Manual Calibration

Manual calibration is an iterative process that improves with experience and is a very important step in modelling. However, it can be cumbersome. One needs to run the model, extract and evaluate results by plotting and calculating indices, and then go to change parameter and repeat the process untill results are satisfactory.

SWAT+ Toolbox streamlines this process by automating some of the most cumbersome parts.

1. Setting Parameter Values

Use the parameter table on the left to select and modify parameters. Click on the value column of the parameter row to type in your desired value. You can enable or disable parameters using the checkbox in the first column to compare the 'with' and 'without' results for that parameter.

Saving and Loading Parameters

• Save Current Parameters: Enter a name in the 'Parameter Set Name' field and click 'Keep Current Parameter Set' to save your current parameter values.
• Load Parameters from File: Click 'Load From calibration.cal' to load parameters from an external file.
• Manage Saved Parameters: Use the 'Saved Parameter Sets' dropdown to View, Load/Set, or Delete previously saved parameter sets.

2. Running with New Parameters

After making parameter adjustments, click 'Run SWAT+' to apply the parameters and run SWAT+. Monitor the model run progress in the global progress bar at the bottom.

3. Evaluating Model Performance

Click on the 'Refresh Metrics' button to calculate new performance indices. The performance panel displays multiple metrics including NSE, KGE, MSE, RMSE, PBias, and MARE for each observation.

In the case that you have more than one observation, you can go to the graphical evaluation of each observation by clicking on it in the 'Performance Index Panel'. Use the arrow at the bottom left to show or hide the graphs panel.

The graphical view displays: - Time Series: Observed vs Simulated data comparison - FDC/LDC Toggle: Switch between Flow Duration Curve and Load Duration Curve - Month Plot: Toggle to view monthly aggregated data - Export Graph: Export the current graph in various formats

Remember to do a model check once you get satisfactory results to see your model's potential issues.