Quantifying uncertainty in future sea level projections downscaled from CMIP5 global climate models

S. Sithara,S. K. Pramada,Santosh G. Thampi
DOI: https://doi.org/10.1007/s00477-024-02669-4
IF: 3.821
2024-02-22
Stochastic Environmental Research and Risk Assessment
Abstract:Sea level projections for the future indicate a likely increase, raising concerns in the community due to its detrimental consequences. It has been reported by several researchers that there is considerable uncertainty in the future climate projections of Global Climate Models (GCMs), the primary tools for projecting future climate. In this study, the support vector machine (SVM) was employed to downscale sea level projections for the future from the projections of CMIP5 (Coupled Model Intercomparison Project Phase 5) GCMs. Quantile regression was employed to examine the predictors' uncertainty, and it was found that sea surface salinity (halosteric component of sea level change) is the highly uncertain variable among the three predictors, followed by sea surface temperature and mean sea level pressure. The uncertainty associated with downscaled future sea level projections under Representative Concentration Pathways (RCPs) 4.5 and 8.5, stemming from GCM structure, was investigated using Normal distribution and non-parametric kernel density estimation. Kolmogorov–Smirnov (KS) test was performed to assess the goodness of fit and found that both normal distribution and kernel density estimation satisfactorily represent the probability density function (PDF) of sea level projections for the future. The uncertainty bounds in the sea level projections under both RCPs were estimated using the bias-corrected and accelerated bootstrap algorithm and found that the lower and upper bounds of sea level projection during December 2050 are 0.529 m and 0.604 m under RCP 4.5 and 0.535 m and 0.700 m under RCP 8.5. Results of the study revealed that uncertainty is comparatively high under RCP 8.5.
environmental sciences,engineering, environmental,water resources, civil,statistics & probability
What problem does this paper attempt to address?