Show/Hide Toolbars

Storm Injector Help

RAFTS 2018 projects were created for the Sorell Creek catchment using a CatchmentSIM result export script. After importing the XPX file into RAFTS, two projects were created. To minimise the run time and number of model runs required, only the 6 hour duration was included. The versions created were:

1.ARR16 6 hour 1% AEP ensemble with an ARF but no preburst adjustments (Sorell-Creek-single-storm-with arf - no pre burst.xp)

2.ARR16 6 hour 1% AEP ensemble with an ARF and preburst adjustments (Sorell-Creek-single-storm-with arf - with pre burst.xp)

 

Storm Injector was then run to try and match the results generated by RAFTS internal ARR16 methods. In both cases, the results matched closely. However, there are 2 points worth noting that apply to any RAFTS modelling with Storm Injector:

RAFTS can be sensitive to the routing increment and this should be tested. The default routing increments proposed by Storm Injector may not be low enough to produce accurate results. For these verification projects, a 1 minute routing increment was enforced.

RAFTS applies any pre-burst depth to one (or more) additional timesteps at the start of the rainfall timeseries whereas Storm Injector typically subtracts the pre-burst depth from the Global Initial Loss (with the option to add any excess to the first timestep). One of the differences between these approaches is that, for the RAFTS method, the ARF and continuing losses may reduce the amount of rainfall in the pre-burst timesteps prior to the model running. Storm Injector has an option to apply the ARF to the pre-burst depth which should be turned on in the Settings tab to best match RAFTS results.

 

More details are provided below:

 

ARR16 6 hour 1% AEP ensemble with an ARF but no preburst adjustments

The results from the first test are shown below for Node 1.09.

 

RAFTS_noPB_table

RAFTS results for 'Sorell-Creek-single-storm-with arf - no pre burst.xp'

 

Storm Injector was able to match these results exactly. To reproduce the project in Storm Injector, follow the steps below:

1.Import the sorell-crekk.XPX file using the 1) Import Model >> RAFTS option. Click Yes to calculate lat/long coordinates and select Map Grid of Australia 1994 >> Zone 55.

2.Import the data hub information from file using the drop arrow on 2) Get Hub Data >> load text file and selecting 'hub-data.txt'

3.Import the temporal patterns by right click on the Point (2016) tab of the Temporal Patterns panel and selecting 'Import TPs'. Then select 'SStas_Increments.csv'

4.Import the IFD data using the drop arrow on 3) Get IFD Data >> Import IFD(s) from csv and select 'IFD.csv'

5.Select the 1% AEP / rare combination in the Selected Events panel. Change to the Duration tab and right click >> uncheck all and then check the 6 hr duration only

6.In the Settings tab, change the Initial Loss setting to 'Global initial loss' as we are not considering pre-burst depths in this example.

7.In the Storm Generator tab, Click 4) Create Storms

8.Open the Routing Increments panel. Testing has shown that RAFTS results can be quite sensitive to the Routing Increment. Change it to 1 min to be consistent with the RAFTS project. Change the Number of Increments to 1000 to ensure a long enough model duration.

9.Click 5) Prepare Model Runs and select the 'sorell-creek.dat' file

10.Click 6) Run Models (ensure your paths to the RAFTS engine file is correct and your RAFTS license is accessible).

11.When the model runs are complete, Click 7) Process Results and you should see an Average Q for Subcatchment 1.09 of 64.41 which matches the RAFTS results.

 

RAFTS_SI_noPB

Storm Injector results which match the RAFTS results.

 

ARR16 6 hour 1% AEP ensemble with an ARF and preburst adjustments

To verify that Storm Injector also matches the results when a pre-burst adjustment is used, a new RAFTS project was created called 'Sorell-Creek-single-storm-with arf - with pre burst.xp'

 

The results from the second test are shown below for Node 1.09.

RAFTS_PB_table

RAFTS results for 'Sorell-Creek-single-storm-with arf - with pre burst.xp'

 

To match these results, continue from the Storm Injector project created previously and make the following changes:

1.In the Settings tab, change the Initial Loss method to 'Global initial loss minus Median pre-burst depth' and check the (Apply ARF) checkbox.

2.In the Storm Generator tab, click 4) Create Storms

3.Again, change the Routing Increment to 1min and Number of Increments to 1000

4.Repeat steps 8 through 11 from above

5.You should now see an Average Q for Subcatchment 1.09 of 79.99 which matches well with the RAFTS result of 79.96

 

RAFTS_SI_PB

Storm Injector results which closely match the RAFTS results.