a horizontal mobile measuring system for atmospheric quantities … · 2020. 7. 16. · received:...

Atmos Meas Tech 7 2967ndash2980 2014wwwatmos-meas-technet729672014doi105194amt-7-2967-2014copy Author(s) 2014 CC Attribution 30 License

A horizontal mobile measuring system for atmospheric quantities

J Huumlbner1 J Olesch1 H Falke2 F X Meixner3 and T Foken14

1University of Bayreuth Department of Micrometeorology 95440 Bayreuth Germany2Gesellschaft fuumlr Akustik und Fahrzeugmeszligwesen mbH 08058 Zwickau Germany3Max Planck Institute for Chemistry Biogeochemistry Department 55020 Mainz Germany4Member of Bayreuth Center of Ecology and Environmental Research (BayCEER) University of Bayreuth95440 Bayreuth Germany

Correspondence toJ Huumlbner (joerghuebneruni-bayreuthde)

Received 25 September 2013 ndash Published in Atmos Meas Tech Discuss 7 May 2014Revised 30 July 2014 ndash Accepted 30 July 2014 ndash Published 16 September 2014

Abstract A fully automatic horizontal mobile measuringsystem (HMMS) for atmospheric quantities has been de-veloped The HMMS is based on the drive mechanism ofa garden railway system and can be installed at any loca-tion and along any measuring track In addition to meteo-rological quantities (temperature humidity and short-long-wave downupwelling radiation) HMMS also measures tracegas concentrations (carbon dioxide and ozone) While suffi-cient spatial resolution is a problem even for measurementson distributed towers this could be easily achieved withthe HMMS which has been specifically developed to ob-tain higher information density about horizontal gradients ina heterogeneous forest ecosystem There horizontal gradi-ents of meteorological quantities and trace gases could beimmense particularly at the transition from a dense forestto an open clearing with large impact on meteorological pa-rameters and exchange processes Consequently HMMS wasfirstly applied during the EGER IOP3 project (ExchanGEprocesses in mountainous Regions ndash Intense Observation Pe-riod 3) in the Fichtelgebirge Mountains (SE Germany) duringsummer 2011 At a constant 1 m above ground the measur-ing track of the HMMS consisted of a straight line perpen-dicular to the forest edge starting in the dense spruce for-est and leading 75 m into an open clearing Tags with barcodes mounted every metre on the wooden substructure al-lowed (a) keeping the speed of the HMMS constant (approx05 m sminus1) and (b) operation of the HMMS in a continuousback and forth running mode During EGER IOP3 HMMSwas operational for almost 250 h Results show that ndash dueto considerably long response times (between 4 and 20 s)

of commercial temperature humidity and the radiation sen-sors ndash true spatial variations of the meteorological quanti-ties could not be adequately captured (mainly at the forestedge) Corresponding dynamical (spatial) errors of the mea-surement values were corrected on the basis of well-definedindividual response times of the sensors and application ofa linear correction algorithm Due to the very short responsetimes (le 1 s) of the applied commercial CO2 and O3 analy-sers dynamical errors for the trace gas data were negligibleand no corrections were done

1 Introduction

The heterogeneity of terrestrial surfaces requires high spa-tially resolved measurements of atmospheric quantities toquantify their heterogeneous structure close to the surfaceForest ecosystems which cover a large portion of the Earthrsquossurface are particularly affected by increasing heterogeneitydue to wind throws pests and human activities such as de-forestation High spatial resolution can be achieved by sam-pling at various locations (Oncley et al 2009) This can berealised either by a large number of locally fixed sensors orby (fast moving) mobile measuring systems

Application of mobile measuring systems started in themiddle of the last century They are used when the spa-tial representation of the investigation site is required tobe as exact as possible and the number of available lo-cally fixed sensors is insufficient In former studies mobilemeasuring systems mostly carried radiometers above and

Published by Copernicus Publications on behalf of the European Geosciences Union

2968 J Huumlbner et al A horizontal mobile measuring system for atmospheric quantities

under forest canopies to measure the areal distribution ofupwelling andor downwelling radiation (egLeonard andEschner 1968 Mukammal 1971 Brown 1973 Baldocchiet al 1984a b Peacutech 1986 Lee and Black 1993 Chenet al 1997 Privette et al 1997 Blanken et al 2001)Comparable measurements above grasslands were taken byRodskjer and Kornher(1967 1971) To study wind veloc-ity and tower-induced errors of its measurementsDabberdt(1968) built a mobile system which carried an anemometerOumlrlander and Langvall(1993) andLangvall and Loumlfvenius(2002) describe their ldquoAsa-Shuttlerdquo a mobile system whichmeasures in addition to radiation the air temperature alonga horizontal transect of decreasing shelterwood density ofNorway spruce The system ofSingh et al(2008) is able tomove in two dimensions the vertical and horizontal Theyused their system for the study of river and lake aquaticsystems The mobile system ofGamon et al(2006) car-ries a dual-detector spectrometer to measure ecosystem spec-tral reflectance The TRAM (Transect Measurement) systemof Oncley et al(2009) runs in a loop measuring wind ve-locity (ultrasonic anemometer) the CO2 concentration airtemperature and humidity They were able to create a mea-surement track through a forest ecosystem and over a creek(AmeriFlux site at Niwot Ridge) through the use of steel ca-bles and masts which allowed mobile measurements overa long horizontal distance and also changing of the verti-cal measurement height Besides the mobile platforms men-tioned above there are investigations with mobile measure-ments in aircrafts balloons lifts and other mobile plat-forms (egLenschow 1972 Kaimal et al 1976 Ogawa andOhara 1982 Balsley et al 1992 1998 Friehe and Khelif1992 Muschinski et al 2001 Mayer et al 2009)

Focusing on energy and matter exchange at a forest edgethe EGER IOP3 project (ExchanGE processes in moun-tainous Regions ndash Intense Observation Period 3) was con-ducted in a disturbed forest ecosystem where heterogeneitieswill substantially impact exchange processes between at-mosphere vegetation and soil Forest edges in particularmay have greater effects on exchange processes in a het-erogeneous forest ecosystem Driving forces at the forestedge are spatial differences of up- and downwelling radia-tion temperature moisture and the resulting wind regime(Murcia 1995 Matlack and Litvaitis 1999 Davies-Colleyet al 2000 Klaassen et al 2002) Forest edges and prevail-ing gradients were targeted in a variety of micrometeorolog-ical studiesChen et al(1993 1995) focused on a Douglas-fir forest (Pacific Northwest USA)Dawson and Sneddon(1969) McDonald and David(1992) as well asDavies-Colley et al(2000) investigated the edge of rain forest nextto a clearing in New ZealandNewmark(2001) studied fourforest edges in the West and East Usambara Mountains (Tan-zania) Even flux measurements at forest edges and numeri-cal studies have been performed (seeDupont et al 2011 fora detailed overview) First results of the EGER IOP3 projectare presented inEder et al(2013) showing an increased

number of coherent structures at the forest edge which mightbe an indication of a quasi-stationary secondary circulationA detailed description of secondary circulations can be foundin Glickman(2000) and for site-specific conditions seeEderet al(2013)

The horizontal mobile measuring system (HMMS) pre-sented in this paper is a fully automatic system movingon rails The chosen measuring transect for EGER IOP3was a straight line installed perpendicular to the forestedge However a greater variation in the measuring heightalong the track is limited because of the wooden substruc-ture and also because of the limited climbing ability of theHMMS (maximum climbing anglelt 10 ) The HMMSwas equipped with long-wave and short-wave radiation sen-sors (individual sensors for up- and downwelling radiation)a sensor for air temperature and relative humidity and analy-sers for CO2 and O3 concentrationsFoken et al(2012) hadshown that trace gas concentrations are good tracers for co-herent structures and coupling regimes and we expected tofind sources or sinks in the horizontal structures which mightbe an indication for the influence of vertical structures suchas coherent structures or secondary circulations The selec-tion of appropriate sensors was a major challenge because of(a) the limited power of the drive mechanism leading to pay-load restrictions for sensors and electronics (b) the space onthe carrier platform restricting the size of sensors analysersand electronics and (c) the requirement for short responsetimes in order to capture the true spatial variations of mete-orological quantities and trace gas concentrations (with thehighest variations occurring mainly at the forest edge) withinan acceptable time Besides the technical description of theHMMS (Sect21) we present the results of laboratory testsfor the individual sensor response times some short time se-ries of the measurements and the application of a correctionalgorithm (Sect22) to correct the measurements for the re-sponse time induced dynamical error

2 Materials and methods

21 Description of HMMS

The technical description of the HMMS comprises details ofthe drive mechanism and the design of the carrier the soft-ware and the sensor system The front and lateral view of theHMMS are shown in Fig1 with the corresponding descrip-tions listed in Table1 Table2 contains the sensorsrsquo detailsFor more technical details of the HMMS including technicaldrawings pictures and the detailed wiring plan seeHuumlbneret al(2011)

211 Drive mechanism and carrier design

The drive mechanism of the HMMS is from a commer-cial garden railway system (Lehmann-Groszlig-Bahn LGBsince 2007 part of Gebr Maumlrklin amp Cie GmbH

Atmos Meas Tech 7 2967ndash2980 2014 wwwatmos-meas-technet729672014

J Huumlbner et al A horizontal mobile measuring system for atmospheric quantities 2969

Figure 1 Front and lateral view of the HMMS Specifications of the numbers are in Table1

