Long-Term Dynamics of Production, Respiration, and Light-Use Efficiency in Two Sagebrush Steppe Ecosystems of the U.S. Intermountain West in Relation to NDVI: Scaling-Up CO2 Fluxes

We continuously measured the net ecosystem CO2 exchange (Fc) of two sagebrush steppe ecosystems in the USDA-ARS AgriFlux network (Burns, OR, 1995-2001 and Dubois, ID, 1996-2001) using the Bowen ratio - energy balance technique. For all sites and years, photosynthetically active radiation (Q) was the most important driver of the CO2 flux during daytime, followed by soil temperature (Ts), as described by ecosystem-scale light-response functions Fc(Q, a1,...,am) and Fc(Q,Ts, a1,...,an), where {ai} are empirical parameters. Using nonlinear identification of the light-response functions parameters, we partitioned Fc into two components, gross primary productivity (Pg) and total ecosystem respiration (Re), thus Fc = Pg - Re (physiological sign convention). For years when winter-time measurements were not available, winter-time fluxes were gap-filled using relationships established from data collected during 1999/2000 and 2000/2001 winter periods. Seven year average annual gross primary production (GPP) at the Burns site was 1111 (range 475-1716) g CO2/m2/yr, average ecosystem respiration (RE) was 1029 (range 581-1544) g CO2/m2/yr, and average net ecosystem CO2 exchange (NEE) was 82 (range -424 to +621) g CO2/m2/yr. At the more productive Dubois site, six year average GPP was 1602 (range 963-2162) g CO2/m2/yr, RE was 1349 (range 786-1778), and NEE was 273 (range -118 to +867) g CO2/m2/yr (Table 1).

Table 1. CO2 budget (GPP,RE,PDAY,NEE), light-use efficiency (LUEwk), and hydrologic year precipitation (PCPNhyd) in two sagebrush steppe ecosystems in different years ------------------------------------------------------------- Year GPP RE PDAY NEE LUEwk PCPNyd g CO2/m2/yr mmolCO2/mol quanta mm ------------------------------------------------------------- Burns, OR 1995 1147 1074 352 73 6.0 360 1996 1120 1544 272 -424 8.3 336 1997 1343 918 833 426 6.8 334 1998 1716 1095 1111 621 10.4 569 1999 1152 924 660 227 5.8 394 2000 828 1069 169 -245 6.3 268 2001 475 581 262 -106 4.4 205 Average 1111 1029 523 82 6.9 352

Dubois, ID 1996 1185 1176 623 9 6.6 244 1997 1892 1778 903 114 12.6 361 1998 2162 1690 1323 472 13.6 396 1999 2075 1209 1439 867 14.3 307 2000 1335 1453 471 -118 12.5 210 2001 963 786 598 177 6.8 181 Average 1602 1349 893 253 11.1 283 ----------------------------------------------------------- Note: GPP, RE, PDAY, and NEE are annual integrals of instantaneous fluxes Pg, Re, Pd, and Fc, respectively.

Over the whole observation period, both ecosystems were net sinks for atmospheric CO2: at the Burns site net sequestration during 1995-2001 period was 574 g CO2/m2, and at the Dubois site net sequestration during 1996-2001 period was 1518 gCO2/m2.

At both sites, period of measurements included both wet and dry years, and we found precipitation as important driver of primary productivity: highest GPP was obtained in 1998 when the precipitation total for the hydrologic year (Oct. 1 1997 to Sep. 30 1998) was highest, whereas the lowest GPP occurred in 2001 when the precipitation for hydrologic year was lowest (Table 1). Coefficient of correlation r(PCPNhyd,GPP) between hydrologic year precipitation and gross primary production was 0.93 for Burns and 0.91 for Dubois.

Maximum daily brutto light-use efficiency LUEday = Pgday/Qday (where Pgday is daily gross primary productivity, and Qday is incoming daily photosynthetically active radiation) at the Burns site was 20.9 mmol CO2/mol quanta, and maximum average weekly light-use efficiency, LUEwk, was 10.4 mmol CO2/mol quanta. For the Dubois site, maximum LUEday = 24.4 mmol CO2/mol quanta, and maximum LUEwk = 14.3 mmol CO2/mol quanta. We have found a clear seasonal pattern of the light-use efficiency of the sagebrush steppe ecosystems during growing season with distinct end of May maximum which has an important implications for light-use efficiency-based algorithms of scaling-up CO2 fluxes.

CO2 exchange components in different years were significantly correlated with normalized difference vegetation index (NDVI): for 1995-1997 period, AVHRR NDVI composites with 14 day time step were used, while for 1998-2001 period, SPOT VEGETATION NDVI composites with 10 day time step were used. Over all 13 site-years, average periodic (14 day average for AVHRR NDVI and 10 day average for SPOT VEGETATION NDVI) gross primary productivity (Pg) obtained from flux partitioning procedure was systematically more highly correlated to NDVI (R2 values from 0.67 to 0.87 for Burns and from 0.76 to 0.87 for Dubois) than daytime CO2 flux (Pd) directly provided by tower flux records (R2 values for Pd-NDVI relationships were 0.47 to 0.84 for Burns and 0.64 to 0.87 for Dubois). The important practical implication of this result is that in looking for CO2 flux characteristics related to remotely sensed variables for scaling-up algorithms, gross primary productivity, Pg, derived from flux observations using light-response analysis is a more appropriate variable than total daytime flux, Pd, directly obtained from measurements. These findings are in agreement with observations of higher Pg-NDVI correlations compared with Pd-NDVI correlations in mixed prairies of the Northern Great Plains (Gilmanov et al., 2004) and in a true steppe ecosystem in Northern Kazakhstan (Gilmanov et al., 2004; Wylie et al., 2004).

Patterns of relationships of Pg and Re to NDVI and other factors indicate possibilities for establishing multivariate functions Pg(NDVI, X1, ..., Xn) and Re(NDVI, Y1, ..., Ym) allowing scaling-up local fluxes to larger areas using GIS data, temporal NDVI, and other environmental factors.

References

Gilmanov, T.G., Johnson, D.A., Saliendra, N.Z., Akshalov, K., and Wylie, B.K. (2004). Gross primary productivity of the true steppe in Central Asia in relation to NDVI. Environmental Management, doi: 10.1007/s00267-003-9157-7.

Gilmanov, T.G., Tieszen, L.L., Wylie, B.K., Flanagan, L.B., Frank, A.B., Haferkamp, M.R., Meyers, T.P., and Morgan, J.A. (2004). Integration of CO2 flux and remotely sensed data for primary production and ecosystem respiration analyses in the Northern Great Plains: Potential for quantitative spatial extrapolation. Global Ecology and Biogeography (submitted).

Wylie, B.K., Gilmanov, T.G., Johnson, D.A., Saliendra, N.Z., Akshalov, K., Tieszen, L.L., Reed, B.C. and Laca, E. (2004). Intra-Seasonal Mapping of CO2 Flux in Rangelands of Northern Kazakhstan at One-Kilometer Resolution. Environmental Management, doi: 10.1007/s00267-003-9156-8.