Measuring soil respiration in the field different chamber designs

Measuring soil respiration in the field different chamber designs

Introduction Earliest studies on soil respiration were conducted by incubating soil cores in laboratory (Lundegårdh 1922). Later a need for undisturbed and continuous measurements. Chambers placed over the soil is the most direct way of measuring gas exchange within the soil surface. A variety of chamber systems has been developed. No single method has emerged as preferable.

Open dynamic chambers (=Steady-state flowthrough chambers) q1 q2 C 1 C 2 The flux is calculated from the difference in concentrations between the air flowing at a known rate through the inlet and outlet of the chamber after the chamber headspace concentration has reached an equilibrium. CO2 efflux (g m -2 h -1 ) 1.2 1.8.6.4.2 39 385 38 375 37 365 36 5 1 15 2 25 3 35 Time (s) CO2 (µ mol mol -1 ) q2c2 q1c A 1 F = where F is gas flux q 1 is gas flow into the chamber q 2 is gas flow out of the chamber C 1 and C 2 are gas concentrations in the outflow and inflow respectively A is surface area

Closed dynamic chambers (=Non-steady state flow-through chambers) Gas analyser The flux is calculated from the concentration change within the chamber headspace. Saturation F = d( V C) dt A CO2 conc. (µmol mol -1 ) 14 12 1 8 6 4 2 25 5 75 1 125 where F is gas flux C is gas concentration V is volume of the chamber A is surface area dt is incubation time Time (minutes)

41 45 Compensation air (C) Chamber air (Ci) C a Q 1 Q 2 CO2 (ppm) 4 395 39 385 38 5 C i C Steady state F Soil C s Q 3 Q 4 F = Q2 Ci QC 1 + Q3C s Q4C i Q5C i + Q6 5 1 15 2 25 Time (seconds) Q 5 Q 6 C a C a 45 C i CO2 (ppm) 4 35 Soil F C s Q 3 Q 4 3 5 1 15 2 25 Time (seconds) ( V C ) ( t) c i F = + Q3C s Q4Ci Q5Ci + 6 Q C a Figure 2.1. (Upper panel) The CO 2 concentration in an open chamber depends on CO 2 concentrations and flow rates of the incoming and outgoing air flows. In addition, possible air flows (Q 3 and Q 4 ) between the soil air space and the chamber as well as between the ambient air and the chamber (Q 5 and Q 6 ) can generate additional mass flow of CO 2 in an out of the chamber. When a steady-state concentration in the chamber has been reached, the CO 2 efflux from soil (F) can be determined from the mass balance equation shown. (Lower panel) In a closed chamber (dynamic or static) the CO 2 efflux rate can be calculated from the slope of the CO 2 concentration increase within the chamber. Similarly, possible air flows between the soil air space and the chamber (Q 3 and Q 4 ) as well as between the ambient air and the chamber (Q 5 and Q 6 ) can generate additional mass flow of CO 2 in an out of the chamber. When designing both chamber types, air flows of type Q 3 -Q 6 should be avoided.

Static chambers (Non-steady-state non-flowthrough chambers) The CO 2 is trapped with chemicals (NaOH and soda lime) The concentration within the chamber remains quite stable. The CO 2 efflux can be calculated from the amount of CO 2 in the trapping solution. In some chambers, the CO 2 concentration is determined by air samples drawn into syringes and analysed separately with IR or GC. This principle is very close to that of closed dynamic chamber.

Gradient method CO 2 concentration in the soil air space is often an order of magnitude higher than in the atmosphere. According to Fick s first law the gas flux is dependent on the concentration gradient and the diffusivity of the soil.

