Date of Award

8-10-2022

Publication Type

Thesis

Degree Name

M.A.Sc.

Department

Civil and Environmental Engineering

Keywords

Analysis of Variance;bias correction;Regional Climate Models;SWAT model;uncertainty

Supervisor

Tirupati Bolisetti

Abstract

Climate change impact modelling studies utilize an ensemble of climate model projections obtained from various Regional Climate Models (RCM). These climate projections, extracted under different emission scenarios (representative concentration pathways) are then bias-corrected using observed meteorological data and the resulting bias-corrected projections are forced through a hydrological model in order to assess the climate change impacts on future streamflow. Uncertainties arise from various sources like inputs, model structure, model parameters, choice of hydrological model, choice of climate models, bias correction, etc. along every step of the impact assessment study. Quantifying such uncertainties and further assessing the contribution of each factor under consideration leads the modellers to draft robust climate change adaptation and management policies. The aim of this Thesis is to quantify and decompose the uncertainty contribution of three factors namely choice of Climate Model (CM), Representative Concentration Pathways (RCP) and Bias Correction Methods (BCM), and the uncertainties arising out of their factor interactions toward the total uncertainty using Analysis of Variance (ANOVA). To this extent, five sets of Climate Models (CM), two different emission scenarios (RCP 4.5 and RCP 8.5) and two non-linear bias correction methods were used in conjunction with the Soil & Water Assessment Tool (SWAT) hydrological model for the headwater catchment of Little River Experimental Watershed in Georgia, USA. The results indicate the overall and seasonal uncertainty decomposition, and it is evident from the results that Climate Model choice (CM) is the biggest contributor toward the total uncertainty in this modelling process.

Download

Share

COinS