Modelling Betula utilis distribution in response to climate-warming scenarios in Hindu-Kush Himalaya using random forest

Abstract

Globally, the increase in the climatic variability has led to adverse effects on the treeline species in the high-elevation mountain landscapes. Identifying the geographical space that supports the treeline species survival over time is essential for conservation biogeography. Increase in the global warming and snowmelt has made available the treeline species favourable niches in the higher elevations. Random Forest algorithm assuming non-parametric distribution was employed to predict the potential distribution of Betula utilis niche in the Hindu-Kush Himalayan (HKH) region. The potential distributions were simulated in the Last Inter-Glaciation (LIG), present (the year 1970–2000) and future (the year 2061–2080) environmental conditions. The actual distribution of the species in the current time was modelled and evaluated. The model sensitivity with reference to independent evaluation dataset for highly suitable B. utilis niche was 0.78. The model statistics of the current time was further applied to both the LIG and future (2061–2080) scenarios in order to get a fundamental niche of B. utilis. The treeline species, B. utilis was projected to become vulnerable to 21st century climate changes. The high suitability of B. utilis occurrence in the LIG, current and the future scenario were more likely in the elevation ranges 2601–2800 m, 3801–4000 m, and 4201–4400 m, respectively. The magnitude of advancement was relatively more along elevation and longitude, compared to the latitudinal gradient. The present study provides scientific evidence to conclude that the treeline species potential distribution in HKH is climate driven.