期刊
ECOLOGY AND EVOLUTION
卷 13, 期 5, 页码 -出版社
WILEY
DOI: 10.1002/ece3.10104
关键词
CLIMEX; environmental variable; random forest; spatial distribution; western conifer seed bug
In this study, the potential distribution of the western conifer seed bug (WCSB) was projected using CLIMEX modeling and random forest analysis. Land cover was found to be the most important variable for classifying the presence/pseudo-absence of WCSB, with an accuracy of 77.1%. The WCSB was predicted to have higher climatic suitability in Southern Europe compared to Western Europe, and primarily occur near coniferous forests. The combination of CLIMEX and random forest analysis can improve the reliability of predictions.
The western conifer seed bug (WCSB) Leptoglossus occidentalis (Heidemann) (Heteroptera: Coreidae) is a pest insect that causes significant losses of coniferous trees worldwide. In this study, we sought to project the potential distribution of the WCSB based on dual CLIMEX modeling and random forest (RF) analysis to obtain basic data for WCSB monitoring strategies. The CLIMEX model, a semimechanistic niche model that responds to climate-based environmental parameters, is a species distribution model that focuses on regional climatic suitability. Given that this model can be used to select areas that are likely to reflect the climatically favorable spread of species, which we initially used CLIMEX to evaluate the potential distribution of the WCSB. The RF algorithm was used to predict the potential occurrence of WCSB and to evaluate the relative importance of environmental variables for WCSB occurrence. Using the RF model, land cover was found to be the most important variable for classifying the presence/pseudo-absence of the WCSB, with an accuracy of 77.1%. Climatic suitability for the WCSB was predicted to be 2.4-fold higher in Southern Europe than in Western Europe, and the WCSB was predicted to occur primarily near coniferous forests. Given that CLIMEX and RF analyses yielded different prediction results, using the findings of both models may compensate for the shortcomings of these models when used independently. Consequently, to ensure greater prediction reliability, we believe that it would be beneficial to base predictions on the combined potential distribution data obtained using both modeling approaches.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据