Table 1Specification of numbers from Fig1

No Description No Description

1 Short-wave radiation sensors on a 04 m long boom 11 Pump for CO2 analyser2 Long-wave radiation sensors on a 04 m long boom 12 Edinburgh Instruments Ltd Gascardreg NG CO2 analyser3 Code 39 bar code 13 National Instruments DAQ device4 Makrolonreg cover to protect the HMMS from rain and dirt 14 7rdquo TFT monitor5 Enviscope O3 analyser 15 Micro PC6 Fan for ventilation of the HMP 155 16 LGB Analog throttle with potentiometer7 Inlet for HMP 155 double shielded 17 Fan for cooling the entire system8 Inlet for O3 analyser made of PTFE 18 On-board storage battery9 Inlet for CO2 analyser made of aluminium 19 Lateral holder to protect the HMMS against falls10 Pump for O3 analyser 20 Sick CLV412-1010 bar code scanner

Goumlppingen Germany) The drive mechanism (two DC en-gines and wheels) were removed from a replica Diesel hy-draulic locomotive and mounted under an aluminium plate(300 mmtimes 1000 mm) incorporating a downward-facing U-profile to increase mechanical stability This plate serves asthe carrier platform of the (a) sensors (b) data acquisitionand (c) control system for speed and position The HMMSmoves on original LGB rails with a track gauge of 45 mmThe rails are made of brass due to its high electrical con-ductivity electrical power for the entire HMMS system wassupplied through the rails (infeed 24 V DC 5 A) This infeedcovers the total requirement of the HMMS which is approx3 A for all devices and the rest is used for the drive mech-anism differing depending on the speed andor the inclina-tion The power was fed in continuously every 25 m startingin the middle of the HMMS track (forest edge) and goingin both directions (feed points were at positions 25 50 75100 125 m) to avoid power loss over long distances Spring-mounted pick-up shoes (sliding contacts) on both sides of theengines tapped the power from the rails

Speed and the driving direction of the HMMS are con-trolled by an analogue throttle (provided by LGB) Its poten-tiometer is controlled via the HMMS software and a data ac-quisition (DAQ) device with an analogue output (for a moredetailed description of the DAQ device see Sect212)The output of the analogue throttle can be set ranging from0 V DC (motor standstill) to 24 V DC (full throttle) with thepolarity determining direction (+ forwardsminus backwards)

The power supply of the engines with maximum 24 V DCis the first electrical circuit on the HMMS but the sensorsand Micro PC of the HMMS need a 12 V DC supply withconstant polarity This is provided by a DCDC-converter inthe second electrical circuit of the HMMS which has a pro-grammable and stabilised output Occasional breakdowns ofelectrical power are buffered by an on-board storage battery(12 V DC 08 Ah)

For determination of position and speed as well as forchanging the driving direction the HMMS was equippedwith a commercial raster bar code scanner (CLV412-1010SICK Vertriebs GmbH Duumlsseldorf Germany) The appliedldquoCode 39rdquo bar code contains information of the distancefrom the starting point Besides determination of the positionand speed it is possible to programme distinct speed profilesalong the measuring track The position of each bar code canbe used as a turning point for reversing the driving directionof the HMMS

A transparent cover made of Makrolonreg was attached tothe carrier platform to protect HMMS electronics and sensorsagainst dirt and moisture (ie drizzle and short showers butnot prolonged wet conditions) Four lateral holders at eachcorner of the carrier platform (No 19 in Fig1) prevent theHMMS from falling (measuring height during EGER IOP31 m)

wwwatmos-meas-technet729672014 Atmos Meas Tech 7 2967ndash2980 2014


Table 2Sensors mounted on the HMMS are two pyranometers (down- and upwellingKdarr Kuarr) two pyrgeometers (down- and upwellingLdarr Luarr) a temperature (T ) and relative humidity sensor (RH) a CO2 analyser (CO2) and an O3 analyser (O3) The accuracies are takenfrom the manufacturersrsquo information

Parameter Sensor Accuracy Remark

Kdarr Kipp amp Zonen CMP3 lt 15 W mminus2 Amplifier used (Factor 50-fold)Kuarr Kipp amp Zonen CMP3 lt 15 W mminus2 Amplifier used (Factor 100-fold)Ldarr Kipp amp Zonen CGR3 lt 15 W mminus2 Amplifier used (Factor 500-fold) optional PT-100 temperature sensorLuarr Kipp amp Zonen CGR3 lt 15 W mminus2 Amplifier used (Factor 500-fold) optional PT-100 temperature sensorT Vaisala HMP155 plusmn 01 K Radiation shield and ventilated with 4 m sminus1

RH Vaisala HMP155 plusmn 1 Radiation shield and ventilated with 4 m sminus1

CO2 Edinburgh Instruments Gascardreg

NG CO2 1000 ppmplusmn 40 ppm Vacuum pump DC2416F (Flow rate 12 L minminus1)

O3 Enviscope O3 analyser sim 009 ppbvlowast Vacuum pump DC2480L (Flow rate 30 L minminus1)

lowast Accuracy at a measuring frequency of 1Hz and a mixing ratio of 50ppbv(1ppbv= mixing ratio of 10minus9)

212 Software and data acquisition system

The HMMS software (Gesellschaft fuumlr Akustik undFahrzeugmeszligwesen mbH (GAF) Zwickau Germany) wasinstalled on a Micro PC (Quanmax QBOX-1010) This isa very small light-weight and low-power BOX PC (powerconsumption approx 1 A) with a sufficient number of portsto connect all parts of the HMMS All measured datawere stored on the internal hard disk drive (HDD) of theQBOX-1010

Through a 16 bit data acquisition (DAQ) device(USB-6211 National Instruments Austin USA) theHMMS software manages the control of speed the de-termination of position the driving direction and the dataacquisition of the HMMS The DAQ device a very compactbox with 16 analogue input and two analogue outputchannels (with an additional four digital inputsoutputs)handles the two pyranometers the two pyrgeometers twoextra PT-100 temperature sensors for the pyrgeometers andthe air temperature and humidity sensor The CO2 and O3analysers and the bar code scanner are connected via serialports directly to the QBOX-1010 For a detailed overview ofall applied sensors see Sect213

The sensorsrsquo sampling rate can be set in the range of 1to 5 Hz with a 200 times higher internal oversampling rateto improve resolution and reduce potential (although highlyunlikely) aliasing effects caused for example by noises ofthe power frequency (Bentley 2005) Oversample mode isa common practice also used in data loggers Because of themovement of the HMMS we used only the first 10 of theinternal sampling rate for linear averaging in order to allo-cate the measured values to an exact position without havingblurring effects Depending on the chosen sampling rate datafiles are updated on the HDD of the QBOX-1010

213 Sensor system

Due to the demands placed on the sensor system mentionedin the Introduction the final selection of sensors and analy-sers resulted in a compromise between their weightsize andacceptable response times The latter may be described by thetime constant which is a sensoranalyser specific quantityGenerally the smaller the time constant the smaller are thedynamical errors and consequently the spatially smoothingeffects of the measurements An overview of the HMMS sen-sors and analysers is given in Table2 and more detailed in-formation particularly of their individual time constantsτ63ndash which have been determined in our own laboratory tests(see Sect31) ndash are given in Table3

For air temperature and relative humidity measurementsthe HMMS was equipped with a HMP155 temperature andhumidity (HUMICAPreg 180R) probe (Vaisala Oyj VantaaFinland) Both sensors of the HMP155 probe had the high-est time constants of all HMMS sensors A common prac-tice to reduce these time constants is to place the sensors inan aspirated radiation shield tube To achieve that the twosensors were detached from the original sensor housing andthe sintered PTFE filter and were then mounted in a doublethermal radiation shield tube after Assmann (Assmann 18871888) and Frankenberger (Frankenberger 1951) Aspirationhas been facilitated by a small PC fan maintaining a constantair flow of approx 4 m sminus1 within the tube According to themanufacturerrsquos information the time constant of both sensorsof the original HMP155 probe islt 20 s While our modifi-cations reduced the time constant of the temperature sensorto approx 12 s there was no change of the relative humiditysensorrsquos time constant Both time constants have been deter-mined in our laboratory tests (see Sect31) Since the sin-tered PTFE filter was removed from the HMP155 probe wechanged the relative humidity sensor approx every 70 oper-ating hours during EGER IOP3 to minimise potential prob-lems of humidity measurements due to the accumulation of



Table 3Results of calculation of the sensor response time with the given signal differences1Xi in the laboratory test and its consequenceson the location of the measured data The spatial relocation is shown for a positive or negative signal difference with the possible effect ofhysteresis For the sensors used see Table2

Parameter Signal difference1Xi τ63 Delay time Hysteresis Spatial relocation for 05 m sminus1

+1Xi minus1Xi

Kdarr Kuarr 350 W mminus2 4 sa ndash no 12 m 12 mLdarr Luarr 150 W mminus2 4 sa ndash no 14 m 14 mT 5 K 12 sb ndash yes 21 m 19 mRH 15 20 sa ndash yes 23 m 20 mCO2 480 ppm lt 1 sa 02 sd no 3 m 3 mO3 ndash lt 01 sc 05 sd no 05 m 05 m

a In agreement with the data given by the manufacturerb 8s shorter than data given by the manufacturer due to sensor modificationc Developed for high-frequency measurements (50Hz) τ63 is negligible (no laboratory test conducted)d Delay time caused by the inlet length

dust on the sensor In comparison with measurements by lo-cally fixed (aspirated) sensors close to the HMMS track wedid not observe an increasing deviation in the humidity mea-surements between the HMMS and the locally fixed mea-surements even after more than 250 operating hours