The gradient method: an example Soil can be divided into distinct soil layers according to podzolic soil horizons. Gas transport in the soil is driven mainly by diffusion. CO CA J AO = DAO ( l + l ) / A 2 D is the diffusion coefficient of the soil, C O and C A are the amounts of CO 2 in the soil air space, L is the distance between soil layers. Soil porosity affects the gas movement within the soil. D Eg u h = ( ) D 1 u o E g is the air filled porosity of soil (m 3 m -3 ) and u and h are empirical parameters, D is the diffusion coefficient of CO 2 in soil and D, the diffusion coefficient in air. O v O T O v A T A rr rm rr rm Atmosphere CO 2 J OA O-horizon, (litter+humus) CO 2, (C O ) J AO A-horizon CO 2, (C A ) B-horizon CO 2, (C B ) l O E o O l A E o A C-horizon CO 2, (C C )

Gradient method We installed Vaisala GMP343 CO 2 sensors in different soil horizons for determining the concentration gradient Soil moisture was measured with TDR and soil temperature with thermistors. Humus.5 m Diffusion in the soil was calculated with Fick s first law of diffusion..12 m A-horizon B-horizon.54 m.11m J H _ ATM = D H C ATM l H C / 2 H.1 m C-horizon.1 m GMP343 CO 2 probe TDR Temperature sensor

CO2 concentration (µmol mol -1 ) 1 9 8 7 6 5 4 3 2 1 Gradient method Soil air CO 2 concentration had a clear diurnal and seasonal pattern following soil temperature and soil moisture. The CO 2 effluxes calculated with the gradient method matched reasonably well with the fluxes measured with chambers. 18-Jun 12: CO2 concentration in air CO2 concentration in humus Gradient based efflux Measured efflux with transparent chamber 19-Jun : 19-Jun 12: 2-Jun : 2-Jun 12: 21-Jun : 21-Jun 12:.9.8.7.6.5.4.3.2.1 CO2 efflux ( g m -2 h -1 ) Temperature ( o CO2 efflux (g CO2 m -2 h -1 ) Volumetric water content (m 3 m -3 ) C) CO2 concentration ( µmol mol -1 ) 4 3 2 1 2 16 12 8 4.6.5.4.3.2.1 1.8.6.4.2 31-Mar 31-Mar 31-Mar 31-Mar 1-Apr 1-Apr 1-Apr Humus A-horizon B-horizon C-horizon Humus A-horizon B-horizon C-horizon Humus A-horizon B-horizon C-horizon 1-Apr 2-Apr 2-Apr 2-Apr 2-Apr 3-Apr 3-Apr 3-Apr 3-Apr 1-May 1-May 1-May Gradient based CO2 efflux Night time efflux measured with chamber Predicted CO2 efflux in chamber Spatial measurements 1-May 2-May 2-May 2-May 2-May 3-May 3-May 3-May 3-May 9-Jun 9-Jun 9-Jun 9-Jun 19-Jun 19-Jun 19-Jun 19-Jun (a) (b) (c) (d) 29-Jun 29-Jun 29-Jun 29-Jun

CO 2 concentration, ppm N 2 O concentration, ppm 6 1 2 3 4 5 6 7 4.348.352.356.36.364.368 29 Aug 22 4 2 2-2 Soil depth, cm -2-4 -6 Efflux Conc -8..1.2.3.4.5.6 Soil depth, cm -4-6 -8 4 2-2 -4-6 Flux Conc 7 Oct 22 CO 2 efflux, g m -2 h -1 Mean soil CO 2 concentrations and CO 2 fluxes from each soil layer over all four pits and the whole measurement period 7th June 22 31st July 23. -8 4 2-2 1 Nov 22-4 Nitrous oxide concentration in the soil profile and N 2 O fluxes from the soil in autumn 22. -6-8 -4-2 2 4 6 8 1 12 N 2 O flux, µg N m -2 h -1 Pihlatie et al. 27

Chambers; advantages and drawbacks + Relatively easy to use + Spatial coverage + Low cost? + Fast measurements + Can be used for measuring photosynthesis and evapotranspiration simultaneously with respiration. + Several gases can be measured simultaneously Chambers affect the flux being measured (collar problem, pressure problem, saturation) Chambers may change the conditions in the soil if left at the same place for a longer period of time Difficult to use in winter Differences between chamber types requires calibration

Gradient method; advantages and drawbacks + As soon as the soil has been stabilized after the installation, the measurement itself does not disturb the CO 2 fluxes significantly. + The source of CO 2 efflux can be estimated based on the concentration in different layers. + Gradient method has a good potential for wintertime measurements Soil porosity is a critical factor when using the gradient method. The diffusion coefficient of the soil is affected by the air-filled pore space as well as the continuity and shape of the pores (Glinski and Stepniewski 1985). Because the diffusion of CO 2 in water is about 1 times slower than in air, soil water content has a substantial effect on CO 2 movement. Continuous soil CO 2 concentration measurements have been difficult due to the lack of robust sensors.

