampling methods were the same as described for the pig buildings.

Further detail of the concentrations within the different types of English poultry buildings, and how these varied with time of day and season, are given in Wathes et al, 1997.

37

Projecttitle



Broilers had consistently higher levels of both inhalable and respirable dust, than the other units, having mean levels of 2 – 3 times those found in cages. Percheries had intermediate dust levels. The mean levels of inhalable dust in broiler houses ranged from 7 mg.m-3 (summer, night) to 12 mg.m-3 (winter, night). In cages they ranged from around 1 mg.m-3 (night-time, winter and summer) to around 3 mg.m-3 (winter, daytime). Respirable dust levels showed much less diurnal variation, presumably because the finer particles are not removed from air by deposition, so remain suspended for long periods after disturbance. Differences between different types of units are attributed to the presence of deep litter in broilers, small amount of litter in percheries, and no litter in cages; cages restrict the movements of the birds and so their potential to generate dust.

(ii) Whyte, 2000

The interim results of a recent study of occupational exposure of stockmen working in 33 barns (percheries) in 10 farms (and also some cage units), gave mean concentrations of 9.8 mg.m-3 in percheries, and 4.2 mg.m-3 in cage systems. This compares favourably with an earlier study by the same author, cited, where exposure to cage system stockmen averaged 7.1 mg.m-3. There is a considerable amount of uncertainty with these results because of the small sample size.

About 25 % of the data are still to be analysed, but no great changes in mean exposure levels are expected (R. Whyte, pers. comm.). The barns studied are considered representative of current practice in the UK, and include a variety of litter storage systems.

Measurements were made of operator exposure during specific tasks in barns: concentrations varied from 5 to 63 mg.m-3. Most of the daily routine tasks gave below 10 mg.m-3 (maintenance, nestbox, bird and dead bird check). However, some less frequent tasks (litter spreading, cleaning and sweeping) showed much higher exposures from 35 to 63 mg.m-3.

Measurements were made by samplers carried by 12 stockmen working in a total of 33 barns on 10 farms, as they performed their duties. The sampling period was a working day, and samplers were standard Institute of Occupational Medicine type, with a flow rate of 2 l/min. The number of stockmen in the sample was fairly small: 10 measurements in the barns, 6 for the cage system, while in the previous study cited there were 49 measurements.

These concentrations are all somewhat higher than those found in Takai et al’s study previously described.

Reasons for the differences could include:-

(a) uncertainty in results(b) measurements during the day, when the birds are more active and so generate more dust(c) measurement of personal exposure could differ from mean building concentration(d) inclusion of tasks such as cleaning, which generate high levels of dust, and were not performed during

the study by Takai et al.

(iii) Berdowski et al, 1997

This TNO report cited a paper by Evers, 1995, which measured dust concentrations in chicken farms in The Netherlands. Concentrations of 5 - 20 mg.m-3 were quoted, of which 70 % was said to be particulate matter smaller than 10 m diameter.

38

Projecttitle



(iv) Jones et al, 1984

Though rather an old study, this is still a useful measure of dust levels in broiler units in the USA. Low levels of inhalable dust were measured in the building with young (7 day old) chicks on new litter (average 2 mg.m-3), higher levels above 7-day old chicks on old litter (mean 3.7 mg.m-3) but much higher levels were found in the building with 30 day old broilers (around 9 mg.m-3). Measurements of respirable dust levels were also made. These were around 0.1, 0.2 and 0.5 mg.m-3 for the 3 types of units, respectively.

Measurements were made at 3 locations in each building, 1.8 m above the floor, using a pump of flow rate 2 l/min to sample air for total dust collection, and using a cyclone collector. Particle size analysis was performed using a cascade impactor. Sampling durations were between 4 and 6 hours.

(v) Louhelainen et al, 1987a

A study of 4 poultry units in Finland in the early 1980's found mean concentrations of 5 - 7.2 mg.m-3 for "poultry yards with floors" (presumed broilers) and 6 - 13 mg.m-3 for "poultry yards with coops" (possibly a mistranslation of "cages"). The values are those measured by stationary sites and breathing zone samplers, as described for the same study of pig units. The sampling durations were short, 30 - 120 minutes, though repeated samples (between 6 and 13) were taken. Most measurements were done in the winter.

The lack of detail, the short sampling durations and the age of the study reduce its reliability in predicting emissions from UK poultry houses, but the findings are within the range of other studies.

(vi) Cravens et al, 1981

A study of dust levels in turkey buildings in the USA, found mean daytime levels of total dust of 12 and 17 mg.m-3 (for human breathing zone and bird breathing zone, respectively). Dust levels before dawn in the bird breathing zone were 6.5 mg.m-3. Identical barns were also measured which were using water spraying to control dust – the mean levels were reduced to between 4 and 5.5 mg.m-3. Respirable dust was also measured after dawn in the human breathing zone: the mean level was 2.6 mg.m-3 in the dry barn, and 0.8 mg.m-3 in the sprayed barn.

Sampling time was 1 hour, repeated 12 - 15 times for each reading, using samplers which collected 1 l/min of air. Particle size distribution was calculated from measurements.

(vii) Nieuwenhuijsen et al, 1999

A small number of measurements of the personal exposure of farmworkers in California were made, using personal IOM samplers to measure inhalable dust and cyclones to measure respirable dust. Feeding and handling birds exposed workers to 3 - 5 mg.m-3 inhalable dust, and around 0.28 mg.m-3 respirable dust (based on 5 measurements). Scraping poultry buildings gave much higher levels of inhalable dust (10.5 mg.m-3) which is consistent with other studies: surprisingly, respirable dust levels were recorded as below 0.2 mg.m-3.

(viii) Leistikow et al, 1989

This paper, which reported "walkthrough measurements" in poultry houses is not covered in detail here, as the sampling period is considered too short to be reliable. However, most measurements were between 0.1 and 6 mg.m-3, with a small number of much higher dust levels. These measurements again are surprisingly close to more recent and representative findings.

39

Projecttitle



The author also refers to older studies, where lower dust levels (below 5 mg.m-3) had been found in layer houses during routine operations such as feeding and watering, and also in broiler houses with young birds (up to 25 days old). Higher levels were noted in broiler houses with older birds, and while moving birds in a house with younger broilers and layer units.

TABLE B.5. Comparison of different surveys of dust concentrations in poultry buildings

Reference country Mean dust concentrations (mg.m-3)

Sample position

Sampling time / season

Other comments

Total dust

Respir-able dust

Takai et al, 1998


24h mean, summer + winter

Included layers in cages, broilers and percheries2-3 times level of dust in broilers as in cages, percheries intermediate

England, Netherlands, Germany, Denmark

3.60 0.45 As above, in 81 buildings

As above

Whyte, 2000 England 9.8 Personal exposure of stockmen

Overall mean during working day

Barns

4.2 Cages 5-10 Personal

exposure of stockmen

During routine tasks Barns

35-65 Personal exposure of stockmen

During cleaning, sweeping, litter spreading

Barns

Evers, 1995Cited in Berdowski et al, 1996

Netherlands 5-20

Jones et al, 1984

USA 1-5 0.05-0.3 3 locations in building 1.8m above floor

Sampling for 4-6 hours

7 day old broilers

7-11 0.4-0.6 30 day old broilers

Louhelainen et al, 1987a

Finland 5 – 13 Short measurement duration, most in winter

"Poultry yards with floors" and "poultry yards with coops"

Cravens et al, 1981

USA 12 2.6 Human breathing zone

Daytime, no misting Turkeys

12-15 short samples (1 hour) with low flow (1 litre/min) sampler

4 0.8 Human breathing zone

Daytime, misting to suppress dust

6.5 Bird breathing zone

Before dawn

Nieuwen-huijsen et al, 1999

USA, California

3-5 0.3 operators During handling and feeding

10 0.2 operators During "scraping houses"

Leistikow et al, 1989

Mostly 0.1 – 6

stockmen Walkthrough measurements

40

Projecttitle



To sum up, as with pig buildings, a wide range of concentrations has been reported, though most mean levels lie between 1 and 15 mg.m-3. Considerable variation occurs according to time of day, season, type of bird building (with broilers on litter being the dustiest, and layers in cages being the least dusty). Dust levels are also considerably increased during operations such as moving birds and cleaning out units. In broiler units, the dust levels were found to increase with age of bird, and the age of the litter was also found to be a significant factor.

As no other study has measured poultry buildings over such long sample times as did Takai et al’s study, it is suggested that these should be used to estimate levels. However, all other studies found higher levels of particulate concentration in buildings. Nature and time of measurements certainly account for much of the difference, but the absence of dusty processes such as moving birds, and cleaning buildings, suggest that the Takai et al results could be regarded as an underestimate of total annual particulate levels.

Respirable dust levels were only measured in 3 other studies, and results were within the 0.2 - 0.6 mg.m-3 for chickens/hens, very close to Takai et al’s mean levels of around 0.5 mg.m-3. Higher levels were found in the turkey houses studied by Cravens.

Cow Buildings

There have been very few studies on particulate levels in cow buildings, presumably because particulate levels are lower than in pig and poultry building, so these present less cause for concern. The findings are presented below.

(i) Takai et al, 1998

Overall mean concentrations of 0.22 mg.m-3 inhalable dust were found in English buildings. This is below the concentrations measured in buildings in The Netherlands, Denmark and Germany, which ranged from 0.30 to 0.65 mg.m-3. Overall mean respirable dust concentrations, however, were higher in English buildings (0.15 mg.m-3) than in the other countries (0.04 to 0.09 mg.m-3).

One major difference in practice between England and the other countries is that cows are only housed in the winter (6 months) and so only buildings with calves in were measured in the summer. In the other countries, measurements were taken in both summer and winter.

Measurements in England were made on 16 buildings in the winter, which were: dairy cows in tie stalls or on litter, dairy cows in cubicles (slurry system), beef cattle on litter, and calves on litter.

The sampling procedure was the same as for pig buildings in the same study.

(ii) Louhelainen et al, 1987a

Measurements at 8 dairy farms in Finland found mean dust concentrations of 1 mg.m-3 at the stationary sites, and 5.6 mg.m-3 in the breathing zone of farmworkers. Sampling procedures were as described in the section on pig buildings. Sample times were short (40 – 135 minutes), though multiple samples were taken (10 for the stationary sites, and 30 for the breathing zone).

The concentrations measured are considerably higher than those found in England, Denmark, Germany and The Netherlands. The higher farmworker breathing zone samples may be explained by workers performing tasks which generate dust for a relatively short duration. Also, short sampling durations, locations of samplers, and only taking measurements in the daytime could explain some of the differences. Also important are differences in practices and conditions between Finland and England. Other articles (Kotimaa et al, 1991;

41

Projecttitle



Kotimaa et al, 1987) describe the opportunity for hay, commonly used as fodder, to be contaminated with moulds in Finland, which greatly increases particulate levels. Furthermore, the indoor season is longer (8 months) and the colder winter temperatures are likely to cause ventilation to be more restricted than in England.

(iii) Californian measurements

Measurements of personal exposure to farmworkers during a range of tasks were made with personal samplers, a small number of which involved work with dairy cows. Further details of sampling procedure are given for the same study in Section B.3, below, on Arable Farming. In the study of Nieuwenhuijsen et al (1999), mean exposure to during feeding, moving, handling and milking animals were all between 0.4 and 0.75 mg.m-3

inhalable dust, and 0.1 - 0.3 mg.m-3 respirable dust. These are the arithmetic means of between 1 and 6 measurements.

In an earlier study by Nieuwenhuijsen et al, (1998) farmworker exposure was measured during milking, feeding and manure removal from dairy cows. Measurements of total dust were made using different samplers to the 1999 study, and showed much higher results: 1 - 4 mg.m-3 for milking and manure removal respectively, and 26 mg.m-3for feeding. Details of feeding procedure are not given, and all are arithmetic means of 3 measurements. Mean PM10 levels were 0.1 and 0.5 mg.m-3 for milking and manure removal, and 2.9 mg.m-3

during feeding. Respirable dust levels were around 0.1 mg.m-3 for milking and manure removal, and 0.7 mg.m-3for feeding. All levels are arithmetic means of 3 measurements.

(iv) Chopping of bedding

Two articles describe measurements of levels of dust while bedding is being chopped. Olenchock et al (1990) measured total and respirable dust levels in 3 barns in the USA. Mean levels for each barn ranged from 8 to 41 mg.m-3 total dust, and 1.6 to 2.5 mg.m-3 respirable dust. 2 samples were made, by gravimetric methods, at each barn for respirable dust, and 4 for total dust. Worst case conditions were simulated by closing all doors and windows, as is usual practice in winter.

Another study of 8 barns during bedding chopping gave a mean concentration of 5.2 mg.m-3, with a maximum level of 8 mg.m-3 (Bauer and Coppolo, 1993), though, as it refers to unpublished data (by R. Jones, J. Dennis and J.J. May), the validity of this measurement cannot be checked.

Chopping and spreading, or blowing, bedding is considered to be the dustiest process in cow husbandry, so these measurements probably represent peak concentrations during the day. The level of total particulate generated would be expected to rapidly decline as larger particles are deposited; respirable particles are likely to remain airborne for much longer unless removed by ventilation.

42

Projecttitle



TABLE B.6. Comparison of different surveys of dust concentrations in cow buildings

Reference Country Mean dust concentrations (mg.m-3)

Sample position Sampling time / season

Other comments

Total dust

Respir-able dust

Takai et al, 1998


24h mean, summer only (except calves on litter, measured winter and summer)

Included layers in dairy cows in tie stalls or on litter, dairy cows in cubicles (slurry system), beef cattle on litter and calves on litter

England, Netherlands, Germany, Denmark

0.38 0.07 80 buildings in winter, 46 in summer

24h mean


Finland 1 Air sampling 40 – 135 minutes, 10 samples

Mouldy hay may cause high dust levels

5.6 Personal sampling

40-135 minutes, 30 samples

Nieuwenhuij-sen et al, 1999

USA, California

.74 0.14 Personal sampling

Feeding Institute of Occupational Medicine sampler for inhalable dust.69 0.13 Milking

.51 0.31 Moving/handling0.36 0.88 Scraping stalls

Nieuwenhuij-sen et al, 1998

USA, California

26.1 (PM102.9)

0.59 Feeding Cascade samplers for total and PM10, (indicated in brackets in total dust column)0.8

(PM100.1)

0.06 Milking

3.7 (PM100.5)

0.15 Manure removal

Olenchock, et al, 1990

USA 8 to 41 1.6 to 2.5

Area sampling (2 for respirable dust, 4 for inhalable)

Worst case conditions during bedding chopping

B.2.1.3. Studies referring to dust levels measured during hay making

Batel, 1979, (Germany) found mean dust levels of 15 mg.m-3 (at tractor driver level). Measurements of up to 100 mg.m-3 were noted.

Norén, 1985 (Sweden) found mean dust levels of 5.6 and 5.7 mg.m-3 at the locations above and on the outside of the tractor. Individual measurements ranged from 1 to 13 mg.m-3. These were based on measurements over 9 days.

Louhelainen el al, 1987a, (Finland) reported dust levels of 4.4 and 9.9 mg.m-3 for haymaking and hay baling, respectively. These were derived from relatively short measurements of 35 - 140 minutes, repeated 4 times for haymaking and 9 times for baling.

These measurements are fairly close to each other (4 – 15 mg.m-3) and could be relevant to UK practice at that time in terms of climate etc, but changes in types of bales produced could alter particulate levels generated. (Further details about sampling methods used by these authors are given in Section B.3, on arable farming.)

43

Projecttitle



B.2.1.4. Harmful substances contained within dust in some livestock buildings

Some components of dust can be harmful even in small amounts. They include:

(i) endotoxin: a substance found in cell wall of many Gram-negative bacteria, originating largely from faecal material. Endotoxins are powerful lung irritants, even in small doses.

(ii) fungal spores, some of which are allergenic. Particularly potent are spores of Aspergillus fumigatus, Faenia rectivirgula (which is also referred to as Micropolyspora faeni) and Thermoactinomyces vulgaris, which are all small enough to be inhaled, and are believed to contribute to the condition "Farmers' Lung". Bronchial symptoms (such as persistent cough, wheezing and/or shortness of breath) can be experienced by sensitised (or immunosuppressed) individuals (or by previously unsensitised individuals following a massive dose, or long term exposure.)

(iii) storage mites, which can infest stores of hay and grain, and are also allergenic; parts of dead mites or their faecal material can be inhaled and are believed to cause bronchial symptoms ("barn allergy" e.g. Cuthbert and Jeffrey, 1993) in some farmworkers in the UK and other European countries.

(iv) other gases present in livestock buildings (primarily ammonia) can also cause lung irritation, and can act synergistically with other substances inhaled.

(v) odours. Dust is believed to be important in carrying odours, as odorous compounds can be absorbed onto fine particles. It is believed that these odorants can be released in the nasal cavity, leading to a strong perception of odour. However, the relationship between dust levels and odour intensity is not simple, and further work is needed in this area (Takai et al, 1998, Heber et al, 1988, citing others).

B.2.1.5. Studies which give information about the character and sources of particulates in animal buildings

Pig and poultry buildings

Heber et al (1988) conducted a study of dust collected from 8 samples from 11 pig buildings, which were also being monitored for environmental conditions. Over 8000 particles were examined by light microscopy, and over 1500 by scanning electron microscopy. Grain meal was the largest component, followed by starch, for particles in size categories between 5.4 and 21.6 m. Less than 1 % of particles were identified as skin. Many particles were not identified.

Across the 11 buildings, the concentrations of dust varied considerably, but the size distribution was very close. The mass median aerodynamic diameter (MMAD) was 18.9 m, and respirable fraction by mass was 3.7 %. The median diameter by particle number was 2.6 m. The geometric mean diameter of starch particles was 12.4 m, and for grain meal was 8.6 m.

The authors also refer to other studies which found feed to be the main component of dust, and skin flakes as a minor component in the 5 - 10 m size range.

Donham et al (1986a) conducted a comprehensive study of dust from 21 pig buildings in Iowa (USA), and also reports on a similar study in Sweden. Light microscopic analysis found the main constituents to be feed and faecal material. Most of the particles above 5 m diameter consisted of feed, but very few feed particles were found below this size. Particles of faecal material were present across the whole size range, to as small as 0.5 m, and these were the major components of dust in the size range 1 - 2 m diameter. Relatively more feed: faecal was found in buildings with larger animals. Other components included skin flakes, mould, pollen, insect parts and mineral ash. The overall proportion of respirable dust by mass (taking cassette measurements, as these were considered more reliable) was 8.4 %, and was slightly higher in the farrowing unit (10.7 %),

44

Projecttitle



compared to the nursery (7.1 %) and finishing unit (6.0 %). The overall MMAD was 9.6 m for all units, and was slightly larger in finishing units (10.7 m), compared to nursery units (9.8 m) and farrowing units (8.95 m).

