A novel index combining meteorological, hydrological, and ecological anomalies used for ecological drought assessment at a grassland-type basin scale

• A new drought index NSEDI was constructed based on a NDVI-fitting model incorporating hydro-meteorological anomalies. • NSEDI can reasonably depict the ecological drought posed by hydrothermal stress within a basin. • The grassland ecological drought shows a strong seasonal dependence, seemingly more serious in the rapid growing stage. Under changing climate, grasslands in arid and semiarid areas have experienced severe degradation due to the increased frequency and intensity of droughts. Quantifying the anomalies in ecosystem caused by water limitation is challenging due to the complex interactions of ecological and hydrological processes at a basin scale. Here, a new comprehensive drought index, the Non-stationary Standardized Ecological Drought Index (NSEDI), was proposed to identify and assess the ecological drought. The NSEDI is calculated based on a non-stationary NDVI-fitted model, which incorporates the highly correlated precipitation, temperature, runoff, and soil moisture as covariates. The performance of the NSEDI method was evaluated by an application on a grassland-type inland river basin, as well as comparations with that of other widely-used drought indices. The results show that the non-stationary model performed better than a stationary model and a time-dependent model, indicating the preferable explanatory ability of the concurrent and 1-month preceding hydrothermal variables to the NDVI changes. Consequently, the NSEDI can reasonably depict the vegetation anomalies posed by hydro-meteorological drought stresses within a basin. The standardized feature employed in NSEDI makes it feasible to compare ecological drought over different regions and periods, and with the other drought types. It is demonstrated that the grassland ecological droughts largely depend on seasonality, expected to be more serious in the rapid growing stage (July to August). Moreover, the ecological drought is strongly determined by the meteorological drought with high temperature and less precipitation, and directly aggravated by the surface soil dryness. Hydrological drying condition seemingly shows an inhibiting effect on the ecological drought, implying the restrictive relation between vegetation dynamics and hydrological processes. It is expected that the proposed NSEDI approach can effectively monitor drought for an eco-hydrological system, further providing scientific guidance for regional ecological management.