EC/FAO Programme on Information Systems to Improve Food Security Decision-Making in the European Neighborhood Policy (ENP) East Area

Proposal to improve crop forecasting in Armenia

Workshop, October 13, 2010 Armstatehydromet Service, Yerevan, Armenia

- WORSHOP REPORT -

1. Background

This half day Workshop was jointly organized by Armstatehydromet (hereinafter: Hydromet) and FAO under the “EC/FAO Programme on Information Systems to Improve Food Security Decision-Making in the European Neighborhood Policy (ENP) East Area” on October 13, 2010. The Programme is financed by the European Commission and implemented by FAO. The Programme aims at improving food security by enhancing the national capacity to generate, analyze, communicate and mainstream more relevant and reliable information into policies and programmes. The Workshop took place in Hydromet’s meeting room.

2. Workshop Objectives

The objective of the workshop was to build consensus on the priorities to be addressed under the Programme to improve crop forecasting and the dissemination of information. This was achieved by presenting and discussing with stakeholders the options identified by an international consultant to improve crop forecasting and the dissemination of information to Marz Regional Centers. The Workshop also provided an opportunity to present the potential of improved crop forecasting for food security policies and programmes to concerned staff from the Ministry of Agriculture (MoA). The Workshop was organized for the stakeholders of the Programme who are interested in agro-meteorological forecasting, i.e. producers and users of agro-meteorological data, namely staff from the Ministry of Agriculture, Hydromet and the National Statistical Service (NSS). The Workshop was designed in such a way as to engage all participants in discussions and reflect on appropriate recommendations for all state institutions involved in crop forecasting activities in Armenia. The international consultant, Bernard Tychon, was on a mission in Armenia for 10 days, from 6 to 15 October 2010. His tasks were the following:

Review current activities related to agrometeorological forecasting and operational service in Hydromet and the dissemination of information (including through information products) to MoA;

2 

Identify weaknesses and limitations of current agro-meteorological forecasting and dissemination of information;

Propose options for addressing users’ demand and needs, taking into account the resources available under the Programme to improve capacity of Hydromet and MoA;

Organize a Workshop to present and discuss with Hydromet and the users of agrometeorological information the options identified to address users’ demand and needs;

Finalize a proposal to provide support under the Programme with different options, integrating the comments received during the Workshop, with a budget and an agenda for implementation.

The agenda of the Workshop was established in consultation with the international consultant and Hydromet staff. It is provided in Annex 1. All participants were provided with a package of documents:

list of participants

agenda

brief description of the EC/FAO Programme

hard copy of the presentation delivered by Nelly Arakelyan (HYROMET)

hard copy of the presentation delivered by Bernard Tychon

agrometeorological bulletin (produced by Hydromet)

3. Participation

The Workshop was attended by 26 participants, including the international consultant and the Country Coordinator. Participants belong to a number of national institutions: Ministry of Agriculture (represented by the Crop Production and Agricultural Planning Departments), National Statistical Service (Food Security Unit), Hydromet and the Armenian state Agricultural Academy. The list of Workshop participants is presented in Annex 2.

4. Process

The Workshop was opened by a Welcoming speech of the Hydromet Director (presented by his Deputy, H. Melkonyan). He noted that the objectives of the Programme were similar to the activities of their Service. He also underlined the importance of the improvement of crop forecasting in Armenia.

The first presentation was devoted to the EC/FAO Programme, its objectives, areas of intervention, main activities and expected results (M. Tapaltsyan, Country Coordinator). The Country Coordinator also thanked the Director of Hydromet for all the support provided to organize the Workshop.

The next presentation on the use of agrometeorological data and delivery of information to the users was delivered by N. Arakelyan from Agrometerological Forecasting Division. The presentation focused on the observation, products, problems and proposals on improvement of agromet forecasting. She presented the activities of their Division, the list of users and the ways of delivering the information to the users.

3 

The last presentation on options to improve crop forecasting in Armenia was made by B. Tychon. The presentation comprised 3 main parts: introduction to operational crop yield forecasting techniques (a new crop-forecasting system to be adjusted to Armenia, present status of the crop yield forecast in Armenia and proposed improvements).

The presentation of the international consultant is provided in Annex 3.

Each presentation was followed by an active discussion among Participants. The most active discussion followed the presentation of B. Tychon.

5. Main results and follow-up activities The main options proposed by the consultant to improve crop forecasting and the dissemination of the information were the following:

Organize an Interinstitutional Working Group, including MoA, Hydromet and NSS representatives, to strengthen the collaboration between the national institutions; this will allow improving communication and collaboration between Hydromet, MoA, farmers and other users of agrometeorological information.