Louhelainen et al (1987b) measured dust from 15 pig farms, including sow and fattening units. Of the mass of the dust 14 % was respirable (PM5), and the MMAD was determined to be 11 m. The MMAD of dust collected from fodder grain was also 11 m, which the authors take as an indication that feed is a major component of dust in pig buildings, though this of course, is far short of proven.

Cravens et al, 1981, measured dust in turkey buildings, with and without misting to control dust levels. The proportion of dust below 10 m diameter (which the author terms "respirable" in this article) was 15 % and 16 % in control and misted barns respectively. The calculated MMAD in the barns were in the range of 20 -40 m.

Jones et al, 1984, studying broiler buildings in the USA, found the MMAD of airborne dust was 15 - 16 m. Considerable differences in dust levels were found between buildings e.g. new/old litter, but the size distributions of dust from the buildings were very close. So, to sum up, several studies find the mass median aerodynamic diameter for dust collected from pig and poultry buildings to be in the range 11 - 19 m. Where light microscopy was used (on pig unit dust), feed was found to be the predominant component, followed by faecal material. Minor components included skin, hair, mould, pollen grains and insect parts. No mention was made of bedding, which was evidently not used in these buildings.

Cow buildings

A mention of particle analysis of dust in a Finnish cow building (referring to unpublished data) gave a smaller MMAD of 7 m. The authors (Louhelainen et al, 1987a) assume this is from large presence of small fungi in fodder. This assumption agrees well with other published work. A study of Swedish dairy farms (Karlsson and Malmberg, 1989) found a predominance of spores such as fungi and actinomycetes: this and another Finnish study of dairy farms (Kotimaa et al, 1991) describe the various microbes found growing on feed and bedding materials. Further discussion is given to particles from hay and straw in the following section.

B.2.1.6. Materials which contribute to dust in animal buildings

Pig and poultry feedPigs and poultry are fed on similar material, a formulated feed based on ground cereal, with small amounts of additives of micronutrients, and in some cases antibiotics. Different compositions of feed are formulated for pigs/birds of different ages. This material, in the ground form, is extremely dusty. Studies of dust collected from pig buildings in the USA (Donham et al, 1986; Heber et al, 1988) found that feed particles make up a large amount of the dust. It is assumed (though not explicitly stated) that these pigs were fed milled feed, which is cheaper but much dustier than pelleted feed.

Most of the commercially available feed in the UK is in pelleted form, though some pig units make up their own feed, which would mostly be ground. The pellets are still fairly crumbly, and a source of dust, though a great improvement on (cheaper) ground material. Most of the pig buildings measured in the USA were using ground material, so dust levels are higher than in many UK buildings. Some studies in the USA documented the beneficial effects of adding fat or oil to the ground feed to reduce dust levels (e.g. Heber and Martin, 1988).

The method of food conveyance is also important, with enclosed systems giving least opportunity for the evolution of dust. Crook et al (1991) found significantly lower levels of dust in buildings with enclosed (unrestricted) feed systems compared to buildings where the pigs were fed several times a day. "Wet systems"

45

Projecttitle



where feed is mixed with water and pumped in a liquid state, are also unlikely to evolve dust. Crook et al (1991) found some of the buildings using this system were relatively dusty, but added that these were older buildings which were continually stocked and seldom cleaned, which could explain the surprising findings.

Ruminant feed

(i) Ensiled feedstuffs

Cows and sheep which are kept indoors in the UK (usually during the winter or part of the winter) are fed principally on silage. This is produced from the lactic fermentation of green fodder crops, in the absence of air. The product is moist (moisture content up to 80 %) and not considered to be a significant source of dust. If the packing is damaged or not sealed properly, undesirable moulds can grow, which can lead to some emissions of spores. However, this is not considered to be a major source of emissions. A Finnish study comparing different feed and bedding materials used for dairy cattle in Finland, (Kotimaa et al, 1991) found that silage had a very low content of microbes.

Horses are fed similar material, though the silage is of a lower moisture content (of around 30 - 60 %) than silage which is fed to cows and sheep, and is generally termed "haylage", rather than silage. This is produced in a similar way to ordinary silage, but the grass is partially dried before processing. The manufacturers Dobson and Horrell Ltd, and Marksway Horsehage, recommend buying bales of a size which can be used within a few days, to avoid moulds developing after opening. This is critical for race-horses, as these moulds can cause lung problems (chronic obstructive pulmonary disorder, or COPD, which is analogous with "farmers' lung") in sensitive individuals.

Bad silage/haylage, i.e. that which has been in contact with air, can contain large numbers of microbes, as described by May et al (1989). In the 5 silos studied, though the bulk of the material was held in the (correct) anaerobic conditions, the top portion, up to 1 m, was exposed to air and often heavily overgrown with micro-organisms. Removing this "cap" material was reported to generate large amounts of dust, and cause suffering various symptoms in workers following exposure. In 3 of these silos, total dust levels (measured within the silo at 2 locations) were under 3 mg.m-3, but up to 7 mg.m-3 was found at the base. The 2 other silos measured had much higher dust levels, of 20 and 113 mg.m-3 (each a mean of 6 measurements). Though the number of samples is small (3 at each location in each silo) this indicates that silage, if allowed to be in a bad condition, can be emit many particles.

(ii) Hay

The traditional winter feed material, hay, is also used, though these days in smaller quantities than silage. Photomicrographs of "good quality hay" show that fine particles consist largely of fragments of plant tissue, plant hairs and pollen grains, with little fungal contamination (Clarke, 1987). However, examination of poor quality hay shows that the majority of particles are fungal spores, in high concentration, many of which are respirable (this also applies to bedding) (Clarke, 1987).

The conditions in which hay is made, collected, and stored, are of prime importance. Clarke (1987) reports on the importance of the moisture content of hay when it is baled. Hay baled at low moisture content (15 - 20 %) generally suffers little fungal contamination, apart from the small numbers of "field fungi" (e.g. Cladosporium and Alternaria) which are always present on plant material. Other authors also recommend that hay be stored at moisture content below 20 % or 16 % (Charlick et al, 1980, and Gregory and Lacey, 1963, respectively; cited by Kotimaa et al, 1991). Baling hay at 20 - 30 % moisture content can result in some heating from microbial activity, and moderate contamination with "storage fungi" including the allergenic Aspergillus fumigatus. Hay baled at higher moisture content is likely to suffer heavy contamination with spores of fungi and actinomycetes (the latter being around 1 m diameter, and able to deeply penetrate into the alveoli).

46

Projecttitle



Cuthbert and Jeffrey (1993) report a succession of spore types during the months of storage in Scottish barns, with initial "field fungi" being replaced by "storage fungi", which later declined as the third group, thermophilic actinomycetes increased in number. The number of colony forming units (cfu) could be over 100 million per g of hay (similar results to Gregory's (1973) examination of mouldy hay). Lacey and Lacey (1964) report even higher concentrations, and levels of over 1000 million spores per m3 of air in a barn, following vigorous shaking of mouldy hay. Hay can also be infested with large numbers of storage mites. Whether they persist or disappear after a few months, allergenic particles of mite parts and faeces will remain in the hay. A Finnish study comparing microbial contents of various feeds (Kotimaa et al, 1991) found storage dried hay had the lowest microbial content (around 1000 cfu/g): hay dried on the ground and collected loosely, and hay dried in the traditional Finnish way (on poles), gave higher, but satisfactory results, but baled hay had significantly higher concentrations of microbes, presumably because it was not thoroughly dried.

(iii) Grains

Whole grains (horse feed, studied in vitro by Vandeput et al, 1997) were found to have a moderate respirable dust and microbe content, compared to other feed materials, but rolled grains were much more dusty. The Finnish study previously cited (Kotimaa et al, 1991) found that acid treated grain had a low microbe content; grain dried with heated air contained more microbes, and grain dried with unheated air had the highest microbe content (though at just over 3000 cfu/g, this was comparable with hay dried on the ground/on poles, and not a great cause for concern).

Vandeput et al (1997) carried out an in vitro comparison of various feed materials. All measurements were performed on samples of 100 g of material, subjected to a constant flow of air (200 l/min), with each measurement repeated 17 to 22 times. Results are quoted as numbers of particles, in the size range 0.5 to 5 m diameter, per litre of air, as identified with a light scattering technique. Silages and whole grains were found to liberate the smallest amounts of respirable dust (4,000 to 9,000 particles from 100 g samples per litre of air). Good hay liberated significantly higher numbers of particles (63,000 30,000) as did rolled grains (120,000 30,000). Dusty hay liberated over 10 times more respirable particles (mean 688,000 144,000). The relative quantities of viable spores of 3 known allergens, (Aspergillus fumigatus, Faenia rectivirgula, and Thermoactinomyces vulgaris) generally followed similar trends to respirable dust levels. Measurements could not be made on the dusty hay as the Petri dishes were overloaded.

(iv) Feed supplements

It is usual for cows (both dairy and beef) to be fed small amounts of supplements, to increase the vitamin, mineral and protein content of the diet, and thus improve productivity. Some sheep, particularly lambs, are also fed supplements. A range of materials are available, some of which are powdery, and potentially a source of particulates, though a minor one, because of the small quantities involved. In general, the more expensive products are less dusty, though economic pressures are probably a more important consideration to many farmers. Supplemented feeds can be made up by farmers themselves, by mixing the supplements with oats or similar. Addition of molasses is a cheap and effective way to suppress dust, and is practised by some farmers.

47

Projecttitle



Bedding

(i) Straw

Straw is the most common bedding material, being generally suitable and also available as a by-product of cereal production. However, it is also a source of dust. In cow-sheds, straw is generally cut, and transferred to the building by blowing or conveying. These operations can cause high levels of suspended dust, as discussed above (Olenchock et al, 1990).

As with hay, good quality straw liberates some amount of dust consisting largely of fragments of plant material, but poor quality straw (which has been kept in a damp condition) can liberate much higher numbers of spores, many of which are respirable (Clarke, 1997). Vandeput et al’s (1997) in vitro comparison of various feed and bedding materials for horses, found that "good straw" liberated moderate numbers of respirable particles and spores, as compared with samples of other bedding and feed materials. Kotimaa et al (1991) found that straw liberated moderately high numbers of microbes.

(ii) Alternative bedding materials

Wood shavings are generally considered to release less dust and fewer spores, though measurements show mixed results, depending partly on the age and depth of litter.

Rubber mats are also used, particularly in the west of Britain, where straw is more costly.

A "no bedding" system (e.g. slats or mesh, slurry system) is used for many animals.

A new product, made from corrugated cardboard, has been sold in the UK since 1995, as "dust-free" bedding for horses, pheasants and partridges. In vitro trials have shown that this product compares very favourably with straw and shavings, in terms of low release of respirable dust and spores of 3 allergenic fungi. (Roberts et al, 1999; Roberts, personal communication, 2000) The manufacturers report increasing sales, state that their product can compete economically with straw in many regions of the UK, and are planning to market this product as a commercial agricultural bedding later this year (Seaman and Rush, pers. comm., 2000).

Faecal materialIf manure dries in the house, it has the potential to be a source of dust, from being disturbed directly, or after being transferred to the animals. A study of dust apportionment in a pig house (which contained no bedding) (Donham et al, 1986) found that faecal material (gut epitheleal cells, bacteria and undigested feed) was the predominant component of very fine particles (1 - 2 m), but not for larger particles, where feed was the main material. Particles of faecal material can be particularly harmful as they may contain significant amount of endotoxin.

The design and conditions of the house determine whether manure is removed from the animals or not, whether it has the opportunity to dry, and thus the likelihood of its disturbance to become airborne.

Skin flakes, hair/feathers/woolLittle mention has been made in the literature of the contribution of particles directly lost from the animals themselves, to dust in buildings. It is assumed that sheep wool (shearing etc.) is not a source of dust, because the lanolin which impregnated it binds any material adhered to it. Hair clipping has been mentioned as a source of dust and thus occupational exposure to workers, because of the material on the hair which is disturbed. Hair itself is not allergenic, but urine, faeces and saliva which may be on it, contain proteins which can trigger allergic reactions. Flakes of animal skin are also known to be allergenic.

48

Projecttitle



A study of dust found in a pig house in the USA (Heber et al, 1988) did identify skin flakes: however these were very small in number compared to other particle types (5 % of the particles size 7 - 9 m, and 10 % of the particles 11 - 16 m). Hair was found in smaller quantities (1 % of particles 11 - 16 m, and not reported in smaller particles). In short, these are not considered an important source of particulate emissions from livestock, though allergenic properties may trigger a reaction in sensitised persons (particularly farm workers) in very small amounts.

Though no report of source apportionment of poultry building dust has been identified, feathers are presumed to be a component, as young chicks lose their original down when they mature and grow feathers, and feather wear is noted on many layer birds.

B.2.1.7. Influence of environmental factors on particulate levels

Many factors influence the emission rate of dust from animal buildings. The type of housing, and extent and type of heating and ventilation, will influence whether particles are likely to settle out, remain suspended in air, or be removed by ventilation. The humidity of buildings is important, as damp floors and other surfaces will generally bind dust particles. Damp bedding may bind particulates and reduce dust levels, though it can be a source of particulates itself. Temperature is important, because it affects evaporation rates and thus relative humidity; it can also affect animal activity, animals generally becoming less active at higher temperatures. The system chosen for delivering feed, and removing manure, also affects the likelihood of these materials becoming suspended in the air as particulates.

The choice of feed and bedding (where used), and of system by which this is distributed, have also been shown to affect concentrations of particulates. The quality of feed and bedding is also important, in particular, whether it is stored in a damp condition where moulds can proliferate. Other factors mentioned in the literature, and during visits to units, are "house-keeping measures", such as how often (if at all) the buildings are cleaned, and how long litter remains in place between changes. How cleaning or mucking out is done is also very important: sweeping can generate a lot of dust, but vacuum cleaning or washing down would be unlikely to.

B.2.2. Outdoor livestock

UK sheep are outdoors for most or all of the year, and cows for approximately 6 months. A significant proportion of pigs (up to 30 % of sows and young piglets, though few fattening pigs) are kept outdoors all of the year. A small but significant proportion of poultry (around 15 %) is "free range", which has access to the outside, although shelter is also provided. However no dust data for outdoor livestock have been found which are relevant to the UK.

Levels of dust have not been reported or measured, presumably because this has minimal potential to affect the health of farm workers. For cows and sheep any emissions are expected to be minimal, with the exception of soil/manure dust from overgrazed areas in dry conditions. (A study of particulates emitted from cattle feedlots in Texas discusses their composition (mostly dried manure) and size distribution; however, this is not considered relevant to UK conditions, particularly as cows and sheep are chiefly farmed in damper areas of the country.) Re-suspension of soil is more likely on pig farms which are on soils which are prone to wind erosion (discussed below).

Ruminants feed on pasture, which is not a significant emitter of particulates. (In fact, grazing will reduce the amount of pollens and field fungi which may be liberated from grazing land, and may also aid control of bracken, which itself releases spores which are considered possible carcinogens.) Some feed is put out for cows and sheep on some farms to supplement the pasture (e.g. mineral supplements) but the potential for release of particulates from this material is thought to be minimal.

49

Projecttitle



Pigs are given the same feed outdoors as indoors. In wet conditions this is unlikely to release any particulates; the quantity which may be blown away in dry weather is not known, but may be significant.

B.3. Arable Farming

B.3.1. Emission factors

B.3.1.1. Land preparation

The USEPA has published a general empirical formula for dust emissions due to land preparation, which can be used to estimate the amount of particulate matter (PM10) generated per acre-pass. For California, the formula is as follows:-

Emission factor (in lb PM10 per acre-pass) = k * 4.8 * S0.6

where k is the proportion of suspended dust which is under 10 m diameter (in California, k = 0.148), andS = % silt content of the soil (in California, taken as 18 %, but this varies from area to area)Thus the emission factor in California is 4.02 lb PM10 per acre-pass.

Using this approach, for the total cultivated area for the UK (taken as 4,721 thousand hectares), this gives a PM10 emission of 21 t per pass by a cultivating machine.

This must be taken as a worst case, as no allowance has been made for the soil being moist during part of the year. (In California, a reduction factor of 50 % is applied during the wettest months.) However, some operations can only be performed when the soil is dry. Furthermore, this figure is only an approximation, because values of silt content and soil dust size distribution may be different in the UK to in California, (and probably vary between regions in the UK) and relevant data for the UK ought to be used. Nevertheless, this gives an indication of the magnitude of emissions.

B.3.1.2. Cereal harvesting

The USEPA has developed an empirical formula for calculating the emission of dust from harvesting wheat. Its results are only given a quality rating of D ("not very reliable"), and furthermore it gives an answer in PM7 (as it is derived from measurements which pre-date the PM10 standards), so an appropriate conversion factor must be employed to estimate emissions of other size fractions, which would further reduce the quality rating of the result.

The emission factor in g PM7 km-2, is 170 from the combine itself, 12 from truck loading, and 110 from field transport, i.e. 292 g PM7 km-2 in total.

This equation has been modified for use in the present inventory (Section 3 and Annex D), to allow for the 4times higher yield per hectare in the UK compared to the USA. The components for combine and truck loading have been multiplied by 4, while the component for field transport has been left unchanged. This gives a UK emission factor of 838 g PM7 km-2. Applied to the UK area of cereals, with a total area of 2791 thousand hectares (winter sown cereals in 1999, from MAFF statistics, 2000) the predicted UK emissions of PM 7 are 23.3 tonnes per year.

B.3.2. Measurements of concentrations of particulate matter in air during arable farming activities

A number of studies have been performed which give some measurements at points at or near to the harvesting, tilling etc. The majority of measurements were made with the aim of calculating worker exposure to dust, and

50

Projecttitle



thus are taken at positions to represent this, such as at tractor driver level. Thus they are not, on the whole, representative of the concentration in the cloud of particulates generated by the activity. Furthermore, studies where concentrations were measured at several points on or around a tractor simultaneously, report wide variations in mean dust levels between the points (Norén, 1985, and Burg, 1981). Even studies measuring similar processes in similar fields, but using different types of samplers, have obtained different results (Nieuwenhuijsen et al, 1998 and 1999). A number of older studies are quoted here, because in more recent years, the drivers have been enclosed in a cab to protect from the dust, so measurements of their exposure no longer indicate concentrations in the field.