A total of four radiation sensors were mounted on theHMMS two pyranometers (CMP3 short-wave radiation03ndash28 microm) and two pyrgeometers (CGR3 long-wave ra-diation 45ndash42 microm) to measure down- and upwelling radi-ation (Kipp amp Zonen Delft The Netherlands) The sensorhousing temperature of the pyrgeometer was measured witha Pt-100 temperature sensor (replacing the original thermis-tor sensor) Whereas the sensors are installed with no modifi-cations the time constants of the sensors are consistent withthe manufacturerrsquos information (verified in own laboratorytests Sect31) Since the output voltage of the thermopile-type radiation sensors are in the microV-range home-built ampli-fiers (electronic workshop University of Bayreuth amplifi-cation factors in Table2) have been used to match the inputsensitivity of the applied National Instruments DAQ device(range 0ndash1 V) Each pair of radiation sensors is mounted on40 cm long booms which in turn are fixed on the front side ofthe HMMSrsquos carrier platform pointing in opposite directionsbut perpendicular to the driving direction (see Fig1) Thislateral installation minimises ldquoshadowingrdquo effects caused bythe HMMS itself andor by the trackrsquos construction elementsand guarantees a rather free hemispheric view

To measure the carbon dioxide concentration the HMMSis equipped with a closed-path single cell non-dispersive in-frared gas analyser (OEM Gascardreg NG CO2 EdinburghInstruments Ltd Edinburgh UK) with a fixed measure-ment range of 0ndash1000 ppm The Gascardreg NG is a verylightweight (300 g) and small (160 mmtimes 100 mmtimes 40 mm)sensor an OEM version which (a) has no housing (shel-tered by the HMMSrsquos Makrolonreg cover only) and (b) issupplied by a small vacuum pump (12 V DC 12 L minminus1model DC2416F Fuumlrgut GmbH Tannheim Germany) TheCO2 analyserrsquos inlet tubing consists of an aluminium tube

(inner dia 3 mm outside of HMMS) and a flexible tubeof polyethylenendashaluminium composite structure (Dekaboninner dia 3 mm inside of HMMS) The overall length ofinlet tubing is 50 cm which translates into a time delaytD = 02 s This has to be added to the delay caused bythe time constant of the sensor According to the manufac-turer the Gascardreg NGrsquos standard (optimised) response timeshould be 10 s (τ90) In our laboratory tests we have deter-mined a time constantτ63 with approx 1 s for the Gascardreg

NG including the delay time caused by inlet tubing lengthThe ozone analyser (Zahn et al 2012) applied on the

HMMS has been developed by a cooperation between Karl-sruhe Institute of Technology (Karlsruhe Germany) and thecompany enviscope GmbH (Frankfurt Germany) The anal-yser contains a fast response (up to 50 Hz) photomultipliermeasuring the chemiluminescence originating from a surfacereaction of O3 with a coumarin dye layer on a small alu-minium disc which is in contact with the sample air streamEnviscope GmbH provided an OEM version (without hous-ing and front panel control elements) resulting in a 53 re-duction of weight and 81 of volume without any loss ofaccuracy and performance The second electrical circuit ofthe HMMS (see Sect211) provides the constant operat-ing voltage for the analyser (12 V DC) while 24 V DC powerfor heating of the reaction chamber (nominally 15ndash60 V DC)was tapped from the engines of the HMMS A small vac-uum pump (12 V DC 3 L minminus1 model DC2480L FuumlrgutGmbh Tannheim Germany) is used to draw ambient airthrough the O3 analyser Inlet tubing consisted of a 30 cmlong black PTFE tube (inner dia 635 mm) similar to thefirst 10 cm of the outlet tubing Another 20 cm of Dekabontube (inner dia 3 mm) leads the output of the O3 analyser outthrough the bottom of the HMMS In- and outlet tubing haveto be black (lightproof) otherwise the (chemiluminescence)detection of O3 in the reaction chamber would be affectedby incident sunlight The chemiluminescence analyser wasdeveloped for aircraft and eddy covariance measurements atup to 50 Hz consequently its time constant is negligible and



only the delay time of 05 s caused by the inlet tubing has tobe considered

The (multiplier) raw signal of the enviscope O3 analyser(counts sminus1) can be calibrated by a suitable O3 source How-ever the sensitivity of the O3 analyser has a considerabletemporal variability caused by the aging of the sensorrsquos discwithin 2 days (Zahn et al 2012) Therefore the signal ofthe HMMS O3 analyser was adjusted in situ by simultaneousO3 concentration measurements performed at a fixed loca-tion 05 m lateral to the HMMS track at the turning point ofthe HMMS in the forest (see Fig2 label ldquo49irdquo) Every timethe HMMS passed this point the fast-response O3 analyserat the HMMS was adjusted to the O3 monitor (approx ev-ery 10 min) These measurements were performed by a com-mercial state-of-the-art UV-absorption O3 analyser (ThermoInstruments model 49i MLU GmbH Moumldling Austria)

22 Correction of sensor response timeinduced spatial errors

The input signal of most sensors will follow a very sud-den change in the measured quantity with certain delay anddamping effects These effects are commonly characterisedby an individual sensor response time or time constant Thebias is usually due to (a) the sensor itself (b) the sensorrsquoshousing andor (c) the sensorrsquos inlet system particularly forgas analysers (seeMayer et al 2009) The signalrsquos tem-poral adaptation of so-called first-order measuring systemsto a sudden temporal change of the quantity to be mea-sured can be described in the following form (egBrock andRichardson 2001 Foken 2008)

Xo(t) = Xi(t) + τdXi

dt (1)

whereτ is the time constant andt stands for the time Thetime constantτ characterises the dependency between the in-put signalXi and the output signalXo in other words the in-ertia of the sensor andor the measuring system respectivelyEquation (1) has the following exponential solution

X(t) = Xinfin

(1minus eminus

tτ

) (2)

whereXinfin is the final value ofX(t) after final adaptation tonew conditions

The time constant of a sensor is of much greater impor-tance for mobile measuring systems than it is for standardtower and other stationary measuring systems The move-ment of the measuring system means that the time constantcan result in an error in time and also in space The total errorof a measuring system is referred to as the dynamical errorA linear change in a measured quantity is achieved when

Xi(t) = 0 for t le 0 (3)

and

Xi(t) = a middot t for t gt 0 (4)

with a as a constant andX(0) = 0 as the initial conditionsThis leads to a change of Eq (1)

a middot t = Xi(t) + τdXi

dt (5)

The exponential solution is

X(t) = a middot t minus a middot τ(1minus eminus

tτ

) (6)

The dynamical error can be defined as the second term inthe right-hand side of Eq (6) There are possibilities to cor-rect the dynamical error mathematically and these correc-tions were done mainly for temperature sensors in aircraft(egRodi and Spyers-Duran 1972 McCarthy 1973 Frieheand Khelif 1992 Inverarity 2000 Saggin et al 2001) Cor-rection algorithms also exist for the temperature and humid-ity measurements with radiosondes (egMiloshevich et al2004) while another linear correction algorithm for tem-perature sensors in a vertical moving platform was devel-oped byMayer et al(2009) Comparable corrections formeasurements with horizontal measuring systems (named inSect1) have not yet been made because of (a) an averag-ing along the measuring track (especially for radiation mea-surements in more homogeneous measuring sites) (b) a non-moving system while measurements are conducted to guar-antee a full adaptation of the sensor input signal to the pre-vailing environmental conditions and (c) the application ofhigh-frequency sensors for which the dynamical errors arenegligible The measured values of the HMMS had to becorrected because we (a) had a heterogeneous measuringsite with the focus on a forest edge (not suitable for averag-ing) (b) would have lost too many details with non-movingmeasurements and (c) had weight and money restrictions sowe had to abandon the application of high-frequency sensors(apart from the O3 analyser)

23 Site description and application of HMMS

The first field application of the HMMS was duringthe EGER IOP3 project in June and July 2011 at theWaldstein-Weidenbrunnen site (5008prime31primeprime N 1152prime01primeprime E775 m asl) It is a research site of the Bayreuth Center ofEcology and Environmental Research (BayCEER) and is lo-cated in theLehstenbachcatchment part of the Fichtelge-birge Mountains (northeastern Bavaria Germany) The siteis also a FLUXNET site (DE-BayBaldocchi et al 2001)where CO2 flux measurements above a dense spruce foresthave been conducted since 1996 Detailed information aboutthe site can be found inGerstberger et al(2004) as well asin Staudt and Foken(2007) The research site is dominatedby Norway Spruce trees of 27 m canopy height (as of 2011)On 18 January 2007 the ldquohurricane likerdquo low-pressure sys-tem ldquoKyrillrdquo destroyed large areas of the dense spruce for-est including areas close to the research site (in a southerlydirection) Meanwhile the clearingrsquos vegetation is similar



