The deep uncertainties of designing scenarios to assess vulnerabilities under deep uncertainty
Abstract
Planning under deep uncertainty, when probabilistic characterizations of the future are unknown, has long challenged water resources planners. Many planning frameworks advocate for bottom-up, scenario-neutral analyses in which alternative policies are evaluated under plausible future scenarios with no assessment of their likelihood. Instead, these frameworks use sensitivity analysis to discover which uncertain factors have the most consequential impacts. This knowledge can be used to design monitoring programs and adaptive policies that respond to observed changes in the critical uncertainties. However, most sensitivity analyses require explicit or implicit assumptions about the range and distribution of the uncertain factors, which are by definition unknown. In an attempt to be neutral, most methods assume uniformity and independence of the uncertainties, but these assumptions could have decision-relevant consequences, biasing which factors are found to be most important and which policies most robust. Consequently, designing appropriate scenarios for vulnerability assessments under deep uncertainty is itself deeply uncertain.
This study illustrates these challenges on a multi-stakeholder planning problem in the Colorado River Basin where municipal, industrial, agricultural, and environmental rights holders all vie for the basin's limited water under prior appropriations laws. Variance and density-based sensitivity analyses are performed to assess vulnerabilities for over 100 stakeholders in the basin to changing hydrologic and socioeconomic conditions. These analyses are performed over scenarios from a number of different experimental designs: 1) uniformity and independence of hydrologic and socioeconomic factors, centered on current conditions, 2) same as 1 but with different ranges, 3) correlation and range of hydrologic factors informed by paleo-data, 4) correlation and range of hydrologic factors informed by CMIP projections, and 5) uniformity and independence across all other experimental designs. Differences in sensitivities across the frameworks illustrate the challenges of not only planning under deep uncertainty, but also assessing sensitivities under deep uncertainty.- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2019
- Bibcode:
- 2019AGUFM.H23G..02Q
- Keywords:
-
- 1807 Climate impacts;
- HYDROLOGY;
- 1880 Water management;
- HYDROLOGY;
- 1918 Decision analysis;
- INFORMATICS;
- 6309 Decision making under uncertainty;
- POLICY SCIENCES & PUBLIC ISSUES