Hi all,

Not a straight question, more of a knowledge gaining exercise.....

I've found that when undertaking 2D models (rainfall direct onto mesh with gullies modelled) that I find flow tends to get trapped behind a number of properties, and as a result the model predicts flood depths at these properties.

In reality, either the mesh is too coarse and flow wouldn't reach this property, or the private drainage of the properties would convey this flow into the drainage network.

Obviously without high level investment you will not have private drainage details for 100's of properties and can only work with the Lidar resolution data available.

I was wondering what other people have found. Do you review the results and discount properties predict to have a flood depth and no historic evidence? I reviewed the methodology of another consultant which mentioned discounting areas of flooding less than 250m2 to remove the issue mentioned above.

Get arounds I can think of are:
-insert dummy gullies where you no there is no evidence of flooding.
-increase infiltration rate to remove this flood depth
-discount properties manually / based on some form of sql so these properties are not 'at risk'