Hello,
I am trying to delineate a floodplain of a small creek in ArcGIS Pro 2.9 with Arc Hydro.
I have been following the workflow presented in the recent webinar: Arc Hydro: Flooding and Forecasting.
1. Filled sinks in LiDAR-based DEM
2. Flow Direction and Flow Accumulation
3. Flow Distance to create a HAND raster
4. ArcHydro tool: Define HAND-based Flood Depth and Extent
How do you determine what the 'Input Flood Depth' parameter should be?
The webinar mentions it is the vertical threshold, but am unsure what that means. Is it the max depth of the creek?
I only have LiDAR data, the DEM, and the creek line/raster available.
Any advice on this tool is appreciated, thanks!
Solved! Go to Solution.
When using HAND method for "flooding", the "Input Flood Depth" is the depth above stream centerline that you want to "flood" with.
What that specific value is, is up to you. For example, if you use the "max depth of the creek", that is the equivalent of the "bankfull" and most of the resulting flood polygon should be contained within the channel. Anything higher should start spilling into the overbanks and you should get "flooding".
So where do you get that number? Depends on what you are doing. It can come from the observations (e.g. high water mark or a stage from a gage) or through flow modeling and rating curve to convert flow to depth. If you are doing the rating curve approach, make sure that the rating curve is based on the same resolution data you used to derive HAND with (or is adjusted accordingly).
And remember that inherent assumption in HAND flood modeling approach is that your water surface is uniform (constant) and parallel to the bottom of the channel. For each channel segment. So if you expect that the depth will change (significantly) as you go downstream, you should partition the overall stream into multiple segments that have uniform (but variable from segment to segment) depth and then use "Define HAND-based Flood Depth and Extent from Table" tool to get the floodplain. More work but might give you better results.
If you stream segment is short, then first follow the approach you have presented (single reach), see if you like the results, and if not, then use the "table" approach. Note that you will have to reprocess the input data and partition the streams and get their catchments before you can use the "table" approach.
Hope this helps.
When using HAND method for "flooding", the "Input Flood Depth" is the depth above stream centerline that you want to "flood" with.
What that specific value is, is up to you. For example, if you use the "max depth of the creek", that is the equivalent of the "bankfull" and most of the resulting flood polygon should be contained within the channel. Anything higher should start spilling into the overbanks and you should get "flooding".
So where do you get that number? Depends on what you are doing. It can come from the observations (e.g. high water mark or a stage from a gage) or through flow modeling and rating curve to convert flow to depth. If you are doing the rating curve approach, make sure that the rating curve is based on the same resolution data you used to derive HAND with (or is adjusted accordingly).
And remember that inherent assumption in HAND flood modeling approach is that your water surface is uniform (constant) and parallel to the bottom of the channel. For each channel segment. So if you expect that the depth will change (significantly) as you go downstream, you should partition the overall stream into multiple segments that have uniform (but variable from segment to segment) depth and then use "Define HAND-based Flood Depth and Extent from Table" tool to get the floodplain. More work but might give you better results.
If you stream segment is short, then first follow the approach you have presented (single reach), see if you like the results, and if not, then use the "table" approach. Note that you will have to reprocess the input data and partition the streams and get their catchments before you can use the "table" approach.
Hope this helps.
Thank you kindly Dean, this explanation helps a lot. Your webinars and presentations have been very useful for my project and learning!
I am using the Define HAND-based Flood Depth and Extent from Table tool. Below is my flood depth table, and the H column is my downloaded NWM retrospective FF Q values (cfs). However, running this tool with this table results in the entire watershed being flooded, even for the high-water threshold.
I assume the HINT column is the flood depth but those numbers look really large. What am I doing wrong, and how is this column calculated?
H value should be the flood depth for that reach, not the discharge (measured from the bottom of the channel for that reach). So what you need to do is use the Q (coming from NWM in cfs in your case) and convert that Q to H using the rating curve for each reach. Make sure that H is in the same vertical units that HAND is in (meaning DEM vertical units, unless you did some unit conversion somewhere in the process). For most consistent results, the rating curve should be based on the same elevation model (DEM) as HAND is. HINT is generated as an internal computational attribute when deriving the floodplain depth and extent and is just int(H*100) - not something you need to manage explicitly.
Thanks so much, Dr Djokic, for your reply. This has been helpful. My problem right now is how to convert my NWM FF Qs (downloaded using the “Download NWM Retrospective FF Qs” tool in Arc Hydro Tools Pro) to stage (H) given the size of my AOI (HUC4). Do you have any thoughts or suggestions for a workflow on how to create/obtain a rating curve for this purpose? Also, my understanding is that the “Define HAND-based Flood Depth and Extent from Table” tool subtracts the HAND surface from H to create flood depth and extent layers. Is that correct? Thanks.