An attempt has been made to estimate the gross emissions of particulate matter from soil tillage and harvesting, using the concentrations reported, and making various assumptions which are stated in Section 3 and Annex D. These estimations must be regarded as extremely approximate indications of the scale of emissions; however, in the absence of appropriate data it is difficult to make any more realistic estimates of particulate emissions from arable farming.

B.3.2.1. Californian measurements

Numerous measurements have been made in California, where particulate emissions from agriculture can be very significant. In one area, the San Joaquin Valley, it has been estimated that a third of PM 10 may originate from farming operations (California Air Resources Board, 1988, cited in Clausnitzer and Singer, 1996), and a number of studies have been performed.

(i) Clausnitzer and Singer, 1996

These authors reported measurements of dust concentrations near to source. The trials were done with the intention of finding information regarding fugitive dust levels caused by agriculture in this region. The motive for the study was safeguarding both occupational and public health. The authors clearly state that all measurements were made close to source, and do not give information about effects further afield. Nevertheless, the immediate concentrations generated are detailed in Table B.7.

Measurements were made of many different operations. Environmental conditions were noted for each operation. Respirable (not total) dust concentrations were measured, using cyclones running at a flow rate of 2.2 l/min, taking the 50 % cut-off size at 4 m, according to the new international standard for respirable dust. Operations were done on 3 plots, and sampled by 2 closely located samplers, so a mean of at least 5 readings was used in each case. The sampling duration varied with operations: for low or medium levels of dust, the sample was continued for the duration of the operation. For dustier procedures, it was necessary to stop at intervals to change the filter to prevent overload.

Sampling apparatus was positioned approximately 1 m above ground level, depending upon operation. Sources of error reported included facing the sample head into or away from the wind, which led to over- or under-sampling, respectively. The authors report the samplers in this study were at 90o to the wind. Another source of measurement error is variation in density of particles: the cyclones are set up to record particles of the same density (taken as 1 g cm-3): however, real dust particles can have a range of densities (up to 3 g cm-3 for soil particles), so they would behave differently in a sampler.

Results of the various operations are listed in the table below.

TABLE B.7. Concentrations of respirable particulate matter in air, measured close to various agricultural operations (Clausnitzer and Singer, 1996)

Operation No. of Height of sampler Mean respirable Fisher's protected

51

Projecttitle



samples above groundcm

dust concentration in airmg m-3

least significant differenceat 0.05 probability level*

Discing in corn stubble 8 95 0.33 aDiscing in vetch / pea 7 106 0.45 aCultivation and furrow cutting

30 81 0.66 a

Injected fertilising and furrow cutting

14 64 0.79 a

Spring tooth harrowing 6 41 1.01 aCorn seeding 7 52 1.02 aWheat harvest 7 241 1.04 aCultivation (weed removal) 31 62 1.09 aVetch/pea mowing 5 68 1.25 aDiscing-in wheat stubble 8 108 2.49 abWheat straw baling 10 142 2.82 ab1st finish discing 33 77 4.05 bcTomato harvest 12 75 4.34 bcd2nd and 3rd finish discing 20 77 4.94 bcdCorn harvest 19 220 5.63 cdRototilling road 8 58 6.69 dLand planing 20 98 10.29 eRipping soil 12 77 10.34 e

*Values which do not share a common letter are significantly different from each other

The dustiest operations were land planing and ripping soil, with measured concentrations of over 10 mg respirable dust m-3 air. However, other soil tillage operations generated much less particulate: discing in wheat stubble gave 2.5 mg.m-3, and discing in corn, pea or vetch stubble, and cultivation and furrow cutting generated levels of below 1 mg.m-3 respirable dust. Wheat harvesting was not found to be very dusty compared to other operations, giving measured respirable dust concentrations of 1 mg.m-3, but wheat straw baling was almost 3 times as dusty (though this difference is not statistically significant). Tomato harvesting was found to be dustier (over 4 mg.m-3 respirable dust): it is significantly dustier than wheat harvesting, but not than baling of wheat straw.

The relevance of these findings to UK agriculture is open to debate. Differences in agricultural practice are considerable, and some of the operations (such as land planing and ripping) are not or only rarely performed in the UK. Probably more significant are climatic differences. This part of California experiences a Mediterranean, semi-arid climate, where solar radiation and temperatures are higher than in the UK, rainfall is very low during the summer, and irrigation is necessary for many operations. Consequently, one would generally expect similar procedures to generate considerably more dust in California than in the UK. However, parts of the UK, particularly in the south and east, can have dry spells in the summer, and high emissions would then also be possible. Conversely, these dry conditions can retard the growth of moulds on crops, which can be a major source of particulate, particularly during harvesting (as discussed later, in Section B.3.2.5).

52

Projecttitle



(ii) Nieuwenhuijsen et al, 1998, and Nieuwenhuijsen and Schenker, 1998

These studies measured occupational exposure to workers doing various agricultural operations in California. Measurements were made using personal samplers. Cascade impactors (sampling 2 litres air/min) measured total dust, PM10 (with 50 % cut-off level set at 9.8 m) and PM3.5 (50 % cut-off level set at 3.5 m, taken as a measure of respirable dust). Additionally, a personal respirable dust cyclone (with 50 % cut-off at 4 m diameter, running at 2.2 l/min) was used.

To allow for the fact that, with the cascade impactor, the particulate loading in air was sampled with different efficiencies for different size fractions, correction factors were used, as stated in the paper. These ranged from 26 % (for inhalable dust) to 99 % for (respirable dust), i.e. the result for inhalable dust was the measurement obtained divided by 0.26, but for respirable dust was the measurement obtained divided by 0.99; intermediate value was for PM10. The results of this study are in Table B.8, below.

TABLE B.8. Concentrations of particulate matter in air, measured close to various agricultural operations (Nieuwenhuijsen et al, 1998).

Operation Cascade impactor results (mg.m-3, all arithmetic means)

Cyclone measurement of respirable dust

No. of samples

Total dust PM9.8 PM3.5 mg.m-3 No. samples

Land planing 18 82.7 4.6 0.5 0.64 12Discing 16 158.6 8.6 1.1 1.19 16Listing 2 11.0 0.4 0.1 0.38 1Incorporating 2 12.9 0.8 0.2 0.29 3Rolling beds 2 48.4 2.2 0.5 0.56 3Fertilising 3 10.4 0.5 0.1 0.09 1Mowing 2 14.0 0.8 0.2 0.31 3Planting tomatoes 1 9.8 0.4 0 0.09 2Planting corn, wheat 6 23.3 1.9 0.2 0.18 8Cultivating 1 9.3 0.6 0.1 0.57 4Harvesting tomatoes 4 41.0 3.0 0.4 0.69 4Irrigating 4 2.3 0.2 0 0.09 6Hand harvest tree fruit 4 5.2 0.2 0 0.09 5

Note: the dustiest operations: harvesting grain, ploughing, ripping, swathing and baling, were all done by an operator inside a cab, so measurements of exposure do not reflect outdoor levels, and are not included.

Table B.8 shows the following:-

Wide variation between measured values of operations. Soil preparation operations caused exposures ranging from 10 to 150 mg.m-3 total dust. Planting (tomatoes and corn) and harvesting tomatoes, caused exposures of 10 - 40 mg.m-3 total dust.

The 2 methods of measuring respirable dust agree very well. Respirable dust levels measured range from 0.1 to 1 mg.m-3 for most operations. The fraction of dust of different sizes can be elucidated. This shows that about 5 % of the total dust falls

within the PM10 category, and about 10 - 25 % of PM10 particulate is PM3.5 (respirable).

53

Projecttitle



(iii) Nieuwenhuijsen et al, 1999

These authors report inhalable dust being measured by an Institute of Occupational Medicine standard sampler (running at 2 l/min) and respirable dust by a cyclone sampler, as described above. All measurements are for operators not enclosed in a protective cab. Results are in Table B.9, below.

TABLE B.9. Concentrations of respirable particulate matter in air, measured close to various agricultural operations (Nieuwenhuijsen et al, 1999)

Operation Number of samples

Inhalable dust Respirable dust

Arithmetic mean, mg.m-3

Arithmetic mean, mg.m-3

Ground preparation 19 4.83 0.27Planting / seeding 6 2.47 0.16Hand harvesting veg 5 2.61 0.73Machine harvesting veg 6 11.74 1.12Machine harvesting dry harvested field crop

1 4.55 No measurement

Machine harvesting green harvested field crop

1 0.69 0.05

(iv) Comments on the two studies by Nieuwenhuijsen et al, 1998 and 1999

Levels of total inhalable dust spanned a wide range as measured using the cascade impactor (10 - 150 mg.m-3), though the IOM sampler found much lower levels (1 to 10 mg.m-3). The IOM results are actually closer to the PM10 results obtained from the cascade sampler (0.2 to 9 mg.m-3; but mostly in the range of 0.4 to 2 mg.m-3). The paper cites differences in type of sampler, and differences in seasons of measurements (the 1999 paper using all-year-round measurements, and the 1998 paper taking results from the dry seasons) as reasons for the discrepancies. A large part of the difference is probably the correction factor employed in the cascade impactor measurements (1998 paper), which appear to have caused an over-estimation in the inhalable dust fraction. There is evidently error inherent in all measurements, but it is thought that the IOM samplers give the more reliable results (Nieuwenhuijsen, pers. comm.).

Nieuwenhuijsen's results (all papers) find that the respirable dust levels to which farm workers are exposed are generally within the range 0.1 - 1 mg.m-3. This is considerably lower than the measurements of Clausnitzer and Singer, 1996, who obtained measurements of 0.5 - 10 mg.m-3, for measuring similar operations under similar ranges of conditions, with similar samplers. These results show the effect of different sampler position. Clausnitzer and Singer's samplers were approximately 1 m above ground, varying with crop, but Nieuwenhuijsen et al’s measurements were made on the farmworkers themselves. Even though not in a protective cab, they were exposed to considerably lower levels of dust.

B.3.2.2. Batel, 1979

(i) Soil tillage

Measurements performed in Germany were made to estimate the exposure of farmworkers to dust during various operations. Measurements were all taken at tractor driver level; there was no protection from a cab. A very wide range of dust concentrations is reported, ranging from 1 to 630 mg.m-3 for soil cultivation. The 50 percentile value is suggested as the best representative of a mean value, and for soil preparation (using

54

Projecttitle



measurements of ploughing, harrowing and other operations which were proportionally weighed) this value is 40 mg.m-3. The number of measurements is not specifically stated, but this paper uses results from prior studies, and appears to include large numbers of measurements. Total dust levels were sampled using an axial cyclone and determined by gravimetric means. Dust from soil cultivation was found to be 90 % inorganic, with between 3 and 10 % below 5 m diameter (Batel, 1975).

(ii) Cereal harvesting

Measurements of dust levels that a tractor driver (without cab) is exposed to during combine harvesting range up to 80 mg.m-3, but a mean value of 20 mg.m-3 is given. Procedures are as described above.

The majority of this dust (70 - 90 %) was found to be organic, with 3 - 5 % of it being below 5 m in diameter (Batel 1975).

B.3.2.3. Norén, 1985

Measurements during soil tillage in Sweden were made at 3 points around a tractor, on 10 different days. The locations of the samplers were 10 cm above the roof of the cab, at the window or ventilation inlet, and at the breathing zone of the operator. Dust was collected on a membrane filter made of cellulose esters, of 37 mm diameter, with a pore size of 0.8 m. Air was drawn through with a pump: for the measurements made at the driver's breathing zone, a portable diaphragm pump, made by MSA, was used, with an air flow of 1 – 3 l/min; at the other measuring points, piston-type pumps (Reciprotor 506R), with flow rates 8 - 10 times as high were used.

The mean value measured at a reference point 10 cm above the rear of the tractor was very high: 146 mg.m-3

(measurements ranged from 22 to 577 mg.m-3 at this point.). The other locations experienced lower concentrations: at the cab ventilation window, or the intake to air filter leading to the cab, an average of 98.9 mg.m-3 (range: 3.3 to 375 mg.m-3) was recorded, and at the operator breathing zone the dust level was much lower (mean level 24.7 mg.m-3). The median value, for the 2 outside sampling points, which may give a better "average" value, was much lower, at 50 mg.m-3 (from 100 measurements).

B.3.2.4. Louhelainen et al, 1987a

Measurements of farmworker exposure during several agricultural operations in Finland are described. Measurements were made at the breathing zone of the operator, and it is not clearly stated which operations were performed inside a cabin. The best indicator is taken as the mean concentration outside the tractor (for all activities), which was 52 mg.m-3. (This was much higher than the mean inside the tractor.) The dust levels outside of the tractor would have been higher for certain operations.

B.3.2.5. Burg et al, 1982

Measurements were taken at several places around a combine harvesting maize, in the southern USA. The interesting point from this article is not the absolute values (as harvesting maize is a different operation from harvesting wheat) but the large variation of dust concentrations found at different points around the combine. Three sampling points were used, each having a high volume sampler (approx. 2000 litres/min), at the front, side and rear of the combine. The number of samples was small, only 2 to 4 at each location, and sampling time was not stated, but to give an indication of the variability, the mean dust concentrations were 133, 33, and 14 mg.m-3 respectively.

55

Projecttitle



B.3.2.6. Darke et al, 1976

Measurements were made, not of particulate count, but of spore counts, around combine harvesters in England over 3 harvest seasons. Unfortunately the numbers of samples taken is not stated, but the means are reasonably similar between years, and a much wider range of individual measurements is reported. Air was sampled at both the pick-up reel, and behind the machine, collected in a cascade impactor, and examined microscopically.

A certain amount of assumption and extrapolation is necessary to obtain particulate concentrations in mg.m -3, so these figures should be regarded as an extremely approximate indication. The assumptions are as follows:

A mean spore concentration of 60 million spores per cubic metre is chosen (based on measurements for the 3 years).

It is assumed that 50 % of the spores are above 10 m aerodynamic diameter (based on relative proportions of types of spores stated – this would need to be verified) and so not included.

Spores are assumed to have a relative density of 1 (which is probably correct within 20 %). It is assumed that the spores are spheres (which, actually, they are not).

(a) If the mean volume of the spores is that of a sphere of 5 m diameter, this gives a mass concentration in air of 2 mg.m-3 (for 30 million spores smaller than 10 m diameter).

(b) If the mean volume of the spores is that of a sphere of 9 m diameter, this gives a mass concentration in air of 10 mg.m-3 (for 30 million spores smaller than 10 m diameter).

(c) If the mean volume of the spores is that of a sphere of 1 m diameter, this gives a mass concentration in air of 0.02 mg.m-3 (for 30 million spores smaller than 10 m diameter).

B.3.2.7. Atiemo et al, 1978

Measurements of both total and respirable dust during a range of arable operations were made by researchers at the University of Saskatchewan, Canada. Levels of dust, as measured within the plume generated, fell in the range of 34 – 195 mg.m-3, with the majority of measurements in the range 40 - 120 mg.m-3. The wide range was attributed to differences in a range of environmental conditions, such as wind velocity, soil moisture content, relative humidity, temperature, soil type and ground cover. Background levels of dust were also measured and these were often significant: between around 2 and 25 mg.m-3. This is attributed by the authors to wind erosion on prairie lands. These background levels are very high compared with UK levels, showing that the environment in the Canadian prairies (for reasons of climate and topography) can be much dustier than in the UK, so these conditions are probably not representative of the UK.

Measurements were made during seeding, "summer fallow cultivation" (i.e. ploughing the land either during/at the end of a fallow period), spraying and combine harvesting (as well as baling and swathing). A total of 32 sets of dust samples were collected, including inside cabs as well as outside. A number of samples of short period were collected, though the number is not stated in the paper. Short sampling periods were used to aim to avoid problems of cross contamination. Rotating rod type aerosol samplers were used, positioned within the plume generated by the operation, which drew air at a rate of 210 litres/min. Dust levels within cabs were very much lower than outside because the air was filtered, so these results have not been included here.

Analysis of particle size characteristics found that mass median aerodynamic diameter of ambient particles during seeding were between 13 and 14 m, and similar (though slightly lower) values were obtained from particles collected inside cabs.

56

Projecttitle



TABLE B.10. Summary of various studies measuring dust concentrations near to various arable operations

Reference Country Mean dust concentration, mg per m3 of air

Comments

Total suspen-ded dust

Particles less than 10 m diameter

Respirable particles (under 4 m diameter)

Soil preparation

Clausnitzer and Singer, 1996

USA, California 0.5 – 10 Depending on operation. Sampler<1m above ground


USA, California 10-150 0.5 -9 0.1-1 Personal exposure not in protective cab, using cascade impactor for total and PM10, and cyclone for respirable


USA, California 5 0.3 As above, using IOM sampler for inhalable dust

Norén, 1985 Sweden 100, 150 Mean dust levels at 2 points on outside of tractor. Range 3 – 580.


Finland 52 Mean concentration outside tractor

Batel, 1979 Germany 40 Mean level at tractor driver level. Weighted average for all soil preparation operations. Max level 630

Atiemo et al, 1978 Canada 40-100 Mean level not stated, range 35-200. Operations were seeding and summer fallow cultivation. Measurements made in dust plume. High background levels of up to 25 also measured.

Cereal harvesting

Clausnitzer and Singer, 1996

USA, California 1 Sampler 2.4m above ground

Batel, 1979 Germany 20 Up to 80. Measured at tractor driver levelDarke et al, 1976 England 2-10 Approximate mass of spores collected

behind combineAtiemo et al, 1978 Canada 35-90 Range of concentrations in dust plume.

Mean not stated

B.3.2.8. Size distribution and characterisation of dust emitted from arable operations

Different studies have made measurements concentrating on different size fractions, so it is necessary to use a conversion factor, based on size distribution results, to compare these different measurements, and to obtain estimates of PM10 and PM2.5 measurements.