Figure 2 Top view of the investigation site of the EGER IOP3project in summer 2011 with all measuring points and exact po-sitions of the main tower M1 (32 m) the turbulence tower M2(36 m) the forest edge tower M3 (41 m) the turbulence mastM4 (55 m) the mast for the Modified Bowen-Ratio MethodM5 (31 m) the turbulence masts M6ndashM8 (55 m) the mast forchemical measurements CM (2 m) the Laser-Scintillometer SLS-40 the Horizontal Mobile Measuring System HMMS the ozonemonitor ldquo49irdquo SODARRASS miniSODAR and the GFS3000(leaf gas exchange measurements) The map is oriented to thenorth Aerial image taken from Bayerische Vermessungsverwal-tung URLhttpgeoportalbayerndebayernatlasbase=910

to the understorey of the forest grass fern bushes (mainlyblueberries) young spruces (1ndash2 m) and deadwood The sci-entific focus of EGER IOP3 were diel cycles of energywater and trace gases of the soilndashvegetationndashatmospherendashboundary-layer system of this disturbed forest ecosystemTo investigate the forest edgersquos impact on the exchange pro-cesses turbulence flux tower measurements were performedat the clearing at the forest edge and within and above theforest (see Fig2 more details inSerafimovich et al 2011)

In addition to these locally fixed towers the 150 m longmeasuring track of the HMMS was built in a NNE to SSWdirection perpendicular to the forest edge (75 m in the for-est and the clearing respectively) The track is marked bythe dotted line (labelled HMMS) in Fig2 The elevationdifference between start (forest) and end (clearing) pointof the HMMS track was approxminus8 m (minus5 or minus3)Applying a sampling frequency of 1 Hz the HMMS ranwith a quasi-constant speed of 05 m sminus1 approx 1 m aboveground The HMMS was usually started in the forest wherethe O3 concentration was continuously monitored by the O3monitor Thermo Instruments model 49i (see Fig2 label

ldquo49irdquo) For a complete reverse run (300 m forest-clearing-forest) the HMMS needed approx 10 min

Taking into account appropriate weather conditions andthe main project issues almost 250 h of data were collectedduring various daytime and night-time periods within the6 weeks of EGER IOP3

3 Results and discussion

By means of laboratory tests we determined the individualtime constantτ63 for most sensors used on the HMMS Theresults of the laboratory tests are presented in Sect31

Following this we want to present results from28 June 2011 There was an almost cloudless sky and due topreceding dry weather conditions high air temperatures andlow humidities also prevailed Low humidities during night-time (lt 90 ) led to the absence of early morning fog anddew The time series of HMMS were nearly complete during28 June Unfortunately the time series of the ozone concen-tration reveals larger gaps due to connection problems of theO3 analyser We present the application of the correction al-gorithm (Sect32) radiation-induced errors in CO2 concen-tration measurements (Sect33) and diurnal variations in allmeasured quantities (Sect34)

31 Sensor response time determination

A detailed knowledge about the individual sensor responsetime is unavoidable if the response-time-induced dynamicalerrors are to be properly corrected We therefore conductedlaboratory tests re-positioning the HMMS very quickly(lt 1 s) in two different environments possessing differentconditions This was carried out for temperature and humid-ity by relocating between a large cold chamber and constantenvironmental conditions in a room for long-wave radiationabove two water bodies with different temperatures and forshort-wave radiation with and without short-wave light TheCO2 analyser was tested between nitrogen and a CO2 cal-ibration gas A laboratory test for the O3 analyser was notnecessary because of the low time constant

Table3 gives the exact numbers obtained with the suddenchange of the environmental conditions The signal differ-ences were used for the calculation of the time constantτ63which indicates a change of the signal of 63 of the finalvalue (Fig3) For a better comparison the signal differenceis presented as the normalised difference between the initialvalue (00) and the final value (10) The calculation alsoserved for the identification of possible hysteresis effectswhich were found for the temperature and humidity sensorFor both sensors the adaptation to changes in the input signalwas faster for a positive signal difference (T coldrarr warmRH dryrarr wet)

A value close to the final signal is measured after a suddenchange of the input signal beyond five times the time constant



Time [s]

Nor

mal

ised

diff

eren

ce [minus

]

τ63

0 10 20 30 40 50

00

02

04

06

08

10

τCO2

1s τKτL4s

τT12s

τRH20s

CO2 Kdarr Kuarr Ldarr Luarr T RH

Figure 3 Measurement results of the laboratory tests of the timeconstantτ63 The small fluctuations of the temperature and humid-ity signal are caused by the turbulence under room conditions Anoverview of all sensors is given in Tables2 and3

(Foken 2008) This delay causes together with the speed ofthe HMMS (05 m sminus1) a spatial relocation of the measure-ment signal The calculated values for the spatial relocationcan be found in Table3 The relocation for the sensors withan identified hysteresis can vary depending on the way thechange in the input signal occurs This relocation can be cor-rected with the correction algorithm (see section below) ormust be taken into account for the interpretation of the data

32 Application of the correction algorithm

The main focus of this work is besides the technical presen-tation of the HMMS together with the exact determination ofthe individual response times the inspection of the measuredquantities Here the focus lies mainly on the forest edge thetransition area from the dense spruce forest to the open clear-ing (this transition area and not the dense forest or the openclearing is also the focus of all further investigations in theEGER IOP3 project) An abrupt change in the input signalsof all quantities is observable mostly within a few metres dis-tance of the forest edge Because of the individual sensor re-sponse times and the motion of HMMS the measured inputsignals lag behind the ldquotruerdquo input signal

To correct the lag or rather the dynamical errors in themeasurements of the HMMS sensors we use a correction al-gorithm of the recorded signalXi with a linear adjustmentafter Eq (6) a first-order differential equation However weare aware that at the measuring site especially at the for-est edge the prevailing gradients are influenced by turbu-lence with an increased number of coherent structures likesweeps and ejections (Eder et al 2013) and also by small-scale heterogeneities Within a short time these structurescan have an effect on the gradient which leads to an over-lapping of several different functions and to non-steady-stateconditions There is for example a generalised dynamicperformance model (Brock and Richardson 2001) which

1113

15

T [deg

C]

FORWARD DIRECTION BACKWARD DIRECTION

8690

94

Position [m]

RH

[]

FOREST CLEARING

0 50 100 150Position [m]

Rel

ativ

e hu

mid

ity (

) CLEARING FOREST

100 50 0

Uncorrected values Corrected values Forest Edge

Figure 4 Example time series of temperature (top) and relative hu-midity (bottom) for one complete run of the HMMS during night-time of 28 June 2011 The runtime was 0253ndash0258 CET in theforward direction and 0258ndash0303 CET in the backward directionThe forest edge is indicated by the vertical green dotted line Thetime series shows the uncorrected measured values (black solid line)and at the forest edge the corrected gradient (grey solid line)

considers ndash beside a steady-state solution ndash a transient so-lution Nevertheless we confined ourselves to the correc-tion algorithm which includes a linear adjustment since thechanges in the quantities caused by the transition from theforest to the clearing are significantly higher than the changescaused by fluctuations of small-scale structures which havea purely random character

For the correction we examined the time series individ-ually for every single run of the HMMS (forest to clearingor back) and determined the starting point and the endpointof the significant change and the gradient itself near the for-est edge Afterwards we corrected the measured values to theldquotruerdquo measured values After the correction the final valueof the signal will be reached earlier in time and also in spa-tial terms The investigation of many runs showed us thatthis procedure is the best way to correct the measured valuesalong the forest edge

Because of the necessity of an individual determinationof the linear function we decided to correct only that dataof measurements which are used for detailed studies of themicrometeorological phenomena near the forest edge In thefollowing we present two examples of the correction algo-rithm and in Sect34 a daily cycle of all parameters whichhas not been corrected Due to a lower spatial resolution ofthe figures the general information is not disturbed

Figure4 shows a night-time series for air temperature (top)and relative humidity (bottom) of one complete run of theHMMS The run led from the forest to the clearing (ldquoForwardDirectionrdquo ndash 0253ndash0258 CET) and from the clearing backto the forest (ldquoBackward Directionrdquo ndash 0258ndash0303 CET) Inthe forest and in the clearing measured (uncorrected) values(black solid lines) were almost constant while near the forestedge there was an obvious steep gradient For this case the



1400 1500 1600 1700 1800

100

9080

7060

50

Time of day [CET]

Pos

ition

[m]

Kdarr [Wmminus2](a)

0

200

400

600

800

1000

1400 1500 1600 1700 1800

100

9080

7060

50

Time of day [CET]

Pos

ition

[m]

Kdarr [Wmminus2](b)

0

200

400

600

800

1000

1400 1500 1600 1700 1800

100

9080

7060

50

Time of day [CET]

Pos

ition

[m]

Ldarr [Wmminus2](c)

360

380

400

420

440

460

480

500

1400 1500 1600 1700 1800

100

9080

7060

50

Time of day [CET]

Pos

ition

[m]

Ldarr [Wmminus2](d)

360

380

400

420

440

460

480

500

1

Figure 5 Detailed horizontal profiles during daytime of 28 June 2011 at the forest edge (horizontal green dotted line) uncorrected profilesfor short-wave downwelling radiationKdarr (a) and long-wave downwelling radiationLdarr (c) and in both cases response time corrected profiles(b d) Position shows distance from starting point in metres starting in the forest (50 m) and ending at the clearing (100 m)

starting points and the endpoints of the respective indecreaseof the corresponding quantity were determined individuallyby visual inspection

The grey lines in Fig4 indicate the corrected data of airtemperature and relative humidity at the forest edge Theyclearly show that a correction of the measured values is re-quired in order to interpret the results of the measurementscorrectly The relocation of the measured signal comparedto the corrected measured signal is obvious and correspondsto the calculated relocation in Table3 After the correctionthe ldquotruerdquo measured values for air temperature have almostthe same values at the forest edge while the gradient for therelative humidity starts in the backward direction much laterthan in the forward direction Possible reasons could be (a) aneffect caused by the hysteresis or (b) an abrupt change of thegradient affected by turbulent structures or (c) a combinationof both effects

