Climate Change Projections for Sri Lanka for the mid-twentieth Century from CMIP5 Simulations under a High Emissions Scenario

Abstract:

Under the Agricultural Model Inter-Comparison program (AgMIP), climate change projections for Sri Lanka were undertaken from the Coupled Model Inter-comparison Project (CMIP5) archives for five locations covering Sri Lanka. These datasets were first quality checked after removing questionable data entries. The gaps in data were filled using AgMERRA data set for the specific location developed by Alex Ruane and Sonali McDermid at NASA- GISS after applying the necessary bias corrections. Future climate projections for 2040- 2070 are based on projections for high Carbon Dioxide emissions (RCP8.5). Analysis was undertaken on the outputs of 20 General Circulation Models (GCMs). Observed climate datasets (for the period 1980- 2010) for each location were used to generate downscaled future predictions. Future projections for maximum temperature, minimum temperature and rainfall were generated while holding solar radiation constant and changing the CO₂ value up to 499 ppm. Results for 5 GCMs that simulate the monsoon region best were then selected for further analysis. These are CCSM4, GFDL-ESM2M, HadGEM2-ES, MIROC5, MPI-ESM-MR.

All 20 GCM outputs predicted that both minimum and maximum temperature shall rise by around 2 ⁰C throughout the year. This result is consistent across all 5 locations and the uncertainty associated with this prediction was observed to be low compared to that of rainfall. In the case of the rainfall, majority (80- 95%) of GCMs predicted an increment in the annual rainfall by around 0.5 mm/day. Rainfall during September- October- November was predicted to have a high increment (around 2- 7 mm/day) and during February- March a decrement of around 1- 2 mm/day was predicted. The uncertainty of this prediction based on outputs of all 20 GCMs were observed to be high. These results are consistent with the Fourth Assessment Report by the Inter-governmental Panel on Climate Change.