Temporal Downscaling of Soil Carbon Dioxide Efflux Measurements Based on Time‐Stable Spatial Patterns

Abstract

Soil CO2 efflux at a field site is often computed as the average of successive chamber measurements at several points to overcome the effects of spatial variability and microclimatic disturbances. As a consequence, the resulting data set has a coarser resolution in space (one average per site) and time than the raw data set. The deviations between raw measurements and the field average may provide additional insights, however, if they can be decomposed into a time‐stable part, characterizing the spatial pattern of emission strengths, and a dynamic part, characterizing rapid changes in soil CO2 efflux. We evaluated data from several measurement campaigns in an agricultural landscape. First, we determined the persistence of spatial CO2 efflux patterns and found that ≥50% of spatial variance was stable for at least 1 d in all examined crop and field types. For fields where vegetation and gradients in soil properties determined the spatial variation in CO2 efflux, some correlation was still found after 10 d. In a next step, we removed the time‐stable patterns from the raw time series. The resulting estimate of instantaneous area‐average soil respiration closely resembled the conventional spatiotemporal field average on days without rapid changes in meteorologic conditions. On days with fluctuations of radiation and temperature, in contrast, soil respiration reacted on a time scale from instantaneous to about 1 h. Based on a discussion of potential mechanisms underlying these reactions for a wheat (Triticum aestivum L.) and a sugarbeet (Beta vulgaris L. ssp. vulgaris) stand, we suggest that the proposed downscaling methodology, in combination with existing decomposition techniques, may help to examine the short‐term dependence of heterotrophic and root respiration on radiation, temperature, and rain.