Commercially available systems PP-systems SRC-1+EGM Infrared analyser (Technical specifications retrieved from PP-systems home page http:\\www.ppsystems.com) Dimensions: 1 mm Ø x 15 mm Weight: 9 g Voltage: 12V Working principle: Closed dynamic chamber without CO 2 scrubbing. CO 2 efflux is calculated from the concentration change using linear and non-linear fitting. Air mixing: Fan inside the chamber Water vapour correction: No Soil temperature probe: Yes

Commercially available systems Li-Cor 64-9 +Li-Cor LI-64 portable photosynthesis System (Technical specifications retrieved from Li-Cor home page http:\\www.licor.com) Li-Cor Inc. Dimensions: diameter 95 mm, volume 991 cm 3, surface area 76.1 cm 2 Weight: 18 g Voltage: 12V Working principle: Closed dynamic chamber with CO 2 scrubbing. Before each cycle of flux measurement, air in the chamber headspace was scrubbed down 3-4 ppm below the ambient CO 2 concentration (depending on the flux), and was then allowed to rise as a consequence of CO 2 efflux from the tank. Air mixing: Fan is used to push the air through a perforated manifold to distribute the air evenly withing the chamber without causing localized pressure gradients. Water vapor correction: Available in Li-Cor 64 Soil temperature probe: Yes

Commercially available systems Vaisala Carbocap Non-dispersive Infrared sensors (GMP22 series and GMP-343 series) + M7 Measurement indicator GMP-22 series (accurracy 2% of reading) and GMP-343 series (accurracy 1.5% of reading at the calibration points, below 3 ppm ± 5 ppm) Can be connected to any chamber or installed directly in the soil. Water vapour, temperature and pressure compensations online when using external RH or pressure sensor. Oxygen compensation also available. ADC, LCA-2, Analytical Development Company Ltd. (Hoddesdon, UK)

Calibration system The calibration system is modified from the system developed by Widén and Lindroth (23). The calibration system consists of a steel chamber with a layer of quartz sand on top. CO 2 inside the chamber is monitored continuously. CO 2 efflux is calculated from the concentration change inside the chamber. Gas analyser CO 2 Chamber being tested FANS Gas analyser

During the summer 22 we calibrated 2 soil respiration chambers from 13 institutes across Europe and USA. Closed dynamic systems (=non-steadystate flow-through systems) q PP Systems SRC-1+ EGM-1, SRC-1+ EGM-3 and SRC-1+EGM-4 q Li-Cor 64-9 (Weizmann Institute of Science, Maz Planck Institute) q University of Bayreuth (Reth et al.) q Woods Hole Research Center (Savage et al.) q Max Planck Institute (Anthoni et al.) q Finnish Meteorological Institute (Lohila et al.) q University of Helsinki (Kolari, Minkkinen) Open dynamic systems (=steady-state flow-through systems) University of Bayreuth (Subke et al.) q Kutsch (1996) q University of Helsinki (Hybrid system) Closed static systems (=non-steady-state non-flow-through systems) q University of Helsinki (Pumpanen et al.) q University of Joensuu q Agrifood Finland

Calibrations Calibrations were carried out with dry coarse sand (.6 mm), fine sand (.5-.2 mm) and fine sand with.25 m 3 m -3 water content. Total porosities were 47% and 53% for coarse and dry sand respectively. Each chamber was calibrated with 6-7 flux rates 2 replicate measurements were done on 1-3 collars with each flux level.

