The application of predefined weather patterns over India within probabilistic medium‐range forecasting tools for high‐impact weather

Abstract

A method is presented for deriving probabilistic medium‐range (1‐to‐2‐week) weather pattern forecasts for India. This method uses an existing set of 30 objectively derived daily weather patterns, which provide climatological representations for unique states in the large‐scale circulation over India. Weather pattern forecast probabilities are based on the number of ensemble members objectively assigned to each weather pattern. Two summer monsoon case studies illustrate the best use of the forecasting tool within medium‐range guidance, such as highlighting the most likely weather pattern transitions and relating these to the likelihood of weather impacts. Forecast skill is evident out to at least 10–15 days. Winter dry period weather patterns have the highest forecast skill, closely followed by retreating monsoon weather patterns. In contrast, monsoon onset and break monsoon weather patterns have the lowest forecast skill. Finally, a prototype weather pattern forecast climatology application is presented for use in highlighting when extreme rainfall is more likely than normal. This application is based on weather pattern empirical probabilities of threshold exceedances using a high‐resolution regional reanalysis. The transitional pre‐ and post‐monsoon seasons have the greatest variability in rainfall across all possible weather patterns, with a slight dip in variability during the main summer monsoon season. In contrast, very little variability across weather patterns is evident during the relatively dry winter months. This highlights the times of year when a climatology‐based weather pattern forecasting approach may have its greatest benefits over that of a basic daily climatology.

Generating weather pattern definitions over South Africa suitable for future use in impact‐orientated medium‐range forecasting

“…Future work intends to explore whether large‐scale circulation predictability is possible in the medium‐range for South Africa by assessing the prediction skill of different weather patterns using objective verification such as the Brier Skill Score (Brier, 1950), which has been used in recent verification studies (Ferranti et al, 2014; Neal et al, 2016; Neal et al, 2022). Given good predictability at the large scale, future research can investigate the use of these weather patterns in a medium‐range forecasting tool driven by global ensemble prediction systems.…”

“…As a result, weather pattern forecasts would be translated into likelihoods of meteorological thresholds being exceeded (e.g., extreme precipitation, heatwaves). Medium‐range forecasting tools that follow this methodology have previously been explored for Europe (Richardson, Fowler, et al, 2020; Richardson, Neal, et al, 2020), the Mediterranean (Mastrantonas et al, 2021; Mastrantonas et al, 2022), southeast Asia (Howard et al, 2022) and India (Neal et al, 2022). In this instance, knowledge of the assigned weather patterns is not required in theory, and therefore the number of weather patterns is less important.…”

“…A wide set of weather patterns enables an understanding of the variability within a particular regime. Weather patterns are already being utilized for medium‐range (<2 weeks) forecasting tools in Europe (Ferranti et al, 2014; Neal et al, 2016) and India (Neal et al, 2020; Neal et al, 2022). The forecasting tool developed by Neal et al (2016)—which assigns ensemble members to the closest matching weather pattern and calculates daily weather pattern probabilities—has been followed by many forecasting applications for UK/Europe: these include forecasting extreme precipitation (Richardson, Neal, et al, 2020), significant wave heights (Steele et al, 2017; Steele et al, 2018), coastal flooding (Neal et al, 2018), meteorological droughts (Richardson et al, 2018; Richardson, Fowler, et al, 2020), temperature‐based excess mortality (Huang et al, 2020), London bike hire demand (Brown et al, 2019), risks associated with lightning activity (Wilkinson & Neal, 2021) and aviation risks associated with Icelandic volcanic ash (Harrison et al, 2022).…”

A seamless blended multi‐model ensemble approach to probabilistic medium‐range weather pattern forecasts over the UK

“…Finally, the seamless blended multi-model ensemble approach to probabilistic weather pattern forecasts presented here would also be appropriate for similar products under development in other parts of the world, such as India (Neal et al, 2020(Neal et al, , 2022, Southeast Asia (Howard et al, 2022) and South Africa (Ireland et al, 2024).…”

The influence of future weather pattern changes and projected sea-level rise on coastal flood impacts around the UK

“…It would also be interesting to apply this methodology to other locations across the globe. For example, weather pattern definitions have been developed for India to represent precipitation variability (Neal et al 2020(Neal et al , 2022 and Southeast Asia to consider tropical variability and heavy precipitation (Howard et al 2022). Weather pattern approaches such as these and what has been presented here show promising potential in extending the useful skill of the medium-to extended-range forecast period (~ 1 to 7 weeks) for a variety of forecasting and climatic applications.…”