From 1987 to 1995, NASA's Ames Research Center, in collaboration with university and health agency scientists, conducted research on the ecology of the Anopheles albimanus mosquito, a key vector of human malaria in the coastal areas of southern Chiapas, Mexico. The field research focused on the relationship of An. albimanus to environmental variables associated with regional landscape elements, including larval habitats, bloodmeal sources, and resting sites. The results indicated the importance of flooded pastures and transitional wetlands for larval habitat, cattle in pastures for bloodmeal sources, and trees for potential resting sites. The remote sensing research involved identifying and mapping these landscape elements, along with seven others, using multitemporal Landsat Thematic Mapper (TM) data. NASA ER-2 aircraft imagery was used to create a map of human settlements, from which 40 villages were randomly selected. These villages were the focus of a study to examine the relationship between landscape elements and mosquito-human contact risk (i.e., malaria risk).
A geographic information system (GIS) was used to calculate the proportion of each landscape element within a 1-km buffer surrounding each village. This 1-km radius was based on the typical flight range of an adult An. albimanus mosquito; within this flight range, she must find bloodmeals, resting sites, and larval habitat in order to reproduce. All 40 villages were sampled weekly for mosquito abundance throughout the mid- to late wet season when malaria transmission is high.
The relationships between mosquito abundance and the landscape proportions were investigated using stepwise discriminant and regression analyses. These analyses indicated that the most important landscape elements in terms of explaining mosquito abundance were the proportions of transitional wetlands and unmanaged pasture. Using these two landscape elements as predictors, we were able to correctly distinguish villages with high and low mosquito abundance, with an overall accuracy of 90%. The models developed using these functions were blind-tested in another area of the Chiapas coastal plain in order to assess the their accuracy and portability. In the test, the discriminant model was able to predict 79% of the high mosquito abundance villages, while the regression was able to identify 7 of the 10 villages with highest abundance.
The landscape approach, which integrates remotely sensed data and GIS capabilities to identify villages with high vector-human contact risk, provides a promising tool for malaria surveillance programs. In general, this approach could be applied to other diseases in areas where the landscape variables critical to disease transmission are known, and these elements can be detected using remote sensing.
Last updated: 22 Jan 1999