57

Projecttitle



TABLE B.11. Estimates of size distributions of dusts evolved from various sources

Type of dust Proportion under 10 m diameter

Proportion under 5 m diameter

Proportion under 3.5 m diameter

Proportion under 2.5 m diameter

Proportion under 1 m diameter

Californian Air Resources Board (P. Gaffney, pers. comm., April 2000)

Tilling dust 45.43 % 10.07 % 3.50 %

Unpaved road dust 59.43 % 12.60 % 5.16 %

Feed and grain dust 29 % 1 % 0 %

Batel, 1975, Germany

Inorganic dust 3 - 10 %

Organic dust 3 - 5 %

Nieuwenhuijsen et al, 1998, California (measurements primarily intended to find worker exposure, not particle size distribution)

5 - 6 % 1 %

Other measurements from California can also be used to estimate size distributions. Thus Nieuwenhuijsen et al (1998) measured three size fractions: total, under 9.8 m, and under 3.5/4 m. These suggest that the dust from soil disturbance operations contains about 5 - 6 % PM10, and around 1 % respirable dust. This is a very low fraction of finer particles, and may be caused by erroneously high estimations of total dust, as already discussed. The proportion of PM10 which is respirable was found to be around 14 %, which is a little lower than the above estimate for tilling dust. The above estimates are considered the more reliable.

Batel (1975) states that dust from harvesting was 70 - 90 % organic, and 3 - 5 % of this material was respirable (i.e. under 5 m diameter). Dust from tilling was found to be 90 % inorganic, of which 3 - 10 % was respirable. Atiemo (1978) measured what proportion of the dust collected from several operations was combustible (i.e. organic). Seeding, summer fallow cultivation and spraying operations generated dusts which were around 3 - 7 % combustible; combining, baling and swathing dusts were 70 - 80 % combustible. Dusts inside cabs were studied to determine aerodynamic properties, though this is probably not representative of dusts in ambient air. Darke et al (1976) studied spores and other parts of fungi collected from combine harvesting, by microscopic observation, and lists relative proportions of the species identified.

B.3.2.9. Conclusion

Taken altogether, there is considerable range in reported levels of particulates during arable operations. For cereal harvesting, a mean level of 20 mg.m-3 total dust corresponds reasonably well with a mean of 1 mg.m-3

respirable dust, as two reports indicate that the respirable fraction of dust from harvesting is around 5 % by mass. The extrapolations from Darke et al’s spore count are also reasonably close to this amount. However, this is not considered enough data to have great confidence in the dust level, let alone in being able to predict emissions. Soil disturbance is reported by a number of surveys to generate dust clouds with a wide range of concentrations, but most lying within the range of 10 - 100 mg.m-3 for total dust, and 0.5 - 1 mg.m-3 for respirable dust.

There are major problems in drawing conclusions from these results. First of all, individual measurements are extremely variable. The range between maximum and minimum individual measurements is often huge, with

58

Projecttitle



maximum measurements sometimes 50 (or more) times as high as minimum ones (Batel, 1979, Norén, 1985 ). Results change considerably with even small changes in location of sampler, as illustrated by Norén (1985) and Burg (1982). Comparison of results by Clausnitzer and Singer (1996) with those of Nieuwenhuijsen (1998 and 1999) for measuring similar processes in California, also shows considerable differences in respirable dust levels. The former study used samplers intended to measure emissions fairly close to source, whereas Nieuwenhuijsen was measuring personal exposure of workers. Another cause of variation in results is choice of sampler and calculation to obtain results, as comparison of Niewenhuijsen 1998 and 1999 illustrates. The older references are much less explicit than those detailing the Californian studies, as to how results are actually derived. A further complication is that a representative density of particulates must be taken, for setting up the samplers. Particles of different densities may behave differently within the apparatus and cause error in results.

Those are the issues related to sampling methodology. There are also numerous factors which cause real variations in levels of particulates. Environmental conditions, namely wind speed and direction, soil moisture content, amount of solar radiation (which causes convection currents in hot conditions, lifting particulates above the ground and allowing them to be dispersed more readily by the wind) and rain, can all have marked effects on dust levels generated, as can the soil type, and the nature of the process. Some work in California (Nieuwenhuijsen and Schenker, 1998, and Clausnitzer and Singer, 2000) concentrates on this theme. However, reliance on results from California is only considered useful as an initial guide to potential levels generated, because of considerable differences in climate between California and the UK. The studies by Batel and by Norén (and probably also that by Louhelainen et al, 1987a) are assumed to have measured in conditions similar to those in the UK, though, unfortunately, sampling procedures are not described in great detail. In short, further monitoring in the UK would be needed to obtain a realistic figure for likely emissions from these processes in the UK. Large numbers of measurements, to allow for spatial variability, and the influence of weather conditions etc., would therefore be needed to obtain any result which could be useful.

A further factor which has not been discussed, is the need to allow for the proportion of particulate which is deposited within the field, and so is never emitted from the farm. This proportion will depend strongly on weather conditions and field size, but it is clear that any measurements close to source must be regarded as the worst case, or as over-estimates of what may actually leave the farm. Researchers in California are starting work to evaluate these dispersion and deposition effects, and are hoping, in the next 1 - 2 years, to develop a PM reduction factor which could be applied to measurements made close to source (P. Gaffney, pers. comm., April 2000).

B.3.3. Spores and pollen grains relevant to agriculture

Spores of moulds, and pollen grains, can be present in high numbers at certain times of year, and on occasions may outnumber other particles in air in some areas. Some are allergenic or can have other undesirable effects on human health if inhaled (such as Aspergillus fumigatus, as described in Section B.2.2, above). This matter is discussed in more detail in Section 4 of the main report. Spores can be emitted also from some plants, namely ferns and mosses. Microbes specific to animal feed and bedding are also discussed in Section B.2.2.

B.3.3.1. Fungal spores

Spores of fungi and other micro-organisms are ubiquitous in the environment. These organisms can grow on a range of materials, including living and decaying plants, which include agricultural crops. The organisms release spores, often in vast numbers, by a variety of means. It is impossible to generalise about the behaviour of fungi and other microbes as whole, though the behaviour of some individual species has been well characterised. Some are active all year round, others mainly at certain times of year; there are also well established diurnal patterns of release. Many different mechanisms are employed to release spores into the air. Some are blown during dry windy weather, others during heavy rain, and some have elaborate methods of dispersing spores. (See Gregory, 1973, for further information.) Disturbance by intervention, such as

59

Projecttitle



harvesting or shaking bales of mouldy hay, can produce particularly high counts, though for short durations. There are certainly occasions when spores outnumber all other types of particles in air, at certain times and in certain areas. However, quantitative information about the amount of biological particles which may be evolved from crops is extremely sparse.

The most common spores in the air of the UK are Cladosporium; other common types include Alternaria, Aspergillus and Penicillium; these can be found on vegetable matter including crops of cereals, but also non-agricultural plants. It would be extremely difficult to ascertain whether the source of such spores was agricultural or not. However, certain species (Didymella) can be found in high numbers in July, with a sharp decline after harvest time. This was thought to be the cause of an outbreak of asthma in Birmingham in 1983 (Warner, 1998).

Levels of infestation can be high in fields of crops, which can lead to the release of many millions of spores from small areas. Reports of clouds of spores being disturbed from harvesting cereal crops, particularly after a damp season which has allowed moulds such as rusts and mildew to flourish, suggest this is not uncommon. Darke et al (1976) measured concentrations of spores behind combine harvester of over 200 million spores per cubic metre of air, with mean levels between 57 and 75 million per cubic metre of air, as compared with a reported "background" count of up to 1 million spores/m3 of air. The article does not state that any of the 3 years studied were particularly bad in terms of fungal infestation, so it assumed these are "average" conditions.

Measurements of spores at several points above wheat fields were made by Eversmeyer and Kramer (1987). For most of the species identified, concentrations declined with increasing sampling height from the top of the crop canopy up to 14 m. One interesting observation was that the location of infestation within the crop is important in determining releases of spores to air. Spores produced and released near the top of the crop resulted in relatively high concentrations in air at heights between 1 and 3 metres, but infestation lower down the plant did not result in high releases. In general, largest numbers of spores were trapped when the plant surfaces were dry. Wind was found to increase spore releases. The maximum levels noted were not particularly high: Cladosporium, the most numerous species, was present in average concentrations of 10,000 particles per cubic metre of air at canopy level, declining with increasing sample height to around 3000 particles per cubic metre at 14 m. (This corresponds to mass concentrations of, very approximately, 3 and 1 g per m3 of air, respectively, assuming the spores have a mean volume of that of a sphere of 9 m diameter. (This is calculated using the assumptions stated in Section B.3.2.6.)

Some information about prevalence of certain diseases on certain cereal crops is routinely collected, by ADAS/MAFF Central Science Laboratories: see, e.g. Hardwick et al (1998a and 1998b). This could, theoretically, be used to give an estimation of some of the spore releases. However, the relationship between degree of infestation (expressed as area of leaf covered) and release of spores is not simple. There is anecdotal evidence that in controlled conditions, a lesion on a diseased leaf would often produce up to around 4000 spores per square cm of lesion per day (Jeger, pers. comm.). Assuming a leaf area index of up to 6 times the field area, and an infestation of 5 % (disease surveys for 1998/9 for winter barley and wheat reported 2 - 12 % of leaf area was affected by disease at the time of survey) this would indicate a maximum infected leaf area of 0.3 square metres per square metre of ground, or 3,000 m2 per hectare of cereal crop. The above prediction suggests that this area of infected leaf could theoretically produce up to 10 million spores per day per m 2 of field, or 1011 spores per hectare of crop per day. If these spores had a mean volume equivalent to that of a sphere of 9 m diameter, this would mean of the order of 3 mg of spores could be evolved from a square meter of field per day, or 30 g per hectare of crop per day. If this level of infestation was sustained for 3 months, the number of spores evolved would be 109 per square metre, or 1013 per hectare. Assuming the same spore sizes as above, this would correspond to a mass of 300 mg per square metre, or 3 kg per hectare.

There is also anecdotal evidence of a diseased leaf producing up to its own weight in spores over its lifetime (Jeger, pers. comm.). The weight of leaf in a field of cereal at harvest time is approximately equal to the weight

60

Projecttitle



of grain, say, about 8 t per hectare. (Note that this is not the maximum weight, as by harvest time the leaves have dried.) If one again assumes a 5 % infestation rate, and also that all of this mass of the infected leaf mass is evolved as spores (probably an over-estimate), then 5 % of this leaf mass would be converted to spores. This would correspond to 40 kg per hectare, over the whole season. This result is fairly close to that predicted from leaf area (3 kg per hectare, over 3 months). All calculations of this sort are very approximate (selection of a mean spore size which is not representative, could alter calculated results by a factor of 100 or more) and agreement within one or two orders of magnitude is the best that could be expected.

Of course, some of these spores would be much larger than 10 m in diameter, so would contribute greatly to total particulate levels in air, but not to PM10.

The above values are far higher than the concentrations measured above wheat fields by Eversmeyer and Kramer (1987), but similar to (slightly lower than) Darke et al’s (1976) spore counts during harvesting. Much higher concentrations are to be expected during harvesting, as this is a single event, which involves massive disturbance to the crop, aiding the projection of spores into the air.

Actual releases from a field would almost certainly be far smaller, and depend upon such factors as weather and position of lesions. Furthermore, these disease surveys only tell part of the story. Crops are normally surveyed in late spring, in order to assess the potential for crop loss through disease, and hence the amount of fungicide to be applied. Thus, any releases earlier in the year, such as from winter cereals during a mild spell in the winter, would remain unrecorded. More importantly, plant pathologists only study species causing diseases of interest, which are rarely the most numerous species present on the crop. In cereals, by harvest time, a different flora is present, living saprophytically on the dried parts of the plants. Although these organisms are not of interest to plant pathologists, they are probably responsible for the largest releases of spores to air, and include known allergens such as Aspergillus species (D. Jones, ADAS Rosemaund; N. Hardwick, MAFF Central Science Laboratories; A. McCartney, IACR, Rothamsted, pers. comm., 2000).

ADAS also routinely collect information about the incidence of potato blight. Reports of episodes, together with weather reports, are designed to give a strategic forecast of high risk areas where farmers would be advised to apply fungicide more frequently. This information is not quantitative, and not intended to give an estimation of the degree of infestation by blight. However, the numbers of spores actually released from this disease are considered to be small in comparison with releases of spores from combine harvesting a cereal crop after a damp season. (N. Bradshaw, ADAS Cardiff, pers. comm.)

B.3.3.2. Emissions of spores from compost, mulches, and other decaying organic matter

Organic matter is broken down by micro-organisms, including bacteria, fungi and Actinomycetes. The temperature, degree of aeration, and other conditions determine which species thrive, and this can change with time. Although this material, left undisturbed, generally does not emit significant concentrations of spores, operations such as turning compost (one way to allow the necessary aeration) are reported to release clouds of steam, including large numbers of spores. Species isolated include Aspergillus fumigatus and Thermoactinomyces vulgaris (which thrive in warm conditions), and a number of others, and these are believed to be responsible for at least some of the cases of "mushroom worker's lung", a similar condition to "farmer's lung" (Crook and Lacey, 1991, Van den Bogart et al, 1993, Vinken and Roels, 1984). Measurements at a sewage sludge composting site in New Jersey, USA, found maximum levels of A. fumigatus and thermotolerant fungi were 2,000 - 4,000 colony forming units per cubic metre of air, measured 1 m downwind from the pile. These peaks were mid morning, after agitation of the pile (Kothary et al, 1984). This is not a high count, and not believed to be significant to anyone but occupationally exposed workers. A similar conclusion was drawn from the review written by Déportes et al (1995) about composts made from municipal waste. High releases of fungal spores and bacteria can cause exposure of workers to allergens and endotoxins, and a case was described where a man was believed to have been exposed to over 109 cfu m-3 spores, after manipulating farmyard waste

61

Projecttitle



composts (Weber et al, 1993). However, the risk to the general public areas near where compost was placed, from inhaling significant numbers of airborne spores, was considered negligible.

The production of mushroom involves preparing a suitable substrate. Mushroom compost is made from a variety of substances, though straw is commonly used in the UK. The material is turned and watered every few days during its processing, and as this can release clouds of steam containing spores, this can evidently be a source of emissions. Emissions of spores from mushrooms themselves are very low from ordinary "champignon" as they do not release before harvest. Any workers who become sensitised are affected by microbes in the compost, rather than by spores from the mushrooms themselves. However, Japanese mushrooms (Shiitake, Pleurotus) are reported to release large numbers of allergenic spores continually, causing more of an occupational hazard (Sastre et al, 1990).

B.3.3.3. Bracken

Though not an agricultural crop, bracken is mentioned here, because it releases large numbers of spores, and agricultural practices (grazing) are important in its control. Bracken occupied an estimated 975,000 hectares in the UK in 1998, and was reported to be increasing in area by between 1 % and 3 % per annum. The increase is attributed, at least in part, to under-grazing by sheep and cattle in areas prone to encroachment (edges of heather moors). Related factors reported include financial pressures on beef farmers from the BSE crisis etc., and a cut in the bracken grant. Bracken emits spores throughout the year, though the main releases occur from August to early October. Sporulation during this season appears to be triggered by the onset of warm sunny weather, drying and warming the fronds, and allowing release of spores to occur. (Taylor, 1998 and pers. comm.) Bracken will contribute significantly to the local particulate count in areas with high cover, though mature spores are generally larger (by 2 to 3 times) than 10 m in diameter, so they probably do not contribute in a large way to the PM10 count. They are of particular concern because they are believed to be carcinogenic, though research is still in progress, aiming to confirm this.

B.3.3.4. Pollen

Pollen grains are generally much larger than 10 m in diameter: most are over 30 m diameter. As these do not fall within PM10, they have not been considered in detail in this report. Although they are generally far too large to enter the lung, allergic reactions occur in sensitised persons when these particles are inhaled and deposited in the nose and throat. In the UK, grass pollen is the major cause of hay fever, however, in Sweden, birch pollen is the major culprit, and in the Mediterranean region, olive tree pollen affects more people. (Emberlin, Pollen UK website). Some individuals can show allergic symptoms from other types of pollens. Allergenic problems are generally found from pollen from plants which are wind pollinated, rather than insect pollinated, as these release far greater quantities of pollen into the air. However, oil seed rape, which is primarily insect pollinated, can be a source of significant quantities of pollen in air (reported as a result of extensive planting in exposed sites), though these pollen grains are reported to have a relatively short range in air, compared with grains which are air-dispersed (McCartney and Lacey, 1991). Rape pollen is also allergenic. Pollen grains from cereals tend to be collected in high numbers at harvest time, though surprisingly, not during the flowering period. (E. Caulton, pers. comm.)

Pollen grains are known to fragment occasionally, as stated in the Airborne Particles Expert Group Report, 1999. Knox (cited in Emberlin, 1995) states that a single mature rye pollen grain can contain 700 starch granules of the size 0.6 - 2.5 m, which are potential allergens. Outbreaks of asthma in Australia during thunderstorms are considered by some (as Knox proposes) to be the result of this; an alternative hypothesis is that releases of fungal spores in heavy rain trigger these symptoms. The general opinion of biologists and aerobiologists is that pollen grains are extremely robust and resist damage in dry air. (Many have also survived in sediments for thousands of years in a good condition; they can also resist some harsh chemicals.) Thus, fragmentation would be expected to be an extremely rare event, perhaps precipitated by extreme meteorological conditions such as static charges during thunderstorms, and affecting a tiny proportion of grains. However,

62

Projecttitle



others believe that fragmentation may be a relatively common event. This remains a matter of controversy to those in the fields of aerobiology and air pollution.

B.3.3.5. Methods of counting and identifying spores and pollen grains

Pollen counts are made using standard methods of using a Burkhard trap, placed in a suitable location to collect mixed air, and draws air at a standard rate over a sticky tape. The tape moves slowly at a known rate, so it is possible to know the date and time at which any section of the tape was exposed. The tape is changes at weekly intervals, then cut and mounted on slides for microscopic examination (Emberlin, Internet, 2000).

This approach is very time consuming and labour intensive, though it is possible to train a person to recognise the common types of pollens in about a week. However, identifying fungal spores is much more difficult. This is because they are smaller (many 5 - 50 m, some smaller), and particles smaller than about 5 m are difficult to see with an ordinary light microscope. Also, the different species of fungi are far more numerous, and identification can be very complicated. Furthermore, some types (e.g. Penicillium and Aspergillus) look so similar it is difficult or impossible to distinguish them by this method.

Routine measurements using this method are made at the UK Pollen Network sites (two of which also monitor spores and identify the main taxa), and also some other research centres, including the Scottish Centre for Pollen Studies (Caulton), and the National Air Quality monitoring site near Rochester (Burt et al). However, much of this data remains unexploited. The main difficulty is the time needed for a trained aerobiologist to analyse the results of the traps. Some institutions are working on development of automated procedures, using approaches including image analysis of particles, and tagging particles with fluorescent stain (Griffith). However, no alternative method is yet in widespread routine use. There is much interest in this subject by researchers in many fields, including plant pathology and human and veterinary medicine. This is obviously an important research and development need.

