Comparing the predictive power of field survey and multithermal spectral imager (MTI) remote-sensed environmental data for the identification of Anopheles (Diptera: Culicidae) aquatic larval habitats in Kisumu and Malindi, Kenya

Date of Award




Degree Name

Doctor of Philosophy (Ph.D.)


Epidemiology and Public Health

First Committee Member

John C. Beier, Committee Chair


This research evaluates the extent to which utilization of data at 5-meter data from satellite surveys of mosquito larval habitats adjusted for field-based ecological data enhances predictions of mosquito counts. Mosquito larval habitats were sampled and Multispectral Thermal Imager (MTI) satellite data in the visible spectrum at 5-meter spatial resolution were acquired for Kisumu, and Malindi, Kenya, during February and March 2001. All entomological parameters were collected from January to May, 2001, June to August 2002, and June to August, 2003. For Kisumu there was a total of 329 anopheline and 1,022 culcine larvae. Of the 329 Anopheles, all were An. gambiae s.l. of which 88.8% (189 of 213 specimens) were PCR identified as An. arabiensis and 6.6% (14 specimens) as An. gambiae s.s. For Malindi there was a total of 459 anopheline and 4,651 culcine larvae. Of the 459 total Anopheles, 22.1% were identified as An. funestus and 77.9% were identified as An. gambiae s.l., of which 88.3% (95 of 108 specimens) were PCR identified as An. gambiae s.s., 4.3% (5 specimens) as An. arabensis and 1% (2 specimens) as An. merus Poisson and logistic regression models were generated for each urban area and the pooled data for Kisumu and Malindi. To validate the initial models, a set of 54 randomly selected aquatic larval habitats were sampled from June 2003 to August 2003. The R 2 of the final model for Kisumu was 0.35 (p < 0.001). The R 2 of the final model for Malindi was 0.52 (p < 0.001). Additionally, SpaceStat 1.80 spatial analytic tools were used to create Local Indicators for Spatial Autocorrelation (LISA) from field and satellite collected ecological datasets from Kisumu, Malindi and the combination of the two dataset for both urban towns. Both Kisumu and Malindi displayed a weak positive spatial autocorrelation; in Kisumu the Moran's I is 0.075, and for Malindi the Moran's I is 0.295. The spatial model for the combined dataset of Kisumu and Malindi was not appropriate for a linear fit. We conclude from this analysis, that satellite data at 5 meter resolution does not have an additional predictive value for mosquito counts in urban environments for the An. gambiae complex and An. funestus when adjusted for field based ecological data.


Health Sciences, Public Health; Remote Sensing

Link to Full Text