Sensitivity analysis
Although this is simple, it is one of the most powerful techniques available to analysts. Sensitivity analysis involves the running of models with different inputs, allowing their effects on results to be assessed. This provides guidance on the importance of various input factors, allowing you to make decisions on suitable inputs.
As an example, consider the effects of selecting different values of the antecedent moisture condition in a model. Major storm results from one of the demonstration models, the Gymea Pipe Drain example, are shown below. These are obtained with an AMC of 4, established in the Rainfall Data.
Now consider what happens when the AMC in the Hydrological Model is changed to 3, 2 or 1. The results for an AMC of 3 are shown below.
The major storm peak runoff from the sub-catchments changes as follows:
The effect of the AMC and of the percentage impervious (paved + supplementary) areas in sub-catchments is apparent. The information from the various runs can be used to select an appropriate value of parameters, and to point to factors for which more information is needed.
Another example concerns times of flow calculated using the kinematic wave equation. A result from the demonstration example for Penrith is shown below:
As an example, consider the effects of selecting different values of the antecedent moisture condition in a model. Major storm results from one of the demonstration models, the Gymea Pipe Drain example, are shown below. These are obtained with an AMC of 4, established in the Rainfall Data.
Now consider what happens when the AMC in the Hydrological Model is changed to 3, 2 or 1. The results for an AMC of 3 are shown below.
The major storm peak runoff from the sub-catchments changes as follows:
| Pit | % Impervious | Flow for AMC = 4 | Flow for AMC = 3 | Flow for AMC = 2 | Flow for AMC = 1 |
| Pit A.1 | 35% | 0.429 m3/s | 0.396 m3/s | 0.336 m3/s | 0.272 m3/s |
| Pit A.2 | 80% | 0.051 m3/s | 0.051 m3/s | 0.049 m3/s | 0.048 m3/s |
| Pit B.1 | 55% | 0.318 m3/s | 0.299 m3/s | 0.270 m3/s | 0.238 m3/s |
| Pit C.1 | 60% | 0.242 m3/s | 0.233 m3/s | 0.216 m3/s | 0.191 m3/s |
Another example concerns times of flow calculated using the kinematic wave equation. A result from the demonstration example for Penrith is shown below:
Examination of the spreadsheet example for Sub-catchment Cat 1 shows that the grassed area time calculated in a 100 year ARI, 20 minute storm is 7.74 minutes for the surface roughness or retardance of 0.05 assumed in the Sub-Catchment property sheet shown above. If this roughness is varied, the times are:
| Roughness or Retardance, n | Time (minutes) | Peak Flowrate (m3/s) |
| 0.25 | 19.51 | 0.642 |
| 0.1 | 11.47 | 1.01 |
| 0.05 | 7.74 | 1.15 |
| 0.02 | 4.68 | 1.27 |
| 0.01 | 3.26 | 1.28 |
The time of flow, or time of concentration, changes considerably when the roughness is altered, and the calculated flows also change, due to changes in the speed with flows from different parts of the sub-catchment concentrate at the outlet. With this knowledge of the effects of the adopted flow path roughness, and appropriate roughnesses for different surfaces, you can make a judgment on a suitable value.
With the complexity of the results obtained using ARR 2019 rainfall ensembles, sensitivity tests are an essential way to compare different hydrological models and model parameters.