湖沼型血吸虫病疫区钉螺分布空间抽样优化研究

Snail distribution in marshland schistosomiasis epidemic areas features vast extent, strong spatial uncertainty, poor unithood etc, posing a great challenge for snail survey and control. The study is intended to choose several marshlands by environmental types, from which an experimental sector is delimited as a test field from each marshland, and a bulldozer scanning method is used to survey snail count and micro geographic factors Kuang by Kuang (0.33m*0.33m in size). The surveyed data is then transferred into GIS snail grid map and its attribute grid maps of the micro geographic factors, and the Snail Population Truth obtained. Geostatistic method is employed to calculate Snail Clustering Scales and Directions in different marshlands. The Kuang Interval Effects and Kuang Size Effects on snail spatial sampling results in the present Systematic Sampling Method will be examined by GIS spatial analysis, the calculated Snail Density and Area with Snails by different Kuang interval and Kuang size combinations will be compared to the truth, and optimal Kuang interval and Kuang size combination then determinated. Weights of Evidence and Data-driven Geographic Model will be adopted to study the quantitative spatial relations between snails and micro geographic factors. Relationship curve between Environmental Kuangs and Area with Snails is applied by comparison with snail theoretical truth to explore a scientific method and practical operation procedure whereby to pinpoint the exact environmental Kuang points when calculating Area with Snails in the present Systematic Sampling plus Environmental Sampling Method.

湖沼型血吸虫病疫区钉螺分布具有地域分布广、空间不确定性强、分布单元性差等特点，给实地查螺、灭螺带来巨大挑战。拟在中国湖沼型血吸虫病疫区，选择几个典型的环境类型草洲，划分试验样地，用"推扫式"逐框普查法进行钉螺和微地理因子详查，将其转换成GIS钉螺格网数据图层和微地理因子属性数据图层，并获得钉螺总体统计"真值"。用地统计学方法研究钉螺空间聚集的尺度规模与方向。用GIS空间分析技术研究现有"系统抽样法"钉螺空间抽样的"框点间距效应"与"框点尺寸效应"，把不同"框点间距-框点尺寸"组合方式下得到的"钉螺密度"与"有螺面积"与"真值"比较，确定最优的"框点间距-框点尺寸"组合。用证据权重和数据驱动的地理模型研究钉螺与微地理因子间的空间定量关系；通过"环境框点-有螺面积"关系曲线和与钉螺理论"真值"的对比，探索现有"系统抽样+环境抽样法"计算"有螺面积"时确定环境框点的科学方法和实用操作流程。

近年全国血吸虫病疫情下降明显，但有螺面积下降有限，感染风险依然存在。卫生部制订的《血吸虫病防治手册》中湖沼型血吸虫病疫区钉螺系统抽样方案受到了国内外学者的质疑，也没有适当的抽样精度评价指标。为提高钉螺调查的科学性，有必要从经典统计学高度，深入研究我国几十年来一直沿用的钉螺调查方法。本课题基于典型有螺草洲推扫与抽样的钉螺、植物与土壤数据，采用多种抽样方案进行了钉螺调查有关指标的对比研究；为精确定位辅助变量在分层抽样中的作用，用确定性系数法、证据权重模型等进行了钉螺与环境因子的关系研究；用聚集度指标、回归分析和聚集因子指数进行了钉螺分布格局的样点尺寸效应研究；用空间统计学方法进行了钉螺全局与局部自相关和钉螺空间统计特征的样点尺寸效应研究。鉴于钉螺与植物等因子的空间关系，用数量生态学方法进行了植物群落与环境因子的关系研究。主要研究结果：（1）确认了中国一直沿用的钉螺系统抽样方案存在的科学问题，因为草洲钉螺一般具有带状分布特征，钉螺密度排列存在与抽样间隔对应的周期规律，不适合采用系统抽样。结果提示，借助与钉螺有关的辅助变量进行空间分层，可以减少工作量并提高抽样精度。（2）首次研究了基于一个以上辅助变量的钉螺空间分层方法，相对基于先验知识的空间分层可靠性更高，也为将来涉及多个辅助变量的空间抽样分层提供了一种新的途径。（3）确定性系数法的钉螺易发指数模型增加了因子权重；从几何概型角度将证据权重法的二位目标变量扩展应用到具有强度差别的连续变量。（4）深入研究了钉螺空间聚集强度、聚集方式和生物习性导致的聚集行为的样点尺寸效应。(5) 通过推扫式、全覆盖普查钉螺数据，洞察到一般抽样调查无法获得的格点尺度单个样点钉螺的局部细微空间自相关连续变化规律、钉螺空间自相关的纯样点大小块金值及其尺度效应；预示了钉螺空间抽样的最佳样点尺寸范围(2.64m×2.64m - 3.33 m×3.33 m)，与目前血防部门一贯沿用的0.33 m×0.33 m尺寸明显不同。