Scholarlink

<p>Ensemble species distribution modelling was compared to the two best-performing individual algorithms using climatic and non-climatic predictors for the critically endangered plant species <i>Commiphora wightii</i>, looking at habitat suitability, niche overlap, and IUCN categories like Extent of occurrence (EOO) and Area of occupancy (AOO) in India. We selected ensemble, Random Forest, and Support Vector Machine algorithms with current and two future climatic time frames (2050 and 2070) along with aspect and slope predictors based on the model quality tools. Using the ensemble methodology and the SVM technique, we found that the seasonality of precipitation had a stronger influence on the habitat suitability of this species in both the present and the 2070 time frames. The SVM was able to capture the magnitude of their effects better than the ensemble technique. However, for the 2050 climate projection, both the ensemble and SVM imply that the wettest quarter’s precipitation has a greater impact. This was found that Rajasthan’s flood-prone eastern plain, internal drainage dry zone, and irrigated north-western plains were less ideal places for this species to survive under the current climate.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Habitat suitability modelling of a critically endangered plant species -Commiphora wightii (Arn.) Bhandari: comparative assessment using various machine learning algorithms

  • 期刊论文

  • 发表日期：

    2025年3月13日

查看全文
  • Manish Mathur,
  • Preet Mathur

摘要

Ensemble species distribution modelling was compared to the two best-performing individual algorithms using climatic and non-climatic predictors for the critically endangered plant species Commiphora wightii, looking at habitat suitability, niche overlap, and IUCN categories like Extent of occurrence (EOO) and Area of occupancy (AOO) in India. We selected ensemble, Random Forest, and Support Vector Machine algorithms with current and two future climatic time frames (2050 and 2070) along with aspect and slope predictors based on the model quality tools. Using the ensemble methodology and the SVM technique, we found that the seasonality of precipitation had a stronger influence on the habitat suitability of this species in both the present and the 2070 time frames. The SVM was able to capture the magnitude of their effects better than the ensemble technique. However, for the 2050 climate projection, both the ensemble and SVM imply that the wettest quarter’s precipitation has a greater impact. This was found that Rajasthan’s flood-prone eastern plain, internal drainage dry zone, and irrigated north-western plains were less ideal places for this species to survive under the current climate.