Impact of Variational Data Assimilation for Simulating Tropical Cyclones over Bay of Bengal Using WRF-ARW - Monitoring and Prediction of Tropical Cyclones in the Indian Ocean and Climate Change - page 232

A number of studies have reported improvements in intensity and track
forecasts of cyclones over the Indian Ocean region as for example using with
multi-satellite data assimilation for Orissa Super Cyclone-1999, Gonu-2007
and Mala-2006 using WRF (Singh et al., 2008), with use of both conventional
and satellite data in WRF for Nisha-2008 (Srinivas et al., 2010), with data
assimilation in WRF-NMM for Gonu, Sidr (Pattanayak and Mohanty, 2008,
2010) and with satellite winds in WRF (Osuri et al., 2012) etc. Singh et al.
(2008) reported that the assimilation of QuickSCAT and SSMI winds improve
the position of the initial vortex, its strength leading to significant improvements
in the intensity and track predictions for Orissa super cyclone. Govindankutty
et al. (2010) studied the impact of 3DVAR assimilation of Doppler Weather
Radar (DWR) wind data on tropical cyclone predictions using WRF. All the
above studies are related to studying the impact of particular datasets on tropical
cyclone simulation. The operational forecasting agencies use global surface/
upper air observations transmitted through Global Telecommunication System
(GTS) and in PrepBufr format for the assimilation purpose. It is essential to
study the performance of numerical models for the tropical cyclone simulation
using the operationally assimilated observations over the North Indian Ocean
for quantifying the error statistics. However, there are no studies available in
India discussing impact of data assimilation in operational mesoscale model
tropical cyclone forecasts. This gap motivated authors to make an attempt to
examine the impact of assimilation of operationally assimilated observational
dataset on the simulation of few tropical cyclones using 3DVAR with a double
nested WRF-ARW model.
2. Numerical Experiments
The Advanced Research WRF (ARW) model (Skamarock et al., 2008) is
designed with two domains of 27 km and 9 km resolutions with 28 vertical
levels (Fig. 1) with inner domain covering the Bay of Bengal and adjoining
coastal land regions. Ten tropical cyclones (Table 1) which formed during
Fig. 1: (a) Domains used for cyclone simulations; (b) NCEP PrepBufr observations
ingest in WRF assimilation system for JAL cyclone.