Introduce a new crop yield forecasting system; there is high demand and need for improving agrometeorological forecasts. A new system on crop forecasting was presented by the Consultant.

Train relevant staff from MoA, Hydromet and NSS on a new crop yield forecasting system and remote sensing. Probably, 2-3 people from each institution.

Post the Agrometerological Bulletin currently developed by Hydromet on the Hydromet Website (www.meteo.am); this will ensure access to all users; there is currently quite a limited access to agrometeorological bulletins. For instance, Marz support centres do not receive the bulletins, which is missed opportunity.

Increase the interest and capacity of Hydromet in Plant Diseases; the consultant has identified high need for using a prediction model for crop disease development while, according to MoA, 30% yield reductions were observed in years with high disease pressure (in case plant protection measures are not implemented in due course).

Design booklets with recommendations to farmers in emergencies and unfavourable climatic conditions.

Procure required relevant equipment (including both specific meteorological and IT equipment, e.g. a laptop for Hydrometeorological station) for Hydromet to improve agromet forecasting.

The implementation of these recommendations will result in mitigating the negative impacts of climate vagaries and generating positive impacts. It will also assist the decision-makers in better planning their activities to support farmers. The Workshop clearly showed that there is a strong interest from the national institutions in improving crop yield forecasting. The participants agreed with the recommendations of the international consultant. The options proposed were explained in details. Improving crop forecasting is based on a system concept which will have to be adapted to the Armenian context. It will require a strong investment from Hydromet and a partnership with MoA (and with NSS at the beginning at least). All parties are now informed of the proposal and there is a consensus on the options proposed by the consultant. This is the main result of the Workshop.

4 

Based on the mission of the international consultant and the inputs provided by Workshop participants, Programme staff will finalize a proposal, integrating the comments received during the Workshop, with a budget and an agenda for implementation.

5 

Annex 1: Workshop Agenda

1300 – Snacks, sandwiches, beverage

1400- Opening speech, Levon Vardanyan, Hydromet, Director (Presented by Hamlet Melkonyan, Deputy Director)

1415 - Presentation on Programme, Mane Tapaltsyan, EC/ FAO Programme on Food Security Information Systems to Improve Decision Making in East area, Country Coordinator

1440 – Presentation on Use of Agrometerological data and Delivery of Information to the users, Nelly Arakelyan, Hydromet, Agrometereological Forecasting Division

1515 - Proposals on Improvement of Crop Forecasting in Armenia, Bernard Tychon, Liege University, Professor, Agrometeorologist

1700 - Discussions

1745 - Visit to the Divisions of Hydromet

6 

Annex 2. List of Workshop Participants

1. L. Vardanyan, RA Ministry of Emergency, Armstatehydromet, Director 2. H. Melkonyan, RA Ministry of Emergency, Armstatehydromet, Deputy Director 3. Z. Petrosyan, Armstatehydromet Operational Hydrometerological Centre, Head 4. V. Grigoryan, Advisor to the Arstatehydromet, Director 5. D. Avagyan, European Commission (not attended) 6. M. Tapaltsyan, EC/ FAO Programme on Food Security Information Systems, Country Cordinator 7. L. Grigoryan, Agrometeorological Forecasts, Head of Division 8. N. Hakobyan, Hydrometeorological Center of Information and Marketing Division, Head 9. L. Simonyan, Agrometeorology Division, Head 10. N. Arakelyan, Agrometeorological Forecasts Division, Leading Specialist 11. K. Yesayan, RA Ministry of Agriculture, Horticulture Division, Head 12. G. Harutyunyan, RA Ministry of Agriculture, Land Management and Use Division, Head 13. H. Lemberyan, RA Ministry of Agriculture, Agricultural Planning Division, Head 14. A. Petrosyan, RA Ministry of Agriculture, Agricultural Planning Division, Chief Specialist 15. A. Hakobdjanyan, RA National Statistical Service, Food Security Division, Head 16. L. Aleksanyan, Agricultural Support Marz Center of Armavir, Head 17. Kh. Mkrtchyan, Agricultural Support Marz Center of Aragotsotn, Head 18. G. Yeghiazaryan, Agrogitaspyur Department, Armenian State Agricultural Academy 19. B. Zakaryan, Armstatehydromet, Hydrography Division, Head 20. G. Surenyan, Meteorological Forecasts Division, Head 21. A. Hovsepyan, Climate Research Division, Head 22. D. Hovhannisyan, Climatology Division, Head 23. V. Badalyan, Agrometeological Division, Chief Specialist

A few more specialists from Operational Hydrometeorological Centre, Hydrometeorological Center of Information and Marketing, Climate Research Division, Climatology Division of Armstatehydromet were also present at the Workshop.