B.3.4. Wind erosion of UK soils

Some areas of the UK are known to be prone to wind erosion of soils. These areas are principally in East Anglia, the East Midlands, and the Vale of York, where the soils are of the fine peat or fine sandy types, which dry out rapidly. Severe problems are most commonly experienced from March to June for spring planted crops, because the ground is bare at this time. Anecdotal evidence is that levels of "nuisance dust" (i.e. total particulate, most or all of which would be large particles) can be very high during episodes of windy conditions and soil disturbance or harvesting in dry conditions. However, opinion is divided among persons consulted at ADAS Arthur Rickwood Experimental Farm (in Cambridgeshire), on whether or not this could include a significant amount of finer particulates (PM10), as the bulk of the suspended particles are believed to be agglomerates with a much larger size. However, it is reported that the airborne dust penetrates window seals over time, so a proportion of fine matter must be present. (S. Runham, pers. comm.) Measurements of soil erosion have been made in the UK, but with regard to crop damage and loss of fertiliser or seed, rather than to particulate levels in air. Current abatement measures are considered able to control the problem (in terms of crop damage and topsoil loss). Measures to reduce wind erosion are listed in two MAFF booklets: Controlling Soil Erosion (1998), and The Soil Code (1998).

US reports and emission factors consider that it is the silt fraction of soils (and also of unpaved road surfaces) which is the source of PM10 emissions, as these can contain particles of this size. (The silt fraction is defined as the particles between 2 and 20 m diameter, while the sand fraction is the coarser material, and the clay fraction the finer material.) Sandy and peaty soils, though prone to erosion, are comprised mainly of larger particles. Clay soils consist mainly of finer particles, but these are usually bound together in larger agglomerates. Silt-rich soils in the UK do not dry out as rapidly as do sandy or peaty ones, so they are not reported as prone to

63

Projecttitle



wind erosion. However, measurements would be needed to ascertain whether this is likely to be a significant problem in UK soils.

A considerable amount of research has been done on the subject of wind erosion of soils in the USA and Canada. Some areas experience severe problems, because of a combination of dry climate, soil type and large areas without windbreaks. Most of these factors are absent, or present only to a much smaller degree in the UK, so these findings are not considered relevant to UK conditions.

B.3.5. Applications to agricultural land

A host of different substances are applied to agricultural land. Reasons for application include improving soil fertility and/or texture (in the case of manure, fertiliser, lime and sewage sludge), controlling pests (pesticide), and disposing of wastes (including manure and sewage sludge). Applications must not pollute water courses or have other harmful effects on the environment during or after handling and spreading. This includes not causing an odour nuisance. (MAFF Soil Code and Air Code, 1998.) There are limits on the maximum nutrient application permitted, and guidelines also state the time of year and weather conditions relevant to certain applications. Nutrients should be applied only during the growing season, to avoid leaching of nutrients which are not absorbed, contaminating water courses.

B.3.5.1. Manure

Cow, pig and sheep manures normally have a high moisture content (90 - 92 %, as quoted by Dagnall et al, 2000) and the potential for particulate emissions to air are generally considered to be very low if the manures are applied directly down on to the surface of the field. Slurry is applied by three main methods.

(i) Tanker spreading involves depositing slurry directly on to the ground from a tanker which is taken around the field by tractor. This is the method used for the majority of slurry.

(ii) Injection of slurry under the surface is also performed, and is the "cleanest" application method. However, this requires more fuel and is thus more expensive, and is only performed on an estimated 5 % of all UK slurry (Sneath and Phillips, pers. comm.). This method is also impractical for some types of soil. Particulate emissions from the tractor exhausts would also be higher from tractors injecting slurry, compared with those spreading slurry on the land.

(iii) The cheapest application method is by spray gun, which flings liquid slurry through the air, covering a wide area within the field. This causes intense fragmentation of the slurry, and although the bulk of the material is certainly deposited in large droplets, there is the potential for fine particles to be emitted, which may reduce in size because of evaporation. Regulations state that this is not to be performed in windy conditions (wind speed over 6 mph), though this condition may not always be adhered to, particularly during prolonged windy weather. This approach is increasingly discouraged because of odour problems, particularly near to houses, and an odour nuisance can lead to an abatement notice being served. However, in areas far from local residents, spraying can be used. It is estimated that only a small proportion of slurry is now applied by this means (around 5 %, Sneath and Phillips, pers. comm.).

A French study investigated the likely risk of pathogens being dispersed from manure spreading processes (Boutin et al, 1988). Air was sampled at several locations in fields where manures were being applied. Sampling was done directly with several Petri dishes, and air was also collected and separated into particle size fractions using an Anderson sampler, and these particles directed on to Petri dishes. Most results showed an average concentration of around 2000 colony forming units per cubic metre of air (as compared with a background level of 20 - 40 cfu.m-3 air). Not surprisingly, higher counts were found in fields using a gun

64

Projecttitle



spraying method; applying slurry to the ground from a tanker, which was transported around the field by tractor, resulted in lower biological particle counts. Spraying manure with a spray gun could be expected to be a significant source of particulates, as well as of microbes; bacteria and spores attached to particles of faecal material are more likely to survive, as the manure can protect them from damage by sunlight and desiccation. However, the authors considered that the risk to local people from inhalation of bacteria spread by this process was very low, especially considering that this is an infrequent process of short duration. It must be emphasised that this study only measured counts of living and viable microbes, so gives no information about dead biological particles (which can also be harmful, particularly if they contain endotoxins) or other particles.

Poultry litter is also applied to soils (though some is burnt in power stations: Dagnall et al, 2000). This material usually has a much higher dry matter content than animal manure, and is often noted as powdery in texture. (The dry matter content of manure from laying and breeding fowl is quoted as 30 %, increasing to 70 % following partial air drying, which is employed in some units to reduce ammonia emissions. Broiler manure, after mixing with litter, typically has a dry matter content of 70 % (Dagnall et al, 2000.) Thus, this has the potential to be a significant emitter of particulate matter. Anecdotal evidence suggests that spreading litter on fields can be a dusty process. Particulate levels measured in poultry buildings were found to be very high during cleaning (as described in Section B.2.1), and add weight to these observations, though no reports of direct measurements have been identified.

B.3.5.2. Sewage sludge

Sludge from sewage works is also applied to UK agricultural land. This is applied in various forms: as a liquid slurry, with a similar consistency to animal manure; as a de-watered sludge cake; or as thermally dried, lime-treated, or composted sludge (MAFF Soil Code, 1998). It is applied using similar methods to those used for slurry, though a larger proportion (around 20 - 40 % of liquid sludge, Sneath and Phillips, pers. comm.) is injected, to reduce odour problems. The potential for emissions depends upon the application method and water content of the material. There are a number of controls in place (as explained in DETR Code of Practice for Application of Sewage Sludge, 1996) though these are mainly concerned with nutrient, heavy metal and pathogen contents of the sludge, which are less relevant to emissions to air.

B.3.5.3. Fertiliser

Most fertilisers are applied in pellet form, rather than as powders, and are not considered to be significant sources of particulate emissions. The USEPA has published emission factors for emission of nitrate from fertilisers, but there is no published data for emissions of particulate matter from this source. Application of fertiliser is controlled in terms of how much nitrate is applied to the land, because of problems of nutrient leaching into watercourses. (This is described in further detail in MAFF Water Code, and MAFF Soil Code, 1998.)

B.3.5.4. Lime

Lime is applied to neutralise acidity and improve the texture of soils. A number of grades of lime are commonly used in agriculture: ground chalk, screened chalk, ground limestone, and screened limestone. These are defined in terms of size of mesh which the material must pass through. The finest material would appear to be ground limestone: at least 40 % of the material must pass through a 150 micron mesh. None of the standards state a minimum particle size, and so the material can (and generally does) include very fine particles. The application of waste lime (and also gypsum) from industry to agricultural land is also permitted, and is exempt from waste licensing conditions (MAFF Soil Code, 1998). The most common industrial by-product used is sugar beet residue, but water works waste, whiting sand and blast furnace slag, lime sludge from cement manufacture or gas processing, are also applied. Lime is considered to be the applied material most likely to pose a significant contribution to particulate emissions, and further study is to be recommended.

65

Projecttitle



B.3.5.5. Pesticides

Pesticides are routinely applied to arable crops, with the exception of the small amount of organic produce grown in the UK. Spray drift, that is, the failure of the pesticide to be deposited on its target (plant, soil) and instead be carried away by the wind, is a subject of concern for several reasons. As well as causing health implications for workers and any other people exposed, drift can cause ecological damage, reduce yields of crops in neighbouring fields. In extreme cases, farmers may apply larger quantities to compensate for loss, and so increase their costs.

In theory, spray drift may occur in any application procedure which involves the transport of liquids or dusts over some distance to the target surface (Göhlich, 1983, cited in Maas and Krasel, 1988). Spray drift is a complex subject, as it is influenced by many factors, the most important ones being mode of application (which determines the size of droplet or particle, its direction and speed of ejection), and wind speed at the time of application (Ganzelmeier, 1990). Droplet size is particularly important, as smaller droplets are deposited much more slowly (particularly those below about 100 m diameter). Furthermore, with pesticides based on water or on a volatile organic solvent, droplets, if not quickly deposited, usually reduce in diameter by evaporation, and so are increasingly unlikely to be deposited.

Much work has been done to reduce this phenomenon. Measures include suitable design of nozzle and choice of pumping speed (which can reduce the number of small particles prone to drift), covering of booms, using fan assistance to direct the air stream towards its target, and also the development of better products with improved formulations. However, it is agreed that a proportion of the pesticide is likely to escape in windy conditions. Guidelines about application of pesticide include the maximum windspeed during which application should be performed, and that pesticides should not be applied near water in windy weather (MAFF Water Code, 1998).

In the UK, most crops are sprayed using vehicle-mounted boom sprayers, which direct the substance down onto the crop, and long range drift from these is considered to be small in amount, provided recommended procedures are properly followed. However, in the spraying of bush or tree crops (orchards) pesticides are directed to heights of say 5 m, and are then much more likely to be dispersed downwind. Aerial spraying (i.e. from aeroplanes) can cause much larger amounts of pesticide to be widely dispersed in the air. This used to be widely done in the UK, but is now rarely performed except in specific areas where access is limited (e.g. to target bracken in upland areas). With boom sprayers, it is estimated that 1 % of the nozzle output is still airborne 10 m downwind; with orchard sprayers, that 10 % is still airborne 10 m downwind, and with aeroplane spraying, more than 10 % (Miller and Walklate, pers. comm., 1999).

A study in the USA (Simcox et al, 1995) found that house dust in homes near to fruit orchards was likely to have low (below 1 ppm) though detectable levels of organophosphate pesticides as used on the trees, regardless of occupation of the residents. The authors concluded that these substances appeared to be widely dispersed in the environment at low level. Pesticide levels in house dust were much higher in homes where a family member worked in the orchards, even when they claimed to follow prudent work practices, such as using protective equipment, and not bringing this home. The pathway was not established, so this does not prove significant wind dispersal of pesticides, though this is certainly a possibility. A potential route is deposition of pesticide some distance from source, and entrainment in soil or dust, which is carried into homes on the soles of feet; however alternative routes are also possible. The authors concluded that further study was warranted to investigate exposure risks to young children. The area sampled could be regarded as a worst case because of the proximity to orchards; also the number of reference households studied was fairly small (11) so this should only be used as an indication of any potential hazard.

The USEPA have developed emission factors for pesticides in terms of active ingredients or solvents dispersed to atmosphere, but have made no estimations of particulate dispersal.

66

Projecttitle



Long range pesticide dispersal can also occur after pesticides are incorporated into soil, or enter watercourses and are deposited with sediments which later have dried out. This occurs if the contaminated soil or sediments are then subsequently eroded by wind. An extreme example of this is reported in a badly degraded area, the Aral Sea region in Central Asia (O'Hara et al, 2000). However, these conditions are not representative of those in the UK. It would only be a potential issue in areas which are both prone to wind erosion, and where the soils have significant concentrations of pesticides.

B.3.5.6. Other wastes

A number of other waste products can be applied to agricultural land, without waste management licensing, subject to certain conditions including maximum total amount (250 t per hectare per year, except for inland water dredgings which can be applied in higher amounts). These are: waste soil or compost, waste wood, bark or other plant material, waste food or drink, blood and gut from abattoirs, waste hair and effluent sludge from a tannery, paper waste sludge, textile sludge, sludge from biological treatment plants, septic tank sludge, and dredgings from inland waters (MAFF Soil Code, 1998). These are not considered likely to be significant sources of particulate emissions to air.

B.4. Post harvest and miscellaneous processes

B.4.1. Grain storage and transport, and processing into animal feed

Particulates from grain processing are not a new problem. In 1555, Olaus Magnus, a Swedish bishop, wrote: "When separating the grain from the chaff, care must be taken to choose a time when there is a suitable wind which will sweep away the grain dust, so that it will not damage the vital organs of the threshers. This dust is so fine that it will almost unnoticeably penetrate into the mouth and accumulate in the throat. If this is not quickly dealt with by drinking a fresh ale, the thresher may never again, or only for a short period, eat what he has threshed." (Quoted in Malmberg and Rask-Andersen, 1993.)

More recent work has confirmed that workers exposed to grain dust can suffer severe illness following exposure (Malmberg and Rask-Andersen, 1993; Chan-Yeung et al, 1978; Dosman and Cotton, eds, 1980). However, the bulk of the studies identified were epidemiological, and did not measure dust concentrations.

It is accepted that a significant proportion (up to 1 %, Burg et al, 1981) of grain is lost as dust from elevators. A study of exposure in 3000 Canadian grain workers (McDuffie et al, 1991), found that in elevators where dust control equipment was installed, mean dust levels were 6 mg.m-3. In elevators without dust control equipment, mean levels were 14 mg.m-3. The range of individual measurements was wide. This shows a considerable proportion of workers are exposed to unacceptably high dust levels. An older study (Chan-Yeung et al, 1978) found much higher levels of dust (10 - 600 mg.m-3) near to Canadian grain elevators. The age of elevators, ventilation, and pre-treatment of the grain (such as partial cleaning) all influenced concentrations measured. This suggests that conditions in grain elevators had considerably improved in the intervening decade, though possibly by venting the dust more efficiently to air. Though it would be extremely difficult to estimate emissions from these factors, these studies suggest that dust emissions from grain elevators can be significant.

The USEPA has formulated emission factors for various grain handling operations, which are shown in Table B.12, below.

67

Projecttitle



TABLE B.12. USEPA estimates for emission from handling and milling wheat grain

Operation Details Control Emission in g/tMeasure total dust PM10

Grain receiving straight truck None 90 30hopper truck None 17.5 3.9railcar None 16 3.9

Grain cleaning vibrating cyclone 37.5 9.38

Grain drying column dryer none 110 27.5rack dryer none 1500 375

Internal handling none 30.5 17

Grain shipping truck none 43 14.5railcar none 13.5 1.1

ANIMAL FEED MILLS

Grain receiving none 8.5 1.3

Grain milling hammer mill cyclone 33.5 16.8hammer mill baghouse 6 6flaker cyclone 75 38graincracker cyclone 12 6

cyclone 180 90Pelletising pellet cooler high efficiency

cyclone75 38

Feed shipping none 1.65 0.4

It is not known how accurate these emission factors are. Shaw et al (1997) describe trials which measuring emissions from grain unloading and feed loading at animal feed mills, and their results indicated much lower emission factors. This study measured emission from maize, rather than wheat (which is the most widely produced UK grain) and the numbers of trucks used in the final emission factor calculation was relatively small (14 - 20) so it is not known how certain these findings are. However, this certainly suggests that the EPA factors should be treated with caution. This article also explains some possible reasons for differences in dust levels. The speed of loading and unloading, and in particular the distance which the grain is allowed to fall freely in air is very important. Arranging loading and unloading conditions to minimise the drop in air by the grain is reported to prevent much of the dust present in the grain from being entrained in air, and thus can effectively reduce dust emissions.

The above factors indicate that PM10 is up to half of the mass of total suspended particulate. However, other literature suggests an even higher proportion of fine material, e.g. Yoshida and Maybank, 1980, found 40 - 50 % of the dust was made up of particles below 5 m aerodynamic diameter.

In the UK, emissions from mills crushing seeds, and also from mills producing animal feed, are covered by the Environmental Protection Act, (IPC part 2 processes regulations), and are under the jurisdiction of Local Authority Environmental Health Officers. Information about national emissions has not been collated from the

68

Projecttitle



various authorities. Many mills process imported seeds. Dust levels up to 50 or 100 mg.m-3 are permitted (for different mills). The particle size fraction is not specified. The industry report that mills are considerably less dusty than they once were, and thanks to effective containment and ventilation systems, there is now little evidence of dust inside mills or in the immediate surroundings. (M. Lapper, BOCM Pauls, pers. comm.)

B.4.2. Cereal drying

Grain is commonly dried in driers on the farm. This is done by forcing air through cereals, usually at high temperature, is not subject to any regulation, but is considered to be a significant source of particulate emission. (P. Metcalf, ADAS, pers. comm.) this process is not subject to any regulation, so data is not collected. Louhalainen (1987) reported farmworkers emptying grain driers were exposed to an average total particulate concentration of 22 mg.m-3, based on 6 samples of 1 - 2 hr duration. Batel (1979) found operator exposures averaged 7 mg.m-3. This is insufficient data to attempt to estimate emissions to air.

It should be stated that failure to dry cereal thoroughly will also result in emission of particulates, because mould infestation is likely to occur during storage, as described by Kotimaa et al (1991).

B.4.3. Potato grading

A study of potato grading enterprises in the north east of Scotland measured personal exposure of workers (Robertson, 1993). Some very high levels of exposure were found. Inhalable dust was found to range from 2 -148 mg.m-3, with an average level across all farms of 27 mg.m-3. Respirable dust levels ranged from 0.3 -52 mg.m-3, with an overall mean of 1.4 mg.m-3. Potato grading operations lasted from about 7 weeks in small enterprises, to 8 months of the year in large facilities. All samples were based on single visits to 6 venues, and were measured both on 4 operators (inhalable and respirable dust) and airborne dust was measured at 6 static locations on grading equipment. Measurements were according to HSE recommended practice. There must be considerable uncertainty because of the small sample size and variability of results, however, these indicate that significant quantities of particulate, including fine particulate, are generated by some of the steps involved in potato grading.

