Prediction Mapping Through Maxent Modeling Paves the Way for the Conservation of Rhododendron arboreum in Uttarakhand Himalayas

In the middle Himalayas, Rhododendron arboreum is shrinking due to low seed viability, poor regeneration, habitat degradation and fragmentation, habitat distortion, and species invasion. Further, developmental activities, the encroachment of forestland for agriculture, urbanization, and industrial expansion put additional burdens on its natural distribution. The present work focused on the prediction of R. arboreum distribution in Uttarakhand Himalayas using the Maxent model. In total, 1077 geospatial data were recorded, 300 well-distributed geo-coordinates were used to predict and estimate the distribution, while the rest were used to validate the model. The Maxent model generated AUC curve with an accurate and significant value of 0.886 ± 0.023. Bioclimatic variables such as temperature seasonality (Bio 4), annual temperature range (Bio 7), altitude (Alt), annual precipitation (Bio 12), and precipitation seasonality (Bio 15) contributed significantly to predict the distribution, as revealed by the Jackknife test. Within the total geographical area of 617.48 km 2 under R. arboreum distribution as shown over Landsat 8 generated map, 167.48 km 2 was found to be very dense, 320.75 km 2 was moderately dense, and 129.25 km 2 was open. The estimated distributed area was 2733.08 km 2 . The satellite-based mapping and model-based prediction of R. arboreum are of paramount importance to the foresters and researchers for species conservation, management, and utilization in a sustainable manner.