Chikowore Gerald, Dicko Ahmadou H, Chinwada Peter, Zimba Moses, Shereni William, Roger François, Bouyer Jérémy, Guerrini Laure
Tsetse Control Division, Department of Livestock and Veterinary Services, Ministry of Agriculture, Mechanisation and Irrigation Development, Harare, Zimbabwe.
Centre de Coopération Internationale en Recherche Agronomique pour le Développement, UMR ASTRE CIRAD-INRA « AnimalS, health, Territories, Risks and Ecosystems », Campus international de Baillarguet, Montpellier, France.
PLoS Negl Trop Dis. 2017 May 3;11(5):e0005566. doi: 10.1371/journal.pntd.0005566. eCollection 2017 May.
Tsetse (Glossina sensu stricto) are cyclical vectors of human and animal trypanosomoses, that are presently targeted by the Pan African Tsetse and Trypanosomiasis Eradication Campaign (PATTEC) coordinated by the African Union. In order to achieve effective control of tsetse, there is need to produce elaborate plans to guide intervention programmes. A model intended to aid in the planning of intervention programmes and assist a fuller understanding of tsetse distribution was applied, in a pilot study in the Masoka area, Mid-Zambezi valley in Zimbabwe, and targeting two savannah species, Glossina morsitans morsitans and Glossina pallidipes.
METHODOLOGY/PRINCIPAL FINDINGS: The field study was conducted between March and December 2015 in 105 sites following a standardized grid sampling frame. Presence data were used to study habitat suitability of both species based on climatic and environmental data derived from MODIS and SPOT 5 satellite images. Factors influencing distribution were studied using an Ecological Niche Factor Analysis (ENFA) whilst habitat suitability was predicted using a Maximum Entropy (MaxEnt) model at a spatial resolution of 250 m. Area Under the Curve (AUC), an indicator of model performance, was 0.89 for G. m. morsitans and 0.96 for G. pallidipes. We then used the predicted suitable areas to calculate the probability that flies were really absent from the grid cells where they were not captured during the study based on a probability model using a risk threshold of 0.05. Apart from grid cells where G. m. morsitans and G. pallidipes were captured, there was a high probability of presence in an additional 128 km2 and 144 km2 respectively.
CONCLUSIONS/SIGNIFICANCE: The modelling process promised to be useful in optimizing the outputs of presence/absence surveys, allowing the definition of tsetse infested areas with improved accuracy. The methodology proposed here can be extended to all the tsetse infested parts of Zimbabwe and may also be useful for other PATTEC national initiatives in other African countries.
采采蝇（狭义舌蝇属）是人类和动物锥虫病的周期性传播媒介，目前是由非洲联盟协调的泛非采采蝇与锥虫病根除运动（PATTEC）的目标对象。为了有效控制采采蝇，需要制定详尽的计划来指导干预计划。在津巴布韦中赞比西河谷的马索卡地区进行的一项试点研究中，应用了一个旨在协助干预计划规划并帮助更全面了解采采蝇分布的模型，该研究针对两种稀树草原物种，即莫氏采采蝇和淡黄采采蝇。
方法/主要发现：2015年3月至12月期间，按照标准化网格采样框架在105个地点进行了实地研究。利用存在数据，基于从MODIS和SPOT 5卫星图像获取的气候和环境数据，研究了这两种物种的栖息地适宜性。使用生态位因子分析（ENFA）研究影响分布的因素，同时使用最大熵（MaxEnt）模型以250米的空间分辨率预测栖息地适宜性。曲线下面积（AUC）作为模型性能的指标，莫氏采采蝇为0.89，淡黄采采蝇为0.96。然后，我们使用预测的适宜区域，根据一个风险阈值为0.05的概率模型，计算在研究期间未捕获采采蝇的网格单元中采采蝇实际不存在的概率。除了捕获莫氏采采蝇和淡黄采采蝇的网格单元外，分别在另外128平方公里和144平方公里的区域存在较高的出现概率。
结论/意义：建模过程有望用于优化存在/不存在调查的结果，从而更准确地界定采采蝇出没区域。这里提出的方法可以扩展到津巴布韦所有采采蝇出没的地区，也可能对其他非洲国家的其他PATTEC国家倡议有用。
