Unité Mixte de Recherche MIVEGEC, UM1-UM2-CNRS 5290-IRD 224, Institut de Recherche pour le Développement, Montpellier, France.
PLoS One. 2012;7(2):e31843. doi: 10.1371/journal.pone.0031843. Epub 2012 Feb 14.
The question of sampling and spatial aggregation of malaria vectors is central to vector control efforts and estimates of transmission. Spatial patterns of anopheline populations are complex because mosquitoes' habitats and behaviors are strongly heterogeneous. Analyses of spatially referenced counts provide a powerful approach to delineate complex distribution patterns, and contributions of these methods in the study and control of malaria vectors must be carefully evaluated.
METHODOLOGY/PRINCIPAL FINDINGS: We used correlograms, directional variograms, Local Indicators of Spatial Association (LISA) and the Spatial Analysis by Distance IndicEs (SADIE) to examine spatial patterns of Indoor Resting Densities (IRD) in two dominant malaria vectors sampled with a 5 × 5 km grid over a 2500 km(2) area in the forest domain of Cameroon. SADIE analyses revealed that the distribution of Anopheles gambiae was different from regular or random, whereas there was no evidence of spatial pattern in Anopheles funestus (Ia = 1.644, Pa<0.05 and Ia = 1.464, Pa>0.05, respectively). Correlograms and variograms showed significant spatial autocorrelations at small distance lags, and indicated the presence of large clusters of similar values of abundance in An. gambiae while An. funestus was characterized by smaller clusters. The examination of spatial patterns at a finer spatial scale with SADIE and LISA identified several patches of higher than average IRD (hot spots) and clusters of lower than average IRD (cold spots) for the two species. Significant changes occurred in the overall spatial pattern, spatial trends and clusters when IRDs were aggregated at the house level rather than the locality level. All spatial analyses unveiled scale-dependent patterns that could not be identified by traditional aggregation indices.
CONCLUSIONS/SIGNIFICANCE: Our study illustrates the importance of spatial analyses in unraveling the complex spatial patterns of malaria vectors, and highlights the potential contributions of these methods in malaria control.
疟疾媒介的抽样和空间聚集问题是媒介控制工作和传播估计的核心。按蚊种群的空间格局复杂，因为蚊子的栖息地和行为具有很强的异质性。对空间参考计数的分析为描绘复杂的分布模式提供了一种强大的方法，必须仔细评估这些方法在疟疾媒介研究和控制中的作用。
方法/主要发现：我们使用相关图、方向变程图、局部空间关联指标（LISA）和距离指示空间分析（SADIE）来检查喀麦隆森林地区一个 2500 平方公里的区域内，用 5×5 公里网格对两种主要疟疾媒介进行采样的室内静止密度（IRD）的空间模式。SADIE 分析表明，冈比亚按蚊的分布不同于规则或随机分布，而无瓣按蚊则没有空间模式的证据（Ia=1.644，Pa<0.05 和 Ia=1.464，Pa>0.05）。相关图和变程图显示在小距离滞后处存在显著的空间自相关，表明在冈比亚按蚊中有大量相似丰度值的聚类存在，而无瓣按蚊的特征是较小的聚类。用 SADIE 和 LISA 在更精细的空间尺度上检查空间模式，确定了两种蚊种的高于平均 IRD（热点）和低于平均 IRD（冷点）的几个斑块。当 IRD 在房屋水平而不是地点水平上聚集时，总体空间模式、空间趋势和聚类发生了显著变化。所有的空间分析都揭示了依赖于尺度的模式，这些模式不能通过传统的聚集指数来识别。
结论/意义：我们的研究说明了空间分析在揭示疟疾媒介复杂空间模式方面的重要性，并强调了这些方法在疟疾控制中的潜在贡献。
