Subject: Streamflow Underestimation and Model Calibration Posted: 4/28/2025 Viewed: 396 times
Dear All,
My catchment was delineated using the Automatic Catchment Delineation tool, and I delineated each catchment upstream of the reservoirs, demand nodes, and gauging stations.
However, in all sections, the modelled monthly streamflow is underestimated compared to the observed flow data at the gauging stations.
What would be the reason and best solution?
Should I use incremental flow from the observed data, or would calibrating the soil moisture parameters be sufficient to achieve a good fit with the observed data?
Or is there another option you would recommend?
Thank you for your time.
Best regards,
Selam
Mr. Doug Chalmers
Subject: Re: Streamflow Underestimation and Model Calibration Posted: 4/28/2025 Viewed: 390 times
Selam,
Using Catchment Delineation produces your climate and land use data, but the soil moisture parameters need to be manually calibrated. You need to calibrate your soil moisture parameters to match your modeled and observed streamflows. Start the gauge location which has the least amount of human diversions and operations upstream of it in your watershed.
For calibration questions, the best resources are our catchment calibration chapter in the WEAP Tutorial. Particularly pages 221, 222-226, and 212 which show how to adjust your hydrologic parameters to achieve the desired results. Additionally, we have video lessons available on calibration on YouTube: https://www.youtube.com/watch?v=Notx1vfeRCk.
It's hard to give very detailed advice on this topic without having worked on your specific model, but here are the best tips I know:
First, study your modeled versus observed streamflow results. Is your annual observed streamflow more or less than observed? What about the average monthly flows? What about for large precipitation storm event peak flows or base flows during dry seasons? Could there be surface-groundwater interactions that need to be represented?
Also, study your results for soil moisture (z1 and z2)- make sure that the soil moisture levels follow a believable trend and are not consistently at 0% or 100%. If your soil moisture pattern is irregular see the associated calibration advice on page 221 of the Tutorial.
If your model is producing simulated flow less than the observed flow, some parameters to adjust could be to decrease Kc, decrease runoff resistance factor, decrease soil water capacity, increase preferred flow direction, and increase root zone conductivity. The Catchment Calibration Tutorial chapter explains further. You can experiment to see which parameters are most sensitive and start by adjusting the most sensitive one first to achieve the best match to observed flows, then move on to the next most sensitive. You may have to run through this processes several times.
Calibration is challenging and time-consuming for any hydrologic model. I wish you the very best of luck with your model.
-Doug
Topic "Streamflow Underestimation and Model Calibration"
