Evaluation of numerical weather prediction models to assess the predictability of regional extremes over India
Student name: Mr Abhishek Bhatia
Guide: Dr Vinay Shankar Prasad Sinha
Year of completion: 2018
Host Organisation: The Energy and Resources Institute (TERI), New Delhi
Supervisor (Host Organisation): Mr Saurabh Bhardwaj
Abstract: The increase in frequency as well as intensity of precipitation and temperature extremes over South-east Asia has been well established in the literature, although a substantial amount of uncertainty remains owing to the small sample set of observational record. Over the Indian region, the precipitation extremes manifest as extended spells of anomalously high (active) and anomalously low rainfall causing widespread droughts and floods over several parts of India. An assessment of these extremes and their predictability through weather/climate models has vital implications for early preparedness especially at the S2S scale. The present study makes use of numerical weather prediction model (based on fully coupled GCMs) outputs from two operational centres, the (NCEP) CFSv2 and (CMA) BCC CPSv1, made available under the WWRP/WCRP S2S collaboration project on the ECMWF data portal. The skill of these models in forecasting the precipitation anomalies of the ISMR over the MZI region and six homogeneous zones, and also the representation of heatwave events over the Indian region over four zones, is studied at extended range lead times (~ 30 days). NCEP CFSv2 is able to simulate well the climatological patterns of the monsoonal rainfall at all lead times although an increasing dry bias at higher leads is inherent. An inaccurate representation of monsoonal mean climatology is noted in the CMA model with a very heavy dry bias. The NCEP model outperforms the CMA model in the deterministic and probabilistic skill measures; which for both are found to be generally highest for the break category over most of the zones (highest skill over MZI) and is significant up to the P4 lead in some cases. The NCEP model is able to skillfully capture the signature of heatwave over the North Central India at a lead time of 2 weeks with significantly high probabilities even up to the 3rd week. A very hot bias is noted in the 1st week forecast with a trend of a less warm bias and high scatter at increasing leads with Northern Hilly regions forecasted to be unusually cold.
Keywords: NCEP, CMA, ISMR, MZI, active, break
