Predicting the weather is a big task, especially in a country like India with a sub-tropical climate condition. In our country, impact studies in the fields of agriculture and hydrology demand short-to-medium range weather forecasts at finer scales.
Particularly, reliable quantitative precipitation forecasts at fine scales are very critical for assessments of vulnerability, impact, and adaptation. However, the forecasting of precipitable structures at weather prediction time scales by the Numerical Weather Prediction models is highly challenging, thus an active area of research.
Global Circulation Models have been predicting the weather by solving mathematical equations. But at coarser resolutions, mesoscale patterns induced by complex topographical interations and coastlines, have not been adequately resolved in the Global Circulation Models. Because of the drawbacks of these models, limited area models were used. But even with these models severe weather patterns are still under-represented. Thus the convective scale models were used in the contemporary research . Further research is required to evaluate the uncertainties in a systematic way through experiments and appropriate model verification techniques.
The present study aimed at evaluating the Weather Research and Forecasting model at convective-permitting scales (4-km) to understand the model sensitivities in simulating the quantitative precipitation forecasts at 4-day lead-time and thereby design an ensemble approach for dependable performance of Weather Research Forecasting across the seasons.
A total of 120 experiments across four events were designed to understand and quantify the model sensitivity at convective permitting scales (CPS) to the physics parameterizations over the southern regions of peninsular India. A distinctive ensemble approach was formulated from the results and was evaluated for performance in capturing the dynamics of quantitative precipitation forecastings as opposed to a single-member deterministic simulation and a larger member conventional multi-physics ensemble approach. Grid-based statistics, spatial statistics, and object-based statistics are used to quantify the performance of the model simulations against Global Precipitation Measurement Integrated Multi-SatellitE Retrievals for the data.
This study is an early attempt in investigating the performance of a regional weather forecast model at convective scales over the southern parts of peninsular India with a unique climatology. The study points out that multi-physics ensemble could be explored as a way-forward for medium-range precipitation forecasting, at convective-permitting scales and across various precipitation mechanisms.
Dr. K. J. Ramesh, Former Director General of India Meteorological Department, New Delhi, said that the design of the experiment was innovative. The model atmosphere used was very close to the real atmosphere. The Weather Research and Forecasting (WRF) framework gave the experiment more options. The Ockhi cyclone, the Gaja cyclone, the South-West Monsoon, and the Pre-Monsoon season were well recorded. Dr. Ramesh also said that the model used was the closest to reality and worked very well in predicting the distribution of rainfall. This model could be useful in predicting floods in Tamil Nadu and Kerala in advance so that immediate action could be taken. Dr. Ramesh feels that the next step of the model could be to predict hail storms in zones of heavy rainfall.
Article by Akshay Anantharaman
Here is the original link to the scientific paper:
https://link.springer.com/article/10.1007/s12040-021-01556-8