Annex 3. Presentation of the international consultant

Crop yield forecasting proposal for Armenia

Bernard TYCHON(Bernard.Tychon@ulg.ac.be)

FAO ConsultantHydromet, Yerervan, 13 October 2010

Content

• Operational Crop yield forecasting techniques

• Present status of the Crop Yield Forecast in Armenia

• Proposed improvements• Conclusions

7 

Operational crop yield forecasting techniques

Objectives of crop production and crop yield forecasting systems

• Pricing• Market stability• Food security• Control of supply

} data in real time!

8 

EU-15 Cereals balance sheet:Marketing year: 2002/2003Common (Mio t)

wheat Barley Durum Maize Rye Sorghum Oats Triticale Others EUR 15Beginning stocks (01.07.2002)

Market 12.7 7.3 1.0 5.3 1.0 0.0 0.6 0.6 0.1 28.5Intervention 0.5 2.5 0.0 0.0 5.1 0.0 0.0 0.0 0.0 8.1

Total 13.2 9.8 1.0 5.3 6.1 0.0 0.6 0.6 0.1 36.6Usable production 93.9 47.7 9.4 40.0 4.7 0.7 6.8 5.2 0.7 209.2Import 6.2 0.1 0.4 3.0 0.0 0.1 0.0 0.0 0.2 10.0TOTAL AVAILABILITIES 113.2 57.6 10.9 48.3 10.8 0.8 7.4 5.8 1.0 255.8USE - Human 33.0 0.0 7.0 2.5 1.5 0.2 1.3 0.0 0.0 45.5 - Seed 2.9 2.0 0.8 0.2 0.2 0.0 0.3 0.2 0.2 6.8 - Industrial 6.3 7.4 0.0 4.4 0.2 0.0 0.2 0.0 0.1 18.6 -Ultra peripheral islands 0.3 0.2 0.0 0.4 0.0 0.0 0.0 0.0 0.0 0.9 - Animal feed 41.6 31.8 1.0 32.0 2.0 0.7 4.1 4.8 0.4 118.4TOTAL USE 84.1 41.4 8.8 39.5 3.9 0.8 5.9 5.0 0.7 190.0Solde disponible 29.2 16.2 2.1 8.8 6.9 0.0 1.5 0.8 0.3 65.8Export (1) 16.5 9.0 0.9 2.3 * 1.5 0.0 0.7 0.0 0.0 30.9End stocks (30.06.2003)

Market 12.7 6.7 1.2 6.5 1.0 0.0 0.8 0.8 0.3 29.9Intervention 0.0 0.5 0.0 0.0 4.4 0.0 0.0 0.0 0.0 5.0

Total 12.7 7.2 1.2 6.5 5.4 0.0 0.8 0.8 0.3 34.9Change in stocks -0.5 -2.6 0.1 1.2 -0.7 0.0 0.2 0.2 0.2 -1.8Change in public stocks -0.5 -2.0 0.0 0.0 -0.7 0.0 0.0 0.0 0.0 -3.1

(1) Grains equivalent. *) Maize includes 1.8 mio. t processed products and animal feedMaximum W T O: 2002/2003 ESTIMATED EXPORT QUANTITIES 2002/2003

Wheat incl. durum 14.438 mio t +0,5 mio t food aid 17.40 mio t (food aid included and refund-free)Coarse grains 10.8432 mio t.(inclu. 0,4 mio t potato starch) 13.50 mio t (inclu 1.8 mio t maize products, but excl. 0,4 mio t potato starch)

Production = Yield X Area

Basic relation

9 

Yield assessment

Yield variability in cereals

0

1

2

3

4

5

6

7

1961 1965 1969 1973 1977 1981 1985 1989 1993

Yiel

d (to

ns/h

a)

Kyrgyzstan

Egypt

Romania

SAUDI

USA

Gommes, FAO, 2003

10 

Yield factors of variability

30

40

50

60

70

80

90

1960 1965 1970 1975 1980 1985 1990 1995

Yiel

d (a

rbitr

ary

units

)

0

5

10

15

20

25

30

Fact

ors

F1 to

F5

(arb

itray

uni

ts)

YieldF1 (innovation) F2 (policy)F3 (trend)F4 (extreme factor) F5 (weather)

