Future geographical distribution of Aedes albopictus in China under climate change scenarios

PLOS One | https://doi.org/10.1371/journal.pone.0327818 August 6, 2025 5 / 14 The importance of each environmental factor was evaluated through Biomod2, and the accuracy of each individual model was assessed using the Receiver Operating Characteristic (ROC) curve and the True Skill Statistic (TSS). The Area Under the Curve (AUC) values ranged from 0.5 to 1.0, where 0.5 represents completely random classification, and 1.0 indicates perfect classification. An AUC value greater than 0.8 suggests that the model has good or very good perfor- mance, while a value below 0.7 indicates poor predictive results. The TSS value, which ranges from 0 to 1, represents the net prediction success rate on measured samples. When the TSS value was greater than 0.7, it indicated a high predic- tion accuracy, while the value was less than 0.5 suggests a low accuracy. KAPPA is an index for measuring the consis- tency of classification models. It assesses the consistency between model predictions and actual observations. Its values typically range from −1–1, with higher values indicating better consistency. A KAPPA value greater than 0.7 is generally considered to reflect “good” consistency [34]. Based on the evaluation results of the individual models, the five modeling outcomes with the highest TSS, KAPPA and AUC values were selected to construct an ensemble model [22,24]. 3 Results 3.1 Model accuracy When compared to individual models (Fig 2), the ensemble model demonstrated superior performance, achieving an ROC

value of 0.968, a TSS value of 0.81, and a KAPPA value of 0.789, indicating that the ensemble model provides a more accurate and credible prediction of the potential suitable distribution of Ae. albopictus. 3.2 Suitability region of Ae. albopictus under current climate situation The highly suitability regions of Ae. albopictus in China were predominantly concentrated in the warm and humid south- ern and eastern coastal areas, providing an optimal environment for its survival. Specifically, Guangdong, Guangxi, and Hunan possessed the largest areas of highly suitability, measuring 15.61 × 104 km2, 20.84 × 104 km2 and 11.71 × 104 km2, respectively. Additionally, Jiangsu, Anhui, Sichuan, Hubei, Chongqing, Zhejiang, Hunan, Jiangxi, Yunnan, Guizhou, Fujian, Guangxi, Taiwan, and Hainan also had varying degrees of highly suitability regions, ranging from 1.18 × 104 km2 to 6.79 × 104 km2. Collectively, the area of highly suitability regions in China was 121.77 × 104 km2, which constituted a sub- stantial proportion of the total suitability regions (Fig 3). Quantitatively, the area of moderately suitability regions was 143.40 × 104 km2. Notably, Hebei, Shanxi, and Shaanxi provinces exhibited the largest measuring 1.52 × 104 km2, 2.47 × 104 km2 and 9.47 × 104 km2, respectively. Tianjin, Jiangsu, Fig 2. Evaluation index of individual model and ensemble model in Biomod2. https://doi.org/10.1371/journal.pone.0327818.g002