Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Inter and Intra-Annual Spatio-Temporal Variability of Habitat Suitability for Asian Elephants in India: A Random Forest Model-based Analysis

P. Anjali, Deepak N. Subramani

TL;DR

The study addresses the decline and fragmentation of Asian elephant habitats in India and the resulting human–elephant conflict by developing a Random Forest habitat-suitability model that integrates climatic, topographic, LULC, and vegetation predictors (NPP, LAI, NDVI) with GBIF presence data and pseudo-labels to capture inter- and intra-annual variability from 2000–2016. The model identifies NPP, LAI, and elevation as key predictors and demonstrates robust predictive performance (AUC ~0.77–0.78; precision ~0.77; recall ~0.76). Spatio-temporal analysis reveals a general habitat decline with seasonal and regional fragmentation, notably in the Western Ghats and central/north-western India, correlating with conflict hotspots. The authors propose using this RF output as a sub-model within an AI-driven agent-based decision-support tool to inform conservation planning, habitat connectivity, and conflict mitigation strategies.

Abstract

We develop a Random Forest model to estimate the species distribution of Asian elephants in India and study the inter and intra-annual spatiotemporal variability of habitats suitable for them. Climatic, topographic variables and satellite-derived Land Use/Land Cover (LULC), Net Primary Productivity (NPP), Leaf Area Index (LAI), and Normalized Difference Vegetation Index (NDVI) are used as predictors, and the species sighting data of Asian elephants from Global Biodiversity Information Reserve is used to develop the Random Forest model. A careful hyper-parameter tuning and training-validation-testing cycle are completed to identify the significant predictors and develop a final model that gives precision and recall of 0.78 and 0.77. The model is applied to estimate the spatial and temporal variability of suitable habitats. We observe that seasonal reduction in the suitable habitat may explain the migration patterns of Asian elephants and the increasing human-elephant conflict. Further, the total available suitable habitat area is observed to have reduced, which exacerbates the problem. This machine learning model is intended to serve as an input to the Agent-Based Model that we are building as part of our Artificial Intelligence-driven decision support tool to reduce human-wildlife conflict.

Inter and Intra-Annual Spatio-Temporal Variability of Habitat Suitability for Asian Elephants in India: A Random Forest Model-based Analysis

TL;DR

The study addresses the decline and fragmentation of Asian elephant habitats in India and the resulting human–elephant conflict by developing a Random Forest habitat-suitability model that integrates climatic, topographic, LULC, and vegetation predictors (NPP, LAI, NDVI) with GBIF presence data and pseudo-labels to capture inter- and intra-annual variability from 2000–2016. The model identifies NPP, LAI, and elevation as key predictors and demonstrates robust predictive performance (AUC ~0.77–0.78; precision ~0.77; recall ~0.76). Spatio-temporal analysis reveals a general habitat decline with seasonal and regional fragmentation, notably in the Western Ghats and central/north-western India, correlating with conflict hotspots. The authors propose using this RF output as a sub-model within an AI-driven agent-based decision-support tool to inform conservation planning, habitat connectivity, and conflict mitigation strategies.

Abstract

We develop a Random Forest model to estimate the species distribution of Asian elephants in India and study the inter and intra-annual spatiotemporal variability of habitats suitable for them. Climatic, topographic variables and satellite-derived Land Use/Land Cover (LULC), Net Primary Productivity (NPP), Leaf Area Index (LAI), and Normalized Difference Vegetation Index (NDVI) are used as predictors, and the species sighting data of Asian elephants from Global Biodiversity Information Reserve is used to develop the Random Forest model. A careful hyper-parameter tuning and training-validation-testing cycle are completed to identify the significant predictors and develop a final model that gives precision and recall of 0.78 and 0.77. The model is applied to estimate the spatial and temporal variability of suitable habitats. We observe that seasonal reduction in the suitable habitat may explain the migration patterns of Asian elephants and the increasing human-elephant conflict. Further, the total available suitable habitat area is observed to have reduced, which exacerbates the problem. This machine learning model is intended to serve as an input to the Agent-Based Model that we are building as part of our Artificial Intelligence-driven decision support tool to reduce human-wildlife conflict.

Paper Structure

This paper contains 5 sections, 5 figures, 1 table.

Table of Contents

  1. Introduction
  2. Data and Methods
  3. Random Forest Model Development
  4. Spatio-temporal Analysis of Habitat Suitability
  5. Conclusion and Future Work

Figures (5)

  • Figure 1: Importance factor of the predictor variables used in the Random Forest model
  • Figure 2: ROC curve for the Random Forest model training
  • Figure 3: The percentage change in the area classified as probable presence during 2001-2016 when compared with 2001 for the four regional zones in India
  • Figure 4: Time series of the area (in sq. km) classified as a suitable habitat for Asian elephants by the Random Forest model during 2001-2016
  • Figure 5: Estimated change in the habitat suitability of E. maximus for the month of March in the years 2001, 2006, 2011 and 2016 by the random forest model. Panels a, b, c, and d show the binary classification of estimated presence and absence. Panels e, f, g, and h show the probability distribution for classification of a pixel as presence, red indicating high habitat suitability and white indicating low habitat suitability