Potential shifting of climate zones and associated hydrological impacts under changing climate conditions in Sri Lanka

Abstract

Climate change plays a significant role in decision-making related to water resources management. Understanding the future climate of Sri Lanka is crucial for the development of adaptation and mitigation strategies. This study investigated the potential shifting of climate zones in Sri Lanka under changing climate conditions using the Köppen-Geiger Climate Classification system and identified the associated hydrological impacts. The research utilized observed daily precipitation data from 27 meteorological stations. Predictive mean matching (PMM) and normal imputation method (Norm) were employed using the Multiple Imputation by Chained Equations (MICE) algorithm to impute missing data. The performances of 15 Global Climate Models (GCMs) from Coupled Model Intercomparison Project Phase 6 (CMIP6) were evaluated using the Evaluation Based on Distance from Average Solution (EDAS) method. In distributing station data into higher spatial resolution, a linear regression analysis was conducted to develop a relationship between observed station data points with corresponding Climate Hazards Group InfraRed Precipitation with Station data (CHIRPs) grid cells. The calculated gradient values (m) were then utilized to distribute historical and future projection data from GCMs to each CHIRPs cell (0.05˚ resolution). Furthermore, a distributed hydrological model was used with a 0.05˚×0.05˚ grid cell resolution for calculating water balance and identifying hydrological impacts of future climate change on basin hydrology. The results depicted varied performance among the CMIP6 models in simulating the monsoon climate of Sri Lanka. The MPI-ESM1-2-HR, CNRM-CM6-1-HR, and CNRM-ESM2-1 models were identified as the top performers in simulating monsoon rainfall patterns in both the wet and intermediate zones, while the CNRM-ESM2-1, CNRM-CM6-1-HR and MRI-ESM2-0 models emerged as the top GCMs for the dry zone. The CNRM-CM6-1-HR and CNRM-ESM2-1 models were the best-performing models among the selected GCMs, with the high-resolution version of CNRM-CM6-1-HR being well-suited for small countries like Sri Lanka. When the Mean-Based method and the Quantile Mapping (QM) method were compared for bias correction performances, the QM method demonstrated strong relationships between observed data and model projections. The results of the Köppen-Geiger Climate Classification indicated that future climate zone influenced by climate change, particularly in the South-West region and the highland areas of Sri Lanka. Highland climates will be the most affected in all projection scenarios, with Cfb and Cwb climate zones projected to disappear under the SSP5-8.5 long-term (TL, 2070-2100) scenario. The outcomes of these changes in basin level indicated that, in the near-term (TN, 2020-2050) period, basins in the eastern side of the island will experience decreased runoff while the west will show an increase. Analyzing the Wet zone under SSP1-2.6 showed a 10% TN increase in runoff, rising to 15% in TL. Under SSP5-8.5, the runoff increase is more significant at 27% (TN) and 38% (TL) levels. In the Dry zone under SSP1-2.6, the TN projections result a 10% increase in runoff, escalating to 35% in the TL period. The findings of this study highlight that the potential climate shifts associated with global warming scenarios vary across distinct regional climate zones in Sri Lanka. This underscores the necessity for region-specific adaptation strategies to effectively mitigate the multifaceted impacts on water resources. Keywords: Climate Change, Data Imputation, Distributed Hydrological Model, Köppen-Geiger Climate Classification