Figure 5 shows horizontal profiles of short-wave down-welling radiation (Kdarr) and long-wave downwelling radiation(Ldarr) during daytime of the 28 June around the forest edgeThe significant change ofKdarr (top) occurred approximately5 m north of the forest edge (position measured from startingpoint 70 m) because the sun was shining into the first metresof the forestrsquos understorey In contrast the significant changeof Ldarr (bottom) lies exactly on the forest edge (position mea-sured from starting point 75 m) The profiles in Fig5a and cshow the uncorrected measurements forKdarr andLdarr respec-tively Time constants of the sensors and the different drivingdirection of the HMMS led to the zigzag pattern of the uncor-rected measurement data To correct radiation measurementsbetween 1400 and 1800 CET of 28 June 2011 we consid-ered corresponding response times (4 s see Table3) defined

starting and end points of the linear gradient and calculatedthe ldquotruerdquo data assuming a linear gradient of radiation databetween starting and end point The dynamical error cor-rected horizontal radiation profiles shown in Fig5b (Kdarr)and5d (Ldarr) are nearly free of the zig-zag pattern

33 Radiation-induced error in CO2 measurements

Due to the small time constants (τ63 lt 1 s see Table3) thereis no need for a response-time-induced error correction ofthe CO2 and O3 measurements Whereas the O3 measure-ments which ndash with the exception of the mentioned connec-tion problems ndash were of high accuracy we observed substan-tial problems in the CO2 measurements during daytime Be-cause of a late delivery of the CO2 analyser shortly before thestart of the project we had no chance to check the analyseragainst different influences Laboratory tests after the projecthave shown that the CO2 concentration can be kept at a con-stant level as long as the analyser is not exposed to directsunlight and immediately drifts down if the analyser is sub-jected to direct sunlight Because of this we had to discardthe CO2 measurements during daytime whereas the mea-sured values recorded during night-time coincide with otherCO2 measurements near the HMMS track Since larger spa-tial differences in CO2 measurements occur mainly duringnight-time this is not a significant limitation in the applica-bility of the HMMS system However in subsequent projectswe should consider the use of a different analyser

34 Measurement results with HMMS

In Fig 6 results of the HMMS for 28 June 2011 are shownThere are complete horizontal profiles (150 m ndash starting in



Figure 6 Complete horizontal profile for the whole of 28 June 2011 for all measured quantities(a) shows the downwelling short-waveradiation (Kdarr) and(b) the upwelling short-wave radiation (Kuarr) (c) shows the downwelling long-wave radiation (Ldarr) and(d) the upwellinglong-wave radiation (Luarr) Remark the scaling of the four radiation components is different in the four graphs Position shows distance fromthe starting point in metres with the starting point in the forest (0 m) forest edge (75 m horizontal green dotted line) and endpoint at theclearing (150 m)

the forest and ending at the clearing ndash with the forest edgeindicated by a green dotted line in the middle of the HMMStrack) focusing on the daily cycles for each measured quan-tity As we mentioned above we are not able to show a com-plete corrected time series of the measured quantities (re-garding the time constant) because of the large effort re-quired for the individual determination of the linear functionof every single run The profiles therefore show the valuesas measured without correction to take into account the timeconstant For the inspection and interpretation of the dailyvariations along the transect forestndashclearing the uncorrectedprofiles are sufficient

The short-wave radiation componentsKdarr andKuarr (Fig 6aand b) are very low in the forest but sunny spots can be seenat different locations and times In the morning the shadowof the forest edge can be seen beginning at 0530 CET on thecomplete clearing decreasing from run to run At 0830 CETthe sunlight on the clearing is no longer influenced by theforest edge As well the HMMS itself has an influence onthe measurements in the morning hours because of the sen-sor location on the sun-shaded side This behaviour starts at0600 CET and after 0900 CET this effect disappears

During the day the long-wave downwelling radiationLdarr

(Fig 6c) in the forest is almost the same as the long-waveupwelling radiationLuarr (Fig 6d) at the clearing So the sur-face temperature at the clearing is almost the same as thetemperature of the top of forest canopy In the night theclearing shows significantly lower values forLuarr than in theforest Only where the forest canopy has open areas (see thesunny spots in short-wave radiation measurements) isLuarr inthe forest almost the same asLuarr at the clearing

In Fig 6e the complete horizontal profile for the temper-ature is shown The lowest temperatures occur during night-time and the highest during daytime at the clearing The in-crease of temperature in the morning starts at the clearingsignificantly earlier than in the forest And also the decreasein the evening starts earlier at the clearing

The relative humidity (Fig6f) is at night-time higher onthe clearing than in the forest and at daytime this is reversedThe decrease after the sunrise is significantly earlier at theclearing than in the forest Also the increase in the eveningstarts earlier at the clearing

The CO2 concentration (Fig6g) has its maximum in theearly morning and early evening under very stable situationsThis in is accordance with the findings ofFoken et al(2012)and Serafimovich et al(2011) They found a concentra-tion accumulation during uncoupled situations Here thehigher values are at the clearing and also during night-timethe higher values can be found at the clearing The varia-tionslower concentrations around 2100 CET are caused by achange in the wind direction from northerly to easterly windsand the advection of colderCO2-depleted air Because of theradiation-induced errors in the CO2 concentration measure-ments we have shown only measured values during night

Unfortunately there are bigger gaps in the profile of theO3 concentration (Fig6h) due to the mentioned connectionproblems but the increase in the afternoon is still clearly vis-ible with a significant gradient between forest and clearingThis is certainly the result of the good weather conditionsduring this day as the ozone formation is mainly a sun-induced photochemical reaction



Figure 6 (Continued) Complete horizontal profile for the whole of 28 June 2011 for all measured quantities(e) shows the temperature (T )and(f) the relative humidity (RH)(g) shows the carbon dioxide concentration (CO2) and(h) the ozone concentration (O3) Position showsdistance from the starting point in metres with starting point in the forest (0 m) forest edge (75 m horizontal green dotted line) and endpointat the clearing (150 m)

4 Conclusions

Regarding the technical aspects of the HMMS we can af-firm that the drive mechanism with the garden railway sys-tem was uncomplicated in terms of the system integrationand the control by the HMMS software and also the appli-cation during dry conditions was uncomplicated In spite ofthat we had several problems with the rail system duringthe project under wet conditions the climbing ability of theHMMS was lost however measurements had to be stoppedunder these conditions anyway because of the insufficientwaterproofness And the relatively high coefficient of ther-mal expansion of brass caused problems in the smooth driv-ing of the HMMS While during temperatures above 20Cwave formations in the total rail system could be observedleading to a halting system there were during night and earlymorning (cold temperaturesT lt 10C) gaps between the2 m long rail segments observable If the gap occurred inthe first 25 m (Position 0 to 25 m) or in the last 25 m (Po-sition 125 to 150 m) this led to powerless rail segments anda stopped HMMS system due to the chosen feed points forthe power (see Sect211) Those problems led to a highvariability in the durations for a single run of the HMMSDuring a subsequent project with the HMMS which wasconducted in summer 2012 we used modified connectorsbetween the rail segments which resulted in a significant im-provement in the rail system And also a feed-in of powerup to the end was used to avoid powerless rail segmentsRegarding the drive mechanism (here especially the innerworkings of the engines) we were surprised that the engineskept up the 250 h of operating time during EGER IOP3 and

that maintenance was only necessary after the end of theproject But an ageing of the engines was already observ-able over time which resulted in longer run duration timesWe had expected a substantially shortened lifespan of the en-gines because the total weight of the HMMS was almost 10times greater than the original weight of the garden railwaylocomotive resulting in a higher axial load and consequentlymore stress for the engines A more robust drive system likeegOncley et al(2009) used (with cables and steel wheels)is certainly more uncomplicated in terms of consistency anddurability but on the other hand the development effort andfinancial input is substantially larger

The set of devices used for the control of the HMMS aswell as the data acquisition was nearly unproblematic TheHMMS software was together with the DAQ device a per-fect interface between the sensors and the data acquisitionand also between the determination of position (bar codes)and the HMMS speed and driving direction system Onlythe bar code scanner had during humid conditions (morn-ing fog) some intermittent failed readings because of dewon the codes and the sensorrsquos glass and during sunny daysthe scanner had to be shadowed to avoid failed readings

In the case of the sensor system on the HMMS most of thesensors (CMP3CGR3 HMP155 O3 analyser) are frequentlyused in meteorological measurements and as a consequencethe application was quite simple and the results were ade-quate Because of the low accuracy (plusmn 40 ppm) and the radi-ation induced errors in the CO2 concentration measurementswe should consider a change to a more accurate sensor orprotect the sensor with a housing and a cooling system Thiscould in turn lead to high demands being placed on the power



supply of the HMMS Here perhaps the change from a trans-parent to a non-transparent Makrolonreg cover could be a suf-ficient remedy We chose the transparent cover only for aneasy view into the HMMS system

