Use of Soil Moisture Active Passive satellite data (SMAP) and Worldclim2 data to predict the potential distribution of visceral leishmaniasis and its vector Lutzomyia longipalpis (Psychodidae: Phlebominae) in Sao Paulo State and Bahia State, Brazil

Abstract:

Abstract

Visceral leishmaniasis (VL) is a neglected tropical disease that is transmitted by Lutzomyia longipalpis, a sandfly that is widely distributed in Brazil. Despite efforts to strengthen current national control programs for VL, reduction in its incidence and geographical distribution in Brazil has not yet been successful and VL is in fact expanding its range, particularly in newly urbanized areas. Development of ecological niche models (ENM) for use in surveillance and response systems may enable more effective operational VL control programs in Brazil by mapping geospatial risk areas and elucidation of eco-epidemiologic risk factors. ENM for VL and, L. longipalpis, were generated using monthly Worldclim 2.0 data (30-year long term climate normal, 1km spatial resolution) and monthly SMAP L4 soil moisture data obtained through a search of NASA’s EarthData site. SMAP L4 Global 3-hourly 9 km EASE-Grid Surface and Root Zone Soil Moisture Geophysical Data V004 were downloaded for the first image of day 1 and day 15 (0:00-3:00 hour) of each month. Models were developed using MaxEnt software to generate risk surfaces based on a computer algorithm for maximum entropy. Initial models were run in Maxent and the jackknife procedure was used to identify contribution of each variable to model performance so that the most meaningful components were used to generate ENM potential distribution maps using ArcGIS 10.6. Principal component analysis (PCA) was performed to reduce the dimension and collinearity of the dataset. Risk models based on classical climate station data have been described for VL in Brazil based on thermal and hydrological drivers and limiting factors on life cycle development (Nieto et al., 2009). Similar eco-epidemiological patterns of VL and vector distribution in Sao Paulo state and Bahia state were observed using SMAP models as compared to Worldclim2 models based on temperature maximum/minimum and precipitation or water budget (precipitation-potential evapotranspiration). Results indicate direct earth observing satellite measurement of soil moisture by SMAP can be used in lieu of models calculated from classical thermal and precipitation climate station data to assess VL disease risk and to guide control program interventions.