Estrada-Peña A, García Z, Sánchez H Fragoso
Department of Parasitology, Veterinary Faculty, Miguel Servet 177, 50013 Zaragoza, Spain.
Exp Appl Acarol. 2006;38(4):307-16. doi: 10.1007/s10493-006-7251-2.
This paper focuses on the distribution of the cattle tick, Boophilus microplus, in Mexico. The study is aimed to understand the climate factors responsible of the recorded distribution that can statistically define the suitable habitat for the tick. Sites where the tick is recorded display significantly higher values of some climate variables in comparison with those where the tick is absent, namely mean monthly temperature (T) and atmospheric water vapour (W), yearly accumulated T, W and rainfall (R) (p < 0.001 for every variable), with smaller significance for the yearly sum of T/R and T/W ratios (p < 0.05). Interestingly, variables involving the Normalized Derived Vegetation Index (NDVI) do not shown statistical differences between the sites where the tick is present or absent. The best set of habitat-defining variables was integrated into a framework to assess the habitat suitability for the tick in Mexico. We used a point-to-point similarity metric to assign a classification value to a candidate site based on the proximity in environmental space of the most similar record site. A combination of 7 yearly and monthly values for temperature, rainfall and water vapour variables captured the tick distribution. Model performance, as tested with a separate set of distribution tests and defined by the AUC value, was 0.89. Causes of errors as detected with a visual comparison of both known and predicted distribution of the tick may be attributed to the use of a medium resolution, unable to capture locally important features of tick distribution, and to incomplete collections in some parts of the country.
本文聚焦于微小牛蜱在墨西哥的分布情况。该研究旨在了解导致所记录分布的气候因素,这些因素能够从统计学角度界定适合蜱生存的栖息地。与未发现蜱的地点相比,记录到蜱的地点某些气候变量的值显著更高,即月平均温度(T)和大气水汽(W)、年累计T、W和降雨量(R)(每个变量p < 0.001),而T/R和T/W比值的年总和差异较小(p < 0.05)。有趣的是,涉及归一化植被指数(NDVI)的变量在有蜱和无蜱的地点之间未显示出统计学差异。将最佳的一组栖息地界定变量整合到一个框架中,以评估墨西哥蜱的栖息地适宜性。我们使用点对点相似性度量,根据最相似记录地点在环境空间中的接近程度为候选地点分配一个分类值。温度、降雨量和水汽变量的7个年度和月度值的组合捕捉到了蜱的分布情况。通过另一组分布测试进行测试并由AUC值定义的模型性能为0.89。通过直观比较蜱的已知分布和预测分布发现的误差原因,可能归因于使用的是中等分辨率,无法捕捉蜱分布的局部重要特征,以及该国某些地区的采集不完整。