Sub-seasonal to seasonal (S2S) prediction of dry and wet extremes for climate adaptation in India

Abstract

Extreme climatic events have considerable impacts on society, and their prediction is an essential tool for climate change adaptation. A reliable forecast of dry and wet extremes is crucial for developing an early warning system and decision-making in agriculture and water resources. Sub-seasonal to seasonal (S2S) forecasts can be valuable for climate adaptation in water resource and agriculture sectors due to their extended range forecast ability and accessibility of different hydrometeorological products. However, the utility of these S2S models’ forecasting capabilities is limited to a certain lead time, rendering them unsuitable for decision-making. We comprehensively examined the prediction skill of nine global S2S prediction models for precipitation and dry and wet extremes over India during the summer monsoon season (June to September). We find that ECCC, NCEP, and UKMO perform better than the other S2S models in predicting dry and wet extremes during the summer monsoon (June-September) in India. Our findings show that the better-performing S2S forecast models can be used to predict wet and dry extreme events several weeks ahead during the summer monsoon season. The extended range forecast system (ERFS), which is currently operational in India, provides better forecast skills for dry and wet extremes than most of the S2S models. However, S2S models provide an extended lead time forecast compared to ERFS. Therefore, a combination of ERFS and better-performing S2S models can be utilized in the early warning of dry and wet extremes at longer lead times.

Practical Implications

India has witnessed climate-related catastrophes over the past few decades that, include flooding and droughts. There is a strong need to develop tools that can provide early warning of weather and climate extremes and help in climate adaptation. Climate services and climate change adaptation need reliable forecast products at seasonal to sub-seasonal scales. Recently, sub-seasonal forecasts bridged the gap between short-range and long-range forecasts and are critical for informed decision-making in India's agricultural and disaster risk reduction sectors. We utilized S2S precipitation forecasts from various forecasting centers around the world to comprehensively examine their utility in India.

Several critical implications are associated with the findings. First, we evaluated the forecasting skill of S2S models in predicting rainfall at different regions and months of the summer monsoon season. The forecast skill of meteorological forecast varies substantially in different regions and lead times. The forecast skill weakens with the increase in forecast lead time. An improved forecast skill during the summer monsoon onset and cessation could be valuable for planning agricultural activities and water resources. In addition, we identify the regions and times where these models do not perform well and where steps can be taken to improve the model’s performances in the future.

Second, there is a difference in the forecast skills of S2S models for dry and wet extremes for different regions over India. We identify a set of S2S models that provide better forecast skill for both dry and wet extremes and can be successfully employed in India's S2S operational forecast system as an early warning.

Third, we highlight the advantage of using S2S models over ERFS in forecasting dry and wet extremes in India. ERFS provide good forecast skills for both wet and dry extremes for the Indian region, but a few S2S models provide extended lead forecasts that are currently unavailable in ERFS. Therefore, we demonstrate the potential of S2S forecast information to provide early warning systems. As a result, S2S forecast information can be integrated into a “ready-set-go” framework to provide an early warning of an extreme event a few weeks in advance.