Evaluating Outlying Futures for the Indus River Basin

Climate change effects are more pronounced in developing regions such as Pakistan's Indus River Basin (IRB), where limited resources, fragile infrastructure, and heavy reliance on water for agriculture exacerbates vulnerabilities. However, complex intersectoral interactions and uncertainties impede consensus on a single future outcome. This doctoral research first identifies ``outlying futures'' - scenarios that statistically deviate from typical model projections - and then evaluates which among these represent high-impact that could pose significant challenges for water resource management. These outlying futures are identified by exploring a wide range of plausible scenarios from Integrated Assessment Models (IAMs) developed for assessing long-term climatic, socio-economic and technological dynamics. First, I address the lack of a scalable method for identifying outlying futures by developing a novel technique ``Outlier Set Two-step Identification (OSTI)'' and evaluate its performance and robustness across thousands of synthetic datasets. Then, I systematically apply OSTI on a subset of a dataset generated by the Global Change Analysis Model (GCAM), an IAM used in high-profile climate assessments. I focus on irrigation withdrawal patterns across multiple global basins to extract coherent outlying futures and assess which among them represent high impact scenarios for water resource management. Finally, I build a basin-scale framework to assess the robustness of the annual and seasonal water supply-demand balance, accounting for both plausible melt water impacts of climate change and GCAM-projected changes in irrigation demand patterns. Results provide insights into both the plausibility and potential impact of previously identified outlying futures, helping determine whether these scenarios are important to consider in adaptation planning. Conversely, results also demonstrate the potential for IAM-informed seasonal irrigation demand patterns to enrich water resource robustness assessments. Thus, integrating global scope IAMs with basin-scale water resource assessments would benefit both modelling efforts.