Assessing the Roles of Regional Climate Uncertainty, Policy, and Economics on Future Risks to Water Stress: A Large-Ensemble Pilot Case for Southeast Asia
The fate of natural and managed water resources over any basin are controlled to varying degrees by energy, agricultural and environmental systems, which are linked together as well as to the hydro-climate cycles. The growing need for risk-based assessments of impacts and adaptation to regional climate change calls for the quantification of the likelihood of these regional outcomes through the representation of their uncertainty - to the fullest extent possible. A hybrid approach that extends the MIT Integrated Global System Model (IGSM) framework to provide probabilistic projections of regional climate changes is presented. This procedure constructs meta-ensembles of the regional hydro-climate, combining projections from the MIT IGSM that represent global-scale uncertainties with regionally resolved climate patterns from general circulation models. These future climates are generated in tandem with consistent projections of socio-economic pathways. From these, a Water Resources System (WRS) module tracks water allocation and availability across competing demands from irrigation, cooling for thermoelectric generation, hydropower, municipal, and industrial sectors. Coupled together, the IGSM-WRS is an integrated modeling tool aimed to provide quantitative insights on the risks and sustainability of water resources over large river basins.
Figure SEQ Figure \* ARABIC 1: Shown is the spatial focus of IGSM-WRS Southeast Asia pilot uncertainty project using the IGSM-WRS framework. Gray lines denote the boundaries of water basins. The colored shading denotes the headwater basins (green), basins discharging to downstream basins (yellow), basins that flow to the sea (orange), and basins that flow to countries outside the focus area (beige). The thicker red lines denote political boundaries whose water demands are considered via the economic outputs of the IGSM (see Figure 2).
This presented work is an extension of our global IGSM-WRS framework with a particular focus on Southeast Asia (Figure 1). This region presents a number of exceptional challenges toward sustainable development - particularly in the area of water resources - as it encompasses a rich texture of river basins that traverse and interconnect small developing nations as well as large, ascending economies and populations – such as China and India. The pilot project employs the IGSM-WRS framework across a large ensemble of outcomes spanning the aforementioned hydro-climatic, economic, and policy uncertainties. Through this procedure, our approach can generate thousands of possible future hydro-climates and economic pathways for a given policy scenario. Thus, for computational efficiency, we employ a Gaussian Quadrature procedure to sub-sample this large ensemble of outcomes (Figure 2), which are then taken through the WRS module to assess water availability across the agricultural, energy, industrial, and municipal demand sectors. The integrated impacts of these water-demand mechanisms are quantified through frequency distributions of changes in water stress (or availability). Taken holistically, the results then allow for interpretation and further experiments to assess: the effects of policy measures; impacts on food production; and the value of design flexibility of infrastructure and institutions. An additional, active area of our model development and exploration within this framework is the feedback of these water-stress shocks to economic productivity (i.e. GDP) and activity (i.e. land use). We provide further discussion and results (where possible) along these lines as well as for ongoing efforts to: refine uncertainty methods, implement greater basin-level and climate detail, and improve process-level representation of glacial-fed melt-water sources.
Figure 2: Distributions across climate scenarios for the full 6,800-member IGSM ensemble (solid lines) for the 14 variables chosen for the Gaussian Quadrature sub-sampling selection. Plots a) through e) show the Climate Moisture Index (CMI) for 5 regions and the two time slices (2030 in green and 2050 in blue). Plots f) and g) show the GNP in 2050 for China and India as a % change from the year-2000 value and h) and i) show the population of China and India in 2050, also as a percent change from the year-2000 value. The dashed lines are the distributions obtained via the Gaussian Quadrature sub-sampling (~450 members). The agreement of the solid and dashed distributions depict the effectiveness of the sub-sampling procedure to provide a representative assessment of risk (as a surrogate to the 6,800 member ensemble).