Frequently Asked Questions

Please consult the ARR chapter for guidance on the design variable method. Additional questions are answered here relating to use basic implementation of the method using software.

What does the software do?

The software implements the design variable method for joint probability extremes. It integrates the probability density of jointly occurring extremes for specified water level exceedances.

How do I get AEPs for the margins?

You must determine the marginal distribution prior to using this software from available data for your case study. Software such as Tuflow Flike is freely available for fitting univariate probability distributions to extremes.

Sometimes there is a grey line on the plot. Why?

The grey line indicates an extrapolation zone, where the limited size of the hydraulic response table has some impact on the quality of the best estimate. This suggests that the upper AEPs of the hydraulic response table are too low and cause some loss of precision in the estimated output. E.g. if the flood table was limited to 0.5% AEP on both margins, it means that there are combinations of rarer events impacting this estimate, but as the water levels are not specified in the table for this region, an extrapolation assumption was made.

The grey line represents a best estimate which is the average of the most conservative and least conservative options for the water level in this region outside the table.

How critical is the grey line to the best estimate?

For each output AEP we calculate the difference between the complete dependence and complete independence cases. This gives you an overall indication of the potential magnitude of the joint probability effect. A threshold is set to 10% of the range between these two water levels (complete dependence and independence). When the assumptions made in the extrapolation region (conservative assumption and non-conservative assumption) result in estimates that differ by an amount more than the 10% threshold, this is indicated in the plot. The further along the 'grey line' the more critical the issue becomes (larger discrepancy between two estimates).

Can I only apply the software to coastal problems?

The guidance in the ARR chapter is for coastal problems but the design variable method itself is general. The software is ignorant of whether the margins are tide/flow/rain.

Can I model uncertainty with the software?

No. Uncertainty estimates are technically feasible, but complex. It is an advanced application not offered in the software.

Can I model 2D grids with the software?

No. This option is technically feasible, but has multiple logistical complications (e.g. transmitting large files, grid preparation and processing time on the order of hours/days, transmission of results).

It is recommended to identify key locations along the main channel(s) and prepare hydraulic response tables for those grid cells and separately input them to the software.