In Gommes, FAO, 2003

Crops

Soils

Weatherstations

Remotesensing

0

2000

4000

6000

8000

10000

120001975197619771978197919801981198219831984198519861987198819891990199119921993199419951996199719981999

Harvest year

Yiel

d (k

g/ha

)

Vegetationindices

(NDVI/DMP)

Vegetationindices

(NDVI/DMP)

Climate dataClimate data

Crop, soil & management parameters

Crop, soil & management parameters

Crop growthsimulation model(s)

Cultivatedarea

Cultivatedarea

Water balance parameters

Water balance parameters

Yieldprediction

moduleProductionProduction

Country yieldAgricultural statistics

Crop monitoring & Yield forecasting: General Flowchart

Estimates for each YEAR x REGION x CROP on:- Yield = fcal(4 types of Indicators)

Trend, Meteorology, Crop growth model, Remote Sensing- Production = Yield x Area

FAO, modified

11 

Water balance at station level

NDVI or other gridded data

ETA /grid

Agricultural statistics

ETA District/marz

Yield district (Marz)

1987

2002

1985

Updated general methodology

INPUTS

AMS ETP

Actual rainfall

STATISTICA

NOAA GAC

V A S T

Yield prediction model at

departmental level

OUTPUT

Temperature, RR, RH,… Yield prediction

model at national level

W I N D I S P

MATLAB

Initial Water Holding Content

Soil Water Satisfaction Index

Water excess, deficit

Actual ETA

…

NDVI

INDEPENDENT VARIABLES

CROP YIELD DATA

Historical crop yield data at departmental level

Yield agregation at national level

METEO

Starting date

NDVI max

Time peak

…

Cumulated actual rainfall

ExplanatoryVariables

Spot-VEGETATION

AGROMET SHELL (AMS)

12 

Yield and forthcoming weather…

Present status of the Crop Yield Forecasting in Armenia

13 

• Information provided by Hydromet– Field observations

• Meteorological data (3-hourly data)• Phenological data (bi-weekly data)• Dense agromet observation network (36 stations !)

– Crop yield forecasting for major crops• Winter wheat, Barley, Apricots, Grape, Onion +

Pasture/Grassland• Based on agromet models developed a long time

ago (input data = temperature, rainfall, development stage, most of the time)

• Information provided by Ministry of Agriculture (MoA)– Based on agricultural research/advise centers

field observations (Mars Centers)– Crop Yield Forecast based on expertise inside

the MoA

14 

Communication

By email or telephone to- MoA- Private companies (Beer, Sugar,…)

Proposed improvements

15 

Proposal

• Communication improvement:– Set up of a multidisciplinary working group

combining Hyrdromet and MoA expertise in order to provide an improved yield assessment/forecast of major crops (sharing of data and experiences)

• Ten daily meeting/Monthly meeting/Emergency meeting

• Led by Hydromet or another rotating management – Website development (Hydromet activities

show-case with a connection to MoA website)

Proposal• Technical improvements:

– Adopt the Crop Yield Forecasting (CYF) general methodology applied in many countries by adapting it to the Armenian context

– Integrate remote sensing data into the Armenian CYF system

– Improve spatial interpolation through adapted interpolation techniques, especially for crop development stages

– Develop plant diseases model for plant protection activities that are also crucial for crop yield forecasting

– Consider extreme events in the crop yield forecast (early frost, hail, locusts invasion,…)

16 

Methodology4 types of explanatory variables

Agrometeorologicalfactors

Meteorologicalfactors

RS derived factors

Output of an agrometmodel AgrometShell Sum of met. variables

Output of Chronos programme and cumulated NDVI

Chronos metricsNDVI image processing tool

Calculation of metrics of the different time series

Input data:

- 10-daily NDVI time series for the considered period

With a spreadsheet

Sum of some met data that could explain yield: rain, ET0, radiation,…

AgrometShellFAO Tool

Simulation of Crop specific soil water balance and calculation of a set of parameters.

Input data:- Rainfall- Potential Evapotranspiration- Phenological parameters

21

Other factors

Methodology4 types of explanatory variables

Agrometeorologicalfactors

Meteorologicalfactors

RS derived factors

Output of an agrometmodel AgrometShell Sum of met. variables

Output of Chronos programme and cumulated NDVI

Chronos metricsNDVI image processing tool

Calculation of metrics of the different time series

Input data:

- 10-daily NDVI time series for the considered period

With a spreadsheet

Sum of some met data that could explain yield: rain, ET0, radiation,…

AgrometShellFAO Tool

Simulation of Crop specific soil water balance and calculation of a set of parameters.

