Numerical Weather Prediction (NWP) technique uses mathematical models of the atmosphere to predict weather based on current conditions. During the last two decades, weather forecasting all over the world has greatly benefited from NWP models. Significant improvement in accuracy and reliability of NWP products has been driven by sophisticated numerical techniques (advanced data assimilation algorithms, new observation systems, improved modeling techniques, physical parameterisations, advancements in computing power allowing for higher spatial resolution) and by the phenomenal increase in satellite observations.
The objective cyclone prediction system (CPS) is a collective approach to address various components for improving cyclone forecasts (Fig 1). The CPS comprises five forecast components, namely, prediction of cyclogenesis by cyclone genesis potential parameter (GPP); multi-model ensemble (MME) technique for cyclone track prediction; intensity prediction by statistical cyclone intensity prediction (SCIP) model; rapid intensification (RI) prediction by RI-Index; and, prediction of decay of a cyclone after landfall by a statistical model.
A cyclone genesis parameter, termed the GPP for the northern Indian Ocean (NIO) is developed based on four variables—vorticity at lower level, middle tropospheric relative humidity, middle tropospheric instability, and the vertical wind shear. A number of low-pressure systems form over the NIO, but not all intensify into cyclones. Therefore, the objective is to understand the potential for intensification of a system at its early stages of development. The composite GPP value is found to be around three to ﬁve times greater for systems developing into cyclones than for non-developing systems. The analysis of the parameters at early development stage is found to provide useful predictive signals for intensification of the system.
The MME technique is based on a statistical linear regression approach. The predictors selected for the ensemble technique are latitudinal and longitudinal positions at 12 hour intervals for up to 120 hours of five operational NWP models. When there is a wide variation in forecasts arrived at by various models, the consensus forecast generated by the MME technique provides very useful guidance on cyclone track forecast.
The ensemble forecast products from NCMRWF, ECMWF, NCEP, UKMO, Meteorological Service Canada (MSC) and other centres are available for near real-time for the NIO region. These products include ensemble track forecasts and strike probability maps. The strike probability forecast for Phailin based on ensemble of ensembles (super ensemble or grand global ensemble) is shown in Fig 2. SCIP has been used for real time forecasting of 12 hourly intensities up to 72 hours. The method has been providing useful guidance for operational cyclone forecasting since 2008. A rapid intensification index (RII) developed for tropical cyclones over the Bay of Bengal uses large scale characteristics of tropical cyclones to estimate the probability of rapid intensification (RI) over the subsequent 24 hours. The RI is defined as an increase of intensity by 30 knots (15.4 m/s) during 24 hours. An empirical decay model was developed and implemented in 2008 for real time forecasting of decaying intensity of cyclones after landfall. The model predicts wind speed at 6 hour intervals up to 30 hours. Owing to dense populations along Indian coasts, the decay forecast has direct relevance to daily activities over a coastal zone, such as transportation, tourism and fishing apart from disaster management.