Dry-season wetlands derived from Landsat archive for NAWRA

Dry-season wetlands data show the average value derived from each season’s maximum number of consecutive days where surface water has been identified during the dry seasons from 1988 to 2016 from Landsat imagery. This dataset was created for the CSIRO Northern Australia Water Resource Assessment (NAWRA) project.\nThese data are maps of dry-season wetlands derived from the Landsat archive from 1988 to 2016 (i.e. 29 years). The Landsat archive data were extracted from Geoscience Australia’s Surface Reflectance NBAR product available on the National Computing Infrastructure. The method used to derive the surface water maps is described in the NAWRA technical report (Sims et al., 2016). The data are provided as average number (from the 29 years of Landsat data) of the maximum consecutive days of inundation during each dry season. The average number of maximum consecutive days is then converted into shape files: one showing all pixels containing an average value of 50 or more maximum consecutive days of inundation from all the dry seasons; the other showing all pixels containing an average value of 100 or more maximum consecutive days of inundation from all the dry seasons.\nMore information is provided in the supporting attachment.\nLineage: The following steps were performed to derive the dry-season wetland maps:\n•\tThe Landsat maps of surface water (described in Sims et al., 2016) for each dry season (defined as May to October) were used to estimate the maximum number of consecutive days that each pixel is inundated. For pixels obscured by cloud cover, if they were mapped as water before and after a period of cloud cover then the pixel was inferred to be inundated for the full duration. If the pixel was dry at the end of the cloudy period then the pixel was assumed to be dry since the last flooded image.\n•\tThe average value of maximum consecutive days of inundation from each dry season was then calculated.\n•\tTo help reduce commission errors, especially where water is mistakenly mapped due to dark shadows from steep topography, the null layer from Geoscience Australia’s Water Observations from Space (WOfS) dataset was used to mask out these areas. In the Fitzroy region, where areas were still incorrectly mapped as permanent water due to the unique topography of the area, an additional step was taken to remove erroneous water features. In this situation, another mask was created based on the maximum water extent as identified from the WOfS archive. This mask can be used on a dry season product since it is not analysing extreme flood events.\n