I am looking to generate flood plain polygons to represent 20 and 100 year flood zones, in the most accurate and efficient way possible, given the following datasets:

  1. River cross section polylines
  2. River flow polylines (center lines between opposing shorelines)
  3. 20 year and 100 year elevation values for the vertices of each cross section
  4. A digital elevation model (raster with an elevation value for each pixel) of the land surface

Here is my proposed methodology so far:

  1. Form polygons delimiting each section, the river center line and inland areas well beyond the flood plain
  2. For each polygon formed in the previous step, extract the intersecting raster surface (Extract by Mask)
  3. Isolate the pixels of the DEM that are less than or equal to the flood elevation (20 and 100 year) value for the section using the Reclassify tool
  4. Perform a raster (reclassified raster from previous step) to vector (polygons) to form floodplains between each section.
  5. Merge the polygons together to form a continuous floodplain

I am looking for insight and suggestions to produce the most accurate results possible given the available data.

I have access to ArcGIS and the Spatial Analyst and 3D Analyst extensions.

  • I am also working almost in the same area but my focus is to understand the impact of precipitation and city development over the flood frequency and intensity. I try to delineate the watershed with different methods, but frankly speaking no model is too good to get good results, but comparatively archydro tool is not a bad option if you choose outflow (pour point) point/s wisely. Similarly, it also depends on the resolution of your DEM data. Did you try River Bathymetry Toolkit (RBT) and HEC-Ras? my next step is to compare the results of these different tools and still I dont have Cross-secti – user9765 Aug 26 '12 at 2:02
  • Have you considered SAGA GIS for your needs? It has very powerful hydrology modules. Also there is a TauDEM utils that are avilable for ArcGIS (and for other GIS too). General suggestion is to use opensource tools for modelling, because you will know the algorithm used and will be able to find corresponding scientific article to asses its precision. – SS_Rebelious Aug 26 '12 at 12:18
  • Thanks to both of you for your comments. I am looking into Geo HEC-Ras; however, I don't have all of the necessary datasets to effectively use the tools available within Geo HEC-Ras. I will also check out RBT and SAGA. I have scripted my procedure and I am able to generate flood plains based the flood elevation assigned polygons. – Brent Edwards Aug 27 '12 at 14:19
  • 1
    If you can have access to GRASS GIS. see this grasswiki.osgeo.org/wiki/Hydrological_Sciences – rkm Nov 19 '12 at 4:54
  • Clarify. Do the vertices from 1) contain two attributes that basically say 20year and 100year or are they actually numeric values 5ft, 6ft etc. – user17614 Aug 5 '13 at 18:58

I recommend you try the Surface Difference tool in 3D Analyst. The datasets you described are what is used as input for this tool as shown in the Help topic, Floodplain delineation from Lidar points. Lidar is used to create the dem used in the tool but you already have that so can ignore that part.

| improve this answer | |

I did a similar research focusing on the risk due to flooding at various return periods. The best inundation mapping tool that i found was HEC-Ras and HEC-Geo-Ras extension used in ArcGIS helped in visualization of those inundation levels.

| improve this answer | |

Not the answer you're looking for? Browse other questions tagged or ask your own question.