Preparation of other root vegetables is likely to be a similar source of particulates.

B.4.4. Incinerators, fires

Many small farms have their own incinerators, for disposing of items such as dead birds. Incinerators operating at below 50 kg/hour are exempt from regulatory control (MAFF Air Code, 1998). If incinerators are run at high temperature, with a plentiful supply of oxygen, then the material emitted should not be rich in particulates or other undesirable substances, as they should be completely combusted. However, products of incomplete combustion may be rich in particulates and noxious substances. The conditions achieved in small incinerators is likely to be variable, but not known, as they are not subject to regulation, other than the black smoke regulations. These are certainly a source of particulates, though likely emissions cannot be estimated without more data. Any open fires used to dispose of items such as fertiliser bags or hedge trimmings would be likely to emit considerable amounts of particulates, but again there are currently few if any data. (NB. Emissions from forestry are not included in this review.)

B.4.5. Burning of stubble

The practice of burning wheat and barley stubble in the field used to be a major source of particulate emission after harvest time. However, this has not been allowed in the UK for several years. Some stubbles are still allowed to be burnt in the UK, however. These are principally oilseeds and flax, which are particularly difficult to plough into the soil. The areas of these grown in the UK are very small so these are unlikely to be significant

69

Projecttitle



on a national scale, however, locally, the impact could be considerable at the time. Ortiz de Zarate et al (2000) have reported an emission factor for the burning of wheat stubble in Spain of 13 7 g Total Particulate Matter per kg of dry stubble.

B.5. Unpaved roads on farms

When vehicles travel over unpaved roads, they can pulverise aggregated material on the road surface, and entrain fine material on the tyres, which can be injected into the air by turbulence behind the vehicle. Most of this dust is deposited close to the source, but a proportion is transported by the wind. The amounts of emissions depend largely on the quality of the road surface, the amount and type of traffic, and the meteorological conditions (Gillies et al, 1999).

Re-suspended particulate has been measured even from paved roads, and has been found to consist of a complex mixture of materials, including: re-suspended dust carried in on vehicles from unpaved roads/construction sites, etc.; material deposited by wind or water; dust liberated by street repairs, exhaust, tyre and brake dust; and also small quantities of biological material such as pollen grains and animal skin flakes, even in urban areas (Claiborn et al, 1995, Rogge et al, 1993, Miguel et al, 1999). Measurements suggest emissions (per vehicle-mile travelled) from paved roads of the order of 0.5 - 10 % of those from unpaved roads (Claiborn et al, 1995).

Measurements in the USA have been conducted on emissions from unpaved roads, as it is a major source of PM10 emissions in many areas (Claiborn et al, 1995). It is estimated that in the agricultural San Joaquin Valley in California, around one third of the total PM10 emissions are from unpaved roads (Gillies et al, 1999). The UK situation is very different, of course, in both climate and lengths of unpaved roads. Nevertheless, the USEPA approach to estimating these emissions is given below.

The USEPA has a general empirical emission factor for predicting emissions from vehicles driving over unpaved roads. It requires the input of various data, namely:

the silt content (particles smaller than 75 m diameter) of the road surface; the moisture content of the road surface, meteorological data to allow for the number of wet days in an average year; average weight of vehicle fleet (i.e. the equation is for a fleet, not individual vehicles.).

Further correction must be made where the average vehicle speed is less than 15 miles per hour. Emissions are given in terms of emissions of total dust (taken as PM30), PM10 and PM2.5 per vehicle-mile travelled.

Data for vehicle-miles travelled have not been found, so total annual emissions have not been computed. The inventory includes predicted emissions per vehicle mile travelled, for several average fleet weights, corresponding to cars and heavier vehicles.

Two studies of emissions from unpaved roads in different parts of the USA (Claiborn et al, 1995, and Gillies et al, 1999) found measured emissions corresponded reasonably well (within an order of magnitude) with the predictions from this equation. However, measurement data are needed to verify which values of the variables should be used in the computation for UK emissions. These will vary across the regions of the UK according to soil/road surface type, rainfall, and road usage.

70

Projecttitle



The general equation for emissions is as follows:-

k (S/12)a (W/3)b

E = (M/0.2)c

k, a, b and c are constants (which depend upon the particle size of interest)S = % silt content of surface material W = average vehicle weightM= moisture content of the road surface (%)

A "worst case" computation, assuming (a) a fairly high silt content of 10 % (this is between the average values suggested for gravel or limestone surfaced roads, and for dirt roads, of 6.4 % and 11 % respectively), (b) a moisture content of 0.2 %, and (c) no allowance for wet weather or slow vehicle speeds, gives the following results.

Average vehicle weight (tons)

Emission per vehicle-mile travelled (kg)(or t per 1000 vehicle-miles)

W PM2.5 PM10 TSP1 0.096 0.658 2.26810 0.242 1.653 7.17350 0.460 3.147 16.038100 0.607 4.153 22.681

Thus, if one million vehicle-miles are driven on unpaved roads in the UK per year, the potential emissions are considerable (several thousand tonnes of total particulate, of which around a thousand tonnes are PM10).

The real value is likely to be much lower because of the frequent rainfall in the UK. The equation can be amended by multiplying by (1 – pwet/365) where pwet is the number of days per year experiencing rainfall (or for a better estimate, the number of days on which the road surface remains wet).

The equation was found to over-estimate emissions from vehicles moving below about 15 mph (which is likely to be relevant for some agricultural vehicles). The emission factor can be corrected to allow for slow vehicle speeds (below 15 mph), by multiplying by (average vehicle speed/15).

The USEPA equation was formulated under the following range of conditions

Surface silt % 1.2 – 35 %Vehicle weight 1.4 – 260 tonnesVehicle speed 5 - 55 mphNo. of wheels 4 - 7Surface moisture content 0.03 – 20 %

Computations outside of this range (e.g. higher moisture content, smaller vehicles) are less reliable. The USEPA attaches fairly good quality ratings to these emission factors (B for total and for PM 10; C for PM2.5). However, in the absence of relevant data for the variables, the rating must be reduced to a grade E.

Control measures in the USA have included treating the surface with petroleum derivatives and other materials, washing with water and reducing vehicle speed (Gillies et al, 1999). Reductions of up to 80 % have been reported using petroleum resins, and more modest reductions using other approaches.

71

Projecttitle



Further investigation is needed to ascertain whether actual emissions are anywhere near to those predicted, and if any abatement efforts are justified.

B.6. Energy use

The inventory in Section 3 details the emissions from the various categories of fuel and energy used on UK farms in 1998.

Combustion of the 200 kt of straw emits about two thirds of the PM10 from energy and fuel use. This is burnt in on-farm combustors.

Use of petroleum (786 kt in total) is the next largest on-farm source of particulates from energy, causing an estimated 710 t of PM10 to be emitted in 1998. Most of this was from "gas oil" (diesel) used in "power units" (most of which are tractors and other vehicles). A smaller amount was diesel used in heaters and dryers, fuel oil in power units (with a very small amount in heaters and dryers), and "burning oil" (paraffin or petrol).

Natural gas and coal use caused small emissions of PM10: 28 t and 23 t respectively. These resulted from a substantial consumption of gas (46 million therms, of the calorific value of 116 kt of oil), and a very small amount of coal (9 kt, with the calorific value of 6 kt of oil).

Consumption of electricity (3,870 GWh, with the calorific value of 333 kt of oil) was the category which caused the largest release (367 t) of PM10 in 1998. These emissions were released from power stations, and are thus already recorded in a sector outside agriculture, in inventories of particulates, but it is useful to apportion the amount used in agriculture. The emissions since 1998 are likely to have decreased, as the proportion of electricity generated in coal-fired power stations declines, and that in the much cleaner gas-fired stations increases.

Notes are given below on the calculation of energy use on farms. Total use of different classes of fuel and energy use are based on information from the DTI Digest of United Kingdom Energy Statistics 1999 ("DUKES 1999"), which lists consumption of fuels by various users, including "agriculture", using data for 1998, which is the most recent available.

Emission factors for PM10 for the fuels (but not electricity) used in the inventory are the ones used in the UK National Atmospheric Emmisions Inventory, 1999, based on USEPA (1995) emission factors for agriculture.

Emissions from the electricity consumed were calculated using emission factors from the Quality of Urban Air Review Group (1996) report. This used emissions data based on National Power and Powergen statistics for 1994-5, and quotes emissions of PM10 per GWh of electricity produced in coal-, oil- and gas-fired stations. Emissions from nuclear and other sources (mainly renewable and imported electricity) were assumed to be zero. The total emission for 1998 was calculated according to the relative amounts of electricity generated from the different classes of stations in 1998 (using information in "DUKES 1999"). (This may be an overestimate, if some stations had reduced emission of PM10 per GWh between 1994 and 1998.)

B.7. Conclusions

B.7.1. Livestock

1. Particulate emissions from livestock buildings in England and other countries in Northern Europe, have been reasonably well quantified. Results are based on buildings fairly representative of current practices. However, these samples did not include infrequent operations, such as moving animals and

72

Projecttitle



cleaning buildings, which could cause very high emissions, albeit for short durations. Thus, these emissions are considered conservative estimates of total emissions from livestock. These are believed to be the major annual particulate emissions from agriculture.

2. A number of studies of dust concentrations in buildings have been identified. They were all made with the aim of measuring exposure of workers. All measurements found higher ambient levels of dust than did the survey described above. There are many factors which could explain the differences, including sampling methodology, sampling duration, and different practices in older buildings. However, they do suggest that the former study could under-estimate annual average dust levels and thus emissions.

3. The nature of dust in pig buildings in the USA has been closely studied, finding that the bulk of the particles were feed, but the very fine ones were mostly manure. A similar study in UK buildings, where practices are different, would be very useful. A detailed study of the composition of particulate in poultry buildings would also be valuable.

4. Housed cattle and sheep should emit only small numbers of particles if kept under good conditions, and the bulk of these would originate from chopping and blowing straw bedding. However, if straw or hay is kept in a damp condition, very large numbers of microbes will grow over time, many of which are respirable and allergenic. Poor silage (i.e. silage which has been exposed to air) can contain also very many spores, though this is a relatively uncommon occurrence. Care with storage and choice of feed and bedding materials are important in controlling these emissions.

5. No studies of emissions from outdoor livestock were identified.

B.7.2. Arable

1. There is no published information relevant to UK conditions, from which a good estimation of UK particulate emissions can be made.

2. The USEPA have empirical emission factors for cereal harvesting and soil preparation. However these have been measured in a climate which is much drier than te UK’s/

3. A number of studies have been made to assess occupational exposure to dust during farming activities. Some studies in England, Germany, Sweden and Finland give dust concentrations for conditions similar to those in the UK. However, the spread of results is extremely wide, and these are older studies, which do not give a complete account either of the sampling procedure, or of the statistical treatment of results. Thus, they can be taken only as a guide to likely concentrations. Levels of over 100 mg.m-3 (total suspended particulate) in air have been found in several studies, with mean levels of well over 10 mg.m-3 for some operations. Harvesting and soil tillage can cause high concentrations of particulates in air while the operations are in progress.

4. Detailed studies have been made of agricultural particulate emissions in California. These studies include many operations, and give detailed accounts of procedures. However, there are significant differences in climate and agricultural practices between UK and California, and these results can be taken only as a guide to potential emissions in the UK.

5. Obtaining measurements which are both accurate and representative of levels of particulates generated by agricultural processes is very challenging. Several papers state how small changes in sampling location, and direction relative to the wind, can have large effects on measurements. Choice of apparatus can also greatly affect results, as sampler error can be important. Particulate concentrations

73

Projecttitle



are affected by many environmental factors, such as wind speed and direction, soil moisture content, soil type, solar radiation, rain, and operation being performed.

6. No allowance has been made in any of the papers, for the proportion of particulate which would deposit within the field. This could be large, thus greatly reducing the net emissions beyond the farm boundary.

7. Biological particles, such as spores and pollen grains, can outnumber other particulates, particularly around the harvest season, though their contribution to total annual emissions is probably fairly low. It has not been possible to estimate releases of these types of particles, as this is a very complex subject requiring further study.

8. Wind erosion of soils is known to occur in some parts of the UK. Monitoring is carried out, though in terms of crop damage, rather than emissions to air. Control measures can be effective. Further study would be needed to find if the contribution to PM10 in some local areas is large enough to be of concern.

9. Applications of substances to agricultural land may be sources of particulate. Lime is often applied in powdery form, and this may be a significant particulate source, and warrants further investigation. Dry poultry litter can also be dusty and may be also emit large amounts of particulate at the time of application. Applications which are of concern because of their potentially harmful properties are slurries spread by spray gun, and pesticides applied to fruit trees. Good Agricultural Practice should keep these to acceptable levels, though no quantitative information about potential emissions from these has been found.

B.7.3. Post harvest and miscellaneous processes

1. Transporting grain and producing animal feed can generate particulates. These processes are covered by IPC.

2. Cereal drying, commonly carried out on UK farms, is considered a high emitter of particulates, though data are not available, as this is not subject to controls. This is a necessary process to avoid development of moulds, but measurement of emissions is needed.

3. Small incinerators on farms may cause significant emissions, although no data are available. Any open fires would also emit particulate matter to air.

4. Permitted burning of stubbles (oilseeds and flax) is not thought to be a significant source on the national scale, but the particulates generated could be important in local areas.

B.7.4. Unpaved roads on farms

1. No data have been found for the UK, but use of unpaved roads is certainly a potential source of emissions.

2. Using the USEPA emission factor, with best estimates for road information, and worst case conditions (dry) suggest that these emissions could be large. Further study is needed to check if this is the case.

74

Projecttitle



B.7.5. Energy

1. Good quality data are available to estimate PM10 emissions from agricultural use of electricity and fuels in the UK.

2. A large source of emissions on the farm is use of diesel and oil, used for vehicles and heaters and dryers. The burning of straw in small boilers generates significant particulate emissions, although the underlying statistics on the numbers of boilers is now six years old, and so need updating. Emissions from the generation of electricity used on UK farms are also significant, though these are of course generated at the power stations.

3. Small emissions arise from the use of coal and gas on UK farms.

75

Projecttitle



Annex C: Sources of error in measured values of particulate concentrations in air

C.1. Errors inherent in the sampling apparatus

C.1.1. Comparison of results obtained using different types of sampler. Different types of samplers have their own characteristics, and often do not actually sample exactly what they aim to sample. Thus comparisons of studies using different apparatus should be made with caution.

C.1.2. Loss of particulate from apparatus. Many samplers collect dust on a filter, which is weighed before operation, then afterwards removed and weighed again. Loss of material, especially from heavily loaded filters, is one source of error, which can be substantial. In some samplers, dust can be deposited within the sampler, and not reach the filter, which would also result in an underestimation of the dust level. Conversely, if apparatus is taken through an area of higher airborne dust level, contamination with more dust is possible.

C.1.3. Definition of "total" or "inhalable", "respirable" and "PM10". Some authors quote total dust as being particles up to 30 m diameter, others up to 50 m, even 100 m diameter. The larger particles settle relatively quickly in air, and so are removed from the air unless re-suspended by strong currents. (According to Stokes' law, a smooth, spherical particle 100 m in diameter, of density 1 g cm-3, has a terminal velocity in air of about 20 cm s-1, whereas a similar particle 10 m in diameter falls at only about 2 mm s-1, and a 1 m diameter particle has a terminal velocity of around 0.02 mm s-1.) Differences in sampling procedure and apparatus will affect how many of the larger particles are included in the sample, and even a small number of larger particles will have a significant effect on the total mass of the sample. Respirable dust is measured in a sampler designed to simulate the deposition in the alveoli of the adult human lung, and the deposition characteristics of these samplers are well defined. (It is noted that most measurements of respirable dust levels agree reasonably well with each other.) PM10 measurements are defined as 50 % of the material of 10 m diameter collected, with lower collection efficiencies for larger particles, and higher collection efficiencies for smaller ones. However, this does not define the total sampling behaviour, so different PM10 samplers can obtain different readings.

C.1.4. Error in estimation of the flow rate of air through the apparatus. Any error in estimating the flow rate would result in a proportional error in estimated concentration of particulate in air. This can also influence the size selection behaviour of samplers.

C.1.5. Use of correction factors to obtain final results. Some samplers do not sample the particulate matter with 100 % efficiency, and so a correction factor is used to allow for this. However, these correction factors are not necessarily reliable, and can be a major source of error. In some of the older references, detailed explanation of the treatment of the results is not included, so it is not known what correction factors, if any, were employed.

C.1.6. Error in weighing the filter and sample. This can be significant for low masses of particulate.

C.1.7. Humidity. Changes in humidity can alter the mass of collected dust, as much of the dust collected from agricultural activities is hygroscopic, and would naturally absorb water vapour from the air. Measurements should ideally be made after leaving the sample in a controlled humidity atmosphere for enough time to allow it to equilibrate.

C.1.8. Density of particulate matter. In some types of samplers, particles of different densities will exhibit different behaviours. This can be problematic in the case of arable agricultural activities, where dust

76

Projecttitle



can be a mixture of plant fragments, spores, and soil, which can range in density from around 1 to 3 g cm-3.

C.1.9. Difference in treatment of results. All the values quoted are for the arithmetic mean (where both arithmetic and geometric are stated), though most studies simply quote "mean". For most of the distributions encountered by this type of study (which are not normal distributions), the geometric mean or the median may give a better measurement of an average value.

C.2. Errors inherent in position, timing or conditions during the measurement

C.2.1. Location of sampler. The value measured depends crucially upon the location of the sampler, as dust concentrations in air are not usually uniform throughout the plume. Several studies (e.g. Louhelainen et al, 1987) consistently found differences in dust levels measured in the breathing zone of farmworkers, from those from stationary samplers aiming to determine dust concentrations in the buildings (though this could have been partly the result of different sampling equipment). This is particularly important in the case of arable farming, where the range of individual values measured is high, (over a factor or 100) and a small change in location of the sampler has been found to have a marked effect on mean results.

C.2.2. Position relative to the wind. Several papers state that the direction of the wind has an important bearing on the results obtained. Some authors report oversampling when a sampler was facing into the wind, and undersampling when facing downwind. Some studies aim to reduce this error by mounting samplers on wind vanes, which maintain a sampler position of 90 degrees to the wind. The older papers in general do not provide this level of detail about how sampling was performed. This is a further reason why results from outdoor measurements should be treated with caution.

