Biases in Indian summer monsoon precipitation forecasts in the Unified Model and their relationship with BSISO index.

Abstract

This study shows that the Boreal Summer Intraseasonal Oscillation (BSISO) dominates the Indian summer monsoon low‐precipitation bias in the Met Office Unified model. Analyzing a recent 9‐year period (June, July, August only), it is found that the precipitation bias is dominated by break and break‐to‐active transition BSISO phases, while some of the other phases have no bias at all over a 7‐day forecast. Evidence of a link to upstream effects is found, in that there is a delayed reduction in the moisture flux entering India from the west. It is also shown that an increase in the net flow of moisture out of India to the east is strongly linked to the low‐precipitation bias, and there is some evidence that this is related to a lack of low‐pressure systems over India. Most atmospheric models have substantial rainfall biases over India, and these results may indicate the circulation patterns responsible.

Plain Language Summary

The Met Office Unified Model (UM) is widely used worldwide for weather forecasting, climate prediction and environmental research. An important deficiency of the UM, in common with many other weather and climate models, is that it simulates significantly too little rainfall over India, when averaged over the summer monsoon season. Indian monsoon rainfall is important to the livelihoods of hundreds of millions of people, and these errors in the models have knock‐on consequences for weather and climate prediction around the world. This study shows that the UM's rainfall bias is dominated by periods when the general monsoon behavior is in transition from low‐activity to high‐activity, while in other periods, the rainfall forecasts perform much better. These results will help us to better understand the causes of the model bias. A systematic evaluation of the UM moisture flow has also been carried out; this suggests that a key problem in these low to high‐activity transition periods is a replacement of monsoon cyclonic systems with too much purely westerly flow out of India. The results should also be of value in weather forecasting, in identifying weather regimes where we have relatively high, and relatively low, confidence in the forecasts.