Ecosystem responses to climate change, particularly in arid environments, is an understudied topic. This study conducted a spatial analysis of ecosystem responses to short-term variability in temperature, precipitation, and solar radiation in the Qilian Mountains National Park, an arid mountainous region in Northwest China. We collected precipitation and temperature data from the National Science and Technology Infrastructure Platform, solar radiation data from the China Meteorological Forcing Dataset, and vegetation cover remote-sensing data from the Moderate Resolution Imaging Spectroradiometer. We used the vegetation sensitivity index to identify areas sensitive to climate change and to determine which climatic factors were significant in this regard. The findings revealed a high degree of heterogeneity and non-linearity of ecosystem responses to climate change. Four types of heterogeneity were identified: longitude, altitude, ecosystem, and climate disturbance. Furthermore, the characteristics of nonlinear ecosystem responses to climate change included: (1) inconsistency in the controlling climatic factors for the same ecosystems in different geographical settings; (2) the interaction between different climatic factors results in varying weights that affect ecosystem stability and makes them difficult to determine; and (3) the hysteresis effect of vegetation increases the uncertainty of ecosystem responses to climate change. The findings are significant because they highlight the complexity of ecosystem responses to climate change. Furthermore, the identification of areas that are particularly sensitive to climate change and the influencing factors has important implications for predicting and managing the impacts of climate change on ecosystems, which can help protect the stability of ecosystems in the Qilian Mountains National Park.