Mapping the distribution of the main host for plague in a complex landscape in Kazakhstan: An object-based approach using SPOT-5 XS, Landsat 7 ETM+, SRTM and multiple Random Forests

Abstract

Plague is a zoonotic infectious disease present in great gerbilpopulations in Kazakhstan. Infectious disease dynamics are influenced bythe spatial distribution of the carriers (hosts) of the disease. Thegreat gerbil, the main host in our study area, lives in burrows, whichcan be recognized on high resolution satellite imagery. In this study,using earth observation data at various spatial scales, we map thespatial distribution of burrows in a semi-desert landscape.The study area consists of various landscape types. To evaluate whetheridentification of burrows by classification is possible in theselandscape types, the study area was subdivided into eight landscapeunits, on the basis of Landsat 7 ETM+ derived Tasselled Cap Greennessand Brightness, and SRTM derived standard deviation in elevation.In the field, 904 burrows were mapped. Using two segmented 2.5 mresolution SPOT-5 XS satellite scenes, reference object sets werecreated. Random Forests were built for both SPOT scenes and used toclassify the images. Additionally, a stratified classification wascarried out, by building separate Random Forests per landscape unit.Burrows were successfully classified in all landscape units. In the‘steppe on floodplain’ areas, classification worked best:producer's and user's accuracy in those areas reached 88% and 100%,respectively. In the ‘floodplain’ areas with a moreheterogeneous vegetation cover, classification worked least well; there,accuracies were 86 and 58% respectively. Stratified classificationimproved the results in all landscape units where comparison waspossible (four), increasing kappa coefficients by 13, 10, 9 and 1%,respectively.In this study, an innovative stratification method using high- andmedium resolution imagery was applied in order to map host distributionon a large spatial scale. The burrow maps we developed will help todetect changes in the distribution of great gerbil populations and,moreover, serve as a unique empirical data set which can be used asinput for epidemiological plague models. This is an important step inunderstanding the dynamics of plague.