Jefri Permadi, Nia Kurniawan, Diana Arfiati , Agung Pramana Warih Marhendra

Predicting Habitat Suitability of Mahseer Fish (Tor spp.) in Tropical River Systems Using MaxEnt and Google Earth Engine: A Geospatial Modeling Approach

Introduction

Predicting habitat suitability of mahseer fish (tor spp.) in tropical river systems using maxent and google earth engine: a geospatial modeling approach. Predict Mahseer fish habitat suitability in tropical rivers using MaxEnt and Google Earth Engine. Identifies upstream areas (river orders 1-3) as key, aiding conservation and sustainable freshwater ecosystem management.

Abstract

Rivers are vital freshwater habitats that face threats of degradation and climate change. Mahseer fish, a key species, is in decline. This study predicted Mahseer fish habitats in Central Java using the Google Earth Engine and the MaxEnt machine learning algorithm. Environmental predictors, including NDVI, elevation, slope, river order, temperature, and rainfall, were extracted from Sentinel, SRTM, MODIS, and CHIRPS data. The model identified river order as the most influential variable (73%), followed by elevation (18%) and rainfall (8%), with an AUC score of 0.7, indicating fair accuracy. Suitable habitats were located in upstream river orders (1–3), typically at higher elevations. These findings provide spatial guidance for conservation planning, such as identifying critical habitats, prioritizing upstream areas, and establishing seasonal fishing ban. This approach supports biodiversity protection and aligns with the Sustainable Development Goals by offering a scalable tool for freshwater ecosystem management. Using MaxEnt with GEE shows promise for rapid, and cost-effective species distribution modeling in data-limited tropical regions.

Review

This study presents a timely and relevant investigation into the habitat suitability of Mahseer fish (Tor spp.) in Central Java, a critical endeavor given the increasing threats of river degradation and climate change to freshwater ecosystems. The authors employ a robust and increasingly popular methodology, combining the MaxEnt machine learning algorithm with the computational power of Google Earth Engine (GEE) for geospatial modeling. This approach is particularly commendable for its potential to provide rapid and cost-effective species distribution modeling in data-limited tropical regions, offering a scalable tool that directly supports biodiversity protection and aligns with Sustainable Development Goals. The integration of various satellite and environmental datasets demonstrates a sophisticated use of remote sensing for ecological applications. The research effectively identifies key environmental predictors influencing Mahseer distribution, with river order emerging as the most significant variable, followed by elevation and rainfall. The finding that suitable habitats are predominantly located in upstream river orders (1–3) at higher elevations is ecologically intuitive and aligns with typical Mahseer preferences for cooler, well-oxygenated, and less disturbed waters. While the reported AUC score of 0.7 indicates fair accuracy for the model, it provides a solid foundation for initial conservation planning. The study's ability to spatially delineate these critical habitats offers invaluable guidance for practitioners, facilitating the prioritization of upstream areas for protection and the potential implementation of management strategies such as seasonal fishing bans. While the study provides a strong framework for Mahseer conservation, there are areas for further consideration. The "fair" AUC score, while acceptable, suggests room for model refinement, perhaps through the inclusion of additional fine-scale or in-situ variables such as water quality parameters (e.g., dissolved oxygen, turbidity), substrate type, or presence of riparian vegetation, which are often crucial for fish habitats. Distinguishing between different *Tor* species, if sufficient presence data exists, could also yield more specific conservation insights. Nevertheless, the integration of MaxEnt and GEE stands out as a powerful demonstration of how advanced geospatial techniques can inform conservation policy in vulnerable tropical freshwater systems. This work sets an excellent precedent for future ecological modeling efforts in regions facing similar data constraints and ecological pressures.

Full Text

Comments

You need to be logged in to post a comment.