|
|
|
|
|
|
|
duration in both relative and absolute terms. For all of these nodes, predicted flood waters return to the channel at some point before leaving the catchment; under these circumstances, it is not surprising that flood duration is sensitive to the capacity of the drainage channel. |
|
|
|
|
|
|
|
|
14.3.2
Extent and Frequency of Flooding |
|
|
|
|
|
|
|
|
Predicting and interpreting the areal extent of flooding from model output is difficult. Extremely detailed topography is required to predict the surface area of flooding, particularly in an urban environment where small changes in depth may lead to step changes in flooded area. In addition, SPIDA's simple flooding models have limited capacity to manage such detailed information, even if it were available. |
|
|
|
|
|
|
|
|
We adopted an alternative approach of selecting a sample of nodes at uniform intervals along the drainage channel; the extent of flooding can then be gauged by the number of these nodes at which flooding is predicted. The option of counting the total number of flooded nodes was rejected, because some nodes are closely bunched and would therefore effectively count the same body of flood water twice. At the same time, we wished to have a reasonably large sample of nodes that flooded at least once, so that the sensitivity to various levels of solids could be gauged. On this basis, a stationing interval of 55 m was selected, which led to a sample of 15 of a total of 38 channel nodes. |
|
|
|
|
|
|
|
|
Development of meaningful intensitydurationfrequency (IDF) curves for Indore rainfall is problematic for a variety of reasons, including restricted access to continuous rain-gauge data. The regional approach of Kothyari and Garde (1992) was explored, but found to predict rainfall intensities far greater than credible in light of local experience. A 'design storm' approach based on conventional return period analysis was therefore unsuitable, and we adopted a more pragmatic approach using the actual rainfall |
|
|
|
|
|