Can you help me interpret the results of the Urban Stormwater Retention model? In particular, how should the runoff ratio and the retention ratio be interpreted?
In my case, the runoff ratio varies between 58 and 98. I have assumed that these values represent the percentage of water that does not infiltrate into the ground. At the same time, the retention ratio varies between −57 and −97, so I have assumed that these values represent the percentage of water that infiltrates into the ground, or at least does not contribute to runoff.
Is this interpretation correct?
Another question concerns the fact that the values representing the runoff ratio and the retention ratio are identical to the CN_B values in the biophysical table (considering that the entire study area belongs to hydrologic soil group B). Is this simply a coincidence, or is it expected?
With the caveat that i haven’t used this model in the real world, i’d say that you’re right about your interpretations of the runoff and retention ratios. The Interpreting Results section of the User Guide chapter says this:
retention_ratio.tif: A raster derived from the LULC raster and biophysical table RC_x columns, where each pixel’s value is the stormwater retention ratio in that area
runoff_ratio.tif: A raster derived from the retention ratio raster, where each pixel’s value is the stormwater runoff ratio in that area. This is the inverse of ‘retention_ratio.tif’
So yes, the runoff ratio is the proportion of precipitation that does not infiltrate and is likely to run off in a storm, and the retention ratio is the proportion of precipitation that is likely to infiltrate, so not likely to run off.
These values are mapped directly from RC_A, RC_B, RC_C and RC_D values in the biophysical table, so they should be identical to the biophysical table values.
@Bretto since you referred to CN_B, i clarified something with the model author that i was unsure of, and want to pass along. The runoff coefficients (RC_*) in the Urban Stormwater biophysical table are not the same as curve numbers. Some of our other models (Seasonal Water Yield and Urban Flood Risk) do use curve numbers along with the soil hydrologic group raster, but Urban Stormwater is different.
The short version of his explanation: The key difference between a Runoff Coefficient and a Curve Number is time scale - RC is at an annual scale and CN is at the event scale. Runoff coefficients are derived from annual totals of rainfall and of runoff volume (or depth, if normalized over a watershed). Curve numbers are derived from lots of rainfall vs runoff observations, and describe how much runoff is likely to occur over a single large storm as a function of soil infiltration capacity and antecedent dryness/wetness, and uses a formula that makes adjustments for how much runoff is abstracted by the land surface at the start of the storm. This latter adjustment means that curve number ends up being proportional to runoff, but is not quite the same as a runoff coefficient.
So be sure to look for runoff coefficients to populate the biophysical table, not curve numbers.
And we need to add this clarification to the User Guide.
Thank you for your feedback. May I ask where I can find the runoff coefficients for each hydrologic soil group and each land cover type? Thank you again.
Fabio, that is one of the sticky points of this model, there don’t seem to be general default values for RC like there are CN and other model parameters. Instead, they are sometimes reported in watershed studies related to the specific study area, if those studies exist, so it is worth looking.
Since i have not used this model in the real world, all i can do is point you to the Data Sources section of the Stormwater chapter of the User Guide, which provides a little bit more guidance. We will work on updating the User Guide with explicit information about the difference between RC and CN, and we’ll dig around to see if there are any example sets of RC values that can be used for reference, or other more practical guidance.
If anyone out there has experience assigning RC parameters for this model, please chime in!