MODIS Water Mask


The Water Mask product is designed and generated at the University of Maryland, Department of Geography.

The Water Mask product is distributed by the GLCF in GeoTIFF format using an in-house tile scheme based on a modified UTM grid. The standard MODIS tile format data may also be obtained from the LPDAAC (Land Process Distributed Active Archive Center).

The 250m water mask product utilizes the SWBD(SRTM Water Body Data) and complement it with information from 250m MODIS data to create a complete representation of global surface water. The original intent of this product is not for hydrologic modeling, rather for masking water in products. For the purpose of modeling it may be useful to perform an intelligent interpolation to close those gaps in small rivers.

The SWBD was used for the region that it is available and MODIS 250m data was used to supplement for the areas where gaps exist in the SWBD and for the entire region between 60° and 90° N. The SWBD was used because of its fine spatial resolution(90m) and because of its consistent representation of the land surface. Since the SRTM data was collected over a short time step, 11 days, it provides a temporally coherent representation of surface water. Additionally, the cloud penetrating properties of the RADAR offers superior performance over spectral data alone, particularly in cloudy areas such as the humid tropics. Using this remotely sensed data product has the advantage of a single source of information, unlike the vector data sets which are dependent on disparate sets of information to create a single data set.

The SRTM mission did not collect data north of 60°N latitude hence, MODIS 250m composites are used for areas between 60° and 90° N to create the new 250m resolution land/water mask. The data were classified using a decision tree classification. Training data were derived from a subset of the inputs, and the decision tree results from iterative runs with multiple training sets. MODIS 250m composites were then processed through the decision tree to produce water detection per time period. Multiple time periods were combined to produce a clean land water mask. This process has been described in a peer reviewed publication, entitled "A New Global Raster Water Mask at 250 meter Resolution", soon to be published in the International Journal of Digital Earth, December 2009.

The MODIS Mosaic of Antarctica (MOA) shoreline data set was used as a replacement for the current 1km product for Antarctica. It has been reprojected, converted from vector to raster and subsetted into MODIS tiles and included in the beta release of the new 250m water mask.

All data sets mentioned here are available free of charge from various websites and have either been published or used in products that have appeared in peer reviewed publications. Please see acknowledgements.

For more detailed information on bands, products, and processing, please review the associated documentation file.

MODIS Watermask
Quick Links
Copyright © 1997-2016 University of Maryland. All rights reserved
University of Maryland National Aeronautics and Space Administration GOFC_GOLD