The focus of this work was the description of the HMMSsystem and the correction algorithm and to show first re-sults of the measurements Those show a clear gradient nearthe forest edge as we expected The application of the cor-rection algorithm shows a more significant gradient withina shorter distance So the before-and-after comparisons illus-trate the need for the correction algorithm especially for theslow responding sensors The results of the correction algo-rithm are for most cases absolutely sufficient Without thesecorrections it is not possible to investigate the processes nearthe forest edge very well Especially in the turbulence influ-enced quantities (temperature humidity and trace gas con-centrations) we can find significant changes in the gradientsat the forest edge within a short period of time So we can as-sume that these gradients are influenced by turbulence struc-tures near the forest edge (and also the first metres of theforest and the clearing) while the influence decreases bothinto the forest and into the clearing until the measurementsare almost unaffected by the forest edge

Combined with turbulence measurements the HMMS isa good tool for a better understanding of the exchange pro-cesses near the forest edge Besides the investigation of theforest edge we get an overview of this heterogeneous forestecosystem not possible if static tower measurements were tobe used instead of the HMMS As with every mobile measur-ing system the HMMS has the limitation that it cannot mea-sure at every location simultaneously The HMMS can there-fore never replace tower measurements otherwise we wouldlose information about small-scale and short-time eventsOur research now goes in two directions (a) the investiga-tion of the heterogeneity of scalars in combination with theresults found inFoken et al(2012) and (b) the influence ofcoherent structures on horizontal structures according to thefindings ofEder et al(2013)

AcknowledgementsThe full functionality and the fast construc-tion of the HMMS would not have been possible without the sup-port of the technical workshops of the University of Bayreuth Wewould like to thank the company enviscope GmbH and in partic-ular H Franke and C Klaus for their fast ozone analyser whichthey lent us for the EGER IOP3 project in summer 2011 In par-ticular their modifications for a smaller and lighter ozone analysermade it possible to measure ozone gradients with the HMMS Wealso want to thank the company SICK Vertriebs-GmbH in par-ticular M Clement for the gift of the bar code scanners and weare also grateful for his fast support Furthermore we would liketo thank L Heymann from the Chair of Applied Mechanics andFluid Dynamics from the University of Bayreuth who helped usdetermine the flow rate inside the radiation-shielded tube of theHMP155 Qianqian Liu helped us manage the measurements withthe HMMS during the project And we want to thank all other PhDstudents of the Department of Micrometeorology student assistants

and G Muumlller from BayCEER for helping us with the constructionof the wooden substructure of the HMMS track

The EGER IOP3 project was funded by the German ScienceFoundation (DFG) under the contract numbers DFG projects PAK446 (FO 22621-1 ME 21005-1 RA 61723-1 and HE 52144-1)The funding of the HMMS was realised by the University ofBayreuth and the Max Planck Institute for Chemistry but notwithin PAK 446 This publication is funded by the GermanResearch Foundation (DFG) and the University of Bayreuth in thefunding programme Open Access Publishing

Edited by U Friess

References

Assmann R Das Aspirationspsychrometer ein neuer Apparat zurErmittlung der wahren Temperatur und Feuchtigkeit der LuftDas Wetter 4 245ndash286 1887


Baldocchi D Hutchison B Matt D and McMillen R Sea-sonal variations in the radiation regime within an oak-hickoryforest Agr Forest Meteorol 33 177ndash191 doi1010160168-1923(84)90069-8 1984a

Baldocchi D D Matt D R Hutchison B A andMcMillen R T Solar radiation within an oak-hickory forestan evaluation of the extinction coefficients for several radiationcomponents during fully-leafed and leafless periods Agr ForestMeteorol 32 307ndash322 doi1010160168-1923(84)90056-X1984b

Baldocchi D Falge E Gu L Olson R Hollinger D Run-ning S Anthoni P Bernhofer C Davis K Evans RFuentes J Goldstein A Katul G Law B Lee X Malhi YMeyers T Munger W Oechel W Paw U K T Pile-gaard K Schmid H P Valentini R Verma S Vesala TWilson K and Wofsy S FLUXNET A new tool tostudy the temporal and spatial variability of ecosystem-scale carbon dioxide water vapor and energy flux densi-ties B Am Meteorol Soc 82 2415ndash2434 doi1011751520-0477(2001)082lt2415FANTTSgt23CO2 2001

Balsley B B Baisley C L Williams J B and Tyrrell G WAtmospheric research using kites here We Go Again BAm Meteorol Soc 73 17ndash29 doi1011751520-0477(1992)073lt0017ARUKHWgt20CO2 1992

Balsley B B Jensen M L and Frehlich R G The use of state-of-the-art kites for profiling the lower atmosphere Bound-LayMeteorol 87 1ndash25 doi101023A1000812511429 1998

Bentley J P Principles of measurement systems Pearson PrenticeHall Harlow 2005

Blanken P D Black T A Neumann H H den Hartog GYang P C Nesic Z and Lee X The seasonal water and en-ergy exchange above and within a boreal aspen forest J Hydrol245 118ndash136 doi101016S0022-1694(01)00343-2 2001

Brock F V and Richardson S J Meteorological MeasurementSystems Oxford University Press New York 2001

Brown G W Measuring transmitted global radiation withfixed and moving sensors Agr Meteorol 11 115ndash121doi1010160002-1571(73)90055-1 1973



Chen J Franklin J F and Spies T A Contrasting microcli-mates among clearcut edge and interior of old-growth Douglas-fir forest Agr Forest Meteorol 63 219ndash237 doi1010160168-1923(93)90061-L 1993

Chen J Franklin J F and Spies T A Growing-season microcli-matic gradients from clearcut edges into old-growth douglas-firforests Ecol Appl 5 74ndash86 doi1023071942053 1995

Chen J M Blanken P D Black T A Guilbeault Mand Chen S Radiation regime and canopy architecture ina boreal aspen forest Agr Forest Meteorol 86 107ndash125doi101016S0168-1923(96)02402-1 1997

Dabberdt W F Tower-induced errors in wind profile mea-surements J Appl Meteorol 7 359ndash366 doi1011751520-0450(1968)007lt0359TIEIWPgt20CO2 1968

Davies-Colley R J Payne G W and van Elswijk M Micro-climate gradients across a forest edge New Zeal J Ecol 24111ndash121 2000

Dawson J W and Sneddon B V The New Zealand lowland rain-forest A comparison with tropical rainforest Pac Sci 23 131ndash147 1969

Dupont S Bonnefond J-M Irvine M R Lamaud Eand Brunet Y Long-distance edge effects in a pine for-est with a deep and sparse trunk space in situ and nu-merical experiments Agr Forest Meteorol 151 328ndash344doi101016jagrformet201011007 2011

Eder F Serafimovich A and Foken T Coherent structuresat a forest edge properties coupling and impact of sec-ondary circulations Bound-Lay Meteorol 148 285ndash308doi101007s10546-013-9815-0 2013

Foken T Micrometeorology Springer-Verlag Heidelberg 2008Foken T Meixner F X Falge E Zetzsch C Serafimovich A

Bargsten A Behrendt T Biermann T Breuninger C DixS Gerken T Hunner M Lehmann-Pape L Hens K JocherG Kesselmeier J Luumlers J Mayer J-C Moravek A PlakeD Riederer M Ruumltz F Scheibe M Siebicke L Soumlrgel MStaudt K Trebs I Tsokankunku A Welling M Wolff Vand Zhu Z Coupling processes and exchange of energy andreactive and non-reactive trace gases at a forest site ndash resultsof the EGER experiment Atmos Chem Phys 12 1923ndash1950doi105194acp-12-1923-2012 2012

Frankenberger E Untersuchungen uumlber den Vertikalaustausch inden unteren Dekametern der Atmosphaumlre Ann Meteorol 4358ndash374 1951

Friehe C A and Khelif D Fast-response aircraft temperaturesensors J Atmos Ocean Tech 9 784ndash795 doi1011751520-0426(1992)009lt0784FRATSgt20CO2 1992

Gamon J A Cheng Y Claudio H Mackinney L andSims D A A mobile tram system for systematic sampling ofecosystem optical properties Remote Sens Environ 103 246ndash254 doi101016jrse200604006 2006

Gerstberger P Foken T and Kalbitz K The Lehstenbach andSteinkreuz catchments in NE Bavaria Germany in Biogeo-chemistry of Forested Catchments in a Changing Eniviron-ment A German Gase Study Ecological Studies edited byMatzner E Springer Heidelberg 15ndash41 2004

Glickman T S Glossary of meteorology Am Meteorol SocBoston MA USA 2nd Edn 2000

Huumlbner J Olesch J Falke H Meixner F X and Foken TDocumentation and Instruction Manual for the Horizontal Mo-

bile Measuring System (HMMS) Work Report University ofBayreuth Dept of Micrometeorology ISSN 1614-8916 48 882011

Inverarity G W Correcting airborne temperature data for lagsintroduced by instruments with two-time-constant responses JAtmos Ocean Tech 17 176ndash184 doi1011751520-0426(2000)017lt0176CATDFLgt20CO2 2000

Kaimal J C Wyngaard J C Haugen D A Coteacute O RIzumi Y Caughey S J and Readings C J Tur-bulence structure in the convective boundary layer JAtmos Sci 33 2152ndash2169 doi1011751520-0469(1976)033lt2152TSITCBgt20CO2 1976

Klaassen W van Breugel P B Moors E J and Nieveen J PIncreased heat fluxes near a forest edge Theor Appl Climatol72 231ndash243 doi101007s00704-002-0682-8 2002

Langvall O and Loumlfvenius M O Effect of shelterwood densityon nocturnal near-ground temperature frost injury risk and bud-burst date of Norway spruce Forest Ecol Manag 168 149ndash161doi101016S0378-1127(01)00754-X 2002

