Predicting the patterns of plant species distribution under changing climate in major biogeographic zones of mainland India

Conservation efforts have traditionally focused on biodiversity hotspots, overlooking the essential ecological roles and ecosystem services provided by cold spots, the regions that harbour relatively low species diversity. In this study, we used a novel plant species database aggregated at 1˚ grid resolution to predict present and future plant species distribution in major cold spot biogeographic zones of India: Desert, Semi-Arid, Deccan Peninsula, and Gangetic Plain. We employed multiple models: Generalized Linear Model, Generalized Boosted Model, Random Forest, Support Vector Machine, and their ensemble. The results demonstrated reasonable predictive ability, with water and energy variables dominating in all the zones, showing a strong agreement with the field based data. Temperature annual range, annual precipitation, and precipitation of the driest month significantly influenced (r > 0.4) plant species patterns in the Desert and Semi-Arid zone. The ensemble model output improved predictive ability, with reduced root mean square error and enhanced correlation (r = 0.8). Other environmental variables (topography: elevation, and Human Influence Index) showed high correlation in combination with water and energy variables in the Deccan Peninsula. Continuous species loss is anticipated under future climate scenarios across all the zones. Semi-Arid is expected to see the most significant increase, with 69% and 52.5% of grids gaining species in 2050 (RCP4.5) and 69% and 84.7% in 2070 (RCP8.6), mainly attributed to an average precipitation increase. However, the Deccan Peninsula and Gangetic Plain show varying trends from 2050 to 2070, emphasizing the complex interplay of environmental factors influencing biodiversity distribution and dynamics. Our study provides insights into the species richness, potential and future distribution of cold spots in the major Indian biogeographic zones, aligning with climate-driven patterns. Our findings suggest that the ensemble modelling predictions are more accurate than individual models, emphasizing its potential for conservation efforts under rapidly changing climate. The study can provide a guiding tool for developing spatial biodiversity approach in the study region for prioritizing conservation in the face of climate change and help meet sustainable development goals.