For certain dates and region we have binary images (masks) of flooded area (pixel value 1) and non-flooded area (pixel value 0) and we also have the full Digital Elevation Model for the region at the same pixel size. We would like to implement easy calculation (hopefuly automatic) of pixel water column (water depth at every pixel). Thanks for any hint.
Hey Ricardo,
with a bit of boolean algebra the solution seems to me not that difficult. In case you use ArcGIS you can use Raster Calculator to perform the manual analysis. If I undestood correct, you have only the boolean (binary mask) for flooded and not flooded areas. Use this mask to estimate the level of the water table. In raster calculator:
out_raster1 = DEM * binary_mask
the result will give you elevation values of DEM for areas which are flooded and zeros for non flooded areas. Consequently the highest elevation value represents the water table (e.g. for range of values for output raster is [0:240] the elevation value of 240 correcponds to the water table level.
Further use the water table value to substract the DEM to obtain water depths.
water_depths = Value_water_table * binary_mask - out_raster1
This would be the manual way. Of course it can be automated with model builder or a small python script. But in order to be able to obtain the maximum value automatically the Zonal satistics tool should be implemented. Schema of the automatic Workflow could look like this:
1. Calculate out_raster1: DEM * binary_mask
2. Calculate zonal statistics of out_raster1 using Spatial Analyst Tools -> Zonal-> Zonal statistics (Calculates statistics on values of a raster within the zones of another dataset.; for your case MAXIMUM has to be specified as output statistics)
3. Calculate depths: stats_out_raster1 - out_raster1
Repeat the workflow for each date and region.
In case of any additional questions, don't hesitate to ask.
Best regards,
Jewgenij
Thanks a lot Jewgenij
this is exactly what we were applying but I forgot to mention that floodind mask may be fragmented in different water bodies, so that one should calculate maxima DEM values for every single bodie in order to get the contour levels and water column. Any hint to do so automatically?
Hey Ricardo,
I have not tried it yet, but what I could imagine is that if you go with the polygon features into the analysis (each polygon have its own OID) then the Zonal statistics will provide a MAX value of the each respective zone, which can have of course different water levels.
To obtain the polygons from your mask raster you could use Conversion tools -> From Raster of the ArcToolbox.
Important is, of course, that the fragmented water bodies do not have connetions via streams over areas with gradient (should be clearly distinguished). If you could give me a small figure to get an idea of the distribution, I could provide you more hints or even write a small python script to automate it. Should not take to much efforts to do that.
Best regards,
Jewgenij
Thanks again Jewgenij
for your suggestions. The raster layers we automatically generate are also automatically prepared and uploaded for WMS at http://venus.ebd.csic.es/imgs/. Sorry it's in Spanish, but if you tick in the left pannel "Máscara de inundación" you will have the flooding mask. There you can have a look on the kind of data we deal with. The calculation would be applied just over one small portion of the area. There are stream connections but we could still keep the criterium to the topest flooded elevation per patch, because flooding might be very variable, from the marshland completely flooded up to few shallow ponds. Download is not available so I am sending you 1 example mask. The blue pixels are flooded ones, white area is the maximum area able to be flooded and green is outside the study area. Hope it may help.
Hallo Ricardo,
I wrote a small python script that manages the problem. You'll find the toolbox attached.
It works as follows:
First you have to specify the workspace you are working in (is a folder where the intermediate data can be stored), then provide the data DEM and your mask raster (mask raster should be added to the current ArcMap project). Specify a value in the mask that delineates the flooded areas (e.g. 1). And finaly give the child a name (e.g. wat_depth). Run the script.
During the execution following steps are performed:
- based on the value given for flooded areas the mask raster is converted in polygons
- Zonal statistics are calculated for each polygon to estimate the MAX value and thus the water table.
- water_table raster minus DEM provide the water depth
Dependent on the system the output can be added automatically to the current document. If not an additional code (AddLayer) can be included. try it. If some questions or remarks arise, don't hesitate to ask. Also in case of possible errors.
Best regards,
Jewgenij
Jewgenij Torizin,
Have you seen a method to cope with flooded areas connected to streams or that present some degree of slope on the water surface? I'm trying to estimate the flooded water volumes on a catchment in Australia, but my inundated extension does present a gradient, following the river shape on the landscape (I used Modis images to get the inundated area, which is blue in the image).
Regards
- Badges
- Science topic
- Similar topics
- Agricultural Science
- Soil Science
- Soil Analysis