Lee X and Black T A Atmospheric turbulence within andabove a douglas-fir stand Part II Eddy fluxes of sensibleheat and water vapour Bound-Lay Meteorol 64 369ndash389doi101007BF00711706 1993

Lenschow D H The Measurement of Air Velocity and Tem-perature Using the NCAR Buffalo Aircraft Measuring SystemNCAR-TNEDD-74 National Center for Atmospheric ResearchBoulder Colarado 9 3 pp doi105065D6C8277W 1972

Leonard R E and Eschner A R A treetop tramway system formeteorological studies Northeastern Forest Experiment StationForest Service US Dept of Agriculture US Forest Service Re-search Paper NE-92 10 pp 1968

Matlack G R and Litvaitis J A Forest edges in MaintainingBiodiversity in Forest Ecosystems edited by Hunter Jr M LCambridge University Press Cambridge UK 210ndash233 1999

Mayer J-C Hens K Rummel U Meixner F X and Fo-ken T Moving measurement platforms ndash specific challengesand corrections Meteorol Z 18 1ndash12 doi1011270941-294820090401 2009

McCarthy J A method for correcting airborne temperature datafor sensor response time J Appl Meteorol 12 211ndash214doi1011751520-0450(1973)012lt0211AMFCATgt20CO21973

McDonald D and David A N Light environments in temperateNew Zealand podocarp rainforests New Zeal J Ecol 16 15ndash22 1992

Miloshevich L M Paukkunen A Voumlmel H and Olt-mans S J Development and Validation of a Time-Lag Cor-rection for Vaisala Radiosonde Humidity Measurements JAtmos Ocean Tech 21 1305ndash1327 doi1011751520-0426(2004)021lt1305DAVOATgt20CO2 2004

Mukammal E I Some aspects of radiant energy in a pine forestTheor Appl Climatol 19 29ndash52 doi101007BF022434011971

Murcia C Edge effects in fragmented forests implications forconservation Trends Ecol Evol 10 58ndash62 doi101016S0169-5347(00)88977-6 1995

Muschinski A Frehlich R Jensen M Hugo R Hoff AEaton F and Balsley B Fine-scale measurements of tur-bulence in the lower troposphere an intercomparison be-



tween a kite- and balloon-borne and a helicopter-bornemeasurement system Bound-Lay Meteorol 98 219ndash250doi101023A1026520618624 2001

Newmark W D Tanzanian forest edge microclimatic gradi-ents dynamic patterns Biotropica 33 2ndash11 doi1016460006-3606(2001)033[0002TFEMGD]20CO2 2001

Ogawa Y and Ohara T Observation of the turbulent structurein the planetary boundary layer with a kytoon-mounted ultra-sonic anemometer system Bound-Lay Meteorol 22 123ndash131doi101007BF00128060 1982

Oncley S P Schwenz K Burns S P Sun J and Mon-son R K A cable-borne tram for atmospheric measure-ments along transects J Atmos Ocean Tech 26 462ndash473doi1011752008JTECHA11581 2009

Peacutech G Mobile sampling of solar radiation under conifers AgrForest Meteorol 37 15ndash28 doi1010160168-1923(86)90025-0 1986

Privette J L Eck T F and Deering D W Estimating spec-tral albedo and nadir reflectance through inversion of simpleBRDF models with AVHRRMODIS-like data J Geophys Res-Atmos 102 29529ndash29542 doi10102997JD01215 1997

Oumlrlander G and Langvall O The Asa shuttle ndash a system for mo-bile sampling of air temperature and radiation Scand J ForestRes 8 359ndash372 doi10108002827589309382783 1993

Rodi A R and Spyers-Duran P A Analysis of time response ofairborne temperature sensors J Appl Meteorol 11 554ndash556doi1011751520-0450(1972)011lt0554AOTROAgt20CO21972

Rodskjer N and Kornher A Eine Methode zur Registrierungder raumlumlichen Verteilung der Globalstrahlung in einemPflanzenbestand Theor Appl Climatol 15 186ndash190doi101007BF02319119 1967

Rodskjer N and Kornher A Uumlber die Bestimmung derStrahlungsenergie im Wellenlaumlngenbereich von 03ndash07 micro in Pflanzenbestaumlnden Agr Meteorol 8 139ndash150doi1010160002-1571(71)90103-8 1971

Saggin B Debei S and Zaccariotto M Dynamic error correc-tion of a thermometer for atmospheric measurements Measure-ment 30 223ndash230 doi101016S0263-2241(01)00015-X 2001

Serafimovich A Eder F Huumlbner J Falge E Voszlig L Soumlrgel MHeld A Liu Q Eigenmann R Huber K Duarte H FWerle P Gast E Cieslik S Heping L and Foken T Ex-chanGE processes in mountainous Regions (EGER) Documen-tation of the Intensive Observation Period (IOP3) June 13th toJuly 26th 2011 Work Report University of Bayreuth Dept ofMicrometeorology ISSN 1614-8916 47 137 2011

Singh A Batalin M A Stealey M Chen V Hansen M HHarmon T C Sukhatme G S and Kaiser W J Mobile robotsensing for environmental applications in Field and ServiceRobotics edited by Laugier C and Siegwart R 42 125ndash135Springer Berlin Heidelberg 2008

Staudt K and Foken T Documentation of reference data for theexperimental areas of the Bayreuth Center for Ecology and Envi-ronmental Research (BayCEER) at the Waldstein site Work Re-port University of Bayreuth Dept of Micrometeorology ISSN1614-8916 35 35 2007

Zahn A Weppner J Widmann H Schlote-Holubek K BurgerB Kuumlhner T and Franke H A fast and precise chemilumi-nescence ozone detector for eddy flux and airborne applicationAtmos Meas Tech 5 363ndash375 doi105194amt-5-363-20122012



under forest canopies to measure the areal distribution ofupwelling andor downwelling radiation (egLeonard andEschner 1968 Mukammal 1971 Brown 1973 Baldocchiet al 1984a b Peacutech 1986 Lee and Black 1993 Chenet al 1997 Privette et al 1997 Blanken et al 2001)Comparable measurements above grasslands were taken byRodskjer and Kornher(1967 1971) To study wind veloc-ity and tower-induced errors of its measurementsDabberdt(1968) built a mobile system which carried an anemometerOumlrlander and Langvall(1993) andLangvall and Loumlfvenius(2002) describe their ldquoAsa-Shuttlerdquo a mobile system whichmeasures in addition to radiation the air temperature alonga horizontal transect of decreasing shelterwood density ofNorway spruce The system ofSingh et al(2008) is able tomove in two dimensions the vertical and horizontal Theyused their system for the study of river and lake aquaticsystems The mobile system ofGamon et al(2006) car-ries a dual-detector spectrometer to measure ecosystem spec-tral reflectance The TRAM (Transect Measurement) systemof Oncley et al(2009) runs in a loop measuring wind ve-locity (ultrasonic anemometer) the CO2 concentration airtemperature and humidity They were able to create a mea-surement track through a forest ecosystem and over a creek(AmeriFlux site at Niwot Ridge) through the use of steel ca-bles and masts which allowed mobile measurements overa long horizontal distance and also changing of the verti-cal measurement height Besides the mobile platforms men-tioned above there are investigations with mobile measure-ments in aircrafts balloons lifts and other mobile plat-forms (egLenschow 1972 Kaimal et al 1976 Ogawa andOhara 1982 Balsley et al 1992 1998 Friehe and Khelif1992 Muschinski et al 2001 Mayer et al 2009)

Focusing on energy and matter exchange at a forest edgethe EGER IOP3 project (ExchanGE processes in moun-tainous Regions ndash Intense Observation Period 3) was con-ducted in a disturbed forest ecosystem where heterogeneitieswill substantially impact exchange processes between at-mosphere vegetation and soil Forest edges in particularmay have greater effects on exchange processes in a het-erogeneous forest ecosystem Driving forces at the forestedge are spatial differences of up- and downwelling radia-tion temperature moisture and the resulting wind regime(Murcia 1995 Matlack and Litvaitis 1999 Davies-Colleyet al 2000 Klaassen et al 2002) Forest edges and prevail-ing gradients were targeted in a variety of micrometeorolog-ical studiesChen et al(1993 1995) focused on a Douglas-fir forest (Pacific Northwest USA)Dawson and Sneddon(1969) McDonald and David(1992) as well asDavies-Colley et al(2000) investigated the edge of rain forest nextto a clearing in New ZealandNewmark(2001) studied fourforest edges in the West and East Usambara Mountains (Tan-zania) Even flux measurements at forest edges and numeri-cal studies have been performed (seeDupont et al 2011 fora detailed overview) First results of the EGER IOP3 projectare presented inEder et al(2013) showing an increased

number of coherent structures at the forest edge which mightbe an indication of a quasi-stationary secondary circulationA detailed description of secondary circulations can be foundin Glickman(2000) and for site-specific conditions seeEderet al(2013)

