Satellite-Derived Monthly Dataset of Precipitation, Land Surface Temperature, and NDVI for Small Administrative Units in Ecuador (2005–2024)
Description
This dataset provides monthly satellite-derived summaries of three environmental variables for continental Ecuador (January 2005 – December 2024): precipitation (mm), land surface temperature (°C) and the normalized difference vegetation index (NDVI). Data are aggregated at the two finest official administrative levels of the country — 218 cantons and 1,034 parishes — defined by the 2022 INEC boundaries (excluding the Galápagos province). For each canton and parish, four statistical summaries are provided per month: minimum, maximum, average, and standard deviation. In addition, long-term monthly climatologies (2005–2024 average for each calendar month) are included for all variables and at both administrative levels. The dataset comprises 8 Excel files (.xlsx): • 6 files of monthly time series (one per variable and administrative level), each with 4 worksheets corresponding to the statistical summaries. • 2 files of long-term monthly climatologies (one per administrative level), each with 3 worksheets corresponding to the variables. Column headers include the official INEC code and administrative name, followed by monthly value columns labeled var_YYYYMM (monthly time series) or var_MM (climatologies). In addition, two shapefiles updated to the 2022 INEC administrative structure are included (canton-level and parish-level), with the official INEC codes as join attributes. Data were derived from NASA’s Earth Observing System Data and Information System (EOSDIS) satellite products (GPM IMERG for precipitation, MODIS 11C3 for temperature, and MODIS 13C2 for NDVI). All products were processed using TerrSet, reprojected to UTM with 100×100 m pixels, and aggregated at the parish level using zonal statistics. Files are provided in Excel format (.xlsx), one file per variable, with four sheets corresponding to each summary statistic. Column headers include official INEC parish codes, parish names, year, month, and value. This dataset is suitable for environmental monitoring, climate variability analysis, spatial epidemiology, and public health research, particularly for studying the environmental determinants of vector-borne and climate-sensitive diseases.
Files
Steps to reproduce
Three satellite products were downloaded from NASA's EOSDIS for January 2005 – December 2024: GPM IMERG V07B for precipitation (mm/h, 0.1°), MODIS MOD11C3/MYD11C3 V6.1 for land surface temperature (scaled 16-bit integer, 0.05°) and MODIS MOD13C2/MYD13C2 V6.1 for NDVI (scaled 16-bit integer, 0.05°). 2. Variable-specific preprocessing (TerrSet liberaGIS v20.05) • Precipitation (GPM IMERG): leftward rotation to correct spatial orientation; conversion to mm/month by multiplying pixel intensities by the number of hours in the month. • Temperature (MODIS 11C3): missing pixels filled with the Adaptive Box filter; conversion to °C as LST(°C) = (LST_Day_CMG × 0.02) − 273.15. • NDVI (MODIS 13C2): rescaling by 10,000; pixel-wise averaging of Terra and Aqua composites. 3. Reprojection All variables were reprojected to UTM zone 17S at 100 × 100 m resolution using bilinear resampling. 4. Administrative boundaries (2022) Official INEC vector boundaries are publicly available only for the 2012 division. The 2022 canton- and parish-level shapefiles were generated by performing dissolve operations on the INEC census geodatabase using the corresponding administrative attribute fields, resulting in 218 cantons and 1,034 parishes for continental Ecuador. 5. Monthly time series extraction Zonal statistics were applied to the monthly rasters using the 2022 boundaries (Extract tool, TerrSet), producing monthly minimum, maximum, average and standard deviation values for each variable and administrative unit. 6. Long-term monthly climatologies For each calendar month, the 20 annual layers (2005–2024) were grouped, summed and divided by 20 using TerrSet Macro Modeler. For NDVI, a quality control mask (Reclass + Cover) was applied to the resulting climatology rasters to remove invalid coastal pixels. Average values were then extracted for each canton and parish.
Institutions
- Universidad UTEPichincha, Quito