Assessment of vegetation dynamics under changed climate situation using geostatistical modeling.

The estimation of spatiotemporal variability of vegetation health and associated climatic variables is a prerequisite for efficient management of mountain ecosystems. The present study aimed to quantify the changing effect of different climatic parameters on vegetation health in Khyber Pakhtunkhwa, Pakistan, from 2012 to 2021. The association between NDVI and three explanatory variables, i.e., temperature, precipitation, and wind speed, was analyzed using geostatistical model optimization and stationarity index analysis. Results showed that local model GWR presented optimal results owing to its ability to deal with spatial non-stationarity and stochasticity inherent in regression relationships. The most dominant climatic factor influencing vegetation dynamics in KPK was temperature, followed by wind speed and precipitation. However, there was a gradual decline in the influence of temperature in controlling vegetation dynamics after 2018, and precipitation emerged to be the major determinant. Moreover, the stationarity index was calculated at multi-scales to determine the operational scale range for each climatic variable. The scale-dependency of NDVI-T relationship exhibited highest spatial heterogeneity, up to a distance of 600 km. Overall, geostatistical modeling using earth observation datasets enabled insightful understanding of spatiotemporal variations in vegetation response to climate change, necessary for sustainable development and responsive climate action. [ABSTRACT FROM AUTHOR]