Subject: Caliberation of WEAP model Posted: 2/2/2016 Viewed: 2169 times
I want help regarding caliberation and validation of WEAP model. Can anybody tell me which are the potential parameters in WEAP which i can use to calibrate the model or can anyone send me any document related to calibration, validation or sensitivity analysis of WEAP model?
I have streamflow, climatic data and using catchment in model. I am using MABIA method.
lookin forward to hear.
Subject: Re: Caliberation of WEAP model Posted: 2/2/2016 Viewed: 2158 times
In using these catchment methods, we are dealing with parameters that we do not have exact values for. These parameters are all potentially negotiable in the calibration process.
To start calibrating, I would do a few things for clarity's sake:
1) Create a favorite chart that will compare the modeled streamflow with your streamflow records at the gauge in question. This will help you access the results quickly rather than having to make the same chart each time.
2) If you have multiple catchments, consider having single values for some of the parameters, which can be added as key assumptions and adjusted there. For instance, the cloudiness fraction may not vary between different catchments in a single watershed.
3) Whenever you want to change a parameter, make the change in a scenario. For example, have a "initial bucket 2 depletion" scenario so when when you change that variable, you can compare it against its original value. Scenarios can be combined in the scenario tree.
4) It's good practice, when you're modeling, to have a calibration period where you try to match the streamflow as closely as possible, and then another period that you don't include in your calibration activities but you can also use for comparison to see how effective your calibration was for other periods of records. This second period is called the validation period.
5) When you're starting to feel satisfied with your calibration, export the results for streamflow and observed data and compare them for indicators like %bias, r-squared and Nash-Sutcliffe. These statistical quantifications can be included in any reports to help readers understand how accurate your model can be viewed to be.
Also, remember that if you're working in a river with multiple gauges, always start your calibration with the gauge closest to the headwaters.