Abstract: Objective　To investigate the spatial distribution of Aedes albopictus in China at different time periods from 2000 to 2019, so as to provide insights into precise management of Ae. albopictus in China. Methods  Data pertaining to the distribution of Ae. albopictus in China from 2000 to 2019 were collected through literature retrieval with terms of "Aedes albopictus", "monitoring", "survey", "density", "distribution", and "outbreak" in national and international databases. The title and time of the publication, sampling sites, sampling time, mosquito capture methods, and mosquito species and density were extracted, and the longitude and latitude of sampling sites were obtained through Baidu Map. Meteorological element data at meteorological observation stations within China were obtained from the National Climatic Data Center of the United States, and the annual maximum temperature, annual minimum temperature, average temperature in January, average temperature in July, annual temperature range, daily temperature range and relative humidity were calculated and subjected to Kriging interpolation. Monthly cumulative precipitation grid data and monthly average temperature grid data with a resolution of 1 km for China from 2000 to 2019 were obtained from the National Tibetan Plateau Scientific Data Center, and the annual precipitation and annual average temperature were calculated cumulatively. Population density data in China from 2000 to 2019 were obtained from the WorldPop Hub, and the gross domestic product (GDP) in China was obtained from the Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences. The above data were divided into 5⁃year intervals to calculate data during the periods from 2000 to 2004, from 2005 to 2009, from 2010 to 2014, and from 2015 to 2019. Ae. albopictus distribution data were modeled in China from 2000 to 2019 and during each period with the classification random forest (RF) model, to predict the distribution of Ae. albopictus across the country and analyze the distribution of Ae. albopictus based on the seven major climate zones in China. The performance of RF models was evaluated by accuracy, precision, recall, and area under the receiver operating characteristic curve (AUC), and the importance of each feature in the RF model was evaluated with mean decrease accuracy (MDA). Results  A total of 1 191 Chinese publictions and 391 English publications were retrieved, among which 580 articles provided detailed data on the sampling sites of Ae. albopictus and specific sampling years, meeting the inclusion criteria. A total of 2 234 Ae. albopictus sampling sites were included in China from 2000 to 2019, and RF modeling results showed that the overall Ae. albopictus distribution area was mainly found in southeastern and southwestern provinces of China from 2000 to 2019, with scattered distribution in coastal areas of northeastern provinces, such as Liaoning Province. The accuracy, precision, recall and AUC of the RF model were 0.915 to 0.947, 0.933 to 0.975, 0.898 to 0.978, and 0.902 to 0.932 for the distribution of Ae. albopictus at different time periods from 2000 to 2019. Among all features in the RF models, population density was the most contributing factor to the distribution of Ae. albopictus in China, followed by GDP, and all meteorological variables contributed relatively less to the predictive power of the RF model. In China's seven major climate zones, Ae. albopictus was almost entirely distributed in the marginal tropical humid region, the north subtropical humid region, and the warm temperate semi⁃humid region. The combined distribution area of these three zones accounted for 100.0% of the national distribution area from 2000 to 2004, from 2005 to 2009, and from 2010 to 2014, and 99.9% from 2015 to 2019, and the proportion of Ae. albopictus distribution area in the warm temperate semi⁃humid region increased gradually from 20.2% to 30.2%. Conclusions Ae. albopictus is mainly distributed in the southeastern and southwestern provinces of China and is greatly influenced by population and economic factors. The warm temperate semi⁃humid region in China is gradually becoming a hot spot for the distribution of Ae. albopictus.

Key words: Aedes albopictus, Spatiotemporal distribution, Random forest model

摘要： 目的　分析2000—2019年我国不同时期白纹伊蚊空间分布，为精准防控白纹伊蚊提供参考。方法　以“白纹伊蚊”、“监测”、“调查”、“密度”、“分布”、“暴发”等为检索词，在国内外数据库中检索有关2000—2019年我国白纹伊蚊分布信息的文献，提取文献文题、发表时间、采样点、采样时间、捕蚊方法、蚊种及密度等信息，并通过百度地图获取采样点经纬度信息。于美国国家气候数据中心获取我国境内气象观测点气象要素数据，计算年最高气温、年最低气温、1月平均气温、7月平均气温、气温年较差、气温日较差和相对湿度，并进行克里金插值。于国家青藏高原科学数据中心获取2000—2019年我国1 km分辨率的月度累计降水栅格数据和逐月平均气温栅格数据，累加计算年降水量和年平均气温。于WorldPop网站获取2000—2019年我国人口密度数据。于中国科学院地理科学与资源研究所网站获取我国国内生产总值（gross domestic product，GDP）数据。将上述各类数据以5年为1个时间段，计算2000—2004年、2005—2009年、2010—2014年和2015—2019年阶段数据。采用分类随机森林（random forest，RF）模型，对2000—2019年及上述各阶段白纹伊蚊分布数据进行建模，预测全国范围内白纹伊蚊分布情况，并基于我国七大主要气候区进行分析。采用准确率、精确率、召回率和曲线下面积（area under curve，AUC）评估模型效果。采用平均准确率下降（mean decrease accuracy，MDA）法评估RF模型中各特征重要性。结果　共检索到1 191篇中文文献、391篇英文文献，其中580篇文献标注了白纹伊蚊采样点详细信息和具体采样年份，符合纳入标准。累计纳入2000—2019年我国2 234个白纹伊蚊采样点。经建立RF模型发现，2000—2019年我国白纹伊蚊整体分布区域以东南和西南省份为主，在东北部辽宁省等沿海地区呈散在分布。不同时期白纹伊蚊分布RF模型准确率为0.915 ~ 0.947、精确率为0.933 ~ 0.975、召回率为0.898 ~ 0.978、AUC值为0.902 ~ 0.932。RF模型特征中，人口密度为影响我国白纹伊蚊分布最重要的因素，其次为GDP，各气象因素在白纹伊蚊分布模型中所起的作用相对较小。在我国七大主要气候区中，白纹伊蚊分布区域几乎全部集中于边缘热带湿润地区、北亚热带湿润地区和暖温带半湿润地区。2000—2004、2005—2009、2010—2014年和2015—2019年上述3个区域白纹伊蚊分布面积占全国白纹伊蚊总面积的比例分别为100.0%、100.0%、100.0%和99.9%，其中暖温带半湿润地区白纹伊蚊分布面积占比由20.2%逐渐增至30.2%。结论　我国白纹伊蚊主要分布于东南和西南省份，受人口与经济因素影响较大。我国暖温带半湿润地区逐渐成为白纹伊蚊分布热点地区。

关键词: 白纹伊蚊, 时空分布, 随机森林模型