Record Details

GEE-TED: A tsetse ecological distribution model for Google Earth Engine

Harvard Dataverse (Africa Rice Center, Bioversity International, CCAFS, CIAT, IFPRI, IRRI and WorldFish)

View Archive Info
 
 
Field Value
 
Title GEE-TED: A tsetse ecological distribution model for Google Earth Engine
 
Identifier https://doi.org/10.7910/DVN/6JR87X
 
Creator Peter, Brad
Messina, Joseph
 
Publisher Harvard Dataverse
 
Description GEE-TED: A tsetse ecological distribution model for Google Earth Engine



Associated publication forthcoming:


Fox, L., Peter, B. G., Frake, A. N., and Messina, J. P. (Forthcoming). A Bayesian maximum entropy model for predicting tsetse ecological distributions. Journal TBD.



Description


GEE-TED is a Google Earth Engine (GEE; Gorelick et al. 2017) adaptation of a tsetse ecological distribution (TED) model developed by DeVisser et al. (2010), which was designed for use in ESRI's ArcGIS. TED uses time-series climate and land-use/land-cover (LULC) data to predict the probability of tsetse presence across space based on species habitat preferences (in this case Glossina Morsitans). Model parameterization includes (1) day and night temperatures (MODIS Land Surface Temperature; MOD11A2), (2) available moisture/humidity using a vegetation index as a proxry (MODIS NDVI; MOD13Q1), (3) LULC (MODIS Land Cover Type 1; MCD12Q1), (4) year selections, and (5) fly movement rate (meters/16-days). TED has also been used as a basis for the development of an agent-based model by Lin et al. (2015) and in a cost-benefit analysis of tsetse control in Tanzania by Yang et al. (2017).



Parameterization in Fox et al. (Forthcoming):


Suitable LULC types and climate thresholds used here are specific to Glossina Morsitans in Kenya and are based on the parameterization selections in DeVisser et al. (2010) and DeVisser and Messina (2009). Suitable temperatures range from 17–40°C during the day and 10–40°C at night and available moisture is characterized as NDVI > 0.39. Suitable LULC comprises predominantly woody vegetation; a complete list of suitable categories is available in DeVisser and Messina (2009). In the Fox et al. (Forthcoming) publication, two versions of MCD12Q1 were used to assess suitable LULC types: Versions 051 and 006. The GeoTIFF supplied in this dataset entry (GEE-TED_Kenya_2016-2017.tif) uses the aforementioned parameters to show the probable tsetse distribution across Kenya for the years 2016-2017. A static graphic of this GEE-TED output is shown below and an interactive version can be viewed at: https://cartoscience.users.earthengine.app/view/gee-ted.
 
Subject Computer and Information Science
Earth and Environmental Sciences
 
Contributor Peter, Brad