Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
Global warming increases the frequency of compound climatic extremes such as dry-wet abrupt alternation (DWAA), which can affect terrestrial ecosystem productivity. However, the asymmetric impacts and legacy effects of DWAA on global vegetation productivity have not been systematically assessed. Using the Standardized Precipitation Evapotranspiration Index (SPEI), the Standardized Precipitation Index (SPI), and three gross primary productivity (GPP) products, we identified monthly DWAA events, including dry-to-wet (DW) and wet-to-dry (WD) transitions, and assessed their global productivity impacts and post-event dynamics. Our results show that DWAA caused spatially heterogeneous and direction-dependent GPP losses: DW caused greater initial GPP losses, whereas WD losses persisted longer. DW loss exceeded WD loss in five of the six main product–index comparisons, and this predominant magnitude asymmetry was robust to alternative event thresholds but more dependent on accumulation scale. Among vegetation groups, grasslands most frequently showed the largest transition-month loss across growing-season-length classes. Event-aligned GPP trajectories further showed that GPP following DW generally returned to or exceeded the local monthly background by the second post-event growing-season month, whereas GPP following WD generally approached the background by the third month. During the post-event period, DW recovery was characterized by a strengthened positive response to soil moisture and more negative responses to radiation and temperature. Following WD events, GPP responses to precipitation were consistently lower than those during normal months, whereas soil-moisture, radiation, and temperature responses were mixed. These findings highlight the direction- and context-dependent impacts of compound dry-wet extremes on the global carbon cycle and support their representation in ecosystem resilience assessments under climate change.