The pulsed response of soil respiration to precipitation in an African savanna ecosystem: a coupled measurement and modeling approach

Tuesday, 16 December 2014
Zhaosheng Fan, Argonne National Laboratory, Argonne, IL, United States, Jason Caufield Neff, University of Colorado Boulder, Boulder, CO, United States and Niall P Hanan, South Dakota State University, Brookings, SD, United States
Savannas cover 60% of the African continent and play an essential role in the global carbon (C) cycle. To better characterize the physical controls over soil respiration in these settings, half-hourly observations of volumetric soil-water content, temperature, and the concentration of carbon dioxide (CO2) at different soil depths were continually measured from 2005 to 2007 under trees (“sub-canopy”) and between trees (“inter-canopy”) in a savanna vegetation near Skukuza, Kruger National Park, South Africa. The measured soil climate and CO2 concentration data were assimilated into a process-based model that estimates the CO2 production and flux with coupled dynamics of dissolved organic C (DOC) and microbial biomass C. Our results show that temporal and spatial variations in CO2 flux were strongly influenced by precipitation and vegetation cover, with two times greater CO2 flux in the sub-canopy plots (~2421 g CO2 m-2 yr-1) than in the inter-canopy plots (~1290 g CO2 m-2 yr-1). Precipitation influenced soil respiration by changing soil temperature and moisture; however, our modeling analysis suggests that the pulsed response of soil respiration to precipitation [known as “Birch effect (BE)”] is a key control on soil fluxes at this site. At this site, BE contributed to approximately 50% and 65% of heterotrophic respiration or 20% and 39% of soil respiration in the sub-canopy and inter-canopy plots, respectively. These results suggest that pulsed response of respiration to precipitation is an important component of the C cycle of savannas and should be considered in both measurement and modeling studies of carbon exchange in similar ecosystems.