How the effluxes were generated? CO2 concentration (ppm) CO2 concentration (ppm) 77 y = 759.23e -.6x 76 R 2 =.9985 75 74 73 72 71 1 2 3 4 5 6 7 Time (min) 163 162 y = 1616.2e -.9x 161 R 2 =.9979 16 159 158 157 156 155 154 153 1 2 3 4 5 6 7 Time (min) Exponential curve was fitted to concentration and time. Flux can be calculated from the concentration change in time Q = (dc / dt) * V / A Where Q is flux C is concentration in the chamber V is volume of the chamber A is surface area of the chamber CO2 efflux (g CO 2 m -2 h -1 ) 1.8 1.6 1.4 1.2 1.8.6.4.2 y =.2x -.275 R 2 =.9975 2 4 6 8 1 12 CO 2 concentration (ppm)

How the effluxes were generated? Repeatability of the effluxes. The effluxes generated during different weeks at similar temperatures deviated less than 6-7% from each other. The effluxes were also spatially very homogeneous. The standard error between the three collars used in the measurements ranged from.6 to.173 with effluxes ranging from.35 to 1 µmolm -2 s -1 (=.6 to 1.58 g m -2 h -1 ). Reference flux from the calibration chamber (µmol CO2 m -2 s -1 ) 1 8 6 4 2 Weizmann Institute of Science University of Bayreuth Woods Hole Research Center Max Planck Institute University of Kiel 2 4 6 8 1 CO 2 concentration in the calibration chamber (ppm)

Results: non-steady-state flow-through systems (=closed dynamic systems) Li-Cor 64 chamber showed effluxes closed to the effluxes generated by the calibration system. Measured flux (g CO2 m -2 h -1 ) 1,8 1,5 1,2,9,6,3 Coarse sand Dry fine sand Wet fine sand Linear (1:1-line) Linear (Coarse sand) Linear (Dry fine sand) Linear (Wet fine sand) y = 1,81x y = 1.92x y = 1.49x Li-Cor 64, Weizmann Institute of Science,3,6,9 1,2 1,5 1,8 Reference flux (g CO 2 m -2 h -1 ) Measured flux (g CO2 m -2 h -1 ) 1,8 1,5 1,2,9,6,3 Coarse Dry fine Wet fine y = 1,127x y = 1,885x y = 1,883x Li-Cor 64, Max Planck Institue,3,6,9 1,2 1,5 1,8 Reference flux (g CO 2 m -2 h -1 )

Results: non-steady-state flow-through systems (=closed dynamic systems) Measured flux (g CO2 m -2 h -1 ) 2, 1,6 1,2,8,4, Coarse Dry fine Wet fine y = 1.2147x y = 1.2738x y = 1.516x EGM-3 + SRC-1,4,8 1,2 1,6 2 Measured flux (g CO2 m -2 h -1 ) 1,8 1,5 1,2,9,6,3 Coarse Dry fine y =,855x y =,9957x EGM-3 + SRC-1 widened lower part,3,6,9 1,2 1,5 1,8 Reference flux ( g CO 2 m -2 h -1 ) Reference flux (g CO 2 m -2 h -1 ) PP-systems overestimated effluxes on all soil types if a collar was used. If collar was not used, the effluxes were closer to the reference. Chamber with widened lower part underestimated effluxes with coarse sand. Measured flux (g CO2 m -2 h -1 ) 1,8 1,5 1,2,9,6,3 Coarse Dry fine Wet fine y = 1.345x y = 1.1891x y =.9437x EGM-4 + SRC-1 no collar,3,6,9 1,2 1,5 1,8 Reference flux (g CO 2 m -2 h -1 )

Woods Hole Research center (Savage s system) University of Bayreuth (Reth s system)

FMI - system (Finnish Meteorological Institute)

Results: non-steady-state flow-through systems (=closed dynamic systems) Non-steady-state flow-through systems showed contradictory results depending on the design of the chamber, and if a fan was used or not. University of Bayreuth and FMI system showed effluxes close to the reference. Woods Hole and Max Planck underestimated. Measured flux (g CO2 m -2 h -1 ) Measured flux (g CO2 m -2 h -1 ) 1,8 1,5 1,2,9,6,3 1,8 1,5 1,2,9,6,3, Coarse Dry fine Wet fine Linear (1:1-line) Linear (Coarse) Linear (Dry fine) Linear (Wet fine) y =.9641x y =.8855x y =.9632x,3,6,9 1,2 1,5 1,8 Coarse Dry fine Wet fine Reference flux (g CO 2 m -1 h -1 ) y =.827x y =.97x y =.825x University of Bayreuth Woods hole R.C.,3,6,9 1,2 1,5 1,8 Reference flux (g CO 2 m -2 h -1 ) Measured flux (g CO2 m -2 h -1 ) Measured flux (g CO2 m -2 h -1 ) 1,8 1,5 1,2,9,6,3, 1,8 1,5 1,2,9,6,3 Coarse y = 1.337x Dry fine y = 1.72x Wet fine y = 1.8x Finnish meteorological Institute,3,6,9 1,2 1,5 1,8 Reference flux (g CO 2 m -2 h -1 ) Coarse y =.854x Dry fine y =.824x Wet fine y =.7862x Max Planck Institute,3,6,9 1,2 1,5 1,8 Reference flux (g CO 2 m-2 h-1)

