Abstract:Deep uncertainty in future climatic and economic conditions complicates developing infrastructure designed to last several generations, such as water reservoirs. In response, analysts have developed multiple robust decision frameworks to help identify investments and policies that can withstand a wide range of future states. Although these frameworks are adept at supporting decisions where uncertainty cannot be represented probabilistically, analysts necessarily choose probabilistic bounds and distributions for uncertain variables to support exploratory modeling. The implications of these assumptions on the analytical outcomes of robust decision frameworks are rarely evaluated, and little guidance exists in terms of how to select uncertain variable distributions. Here, we evaluate the impact of these choices by following the robust decision-making procedure, using four different assumptions about the probabilistic distribution of exogenous uncertainties in future climatic and economic states. We take a water reservoir system in Ethiopia as our case study, and sample climatic parameters from uniform, normal, extended uniform, and extended normal distributions; we similarly sample two economic parameters. We compute regret and satisficing robustness decision criteria for two performance measures, agricultural water demand coverage and net present value, and perform scenario discovery on the most robust reservoir alternative. We find lower robustness scores resulting from extended parameter distributions and demonstrate that parameter distributions can impact vulnerabilities identified through scenario discovery. Our results suggest that exploratory modeling within robust decision frameworks should sample from extended, uniform parameters distributions.

Assessing fluctuations of long-memory environmental variables based on the robustified dynamic Orlicz risk

Assessing fluctuations of long-memory environmental variables based on the robustified dynamic Orlicz risk

Orlicz regrets to consistently bound statistics of random variables with an application to environmental indicators

Generalized divergences for statistical evaluation of uncertainty in long-memory processes

Environmental management and restoration under unified risk and uncertainty using robustified dynamic Orlicz risk

Mildly Explosive Autoregression with Anti-persistent Errors

A Dynamic Spatiotemporal Stochastic Volatility Model with an Application to Environmental Risks

Toward Scalable Risk Analysis for Stochastic Systems Using Extreme Value Theory

Measuring tail risk at high-frequency: An $L_1$-regularized extreme value regression approach with unit-root predictors

Impact of Uncertainty Parameter Distribution on Robust Decision Making Outcomes for Climate Change Adaptation under Deep Uncertainty

Spatial risk measures and rate of spatial diversification

Modeling spatial extremes using normal mean-variance mixtures

Unleashing the Potential of Mixed Frequency Data: Measuring Risk with Dynamic Tail Index Regression Model

Dynamic risk measures with fluctuation of market volatility under Bochne-Lebesgue space

The Robust Orlicz Risk with an Application to the Green Photovoltaic Power Generation

Distortion risk measures in random environments: construction and axiomatic characterization

A simple model on streamflow management with a dynamic risk measure

Modeling volatility using state space models

Bayesian non-asymptotic extreme value models for environmental data

River environmental restoration based on random observations of a non-smooth stochastic dynamical system

Generalized logistic model for r largest order statistics, with hydrological application