REMOTE SENSING OF POTENTIAL MOSQUITO BREEDING SITES

Abstract

Human-induced changes in hydrology (e.g., creation of stagnant water) are pervasive and are inextricably linked to processes that affect national health and disease control in myriad ways. The causal relationships between mosquitoes, stagnant water and malaria are well appreciated among scientists, however the contribution of neglected swimming pools to the stock of stagnant water is largely unknown in Harare. Although Harare is not a malaria risk area, one of the major fears associated with climate change is the emergence of deadly tropical vector borne diseases (e.g., malaria) to areas previously free from those diseases. In this work we used a combination of field-based survey techniques and object-based image analysis (OBIA) of GeoEye-1 imagery to identify neglected swimming pools in Harare. Our object-based image classification algorithm achieved a very high accuracy (overall accuracy 92% and overall Kappa Coefficient 70%) in extracting neglected swimming pools and showed that thirty-seven percent of swimming pools in Harare are neglected. The average age of the neglect is 9.6 ± 4.2 years. We hypothesize that in the event of climate change-induced emergence of malaria in previously malaria-free areas, malaria vectors (e.g., Anopheles gambiae mosquitoes) may use neglected swimming pools as opportunist stages for oviposition. Our results upgrade the database for potential breeding grounds for mosquitoes, support malaria control efforts and highlights the utility of geo-information and remote sensing science to disease control.