C.3. Reasons for variability in actual particulate concentrations

C.3.1. Activity of livestock or birds is an important variable, as movements have been shown to can generate higher dust concentrations. For this reason, dust levels are normally found to be higher during the day than at night, so estimating total emissions from daytime measurements would in most cases overestimate particulate levels.

C.3.2. Type of livestock building.

C.3.3. For arable farming, climatic factors such as soil moisture, wind, solar radiation (which can cause thermals which re-suspend particulates in hot conditions), all have major effects on the quantities of particulates evolved from operations.

77

Projecttitle



Annex D: Spreadsheets for preliminary inventory

Robert's tables in MS Excel

78

Projecttitle



Annex E: Feasible options for mitigating PM10 emissions from agriculture (i) livestock

Options for mitigating PM10 emissions from agricultureLivestock

A B C D E F G H I Joption description availabilit

yapplicability

magnitude

effectiveness

overall ease of cost of other "knock-on" acceptability

of of (reduction in

magnitude

on-farm on-farm interactions and farmer public

technology unabated emission) of implemen-

implemen-

on-farm effects

emissions

reduction

tation tation production environment

1 feed modification 4 wide high 3 4 3 none none 3 4low dust pellets

2 oil spraying 3 pigs only 2 2 safety increased 2 2CH4

3 bedding modification 4 wide 2 3 3 may need to find new 3 3use for straw

4 spreading bedding 1 mainly broilers 2 3 2 none none 4 4and cattle

5 hay-silage 5 cattle only 4 4 4 little little 4 4

6 free range poultry 5 poultry 2 2 2 reduced unknown 2 3production

7 filters on buildings 3 force vented

4 1 1 none helps! 1 3

bdgs only8 broilers without litter 4 broilers only 3 1 1 reduces 2 1

production helps!9 enhanced milk yield 4 cattle 2 4 2 should help should help 3 210 encourage burning

poultry5 poultry 4

manure in power stas.11 ban splash plates and 5 wide 3 4 3 none positive 3 4

rain guns 112 tethering animals 5 lowish 3 3 none none 2 1

Notes:Quality of rating B D E F G H Ja. very reliable 5: in use 4. High 4. >50 % 4. Substantial 4. Easy 4. Inexpensive 4. Acceptable

4. Technology available 3. 30-50 %e. not reliable 3. R&D conducted 1. Low 2. 10-30 % 1. Insignificant 1. Difficult 1.

Expensive1. Strong

2. R conducted, D required 1. <10 % opposition

79

Projecttitle



Annex E: Feasible options for mitigating PM10 emissions from agriculture (ii) arable

Options for mitigating PM10 emissions from agricultureArable farming

A B C D E F G H I Joption description availability applicability magnitude effectivenes

soverall ease of cost of other "knock-on" acceptability

of of (reduction in

magnitude

on-farm on-farm interactions and farmer public


implemen-

on-farm effects

emissions reduction tation tation production

environment

1 particle trap on exhaust 4 2 80 % 4 3 3 3 4

2 dust collection on 1 all 4 90 % high 3 – 4 3 - 3 3 4harvesters

3 LPG/CNG fuelled 3 all 3 high 2 – 3 2 - 5 2 4tractors / generators reduce gases

4 emissions test for 4 all 4 2 4 1 4tractors reduce

gases5 alternative cereal 3 - 4 cereals 4 variable variable 1 1 change 5 2 3

harvesting (stripping grain) light soils? system?6 low-till farming 4 2 variable 4 3 - 4 4 less 4? 3 3

labour7 redirection of dust 1 - 2 all 4 low / mod mod 3 – 4 4 - 3 4 4

from harvesters8 wind breaks 5 light soil all 5 4 minimal 4 4 5

any crop9 minimum soil-blow 4 some soils

farming methods and crops 2 3 - 4 reduce? 5 3 510 no burning of any crop 5 oils+alt

cropssmall high small 2 – 3 2 - 3 problems? 4 1 4

11 crop covers (poly) 5 high value crops

small 1 1 1 – 2 1 yield up disposal 1

1 1

Notes:Quality of rating B D E F G H I Ja. very reliable 5: in use 4. High 4. >50 % 4. Substantial 4. Easy 4.Inexpens

ive4. Benefit 4. Acceptable

4. Technology available 3. 30-50 % 3. neutrale. not reliable 3. R&D conducted 1. Low 2. 10-30 % 1. Insignificant 1. Difficult 1.

Expensive2. detriment

1. Strong

2. R conducted, D required 1. <10 % opposition1. R required

80

Projecttitle



Annex E: Feasible options for mitigating PM10 emissions from agriculture (iii) other sources

Options for mitigating PM10 emissions from agricultureOther sources

A B C D E F G H I Joption description availabilit

yapplicability magnitud

eeffectiveness

overall ease of cost of other "knock-on" acceptability

of of (reduction in

magnitude on-farm on-farm interactions and farmer public


implemen-

on-farm effects

emissions

reduction tation tation production

environment

1 dust suppressant on 3 2 2 3 2 - 3 3 3 3 3unpaved tracks

2 vegetation to 1 1 - 2 3 2 - 3 4 4 4 4capture dust

3 control on-farm 4 high 3 2 - 3 2 - 3 2 4 3 4incineration

4 filter off gases from 4 high 50 % 3 2 3 3 3 3grain dryers

5 greater control of dust 1 ? ? ? ? 3 3 4 3from animal feed prep

6 composting in-vessel 5 3 ? 4 3 4(mushroom)

7 washing down at farms

5 1 2 3 1 - 2 3 2

8 anaerobic disgestion 5 slurries 1 1 1 3 4 3 3as E source

9 wind power as E 5 safe ? 1 1 - 2 4 1 1 - 2source dependent

10

Notes:Quality of rating B D E F G H Ja. very reliable 5: in use 4. High 4. >50 % 4. Substantial 4. Easy 4. Inexpensive 4. Acceptable

4. Technology available 3. 30-50 %e. not reliable 3. R&D conducted 1. Low 2. 10-30 % 1. Insignificant 1. Difficult 1. Expensive 1. Strong

2. R conducted, D required

1. <10 % opposition

1. R required

81

Projecttitle



Annex F: Other relevant studies in progress

F.1. Bioaerosol-related emission inventories for livestock buildings

Project Leader: Jens. Seedorf,Institute of Animal Hygiene and Animal Welfare,Hannover School of Veterinary Medicine,Bnteweg 17 p,30559 Hannover,Germany

Funding body: The state of Lower Saxony, Germany

Type of project: Experiments, data analysis and dispersion modelling.

Publications: Seedorf, J, 1997Human health and dust related endotoxin concentrations in livestock buildings. In Voermans J A M and G J Monteny, eds "Ammonia and odour emissions from animal production facilities", Vol II, pp 733-740. Published by NVTL 5240 AB Rosmalen, Netherlands.

Hartung, J and J Seedorf, 1999Characterisation of airborne dust in livestock housing and its effects on animal and environment. Proceedings of the International Symposium on Dust Control in Animal Production Facilities, Aarhus, Denmark, 30 May - 2 June 1999, pp 140-153.

Seedorf, J and J Hartung, 1999Measured and calculated bacteria concentrations in the vicinity of a duck fattening unit.Proceedings of the International Symposium on Dust Control in Animal Production Facilities, Aarhus, Denmark, 30 May - 2 June 1999, pp 179-185

F.2. Particulate matter emissions from industrial-scale pig houses

Project Leader: Cary Secrest,US Environmental Protection Agency,Ariel Rios Building, Room 2119,1200 Pennsylvania Ave NW,Washington DC 20460,USA.

Funding Body: US Government

Type of Project: Enforcement of the US Clean Air Act

82

Projecttitle



F.3. Estimation of fine dust emissions

Project Leader: Pieter van der MostInspectorate for Environmental Protection,Postbus 20945,2500 GX Den Haag,Netherlands

Funding Body: Netherlands Government

Type of Project: Drafting a chapter for the Joint EMEP/CORINAIR Atmospheric Emission Inventory Guidebook

(The UNECE Agriculture Panel, under its chairman Prof. U. Dmmgen, Institute of Agro-ecology, FAL, Braunschweig, Germany, is planning to write a sub-chapter of the guidebook to summarise the current state of international knowledge on emissions of agricultural particulates.)

83

Projecttitle



Annex G: Full list of abatement ideas (before any assessment for feasibility) from brainstorm held at SRI on 8 May 2000

G.1. Livestock

Phase out unpelleted feed Liquid feed to be promoted Coating of pelleted feed Oil spraying in building Good care of bedding and feed Eliminate bedding Bedding designed for low dust (cardboard, rubber, shredded newspaper) Ditto feed Preservatives to help feed quality Go from hay to silage Care in making hay "Regulation" of hay making Free range poultry (no dust baths) No "fines" Bio-filters/bio-scrubbers optimised for the duty of reducing dust Low-dust ways to remove muck (broilers and layers) Keep broilers without litter No livestock in UK More productive ) livestock

Fewer ) Keep all surfaces wet Electrostatic collection of dust More animals outdoor Outdoor pigs on heavy soils only Encourage combustion of poultry manure in power stations Distribution of bedding without blowing Wrapping of bales Ban splash plates and rain guns More fish in the human diet, less cereal, fewer mammals or poultry Avoid overgrazing Avoid rooting by outdoor pigs Keep pigs or poultry in the forest Featherless poultry Hard-wearing feathers Tethered animals

G.2. Arable

Electric tractors Gas powered tractors M.O.T. type tests for tractors and diesel generators Particle traps on exhausts from diesel engines and generators Gantries to replace tractors Whole crop harvesting (to allow max processing off field) Only harvest under low dust conditions

84

Projecttitle



Only cultivate under low dust conditions "No-till" farming No farming at all Dust collection on the combine Dust redirection on the combine Low-dust ways to apply lime or FYM or poultry muck Electrostatic collection of dust Modify farming to minimise moulds (with or without G.M. techniques) Harvest earlier Irrigate to control dust Wind breaks No burning of any stubble Minimum period of bare soil Tufts of straw, strip cropping, mulches (to increase roughness) Stripper harvesting Select crops for low dust production Ban rain Greenhouses Covered crops – include outdoor row crops Use GM to reduce pollen amounts Cover crops Burn straw in power stations Burn straw under controlled conditions Conservation headlands Switch from cereals to (organic) orchards or low-dust crops Good management of spray application Eliminate use of non-pellet fertilisers, lime etc Go to larger farms (which have, proportionately, less boundary) Encourage Set-Aside Encourage traditional mixed farming Have Dust-Sensitive Areas (like NSA) Add dust abatement kit to potato lifters, root crop harvesters and all other types of harvester Go to smaller farms (which have proportionately, more length of wind-break) Do dusty operations in middle of farm only Encourage farmers' markets etc, to reduce wastage Don't grow legumes ) Harvest all legumes green ) Because conventional combining of legumes is particularly dusty.

G.3. Other sources

Reduce unpaved-road emissions from farms (cf Good Quarry Practice) Spray old engine oil, to bind dust on unpaved roads Capture dust by vegetation (strategically placed, esp. on farm boundary) Strong ban on on-farm incineration Abatement measures to be applied more widely to grain dryers Ditto on-farm feed milling Covering of waste or produce stores More regulation of cleaning-out of grain stores + disposal of residues Redesign farm to minimise handling of products All farming to be self-pick

85

Projecttitle



Low-dust flails and chippers Raise farmers' awareness of health risks of dusts General washing-down of farms Green Label type scheme for farm equipment (cf NH3 in Netherlands) Encourage best practice esp. for loading and unloading (local extracts) Grain storage in ways to minimise moulds e.g. drying Solar dryers Promote anaerobic digestion with Combined Heat and Power Mushroom compost making – Compost in-vessel Compost disposal – Make compost tea Mushroom growing – More robotics Segregate farming from the general residents More use of hydroponics Encourage use of rice Grant schemes which encourage low-dust farming Check dust arisings from hops Persuade major retailers to encourage low-dust farming Replace diesel-based power with e.g. wind Forestry – low dust harvesting of timber products

– choice of timber type

86

Projecttitle



Annex H: Full reference list

Airborne Particles Expert Group, 1999Source apportionment of airborne particulate matter in the United KingdomReport prepared on behalf of the Department of the Environment, Transport and the Regions, the Welsh Office, the Scottish Office and the Department of the Environment (Northern Ireland) January 1999, 158 pp

Atiemo, MA, et al, 1978Dust measurements in tractor and combine cabsAmerican Society of Agricultural Engineers, Technical paper no. 78-3019Paper for presentation at the 1978 Summer Meeting of the ASAE, Utah, June 1978

Batel, W, 1975Messungen zur Staub-, Lärm- und Geruchsbelastung an Arbeitsplätzen in der landwirtschaftlichen Produktion und Wege zur Entlastung – erster BerichtGrundlagen der Landtechnik Vol 25. No. 5, pp. 135-157(Measurement of incidence of dust, noise and odour at workplaces in agriculture and ways of reducing it. Initial report: SRI Translation No. 409)

Batel, W, 1979Staubbelastung und Staubzusammensetzung an Arbeitsplätzen der landwirtschaftlichen Produktion und daraus abzuleitende Belastungsgrenzen und StaubschutzmassnahmenGrundlagen der Landtechnik Vol 29. No. 2, pp. 41-54(Dust load and dust composition in workplaces in agriculture and load limits and dust protection measures derived therefrom: SRI Translation No. 467)

Bauer, MA, and DP Coppolo, 1993Agricultural lung disease: preventionSeminars in Respiratory Medicine Vol. 14, No. 1, pp.83-89

Berdowski, JJM, et al., Feb 1997Particulate matter emissions (PM10 – PM2.5 – PM0.1) in Europe in 1990 and 1993TNO report TNO-MEP – R 96/472TNO Institute of Environmental Sciences, Energy Research and Process Innovation, Apeldoorn, The Netherlands

Bottcher, R. W. et al., 1999Windbreak walls and wet pad scrubbers for reducing odorous dust emissions from tunnel ventilated swine buildings.Proceedings of the International symposium on dust control in animal production facilities, Aarhus, Denmark, 30 May-2 June, pp 186-193

Boutin, P, Torre, M, Serceau, R and Rideau, P-J, 1998Atmospheric bacterial contamination from landspreading of animal wastes: evaluation of the respiratory risk for people nearby. Journal of Agricultural Engineering Research. 39, 149-160

Burg, WR, et al, 1982Measurement of airborne aflatoxins during the handling of 1979 contaminated cornAmerican Industrial Hygiene Journal Vol. 43, No. 8, pp. 580-586

87

Projecttitle



Burt, P, B Fisher and P Sharma NRI/University of Greenwich, Chatham, Kent.

Caulton, E, Director of the Scottish Centre for Pollen Studies, Napier University, Edinburgh

Choudat, D, Goehen, M, Korobaeff, M, Boulet, A, Dewitte, JD, and Martin, MH, 1994Respiratory symptoms and bronchial reactivity among pig and dairy farmers.Scandinavian Journal of Work and Environmental Health. 20, 48-54

Claiborn, C, et al, 1995Evaluation of PM10 emission rates from paved and unpaved roads using tracer techniquesAtmospheric Environment Vol. 29, No. 10, pp. 1075-1089

Clarke, A, 1987Air hygiene and equine respiratory diseaseIn Practice, Nov, pp 196-204

Clarke, A, et al, 1987The relationship of air hygiene in stables to lower airway disease and pharyngeal lymphoid hyperplasia in two groups of thoroughbred horsesEquine Veterinary Journal Vol. 19 No.6 pp 524-530

Clausnitzer, H, and MJ Singer, 1996Respirable-dust production from agricultural operations in the Sacramento Valley, CaliforniaJournal of Environmental Quality Vol. 25, pp. 877-884

Clausnitzer, H, and MJ Singer, 2000Environmental influences on respirable dust production from agricultural operations in CaliforniaAtmospheric Environment, Vol. 34, pp. 1739-1745

Committee on the medical effects of air pollutants, 1988Quantification of the effects of air pollution on health in the United Kingdom.The Stationery Office, London. 78 pp

Cox, CS and Wathes, CM (Eds), 1995Bioaerosols Handbook. CRC Press, Boca Raton, Florida. 623 pp

Cravens, RL, et al, 1981Characterisation of the aerosol in turkey rearing confinementsAmerican Industrial Hygiene Association Journal Vol. 42 no.4 pp 315-318

Crook, B, and J Lacey, 1991Airborne allergenic mcroorganisms associated with mushroom cultivationGrana Vol. 30, pp. 446-449

Crook, B and Olenchock, SA, 1995Industrial workplaces.In 'Bioaerosols Handbook'. Eds C. S. Cox and C. M. Wathes. CRC Press, Boca Raton. 531-546

88

Projecttitle



Crook, B, et al, 1991Airborne dusts, ammonia, micro-organisms and antigens in pig confinement houses and the respiratory health of exposed farm workersAmerican Industrial Hygiene Association Journal, Vol. 52, No. 7, pp 271-279

Cuthbert, OD, and IG Jeffrey, 1993Barn allergy: an allergic respiratory disease of farmersSeminars in Respiratory Medicine Vol. 14 no. 1 pp. 73-82 (Jan)

D'Amato, G, Spieksma, FThM., Liccard, G, Jger, S, Russo, M, Kontou-fili, K, Nikkels, H, Wthrich, B and Bonini, S, 1998Pollen-related allergy in Europe.Allergy. 53, 567-578

Dagnall, S, et al, 2000Source mapping and analysis of farm livestock manures – assessing the opportunities for biomass-to-energy schemesBioresource Technology Vol. 71, pp. 225-234

Danish Institute of Agricultural Sciences, 1999Proceedings of the International Symposium on Dust Control in Animal Production Facilities; held at Aarhus, Denmark, 30 May-2 June; 349 ppResearch Centre Bygholm, 8700 Horsens, Denmark

Darke, CS, et al, 1976Respiratory disease of agricultural workers harvesting grainThorax Vol. 31, pp. 294-302

Department of the Environment, Transport and the Regions, 1999Digest of United Kingdom Energy Statistics 1999Government Statistical Service, London

Department of the Environment, Transport and the Regions, 1999UK National Atmospheric Emissions Inventory, 1999

Department of Trade and Industry, 1999Digest of UK Energy Statistics 1999

Déportes, I, et al, 1995Hazard to man and the environment posed by the use of urban waste compost: a reviewThe Science of the Total Environment Vol. 172, pp. 197-222

Dobson and Horrell Ltd., Feed specialists (for horses), NorthamptonshireProduct information and personal communication (2000)

