Abstract:
The aim of this study is to summarise the changing behaviour of rainfall in space
and time over the northeast region of Bangladesh. During the study various
rainfall parameters were investigated and mathematical relationships were
established.
Updated isohyetal maps of mean monthly, monsoon (June-September) and annual
rainfall were constructed for the study region. Spatial distribution of mean
monthly, monsoon and annual number of rainy days were investigated. Normal
monsoon rainfall of selected 32 stations within the region have been estimated and
normal annual rainfall chart has been updated. Temporal variability of rainfall and
number of rainy days over the region has been investigated. Moving average and
percent departure of rainfall for every year with reference to the normal rainfall
were worked out for this purpose.
Attempts were made to establish mathematical relationships between number of
rainy days and monsoon rainfall total and between mean daily rainfall intensity
and monsoon rainfall total. Various forms of equations were examined.
Analytical probability distributions were fitted to 1-, 2-, 3-, 5-, 7- and lO-day
maximum rainfall values. Normal, lognormal and Gumbel (EV1) distributions
were examined and all of them were found good for maximum rainfall values.
Gumbel distribution was found to be better. Pattern of Gumbel parameters over
the region for 1-, 3-, 5- and 7-day maximum rainfall were presented on the base
maps. Binomial and Poisson discrete probability models were tested to monsoon
and annual number of rainy days. Normal approximation to Poisson distribution
was found suitable to describe monsoon and annual number of rainy days.
HYMOS (HYdrological MOdelling System), developed at Delft Hydraulics, the
Netherlands, has been extensively used for probability distribution and other
analysis in performing the study.
