| Successive
days of precipitation are known to cause flooding in monsoon-susceptible
countries. The analysis of extreme precipitation trends is important
for the prediction of high precipitation events. Forecasting of
daily precipitation facilitates the prediction of the occurrences
of rainfall and number of wet days. Using the maximum five-day
accumulated precipitation (MX5d) index, we can predict the magnitude
of precipitation in a specific period as it may indicate the extreme
precipitation. Traditionally, a data-driven model is built on a
whole data set describing the phenomenon within the data. This
type of model does not considered the seasonal processes embedded
in the data. Therefore, there is a need to built a model that encompassing
different time scale and localized. One of the ways of doing this
is to discover the different physically embedded relationships
in precipitation process at different seasonal period and to built
separate localized models for each of these seasonal periods. The
use of committee models such as modular and ensemble models in
weather and hydrological forecasting are increasing day by day.
The study uses the modular concept by separating the heavy precipitation
events based on sub-processes which are the seasonal monsoon and
trained the subset of seasonal data using data driven models. Besides
that, the study is carried out to evaluate the influence of global
climate indices on local precipitation. It is interesting to see
the influence of combining global and local predictors on local
precipitation events. The method used past MX5d data and global
climate indices such as Southern Oscillation Index (SOI), Madden
Julian Oscillation (MJO), and Dipole Mode Index (DMI) in Kuantan
and Kota Bharu, Malaysia using modular model trained on subset
of data that represent the seasonal monsoon. The analysis started
with evaluating the local and global inputs (MX5d with SOI, MJO,
and DMI) in order to investigate the concurrent effect of lagged
values of local precipitation data and global climate indices on
seasonal extreme precipitation. Four subset of data are sampled
representing two major seasonal variations in Malaysia. The experimental
data are focused on the east coast area of Malaysia such that the
effect of northeast monsoon season causes heavy precipitation events.
The results show that the combination of local and global modes
in a modular model is favorable than a single local mode. The proposed
modular model is promising an encouraging result when different
subset of data are trained on separate methods with different parameter
values.
|
|
|
|
|
|
|
|