• 论文与研究报告 •

### 外来入侵新害虫刺槐突瓣细蛾在中国的适生区预测

1. 1. 山东农业大学植物保护学院 泰安 271018;
2. 山东省林业有害生物防控工程技术研究中心 泰安 271018;
3. 泰安市园林管理局 泰安 271021;
4. 泰安市徂徕山林场 泰安 271000
• 收稿日期:2017-10-23 修回日期:2018-01-26 出版日期:2019-06-25 发布日期:2019-07-11

### Prediction of Suitable Distribution Regions of a New Invasive Pest: Chrysaster ostensackenella (Lepidoptera: Gracillariidae) in China

Fan Tingting1, Gao Shangkun1,2, Meng Fanling3, Yin Hongzeng3, Li Chao4, Wang Qinghua4, Zhou Chenggang1,2

1. 1. College of Plant Protection, Shandong Agricultural University Tai'an 271018;
2. Engineering Research Center of Forest Pest Management of Shandong Province Tai'an 271018;
3. Tai'an Landscape Administration Bureau Tai'an 271021;
4. Mount Culai Forest Farm of Tai'an Tai'an 271000
• Received:2017-10-23 Revised:2018-01-26 Online:2019-06-25 Published:2019-07-11

Abstract: [Objective] Chrysaster ostensackenella ((Lepidoptera:Gracillariidae) is an important alien invasive pest newly discovered in Yantai, Shandong Province in 2008 and severely damages Robinia pseudoacacia, an important economic greening plant in China. Prediction of the suitable areas can provide a basis for promoting the efficiency of quarantine, supervision and control of the pest.[Method] In this study, data of 11 distribution locations of C. ostensackenella in China were collected, and nine environment variables with correlation coefficients <|0.9|were screened from 19 environment variables downloaded from WordClim between 1970 and 2000 by ArcGIS 10.0. Then the data were converted into the ASCⅡ format data required for MaxEnt, and the model was set to cloglog output format, outputting file type was ASCⅡ and linear features. To improve the accuracy of the predictive effects and reduce the level of uncertainty, a 10-times cross-validation was set up in the model and repeated 10 runs to get the average values. The contribution rate of each environment variable to the potential geographic distribution in this model was analyzed by Jackknife method, and the optimal simulation results were transformed and classified in ArcGIS10.0, that is, the habitat suitability indexes of C. ostensackenella in China were divided into four categories:non-suitable areas, low suitable areas, middle suitable areas, high suitable areas. Finally, the distribution map of different extent suitable habitats of the insect was obtained. The future distribution areas of C. ostensackenella were predicted using RCP 8.5 climate data of 2050 and 2070. The model precisions of MaxEnt were evaluated by areas under the ROC curve and True skill statistics.[Result] Under current climatic conditions, the high and middle suitable areas of C. ostensackenella were mainly concentrated in Shandong and its neighboring provinces (Liaoning, Beijing, Tianjin, Hebei, Shanxi, Henan, Anhui, Jiangsu), as well as in parts of Sichuan and Yunnan; Under the future climate conditions, by 2050, the middle and high suitable regions of C. ostensackenella under RCP 8.5 climatic conditions would be larger than those in present and spread to the southwest. By 2070, the suitable regions also would enlarge obviously and move to the northeast. Compared with the predicted results in 2050, the high suitable area decreased slightly. The Jackknife test indicated that annual precipitation, precipitation of wettest quarter, min-temperature of coldest month had a great contribution to the distribution of C. ostensackenella. The range of the optimum annual average precipitation is 382.08-1 135.81 mm, with the optimum of 753.85. The optimum wet season precipitation is 241.61-693.86 mm, with the optimum of 464.55 mm, and the optimum minimum temperature in the coldest month is -16.96-6.36℃ with the optimum of -5.5℃. The AUC and TSS values are 0.957±0.052 and 0.8±3.05, respectively, which indicates that the prediction accuracy of the model is excellent.[Conclusion] According to the results of this study, it is known that the pest is a major threat to black locust, R. pseudoacacia in China, and a high attention should be paid by the relevant afforestation and plant quarantine departments.