Spatial–temporal variation in soil respiration in an oak–grass savanna ecosystem in California and its partitioning into autotrophic and heterotrophic components

Abstract

The spatial upscaling of soil respiration from field measurements to ecosystem levels will be biased without studying its spatial variation. We took advantage of the unique spatial gradients of an oak-grass savanna ecosystem in California, with widely spaced oak trees overlying a grass layer, to study the spatial variation in soil respiration and to use these natural gradients to partition soil respiration according to its autotrophic and heterotrophic components. We measured soil respiration along a 42.5 m transect between two oak trees in 2001 and 2002, and found that soil respiration under tree canopies decreased with distance from its base. In the open area, tree roots have no influence on soil respiration. Seasonally, soil respiration increased in spring until late April, and decreased in summer following the decrease in soil moisture content, despite the further increase in soil temperature. Soil respiration significantly increased following the rain events in autumn. During the grass growing season between November and mid-May, the average of CO2 efflux under trees was 2.29 lmol m � 2 s � 1 , while CO2 efflux from the open area was 1.40 lmol m � 2 s � 1 . We deduced that oak root respiration averaged as 0.89 lmol m � 2 s � 1 , accounting for 39% of total soil respiration (oak root + grass root + microbes). During the dry season between mid-May and October, the average of CO2 efflux under trees was 0.87 lmol m � 2 s � 1 , while CO2 efflux from the open areas was 0.51 lmol m � 2 s � 1 . Oak root res- piration was 0.36 lmol m � 2 s � 1 , accounting for 41% of total soil respiration (oak root + mic- robes). The seasonal pattern of soil CO2 efflux under trees and in open areas was simulated by a bi- variable model driven by soil temperature and moisture. The diurnal pattern was influenced by tree physiology as well. Based on the spatial gradient of soil respiration, spatial analysis of crown closure and the simulation model, we spatially and temporally upscaled chamber measurements to the ecosystem scale. We estimated that the cumulative soil respiration in 2002 was