The horizontal mobile measuring system (HMMS) pre-sented in this paper is a fully automatic system movingon rails The chosen measuring transect for EGER IOP3was a straight line installed perpendicular to the forestedge However a greater variation in the measuring heightalong the track is limited because of the wooden substruc-ture and also because of the limited climbing ability of theHMMS (maximum climbing anglelt 10 ) The HMMSwas equipped with long-wave and short-wave radiation sen-sors (individual sensors for up- and downwelling radiation)a sensor for air temperature and relative humidity and analy-sers for CO2 and O3 concentrationsFoken et al(2012) hadshown that trace gas concentrations are good tracers for co-herent structures and coupling regimes and we expected tofind sources or sinks in the horizontal structures which mightbe an indication for the influence of vertical structures suchas coherent structures or secondary circulations The selec-tion of appropriate sensors was a major challenge because of(a) the limited power of the drive mechanism leading to pay-load restrictions for sensors and electronics (b) the space onthe carrier platform restricting the size of sensors analysersand electronics and (c) the requirement for short responsetimes in order to capture the true spatial variations of mete-orological quantities and trace gas concentrations (with thehighest variations occurring mainly at the forest edge) withinan acceptable time Besides the technical description of theHMMS (Sect21) we present the results of laboratory testsfor the individual sensor response times some short time se-ries of the measurements and the application of a correctionalgorithm (Sect22) to correct the measurements for the re-sponse time induced dynamical error

2 Materials and methods

21 Description of HMMS

The technical description of the HMMS comprises details ofthe drive mechanism and the design of the carrier the soft-ware and the sensor system The front and lateral view of theHMMS are shown in Fig1 with the corresponding descrip-tions listed in Table1 Table2 contains the sensorsrsquo detailsFor more technical details of the HMMS including technicaldrawings pictures and the detailed wiring plan seeHuumlbneret al(2011)

211 Drive mechanism and carrier design

The drive mechanism of the HMMS is from a commer-cial garden railway system (Lehmann-Groszlig-Bahn LGBsince 2007 part of Gebr Maumlrklin amp Cie GmbH


























213 Sensor system







+1Xi minus1Xi
















dt (1)


X(t) = Xinfin

(1minus eminus

tτ

) (2)



Xi(t) = 0 for t le 0 (3)

and




dt (5)



tτ

) (6)




















Time [s]

Nor

mal

ised

diff

eren

ce [minus

]

τ63

0 10 20 30 40 50

00

02

04

06

08

10

τCO2

1s τKτL4s

τT12s

τRH20s







1113

15

T [deg

C]


8690

94

Position [m]

RH

[]

FOREST CLEARING

0 50 100 150Position [m]

Rel

ativ

e hu

mid

ity (

) CLEARING FOREST

100 50 0









1400 1500 1600 1700 1800

100

9080

7060

50

Time of day [CET]

Pos

ition

[m]

Kdarr [Wmminus2](a)

0

200

400

600

800

1000

1400 1500 1600 1700 1800

100

9080

7060

50

Time of day [CET]

Pos

ition

[m]

Kdarr [Wmminus2](b)

0

200

400

600

800

1000

1400 1500 1600 1700 1800

100

9080

7060

50

Time of day [CET]

Pos

ition

[m]

Ldarr [Wmminus2](c)

360

380

400

420

440

460

480

500

1400 1500 1600 1700 1800

100

9080

7060

50

Time of day [CET]

Pos

ition

[m]

Ldarr [Wmminus2](d)

360

380

400

420

440

460

480

500

1
























4 Conclusions













Edited by U Friess

References
























































































213 Sensor system







+1Xi minus1Xi
















dt (1)


X(t) = Xinfin

(1minus eminus

tτ

) (2)



Xi(t) = 0 for t le 0 (3)

and




dt (5)



tτ

) (6)




















Time [s]

Nor

mal

ised

diff

eren

ce [minus

]

τ63

0 10 20 30 40 50

00

02

04

06

08

10

τCO2

1s τKτL4s

τT12s

τRH20s







1113

15

T [deg

C]


8690

94

Position [m]

RH

[]

FOREST CLEARING

0 50 100 150Position [m]

Rel

ativ

e hu

mid

ity (

) CLEARING FOREST

100 50 0









1400 1500 1600 1700 1800

100

9080

7060

50

Time of day [CET]

Pos

ition

[m]

Kdarr [Wmminus2](a)

0

200

400

600

800

1000

1400 1500 1600 1700 1800

100

9080

7060

50

Time of day [CET]

Pos

ition

[m]

Kdarr [Wmminus2](b)

0

200

400

600

800

1000

1400 1500 1600 1700 1800

100

9080

7060

50

Time of day [CET]

Pos

ition

[m]

Ldarr [Wmminus2](c)

360

380

400

420

440

460

480

500

1400 1500 1600 1700 1800

100

9080

7060

50

Time of day [CET]

Pos

ition

[m]

Ldarr [Wmminus2](d)

360

380

400

420

440

460

480

500

1
























4 Conclusions













Edited by U Friess

References



































































+1Xi minus1Xi
















dt (1)


X(t) = Xinfin

(1minus eminus

tτ

) (2)



Xi(t) = 0 for t le 0 (3)

and




dt (5)



tτ

) (6)




















Time [s]

Nor

mal

ised

diff

eren

ce [minus

]

τ63

0 10 20 30 40 50

00

02

04

06

08

10

τCO2

1s τKτL4s

τT12s

τRH20s







1113

15

T [deg

C]


8690

94

Position [m]

RH

[]

FOREST CLEARING

0 50 100 150Position [m]

Rel

ativ

e hu

mid

ity (

) CLEARING FOREST

100 50 0









1400 1500 1600 1700 1800

100

9080

7060

50

Time of day [CET]

Pos

ition

[m]

Kdarr [Wmminus2](a)

0

200

400

600

800

1000

1400 1500 1600 1700 1800

100

9080

7060

50

Time of day [CET]

Pos

ition

[m]

Kdarr [Wmminus2](b)

0

200

400

600

800

1000

1400 1500 1600 1700 1800

100

9080

7060

50

Time of day [CET]

Pos

ition

[m]

Ldarr [Wmminus2](c)

360

380

400

420

440

460

480

500

1400 1500 1600 1700 1800

100

9080

7060

50

Time of day [CET]

Pos

ition

[m]

Ldarr [Wmminus2](d)

360

380

400

420

440

460

480

500

1
























4 Conclusions













Edited by U Friess

References






































































dt (1)


X(t) = Xinfin

(1minus eminus

tτ

) (2)



Xi(t) = 0 for t le 0 (3)

and




dt (5)



tτ

) (6)




















Time [s]

Nor

mal

ised

diff

eren

ce [minus

]

τ63

0 10 20 30 40 50

00

02

04

06

08

10

τCO2

1s τKτL4s

τT12s

τRH20s







1113

15

T [deg

C]


8690

94

Position [m]

RH

[]

FOREST CLEARING

0 50 100 150Position [m]

Rel

ativ

e hu

mid

ity (

) CLEARING FOREST

100 50 0









1400 1500 1600 1700 1800

100

9080

7060

50

Time of day [CET]

Pos

ition

[m]

Kdarr [Wmminus2](a)

0

200

400

600

800

1000

1400 1500 1600 1700 1800

100

9080

7060

50

Time of day [CET]

Pos

ition

[m]

Kdarr [Wmminus2](b)

0

200

400

600

800

1000

1400 1500 1600 1700 1800

100

9080

7060

50

Time of day [CET]

Pos

ition

[m]

Ldarr [Wmminus2](c)

360

380

400

420

440

460

480

500

1400 1500 1600 1700 1800

100

9080

7060

50

Time of day [CET]

Pos

ition

[m]

Ldarr [Wmminus2](d)

360

380

400

420

440

460

480

500

1
























4 Conclusions













Edited by U Friess

References















































































Time [s]

Nor

mal

ised

diff

eren

ce [minus

]

τ63

0 10 20 30 40 50

00

02

04

06

08

10

τCO2

1s τKτL4s

τT12s

τRH20s







1113

15

T [deg

C]


8690

94

Position [m]

RH

[]

FOREST CLEARING

0 50 100 150Position [m]

Rel

ativ

e hu

mid

ity (

) CLEARING FOREST

100 50 0









1400 1500 1600 1700 1800

100

9080

7060

50

Time of day [CET]

Pos

ition

[m]

Kdarr [Wmminus2](a)

0

200

400

600

800

1000

1400 1500 1600 1700 1800

100

9080

7060

50

Time of day [CET]

Pos

ition

[m]

Kdarr [Wmminus2](b)

0

200

400

600

800

1000

1400 1500 1600 1700 1800

100

9080

7060

50

Time of day [CET]

Pos

ition

[m]

Ldarr [Wmminus2](c)

360

380

400

420

440

460

480

500

1400 1500 1600 1700 1800

100

9080

7060

50

Time of day [CET]

Pos

ition

[m]

Ldarr [Wmminus2](d)

360

380

400

420

440

460

480

500

1
























4 Conclusions













Edited by U Friess

References

































































1400 1500 1600 1700 1800

100

9080

7060

50

Time of day [CET]

Pos

ition

[m]

Kdarr [Wmminus2](a)

0

200

400

600

800

1000

1400 1500 1600 1700 1800

100

9080

7060

50

Time of day [CET]

Pos

ition

[m]

Kdarr [Wmminus2](b)

0

200

400

600

800

1000

1400 1500 1600 1700 1800

100

9080

7060

50

Time of day [CET]

Pos

ition

[m]

Ldarr [Wmminus2](c)

360

380

400

420

440

460

480

500

1400 1500 1600 1700 1800

100

9080

7060

50

Time of day [CET]

Pos

ition

[m]

Ldarr [Wmminus2](d)

360

380

400

420

440

460

480

500

1
























4 Conclusions













Edited by U Friess

References













































































4 Conclusions













Edited by U Friess

References







































































Edited by U Friess

References
































































a horizontal mobile measuring system for atmospheric quantities … · 2020. 7. 16. · received:...

Documents