PDA

View Full Version : Flooding frequency



Iain Scott
December 16, 2016, 03:11 AM
I have run some historical rainfall via a continuous sim, and would like to analyse the results to find the frequency of flooding from a (set of) nodes, and the total flood volume for each flood event.

The stats template function appears to be the best for this: when I set it to look at Flood Volume on nodes with flood type Stored, and the threshold to "0" (or a slightly higher threshold if I want to ignore trivial amounts of flooding), the output successfully gives me a row for each flood event, and I can find the flood volume by checking the peak value. However, when I try this with Volume Lost on nodes with flood type Lost, I get events which start when the node first floods, only ending at the end of the sim, and a peak value of the total volume lost across the period of the entire sim.

This seems to be because the Flood volume is recorded on a per-timestep basis (i.e. each time step records the flood volume at that time step) while the Volume Lost is recorded on a cumulative basis (i.e. each time step records all the volume lost prior to that time step).

Is there a way to look at volume lost on a per-timestep basis, or otherwise get the frequency and event volume results I want?

Using ICM v6.5 (for compatibility with the client), but the behavior seems to be the same across ICM v7 and CS v15.

Andy Bolden
December 19, 2016, 03:57 AM
If it is only the end time you wish to see, set up a statistics template for level in the 'Lost' nodes. Not sure whether it would be best to set 1mm below the cover level or at cover level. Either way it would show each surge in flooding.

Another approach is to define a long weir (so as to avoid excessive head) at ground level and for this to connect to a dummy outfall. Flooding can then be interrogated as a flow through the weir link using the statistics template. Raise the ground level of the 'Lost' nodes so as to make sure the head needed to drive flow over the weir can still be achieved - hence why it's best to use a wide weir as you don't want flooding still to be lost from the node; you want it all to go over the weir you have created.

A third approach is to export a level hydrograph on the 'Lost' nodes. This will of course give you the level trend rather then simply the start and stop which you would get from the Statistcsic template approach (but it won't give flood volumes). With the weir technique you can export a flow hydrograph to show the rates at which water is flowing out of the 'Lost' nodes.

(It is unfortunate that Stored and Lost flooding are two different values in ICM. I believe they should be united because flooding is flooding is flooding as far as victims of flooding are concerned. I have seen instances where the presence of flooding from 'Lost' nodes has not been spotted as only the stored flooding has been interrogated. Innovyze suggested that the way to show flooding of either on the Geoplan is to include a SQL IIF in the theme so that which ever type of flooding is non zero will be used to display as node circles. When you replay an event, 'Stored' flooding will show as progressively more circles and then reduce, whereas 'Lost' flooding will increase to a maximum and then stay at the maximum until the end of the event).