Results: non-steady-state flow-through systems (=closed dynamic systems) Measured efflux (g CO2 m -2 h -1 ) 1,8 1,5 1,2,9,6,3, Coarse Dry fine Wet fine y = 1,38x y =,8484x y =,8746x EGM-3+University of Helsinki, Kari Minkkinen,3,6,9 1,2 1,5 1,8 Reference efflux (g CO 2 m -2 h -1 ) Measured CO 2 (g CO2 m -2 h -1 ) 1,8 1,5 1,2,9,6,3 Fine dry Fine wet Linear (1:1-line) Linear (Fine dry) Linear (Fine wet) y = 1,131x y =,931x EGM-3+University of Helsinki, Pasi Kolari,3,6,9 1,2 1,5 1,8 Reference flux (g CO 2 m -2 h -1 )

Non-steady-state non-flow-through systems(=closed static systems) NSNF-system (Agrifood Finland) NSNF-system (University of Helsinki, Jukka Pumpanen)

Non-steady-state non-flow-through systems(=closed static systems) NSNF-chambers underestimated effluxes on average by 4-13%. The underestimation was smallest with wet sand. Incubation time and headspace concentration affected the results. Measured flux (g CO2 m -2 h -1 ) Measured efflux (g CO2 m -2 h -1 ) 1,8 1,5 1,2,9,6,3 1,8 1,5 1,2,9,6,3 Coarse Dry fine Wet fine Linear (1:1-line) Linear (Coarse) Linear (Dry fine) Linear (Wet fine) y =.982x y =.9378x y =.8451x,3,6,9 1,2 1,5 1,8 Coarse 3 min Dry fine 3 min Wet fine 3 m in University of Joensuu Reference flux (g CO 2 m -2 h -1 ) y =,8452x y =,8462x y =,8991x Agrifood Finland 3 min,3,6,9 1,2 1,5 1,8 Reference flux ( g CO 2 m -2 h -1 ) Measured flux (g CO2 m -2 h -1 ) Measured flux (g CO2 m -2 h -1 ) 1,8 1,5 1,2,9,6,3 1,8 1,5 1,2,9,6,3 Coarse 1 min y =.9625x Dry fine 1 min y =.9563x Wet fine 1 min y =.9492x Agrifood Finland 1 min,3,6,9 1,2 1,5 1,8 Reference flux (g CO 2 m -2 h -1 ) Coarse y = 1.64x Dry fine y =.8197x Wet fine y =.8538x University of Helsinki,3,6,9 1,2 1,5 1,8 Reference flux (g CO 2 m -2 h -1 )

Steady-state flow-through systems(=open dynamic systems) University of Bayreuth (Subke s system similar to Rayment s system) University of Helsinki

Steady-state flow-through systems (=open dynamic systems) Steady-state systems showed effluxes close to the reference. The hybrid system of the University of Helsinki overestimated effluxes with dry fine sand. Measured flux (g CO2 m -2 h -1 ) Measured flux (g CO2 m -2 h -1 ) 1,8 1,5 1,2,9,6,3 Coarse Dry fine Wet fine y = 1.262x y =.963x y = 1.88x,,,3,6,9 1,2 1,5 1,8 2 1,6 1,2,8,4 Coarse Dry fine Wet fine University of Bayreuth Reference flux (g CO 2 m -2 h -1 ) y = 1.83x y = 1.1871x y = 1.47x University of Helsinki (Hybrid system),4,8 1,2 1,6 2 Measured flux (g CO2 m -2 h -1 ) 1,8 1,5 1,2,9,6,3 Coarse y = 1.56x Dry fine y = 1.83x Wet fine y =.9468x University of Kiel,3,6,9 1,2 1,5 1,8 Reference flux (g CO 2 m -2 h -1 ) Reference flux (g CO 2 m -2 h -1 )

