【作者】 洪波；

【导师】 赵惠燕；

【作者基本信息】 西北农林科技大学 ， 计算机应用技术， 2009， 硕士

【摘要】 农业有害生物,即农业病虫害,对农业生产、人类生活与健康造成了巨大的经济损失;特别是外来有害生物因入侵地的适生环境更易爆发成灾,成为国际社会面临的重大科技问题。加强对外来有害生物预防及控制技术的基础研究,发展早期预警与预防、监测与控制的理论与技术是解决这一问题的根本途径。地理信息系统(GIS)是20世纪六十年代开始迅速发展起来的研究技术,运用GIS将有害生物的爆发特征与其分布区的生物学和自然环境特征联系起来,为有害生物预测预报和综合防治提供有效的宏观管理方案。国内外研究者应用GIS技术在有害生物监测方面,特别是外来有害生物入侵风险评估方面取得了较多的进展。然而,在实际应用中存在三方面问题:1.大尺度范围的外来有害生物现有分布数据不易获得;2.大部分入侵物种的预测方法缺少对预测结果进行模型评价过程,没有完整的技术框架;3.无法对入侵物种扩散路线进行历史推演和预测。对此,本研究在ArcGIS Engine组件的基础之上进行二次开发,将R统计平台和GIS平台结合成为一个整体平台,利用生态位模型分析法研究了外来生物在我国的潜在适生性分布,并尝试探索一种新的外来物种扩散路线历史推演方法,力求明确掌握影响外来生物分布和扩散的制约环境因素。所取得的结果如下:(1)根据需求分析,建立了一套外来有害生物空间分布预测系统,应用于有害生物潜在适生性分布预测分析、物种发生与环境关系分析、外来物种入侵扩散路线历史推演,为这些领域的探索和研究提供了可靠的技术平台。(2)使用GARP和MaxEnt两种不同的生态位模型,预测了7种外来有害生物在我国的潜在适生性分布,并对两种模型的预测结果进行了比较。实验表明,两种模型预测结果的精确率分别达到了80%和90%以上,MaxEnt的预测结果相对于GARP精确度更高,但GARP模型的预测结果则更能反映出外来有害生物的生态位特点。(3)使用最低成本距离和主成分分析相结合的方法,对外来生物马铃薯甲虫在我国新疆地区的扩散路线进行了预测,同时,根据其适生环境对新疆的35个县(市)进行了分类,并与实际调查情况进行了对比分析。实验证明,预测结果比较准确,具有良好的实用性。建立的外来有害生物空间分布预测系统,不仅可用于有害生物分布预测、分析大区域物种种群动态与环境关系,同时能够跟踪和预测外来入侵生物在我国的适生区和扩散路线,为外来有害生物生态治理提供了理论依据。更多还原

【Abstract】 Agricultural pests have caused great economic losses to agricultural production, human life and health, especially, alien pests are easy to outbreak as a result of suitable invasive environment, so it becomes a significant technological problem that international society faced with. The basic approach to solve this problem is to strengthen the basic research of pests’prevention and control, and develop the theory and technology of early warning and prevention, monitoring and management.Geographic information system (GIS) is a rapid developing technology from 1960s. By using GIS, a relationship between outbreak characteristics of pests with biological and environmental characteristics of their distribution area will be established, an effective and macro management solution will be provided for prediction and comprehensive control of pests. In pests monitoring, especially in invasive risk evaluation of alien pests, domestic and foreign researchers have made great progress with GIS; however, there are three problems in practical application: Firstly, alien pests’existing distribution data of large scale range are difficult to obtain. Secondly, most prediction results of invasive species’distribution are lack of model evaluation, and have not formed a complete technical framework. Finally, migration route of invasive species are unable to acquire by making a historical deduction and prediction.To this point, the study had done a secondary development based on ArcGIS Engine component, and established a integral platform by binding R statistic platform with GIS platform. It used niche model to analysis and study potential suitable distribution of alien pests in China, tried to explore a new method to make a historical deduction of alien species’migration route, and strived to clear restrictive factors of alien pests’distribution and migration. Achievements of the research are as follows:(1) According to the demand analysis, a GIS-based pest distribution prediction system was established. It could be applied to predict potential suitable distribution of alien pests, analysis the relationship between species occurrence and environment, make a historical deduction for alien species’migration route, and provide a reliable technology platform for exploring and studying these fields.(2) The system used two niche models GARP and MaxEnt to predict potential suitable distribution of seven alien pests in China, and compared two prediction results. The experiments showed that accuracy rating of two prediction results were respectively above 80% and 90%, MaxEnt’s accuracy was higher than GARP’s, but the prediction results of GARP could reflect more niche characteristics of alien pests.(3) The system used lowest cost distance model combined with principal component analysis to predict migration route of Colorado potato beetles in Xinjiang, made a classification of 35 counties according to its suitable environment, and made a comparative analysis with actual investigation. The experiments showed that the system had more accurate prediction results, and had a good performance.The pest distribution prediction system could not only be used to predict pests distribution, analysis the relationship between population dynamics of large area species and environment, but also track and predict the suitable areas and migration route of alien invasive pests, so it provided a theoretical basis for ecological management of alien pests.更多还原