The terrestrial biosphere is the largest net sink of global CO2, but its sink capacity varies considerably from year to year depending on environmental conditions. Recent work has highlighted the importance of semi-arid ecosystems in interannually driving global concentrations of atmospheric CO2. We therefore need to better understand the dynamics and drivers of the CO2 fluxes and their modeling along semi-arid to humid gradients. Respiration is an even more uncertain flux compared to photosynthetic fluxes and its spatially variability is not well understood. Here we focus on terrestrial ecosystem respiration (TER) in Australia, and, specifically, on disentangling the impacts of temperature and soil water on TER. We use nighttime net ecosystem exchange data as a viable proxy for daily TER collected by 40 flux tower stations within the OzFlux network over the last 20 years in Australia. These stations cover a broad range of climatic conditions enabling us to analyze the dependence of TER on soil moisture under varying aridity and temperature conditions. We find that the sensitivity of TER to soil moisture is the strongest in semi-arid regions. In these moisture-limited locations, the TER sensitivity to soil moisture increases strongly with temperature. Soil respiration fluxes at humid stations are large but exhibit low sensitivity to high soil moisture levels indicating that TER at humid stations is not water-limited. Using the dynamic global vegetation model LPJ, we show that common model approaches assuming increasing TER with increasing soil moisture for all soil moisture levels perform poorly in reproducing the observed TER patterns in Australia due to interactions with carbon availability and representation of soil hydrology. Hence a more sophisticated description of the dependence of TER on soil moisture is necessary to capture TER dynamics under different climatic conditions accurately.

Read original article

Read original article