Conclusions Reliability of the chamber system was not related to the measurement principle. Good results can be achieved with steady-state chambers as well as with non-steady-state chambers. General trend was that non-steady-state non-flow-through chambers underestimated systematically by 4-14% whereas significant differences between flow-through chambers were not observed. Soil porosity affected the results, probably due to mass flow within the soil beneath the chamber. Special attention should be paid to the mixing of air within the chamber. Excessive turbulence may cause mass flow of CO 2 between soil and the chamber. However, some kind of mixing is needed in non-steady-state chambers, because the CO 2 concentration has to be evenly distributed within the chamber headspace. Headspace concentration should be as close as ambient possible, to avoid altering the concentration gradient between the soil and the chamber.

Summary table Table 1. Correction factors for different chambers. Each chamber can be scaled to the reference flux obtained from the calibration tank by dividing the measured flux by the correction factor of a specific soil type. Chamber type* Coarse 95% confidence Dry fine 95% confidence Wet fine 95% confidence sand interval sand interval sand interval NSF-1 (Licor 64-9) 1.1.99-1.3 1.1.98-1.4 1.5 1.1-1.9 NSF-1b (Licor 64-9) 1.13 1.7-1.18 1.9.98-1.19 1.9 1.4-1.14 NSF-2 (EGM-3+SRC-1) 1.21 1.17-1.26 1.27 1.15-1.39 1.5.97-1.13 NSF-3 (EGM-3+SRC-1 widened collar).86.82 -.89 1..94-1.5 - - NSF-4 (EGM-1+SRC-1 no collar) 1.3 1.1-1.6 1.19 1.14-1.24.94.86-1.3 NSF-5 (EGM-4+SRC-1 mesh) 1.16 1.12-1.19 1.19 1.11-1.27 1.33 1.2-1.47 NSF-6 (University of Bayreuth).96.91-1.2.89.86 -.92.96.87-1.6 NSF-7 (Finnish Meteorological Institute) 1.3 1.1-1.5 1.7.99-1.15 1..92-1.8 NSF-8 (Woodshole Research Center).83.79 -.86.91.86 -.96.83.8 -.85 NSF-9 (Max Planck Institute).81.79 -.83.8.79 -.82.79.77 -.8 NSF-1 (University of Helsinki) 1.1.96-1.5 1.19 1.14-1.23 1.4.96-1.13 NSF-11 (University of Helsinki) 1..96-1.3.85.81 -.87.87.84 -.89 NSF-12 (University of Helsinki) - - 1.13 1.8-1.18.93.87 -.99 NSF- Average 1. 1.4.99 NSNF-1 (University of Joensuu).98.95-1.1.94.89 -.98.85.81 -.88 NSNF-1 (University of Joensuu with extension).95.86-1.5.98.92-1.3.85.75 -.94 NSNF-2 (Agrifood Research Finland, 1 min)..96.91-1.1.96.76-1.15.95.84-1.6 NSNF-2 (Agrifood Research Finland, 3 min)..85.79 -.9.85.71 -.98.9.8-1. NSNF-3 (University of Helsinki) 1.6.96-1.17.82.63-1.1.85.78 -.93 NSNF-4 (University of Helsinki) - -.65.56 -.74.84.81 -.87 NSNF- Average.96.86.87 SSFL-1 (University of Bayreuth) 1.3 1.1-1.5.96.92-1.1 1.9 1.2-1.15 SSFL-2 (University of Kiel) 1.5.99-1.11 1.8 1.1-1.15.95.8-1.9 SSFL - Average 1.4 1.2 1.2 * NSF = non-steady-state flow-through chamber, NSNF = non-steady-state non-flow-through chamber SSFL = steady-state flow-through chamber

Thank you!