Non-uniform time-lag effects of terrestrial vegetation responses to asymmetric warming

The immediate effects of asymmetric warming (i.e., day- and night-time warming) on terrestrial ecosystems have been well documented, but the time-lag effects remain poorly understood. In this paper, we investigated the global inter-annual hysteretic responses of vegetation to asymmetric warming over the period of 1982–2013. The net primary production (NPP) was employed as the indicator of vegetation growth, and accumulated monthly average daily maximum temperature and minimum temperature (ATmax and ATmin) were used to reflect the asymmetric warming condition. Additionally, partial correlation analyses were conducted to examine the correlations between NPP and ATmax/ATmin on a monthly scale. Furthermore, the best time lags that ATmax/ATmin had on NPP and the optimal correlations between ATmax/ATmin and NPP were analyzed by time-lag analyses. The results showed that (i) vegetation responded to the asymmetric warming with near 12-month delays at a global scale, and vegetation exhibited larger lags in responding to recent warming over temperature-limited areas or semiarid and subhumid regions compared to other places; (ii) compared with ATmin, ATmax had longer time-lag effects on biomes over mid-high latitudes (45 °N–90 °N, 23 °S–60 °S) and high altitudes (i.e., the Tibetan Plateau and the Brazil Plateau) and smaller delay impacts on biomes in other regions; (iii) with hysteretic impacts considered, ATmax correlated positively with vegetation in temperature-limited areas and negatively in heat-sufficient and water-deficit places, and the reverse was mostly true for ATmin. These phenomena may be associated with the intrinsic differences in the mechanisms that day- and night-time temperatures have on vegetation growth. Our paper gives new insights into the non-uniform responses of the terrestrial ecosystem to asymmetric warming. Looking ahead, terrestrial ecosystem models are highly recommended to incorporate such non-uniform time-lag impacts and distinguished correlations so as to improve their performances in future work.