Abstract
The well recognized constraint of non-linear and non-Gaussian
distribution of rainfall observation limits its assimilation in the
high-dimensional numerical weather prediction (NWP) model. In this
study, rain-gauges’ observed rainfall from Indian Meteorological
Department (IMD) over Indian landmass is assimilated in the Weather
Research and Forecasting (WRF) model using particle filter. In the
framework of imperfect weather model, particles (or ensembles) for
rainfall predictions are created with various combinations of model
physics (viz. cumulus parameterization, micro-physics, planetary
boundary layer schemes). With the help of IMD observed rainfall, weights
are provided to various particles using multiple hypotheses, and this is
the step in which IMD rainfall observations are used for assimilation.
Further, a resampling step is performed to generate new particle from
high weight particle using stochastic kinetic-energy backscatter scheme
(SKEBS) method in which dynamical variables are perturbed into the model
physics. Results based on rainfall verification scores suggest that
assimilation of the rain-gauges observed rainfall using particle filter
improved prediction of rainfall over CNT runs (unweighted particle;
without assimilation). Moreover, surface and vertical profile of
temperature, water vapour mixing ratio (WVMR) and wind speed are also
improved in 24 h forecasts.