Prior rainfall pattern determines response of net ecosystem carbon exchange to a large rainfall event in a semi-arid woodland

Semi-arid climate is characterized by long dry periods which are interrupted by rainfall events differing in magnitude. The effect of these rainfall events on ecosystem carbon fluxes has been studied in semi-arid grass and shrublands, but data for woodlands is lacking. In this study net ecosystem productivity (NEP), partitioned ecosystem respiration (Reco) and gross primary productivity (GPP) in a semi-arid eucalyptus woodland were measured using eddy covariance data. Four natural large rainfall events with similar magnitude (35-60 mm), but contrasting previous rainfall patterns (several moderate rainfall events vs. limited or no rainfall) were chosen. NEP, Reco and GPP rates 28 days prior to and 35 days after the three central rainfall events were used. Prior rainfall patterns influenced ecosystem carbon fluxes after the central rainfall event. GPP rates were not affected by the four rainfall events. After four weeks with several medium to large rainfall events, the central rainfall event had little effect on Reco. In contrast, a large rainfall event following four weeks with very little rainfall induced an increase in Reco for about three weeks and thus a decrease in NEP. The strong increase in Reco after the central rainfall event can, at least partly, be explained by an increase in soil respiration upon rewetting. A water addition experiment (30 mm rainfall simulation) conducted in the field following a long dry period (only 4.8 mm rainfall input in 45 days) showed that heterotrophic soil respiration usually decreases rapidly after rewetting. In the sandy soil of the study area, the top 0-5 cm dry within a few days after rainfall, but at 10 cm depth the soil remained moist for several weeks after the large rainfall event that followed a long dry period. Therefore, sustained higher Reco following the large rainfall event could be due to respiration of roots in the moist deeper soil layers. We conclude that the previous rainfall should be considered when interpreting the response of ecosystem C fluxes to rainfall events which could alter ecosystem carbon balance and may potentially affect seasonal or inter-annual variability of ecosystem carbon uptake.