Analyze Model Results

At the completion of a model run, clicking on the Output tab of the OmsEtCompare window shows the output file name of the current run and the file names of previous runs made with this model. Note that they are numbered in sequence, folder 003 contains the results of the last run. Expanding the efcarson.csv file shows the output icon efcarson.

Double click on the efcarson icon. This creates a plot containing all the output variables that were generated by the output component of the model. The x-axis has the dimension of time and the y-axis ranges from the minimum to the maximum value of all the variables plotted. To the right of the plot is the efcarson - Properties window. All the variables on the plot are listed in the window. The check mark next to the variable name indicates that this variable is being displayed on the plot. In the plot below, five variables are being displayed

Deleting the check mark next to a variable name will cause the variable to be removed from the plot. In the example below the basin-tmax (daily basin-average maximum temperature) has been removed from the plot

Another feature of the efcarson - Properties window is the Plot Type selection. Clicking on the down arrow in the Plot Type field opens a list of available plot types. The plot type displayed here is the Combined Scale plot. All variables are on the same plot using a single scale on the y-axis.

The second Plot Type option is Multiple Scale. All variables are on the same plot and each has had their y-axis values rescaled to fit in the full plotting window. The y-axis scale values for each variable are displayed on the right side of the plot using the same color as the related plot trace.

The third Plot Type is the Vertically Arranged Multiple plot. Here each variable is plotted separately and stacked vertically. Each has the same x-axis (time) but each has its own y-axis scale.

The last Plot Type is the Horizontally Arranged Multiple plot. Each variable has its own plot but each uses the same y-axis scale. The plots are arranged side-by-side.

Two other options in the efcarson - Properties window are the Subtitle and Title of the plot. In this case efcarson is shown as the Title and PRMS output is shown as the subtitle.

Another option for output display is a spreadsheet. Click on the Spreadsheet tab to display the output variables in spreadsheet form.

The plotted variables can become difficult to read when a long period of record is plotted. One capability is to zoom in on a section of interest. You can do this by rubber banding around that section of interest. Pick a point above and to the left of the region on interest and move the mouse pointer to that position. Hold the left mouse button down and drag a rectangular box around the area of interest.

When you release the mouse button, the area selected is now displayed.

Another function provided by the mouse pointer is the ability to query any data point on the plot. Below the mouse pointer has been positioned over the peak identified by the blue arrow. When the pointer is on the plot line, the variable name and the x- and y-axis values are displayed in a data field.

If you want to return to the full view of the plot, click anywhere in the plot with the right mouse button. In the options window that appears, select Auto Range and then select Both Axes. This will bring the plot back to full view.

Access to the output files can also be obtained using the Projects window. One of the directories under the project OmsEtTutorial is Model Output. Expanding that directory shows the subdirectory OmsEtCompare. Expanding this directory shows the output folders from all of the previous runs. Opening the most recent folder, in this case 003, you can see the contents of the output folder. Double clicking on the efcarson directory will provide the same plot results as those displayed above when folder 003 was accessed through the output tab of the model run.

However, accessing the output folders in the Projects window provides additional data analysis capabilities. In many cases, you may want to customize the contents of desired graphs or apply additional analysis tools prior to displaying analysis results. In the Projects window, the Data Analysis directory can provide those options. Click on Data Analysis with the right mouse button. In the options window that opens select New and then Data Analysis with the left mouse button.

This will open a Data Analysis window in which you will provide a name for this specific data analysis. For the first analysis we are going to plot only the basin_potet_hamon and the basin_potet_jh values for comparison. Type the name hamon-jh_PEt in the Name: input field and click Finish with the left mouse button.

Looking in the Projects window you will see that the hamon-jh_PET.oad node has been created under the Data Analysis directory.

Click on hamon-jh_PET.oad node with the right mouse button and select Add Plot from the options window using the left mouse button.

A New Plot node is now added under the hamon-jh_PE.oad node.

The next step is to add the desired output variables to the New Plot node. Select the variable date under efcarson with the left mouse button. Holding the left mouse button down, drag the date icon to the New Plot node.

When the left mouse button is released, the date variable will be linked to the New Plot node. Note that the link is made using the path to the date variable in folder 003.

Then using the same drag and drop procedure, select the basin_potet_hamon varible and drag and drop it below the date variable in the New Plot node.

Use the same drag and drop procedure to copy the basin_potet_jh variable to the New Plot node.

Double clicking on the ham-jh_PET.oad node with the left mouse button produces a plot of basin_potet_hamon and basin_potet_jh values with date displayed on the x-axis.

Creating a New Plot node using the Data Analysis tools enables you to create your own graphical analysis presentations. In the example above the output variables from output folder 003 were selected and plotted. If you plan to run the model multiple times and you want to view the results at the end of each run, modifications to ham-jh_PET.oad node need to be made. The current implementation of the ham-jh_PET.oad node uses hard-wired paths to the variables in folder 003. You can modify this path to always use the most current run of the model. Click on the first variable in the New Plot node and the output/OmsEtCompare/003/... window will open to the right of the plot. Click on the selection box at the end of the File Name input field with the left mouse button and an editor window will open showing the path output/OmsEtCompare/003/efcarson.csv.

To make this a generic plot, edit the path by replacing 003 in the path window with the term $current. Then click OK with the left mouse button. This will cause the New Plot node to always use the most current model run for this output variable.

You can see the change in the path displayed in the Data Analysis | ham-jh_PET.oad node analysis.

Edit the paths of the other two elements in the ham-jh_PET.oad node using the same procedure as above to replace 003 with $current. Now each plot variable of ham-jh_PET.oad node will always use the most current model output.

Another type of analysis available can be used to compute the sum of the difference between two variables. In this case we can sum the difference between teh daily values of basin_potet_ham and basin_potet_jh. To add this analysis click on Data Analysis with the right mouse button. Then click New and Data Analysis with the left mouse button.

In the Data Analysis window that opens, enter the name PET-diff in the Name input field.

This will add the PET-diff.oad node to the Data Analysis directory.

Click on the PET-diff.oad node with the right mouse button and select Add Plot with the left mouse button.

Click on the New Plot node with the right mouse button and then Add with the left mouse button. An option menu will open with several plot type options. The options are

• Data -
• Expression -
  
• Error Sum Analysis -
• Flow Duration Analysis -
  
• Residuals Analysis -

Select Error Sum Analysis with the left mouse button.

An Error Sum node is now created under the New Plot node and has two data elements, firsData and secondData. The variable assigned to secondData will be subtract from the variable assigned to firstData each time step. The difference will then be summed each timestep and plotted.

To assign a variable to firstData, click on basin_potet_jh with the left mouse button. Holding the left mouse button down, drag and drop this variable to the firstData icon.

Assign basin_potet_hamon to secondData using the same drag and drop procedure. At the completion of this assignment the plot of the error sum will be displayed. As above, this plot is for output variables in output folder 003 only. If it is desired to conduct this analysis at the end of multiple model runs then the procedure described above to replace 003 in the variable path with $current can be applied.