TN-12 : Methodology for estimation of design storm

Abstract

Estimation of 'Design Flood is a first and vital step in the design process for a large range of water resources development works. The design flood is generally derived from the design storm using some rainfall-runoff procedure. The design rainfall comprises of three components namely the design rainfall magnitude, its distribution in time and areal pattern.

While the 'Design Storm' is derived by both statistical and physical methods, the type and magnitude of the design storm is decided by the type of structure and the magnitude of risk which could be associated with it. The intensity duration-frequency (i-d-f) relationship falls under the category of statistical methods while the probable Maximum Precipitation (PMP) estimation is done by both statistical and physical analysis,. Both methods have certain limitations on account of the subjectivity inherent in the analysis, assumptions made and constraints imposed by data availability.

The methods available and those in practice in different countries are reviewed in this technical note and the advantages and limitations are identified and highlighted. More often, the scepticism in respect of design storm concept was the comprehensiveness of the concept for covering the risk which the engineer assumed it would provide. The other limitations arise out of the uncertainty of an intensity picked up for a certain duration and for a particular recurrence interval providing flood of corresponding recurrence interval This, as is to be expected depends on the rainfall-runoff procedure/ model used for deriving the runoff from rainfall. While the limitations of the rational formula are well known, the usage of synthetic rainfall hyetograph also has its due share in influencing the magnitude of the runoff. The other approximation arises when the intensity-duration-frequency relationships which are developed from point rainfall values are extended to provide areal estimates.

In the case of the physical approach the main limitation arises out of the small data sample available at the disposal of the analyst and the extent of representativeness of the sample to the problem area. Other approximations relate to use of surface dew point temperatures as representative of the precipitable water in the atmosphere above and extending these procedures for use in tropical areas.

The alternatives suggested include storm modelling, which of course would place a demand on data of a number of parameters at a number of locations and at frequent intervals.