Storm design for data-scarce locations.

[cadsync]

Good day,

Please, I reside in Africa where it is almost impossible to acquire reliable highres hydrolometeorologic data, and as an aspiring Civil Engineering professional, I would prefer to do my designs to world-class standards . As is the case with designing for data-scarce locations, I have resorted to hydrological modelling using CAD/GIS and remote sensing data. I've been able to acquire satellite-based daily rainfall estimates (FEWS NET) and I would like the experienced professionals in the house to help me answer the following questions:

1) Is it possible to generate IDF curves from daily (24hr) rainfall data or i have to find a way of disaggregating the values to lesser time intervals (assuming i want to use the rational method for my designs)?

2) I have 11 yrs of satellite-based daily rainfall estimates, how feasible is it for me to perform a frequency analysis with this dataset and estimate my 2, 5, 10, 25, 50, 100 year return periods for the 24hr design rainfall? What statistical distribution would you advise me to fit the annual maximums to, cos quite frankly all the distributions will fit this meagre data.

3) How well would the synthetic rainfall distributions (SCS and Hoff) work in locations outside the United States; especially Africa.

If by some form of over-zealousness i am getting this whole design process wrong, could someone please point me in the right direction...Thank you!!!

Out of the mire...a spectacle of perfection. #TheLotusConstruct

 

[dik]

Do you have any local weatherstations? We have numerous private stations the Environment Canada uses their data as part of their forcasting info... If you do, then can this data be compared to data from weatherstations that may have similar weather/climate?

Look up Gumbel plot/distribution; it may be of interest...

Dik

 

[psmart]

Where in Africa? I know that RSA has a standard set of synthetic rainfall distributions for use with the SCS/NRCS method.

Also note that if you have local IDF curves, they can be used to generate custom synthetic rainfall distributions. This would probably be more accurate than using one of the standard US rainfall distributions.


Peter Smart
HydroCAD Software

http://www.hydrocad.net