This is the authors' abstract. We don't add key points for this paper.
The conservation of Komodo dragons ( Varanus komodoensis ), an apex predator whose top-down regulation of ungulate prey helps sustain savanna ecosystem balance, and an iconic yet endangered species, requires a nuanced understanding of habitat suitability, wildfire susceptibility, and conservation zoning within Komodo National Park (KNP), Indonesia. This study develops a comprehensive multi-indicator Komodo Habitat Suitability Index (KHSI) that integrates the availability of prey for both adult and juvenile Komodos, habitat characteristics, and spatial distribution data. Coupled with machine learning (ML)-based wildfire susceptibility modelling, this research identifies areas critical for Komodo conservation and assesses their alignment with existing zoning regulations. The findings reveal that 36% of KNP exhibits high habitat suitability for Komodos, predominantly in lowland savanna ecosystems, while 11% of the park faces high wildfire vulnerability, particularly in coastal areas with dry savanna vegetation. Notably, 5% of the park combines high habitat suitability with elevated wildfire risk, underscoring the need for targeted management strategies. The habitat model achieved high discrimination (AUC = 0.937). The wildfire ensemble reached a mean AUC of 0.861 across 100 stratified random splits, falling to 0.735 under leave-one-fire-out validation, in which each of the three mapped burn polygons was withheld in turn; the latter figure is the appropriate estimate of transferable performance and the priority maps should be read accordingly. Furthermore, the evaluation of conservation zones highlights discrepancies in zoning effectiveness, with the core zone encompassing a significant proportion of low-suitability areas and high wildfire vulnerability. This study’s novelty lies in its integration of species habitat modelling, wildfire risk assessment, and conservation zoning evaluation using advanced geospatial and machine learning (ML) techniques. The results contribute to improved conservation planning by identifying priority areas for habitat protection, wildfire mitigation, and potential revisions to zoning policies. These insights provide a scalable framework for addressing similar conservation challenges in other biodiversity hotspots.
Answers are grounded in this paper and cite their sources.
No takes yet. Share an insight, caveat, or question.
Sakti et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: