Influence of Sea Surface Temperature on Simulated Future Change in Extreme Rainfall in the Asia-Pacific

The changes in precipitation simulated by the 40-model CMIP5 ensemble for the Asia-Pacific region are assessed, focusing on the two periods 1986–2005 and 2080–2099 under the RCP8.5 scenario. The frequency distributions of both daily and monthly mean rain rates at model grid points in each of four seasons are considered. In spatial averages for the land domain and seven selected regions, there is both an increase in the frequency of dry days and a shift towards heavier rain. Three 20-year seasonal aggregate statistics, the mean, the top decile of monthly amounts, and top percentile of daily amounts are assessed. The percentage changes over land are on average 4% higher for monthly extremes than for the mean, and a further 10% higher for the daily extremes. Over Australia and central Indonesia in some seasons the mean decreases but daily extremes increase. In much of Asia the daily extremes increase by 30% or more. For these statistics, there is a range across the ensemble that can often be linked to a pattern of sea surface temperature change that is quantified by a Pacific-Indian Dipole (or difference, PID) index. Correlations with both mean and extreme rain are highly negative over Australia and Indonesia and moderately positive over parts of south and east Asia in some seasons. The mechanism for this is explored through additional simulations and links with water vapour path. The results provide some understanding of the range of projections of future rainfall change based on the CMIP5 results, with some potential for narrowing it.