Input data:- Rainfall- Potential Evapotranspiration- Phenological parameters

21

Other factors

17 

Crop diseases modelCase of W. Wheat (Septoriosis) – Proculture model

100806040

1/10/99

1/10/99

18 

t°

1/12/99

1/12/99

1/02/00

1/02/00

31/03/00

31/03/00

31/05/00

31/05/00

31/07/00

31/07/00

Données climatiques journalières

102

3020100

% hr

3020100

Pluie

102

101

100

102

101

100

102

101

100

102

101

100

101

100

F5F4F3F2F1

Sensitive variety

Infectionsjournalières

(mm)

Meteo hourly data

Phytophtora Infestans (Mildew) on Potato

Time in hour

No infection

Light infection

Aver

age

Tem

p w

ith R

H>9

0%

Moderate infection

High infection

Very high infection

-Wind-Rain

4 – 10 days-Opt T°: 18 to 22 °C-Min 8h with RH > 90%

Infection-free water-opt T° : 12 to 16°C-min 5°C

-opt T°: 17 to 20°C

Guntz-Divoux curves

Potential Infection

Relative Humidity

DECEMBER 2002

JANUARY 2003

FEBRUARY 2003

MARCH 2003

APRIL 2003

-30

-20

-10

0

10

20

30

40

1-Jan 1-Mar 1-May

°C

19 

T max T min

POLAND - Western

0

5

10

15

20

25

30

35

1-Jan 1-Feb 1-Mar 1-Apr 1-May

mm snowrain

Frost kill risk

Frost Risk monitoringFrost Risk monitoringFrost Risk monitoring

In Genovese, 2005 (JRC)

Example for Armenia

Ararat Marz

0

20

40

60

80

100

120

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35

Time (dekade)

Rain

fall

(mm

)

2006-bad2007-good2003-bad

Ararat Marz

-15

-10

-5

0

5

10

15

20

25

30

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35

Time (dekade)

Tem

pera

ture

(°C)

2006- bad2007-good2003-bad

Example of weather conditions for 3 years in one Marzof Armenia

20 

Erevan

Cold

Temperature

(Average)

In NewLocClim

Agromet explanatory variables

April 2006 dek 3 Rain

Actual evapotranspiration

Wheat 2006Water excess during wheat crop growth period (2006)

21 

SPOT-VEGETATION IMAGERY (Average 10-daily data)

Available at FAO

2007 – Good for wheat

2006 – Bad for wheat

Armenia Marz Vegetation index(2007)

-0.1

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35

Time (Dekade)N

DVI

KotaikAragatsotnArmavirAraratErevanChirakLoriVaiots DzorSiunikGhergharkounikTavouchAverage

Armenia Marz Vegetation index (2006)

-0.1

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35

Time (Dekade)

NDVI

KotaikAragatsotnArmavirAraratErevanChirakLoriVaiots DzorSiunikGhergharkounikTavouchAverage

First trial for potato

• Simplification: – 10 years of data (meteo, RS, agricultural

statistics)– 1 station per Marz– 1 single model for the whole country– Agrometeo and Remote sensing explanatory

variables only• 26 potential explanatory variables• 4 variables retained after statistical analyses:

– WEXv, ETAf, SVAL and SLOP

22 

Potato Yield ForecastModel Cross Validation Results

Trend CY Model

R2 = 0.7796

0.00000

5.00000

10.00000

15.00000

20.00000

25.00000

0.000000 5.000000 10.000000 15.000000 20.000000

Official Statistics (T/ha)

Tren

d (T

/ha)

Series1

Linear (Series1)

Agrometeorological CY Model

R2 = 0.8027

0.000000

5.000000

10.000000

15.000000

20.000000

25.000000

0.000000 5.000000 10.000000 15.000000 20.000000

Official statistics (T/ha)

Agr

omet

Mod

el (T

/ha)

Series1

Linear (Series1)

Potato Yield

0.000000

5.000000

10.000000

15.000000

20.000000

25.000000

2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Time (year)

Yiel

d (t/

ha)

Official figuresAgromet modelSimple trend

Conclusions• Create a CYF working group including both Hydromet

and MoA• Train experts from both institutions in new

agrometeorolgical tools (team building and capacity building)

• Adapt the General CYF flowchart to Armenia in order to provide the first CYF before 20 months

• Equip both teams with hardware and software material• Acquire 3 automatic weather stations for plant disease

forecasting and early warning• Website?

23