Donaldson, AI, 1983Quantitative data on airborne foot-and-mouth disease virus: its production, carriage and deposition.Philosophical Transactions of the Royal Society, London, B, 302, 529-534

89

Projecttitle



Donham, KJ, 1987Human health and safety for workers in livestock housing.In "Latest developments in livestock housing". Proceedings of Commission Internationale du Gnie Rural, Section 2 Seminar. June 22-26 1987, Illinois, 86-95

Donham, KJ et al, 1986aCharacterisation of dusts collected from swine confinement buildingsAmerican Industrial Hygiene Association Journal Vol. 47 no.7 pp. 404-410

Donham, KJ et al, 1986bCharacterisation of dusts collected from swine confinement buildingsAmerican Journal of Industrial Medicine Vol. 10 pp. 294-297

Donham, K, et al, 1989Environmental and health studies of workers in Swedish swine buildingsin: Principles of Health and Safety in Agriculture, 1989, pp 66-68ed. Dosman, JA, and DW Cockcroft, CRC Press Inc., Boca Raton, Florida, USA

Donham, K, Reynolds, S, Whitten, P, Merchant, J, Burmeister, L, Popendorf, W, 1995Respiratory dysfunction in swine production facility workers: dose-response relationships of environmental exposures and pulmonary function.American Journal of Industrial Medicine. 27, 405-418

Donham, KJ, Cumro, D, Reynolds, SJ and Merchant, JA, 2000Dose-response relationships between occupational aerosol exposures and cross-shift declines of lung function in poultry workers: Recommendations for exposure limits.Journal of Occupational and Environmental Medicine. 42, 260-269

Dosman, JA, and DJ Cotton (eds), 1980Occupational Pulmonary Disease. Focus on Grain Dust and Health (Proceedings of the International Symposium on grain dust and health held in Saskatoon, Canada, November 7-9, 1977)Academic Press, New York

Dutkiewicz, J, 1997Bacteria and fungi in organic dust as potential health hazards.Annals of Agricultural and Environmental Medicine. 4, 11-16

Emberlin, J, 1995Interaction between air pollutants and aeroallergensClinical Experimental Allergy, Vol. 25, Supplement 3, pp. 33-39

Emberlin, J, 2000UK Pollen Network internet site: http://pollenuk.worc.ac.uk/

Eversmeyer, MG, and CL Kramer, 1987Vertical concentrations of fungal spores above wheat fieldsGrana Vol. 26, pp. 97-102

90

Projecttitle



Gaffney, PCalifornia Air Resources Board2020 L Street, Sacramento, California 95814, USAhttp://arbis.arb.ca.gov/emisinv/pmnh3/pmnh3.htm

Ganzelmeier, H, 1990Mit neuer Technik – Abdrift vermeidenDLG-Mitteilungen, No. 8, pp. 384-388(New technology to prevent drift: SRI Translation 49 (New Series))

Gillies, JA, et al, 1999Long-term efficiencies of dust suppressants to reduce PM10 emissions from unpaved roadsJournal of the Air and Waste Management Association Vol. 49 pp. 3-16 (Jan)

Gloster, J, Blackall, RM, Sellers, RF and Donaldson, AI, 1981Forecasting the airborne spread of foot-and-mouth disease.Veterinary Record, 108, 370-374

Gregory, PH, 1973Microbiology of the atmosphereLeonard Hill Books, Aylesbury, Bucks.

Griffith, G, University of Aberystwyth

Hardwick, NV, JE Slough and DR Jones, 1998aWinter barley: a survey of diseasesA survey to determine the severity of winter barley diseases in England and WalesADAS/CSL Research and Development

Hardwick, NV, JE Slough and DR Jones, 1998bWinter wheat: a survey of diseasesA survey to determine the severity of winter wheat diseases in England and WalesADAS/CSL Research and Development

Health and Safety Executive, 1993The occupational zoonoses.Health and Safety Executive, Sheffield, 32pp

Health and Safety Executive, 1997Occupational exposure-limits.Health and Safety Executive, London. EH40.97

Health and Safety Executive, 1999Mainstream research projects handbook.Health and Safety Executive, Sheffield, 114pp

Heber, AJ, and CR Martin, 1988Effect of additives on aerodynamic segregation of dust from swine feedTransactions of the American Society of Agricultural Engineers Vol. 31, No. 2, pp. 558-563 (March-April)

91

Projecttitle



Heber, AJ et al, 1988Size distribution and identification of aerial dust particles in swine finishing buildingsTransactions of the American Society of Agricultural Engineers Vol. 31 No. 3 pp 882-887 (May-June)

Jeger, MJ, professor of plant pathology, Imperial College at Wye, pers. comm., 2000

Jones, W et al, 1984Environmental study of poultry confinement buildingsAmerican Industrial Hygiene Association Journal Vol. 45 no.11 pp 760-766

Karlsson, K and P Malmberg, 1989Characterisation of exposure to moulds and actinomycetes in agricultural dusts by scanning electron microscopy, fluorescence microscopy and the culture methodScandinavian Journal of Work, Environment and Health Vol. 15 pp 353-359

Kilpelinen, M, Terho, EO, Helenius, H and Kosuenvuo, M, 2000Farm environment in childhood prevents the development of allergies.Clinical and Experimental Allergy, 30, 201-208

Kothary, MH, et al, 1984Levels of Aspergillus fumigatus in air and in compost at a sewage sludge composting siteEnvironmental Pollution (Series A) Vol. 34, pp. 1-14

Kotimaa, MH, et al, 1991Feeding and bedding materials as sources of microbial exposure on dairy farmsScandinavian Journal of Work, Environment and Health Vol. 17 pp 117-122

Kotimaa, MH, et al, 1987Airborne moulds and actinomycetes in the work environment of farmersEuropean Journal of Respiratory Diseases Vol. 71 (suppl) no.152 pp 91-100

Lacey, J and ME Lacey, 1964Spore concentrations in the air of farm buildingsTransactions of the British Mycological Society, Vol. 47, No. 4, pp. 547-552

Lenhart and SA Olenchock, 1984Sources of respiratory insult in the poultry processing industryAm J Ind Med 6: 89-96

Leistikow, B, et al, 1989Respiratory risks in poultry farmersin: Principles of Health and Safety in Agriculture, 1989, pp 62 - 65ed. Dosman, JA, and DW Cockcroft, CRC Press Inc., Boca Raton, Florida, USA

Louhelainen, K, et al, 1987aTotal concentrations of dust in the air during farm workEuropean Journal of Respiratory Diseases Vol. 71 (suppl) no.152 pp 73-79

Louhelainen, K, et al, 1987bDust exposure in piggeriesEuropean Journal of Respiratory Diseases Vol. 71 (suppl) no.152 pp 80-90

92

Projecttitle



Maas, G, and G Krasel, 1988Direkte Abdrift von Herbiziden bei Verwendung verschiedener Düsentypen und ZusatzstoffeZeitschrift für Pflanzenkrankheiten und Pflanzenschutz, Sonderheft XI, pp 241-247(Effects of nozzle type and additives on direct drift of herbicides: SRI Translation No. 47 (New Series))

Malmberg, P, and A Rask-Andersen, 1993Organic dust toxic syndromeSeminars in Respiratory Medicine Vol. 14, No. 1, pp. 38-48 (Jan)

Marksway Horsehage and Mollichaff, Manufacturers of horse feed, Paignton, Devon. Product information and personal communication (from D. Lucas), 2000

May, JJ, et al, 1989A study of dust generated during silo opening and its physiological effects on workersin: Principles of Health and Safety in Agriculture, pp 76-79ed. Dosman, JA, and DW Cockcroft, CRC Press Inc., Boca Raton, Florida, USA

McCartney, H. A. and M. E. Lacey, 1991Wind dispersal of pollen from crops of oilseed rape (Brassica Napus L.).Journal of Aerosol Science Vol. 22, No. 4, pp 467-477

McDuffie, HH, et al, 1991Respiratory health status of 3098 Canadian grain workers studied longitudinallyAmerican Journal of Industrial Medicine Vol. 20, pp. 753-762

McSharry, C., 1992New aeroallergens in agricultural and related practice. Editorial.Clinical and Experimental Allergy, 22, 423-426.

Mercer, DR, 1993Estimates of the number and types of poultry housing in use in England and Wales.Report to MAFF, ADAS Nottingham.

Miguel, AG, et al, 1999Allergens in paved road dust and airborne particlesEnvironmental Science and Technology Vol. 33, No. 23, pp. 4159-4168

Miller, PCH, and PJ Walklate, Silsoe Research Institute, Bedfordshire, pers. comm., 1999

Millner, P.D., Olenchock, S.A., Epstein, E., Rylander, R., Haines, J., Walker, J., Ooi, B.L., Horne, E. and Maritato, M., 1994Bioaerosols associated with composting facilities.Compost Science and Utilization, 2, 4-57

Ministry of Agriculture, Fisheries and Food, 1998Controlling Soil Erosion (Booklet PB 3280)

Ministry of Agriculture, Fisheries and Food, and Welsh Office Agriculture Department, 1998Code of Good Agricultural Practice for the Protection of Air. Second Edition

93

Projecttitle



Ministry of Agriculture, Fisheries and Food, and Welsh Office Agriculture Department, 1998Code of Good Agricultural Practice for the Protection of Soil. Second Edition

Ministry of Agriculture, Fisheries and Food, and Welsh Office Agriculture Department, 1998Code of Good Agricultural Practice for the Protection of Water. Second Edition

Muir, D.M., ed., 1992Dust and fume control – a user guide (second edition)Institution of Chemical Engineers, Rugby, UK, 162 pp

Nieuwenhuijsen, MJ, and MB Schenker, 1998Determinants of personal dust exposure during field crop operations in California agricultureAmerican Industrial Hygiene Association Journal Vol. 59, pp. 9-13

Nieuwenhuijsen, MJ, et al, 1998Exposure to dust and its particle size distribution in California agricultureAmerican Industrial Hygiene Association Journal Vol. 58, pp. 34-38

Nieuwenhuijsen, MJ, et al, 1999Personal exposure to dust, endotoxin and crystalline silica in California agricultureAnnals of Occupational Hygiene Vol. 43, No. 1, pp. 35-42

Norén, O, 1985Dust concentrations during operations with farm machinesAmerican Society of Agricultural Engineers, paper no. 85-1055Paper for presentation at the 1985 summer meeting of the ASAE, Michigan State University, East Lansing, June 23-26

O'Hara, SL, et al, 2000Exposure to airborne dust contaminated with pesticide in the Aral Sea RegionThe Lancet Vol. 355, pp. 627-628 (Feb 19th)

Olenchock, SA, et al, 1990Presence of endotoxins in different agricultural environmentsAmerican Journal of Industrial Medicine. Vol. 18, pp. 279 – 284

Ortiz de Zarate, I, et al, 2000-07-21 Emission factor estimates of cereal wastes burning in Spain. Atmospheric Environment vol 34, pp 3183-3193

Parbst, K. E., 1998Evaluation of particulate removal methods for controlling odor emissions from swine housing.Unpublished M.S. thesisNorth Carolina State University, Dept. of Biological and Agricultural Engineering, Raleigh, NC, USA, 137 pp

Perkins, SL, et al, 1997Effect of sow and piglet activity on respirable particle concentrationsTransactions of the American Society of Agricultural Engineers Vol. 13, No. 4 pp. 537-539

Phillips, VR, et al, 1998The development of robust methods for measuring concentrations and emission rates of gaseous and particulate air pollutants in livestock buildings

94

Projecttitle



Journal of Agricultural Engineering Research Vol. 70, pp 11-24

Preller, L., Heederik, D., Boleij, J.S.M., Vogelzang, P.F.J. and Tielen, M.J.M., 1995Lung function and chronic respiratory symptoms of pig farmers: focus on exposure to endotoxins and ammonia and use of disinfectants.Occupational and Environmental Medicine, 52, 654-660

Quality of Urban Air Review Group, 1996Airborne particulate matter in the United Kingdom (Third report), May 1996Prepared at the request of the Department of the Environment

Rantio-Lehtimki, A., 1995Aerobiology of pollen and pollen antigens.In 'Bioaerosols Handbook'. Eds C. S. Cox and C. M. Wathes. CRC Press, Boca Raton, 387-406

Roberts, C., Centre for Equine Studies, Animal Health Trust, Newmarket; pers. comm., 2000

Roberts, C., I. Sbai, S. Vandeput, T. Art and P. Lekoux, 1999The use of cardboard bedding (Ecobed™) as part of a minimum-dust stable regime for horses with COPDJournal of Equine Science, Vol. 19, pp. 580. (Abstract only)

Robertson, JF, 1993Exposure to inspirable dust and crystalline quartz during potato handling processesFarm Building Progress Vol. 114, (Oct) pp. 7-11

Rogge, WF, et al, 1993Sources of fine organic aerosol. 3. Road dust, tire debris and organometallic brake lining dust: roads as sources and sinks. Environmental Science and Technology Vol. 27, No. 9, pp 1892-1904

Sastre, J, et al, 1990Respiratory and immunological reactions among Shiitake (Lentinus edodes) mushroom workersClinical and Experimental Allergy Vol. 20, pp. 13-19

Seaman, M, and P Rush, Ecobed Ltd, Norfolk. Personal communication, 2000.

Seaton, A., 1999Airborne particles and their effects on health.In 'Particulate matter: properties and effects upon health'. Eds R. L. Maynard and C. V. Howard. BIOS Scientific Publishers Ltd, Oxford, 9-17

Senthilselvan, A., Dosman, J.A., Kirychuk, S.P., Barber, E.M., Rhodes, C.S., Zhang, Y., and Hurst, T.S., 1997Accelerated lung function decline in swine confinement workers.Chest, 111, 1733-1741

Shaw, BW, et al, 1997Emission factors for grain receiving and feed loading operations at feed millsTransactions of the American Society of Agricultural Engineers Vol. 41, No. 3, pp. 757-765

Sheppard, A, 1998The structure of pig production in England and Wales. Results of the National Survey of Pig Production Systems.

95

Projecttitle



Special studies in agricultural economics Report No. 40, University of Exeter

Simcox, NJ, et al, 1995Pesticides in household dust and soil: exposure pathways for children of agricultural familiesEnvironmental Health Perspectives Vol. 103, No. 12, (Dec) pp. 1126-1134

Sneath, RW, and VR Phillips, Silsoe Research Institute, Bedfordshire, pers. comm., 2000

Stahuljak-Beritic, D., Dimov, D., Butkovic, D., and Stilinovic, L.,1977Lung function and immunological changes in poultry breeders.International Archives of Occupational and Environmental Health, 40, 131-139

Strachan, D.P., 1995Epidemiology of hay fever: towards a community diagnosis.Clinical and Experimental Allergy, 25, 296-303

Takai, H, et al, 1998Concentrations and emissions of airborne dust in livestock buildings in northern EuropeJournal of Agricultural Engineering Research Vol. 70, pp 59-77

Taylor, J, Chairman of The Bracken Advisory Commission (Aberystwyth)Bracken Advisory Commission Mission Statement, July, 1998

Thu, K.M., Donham, K.J., Ziegenhorn, R. and Reynolds, S., 1997A control study of the physical and mental health of residents living near a large scale swine operation.Journal of Agricultural Safety and Health, 3, 13-26

United States Environment Protection Agency (USEPA), Office of Air Quality, Planning and StandardsAP-42, Fifth Edition, Volume I, CHAPTER 9 Food and Agricultural IndustriesAvailable at: http://www.epa.gov/ttn/chief/ap42c9.html

Van den Bogart, HGG, et al, 1993Mushroom worker's lung: serological reactions to thermophilic actinomycetes present in the air of compost tunnelsMycopathologia Vol. 122, pp. 21-28

Vandeput, S, et al, 1997Airborne dust and aeroallergen concentrations in different sources of feed and bedding for horsesVeterinary Quarterly. Vol. 19, pp. 154-158

Vinken, W and P Roels, 1984Hypersensitivity pneumonitis due to Apergillus fumigatus in compostThorax Vol. 39, pp. 74-75

Visschedijk, AJH et al, 1997Abatement efficiencies and technologies for controlled particulate matter emissions in Europe. TNO-MEP Report R96/473

Warner, F, 1998Fungal Spores section of Pollen UK website: http://pollenuk.worc.ac.uk/aero/fungi/fungi.htm#top

96

http://pollenuk.worc.ac.uk/aero/fungi/fungi.htm#top

Projecttitle



Wathes, C.M., 1995Bioaerosols in animal houses.In 'Bioaerosols Handbook' Eds C. S. Cox and C. M. Wathes. CRC Press, Boca Raton, 547-578

Wathes, CM, et al, 1991Air hygiene in a pullet house: effects of air filtration on aerial pollutants measured in vivo and in vitroBritish Poultry Science Vol. 32 pp. 31-46

Wathes, CM, et al, 1997Concentrations and emission rates of ammonia, nitrous oxide, methane, carbon dioxide, dust and endotoxin in UK broiler and layer housesBritish Poultry Science Vol. 38, pp 14-28

Wathes, CM, et al, 1998Emissions of aerial pollutants in livestock buildings in northern Europe: overview of a multinational projectJournal of Agricultural Engineering Research Vol. 70, pp 3-9

Weber, S et al, 1993Organic dust exposures from compost handling: case presentation and respiratory exposure assessment. American Journal of Industrial Medicine Vol. 24, pp. 365-374Whyte, R.T., 1993Aerial pollutants and the health of poultry farmers.World's Poultry Science Journal, 49, 139-156

Whyte, RT, 2000Operative exposure in systems using litterPaper presented at the Spring Conference of the International Egg Commission, London, 3 April

Williams, M., 1999Air quality strategies with respect to particles.In 'Particulate matter: properties and effects upon health'. Eds R. L. Maynard and C. V. Howard. BIOS Scientific Publishers Ltd, Oxford, 159-174

Yoshida, K and J Maybank, 1980Physical and environmental characteristics of grain dust.In Dosman, JA, and DJ Cotton (eds) Occupational Pulmonary Disease. Focus on Grain Dust and Health, (Proceedings of the International Symposium on grain dust and health held in Saskatoon, Canada, November 7-9, 1977) pp. 441-461Academic Press, New York (1980)

Zhang, Y., 1999Engineering control of dust in animal facilities.Proceedings of the International symposium on dust control in animal production facilities, Aarhus, Denmark, 30 May-2 June, pp 22-29

Please press enter

97

research and development - gov.ukrandd.defra.gov.uk/document.aspx?document=wa0802_1… · web...

Documents