Uncertainty Assessment of Future Hydroclimatic Predictions: A Comparison of Probabilistic and Scenario-Based Approaches

Abstract

During the last decade, numerous studies have been carried out to predict future climate based on climatic models run on the global scale and fed by plausible scenarios about anthropogenic forcing to climate. Based on climatic model output, hydrologic models attempt then to predict future hydrologic regimes at regional scales. Much less systematic work has been done to estimate climatic uncertainty and to assess the climatic and hydrologic model outputs within an uncertainty perspective. In this study, a stochastic framework for future climatic uncertainty is proposed, based on the following lines: 1) climate is not constant but rather varying in time and expressed by the long-term (e.g., 30 yr) time average of a natural process, defined on a finescale; 2) the evolution of climate is represented as a stochastic process; 3) the distributional parameters of a process, marginal and dependence, are estimated from an available sample by statistical methods; 4) the climatic uncertainty is the result of at least two factors, the climatic variability and the uncertainty of parameter estimation; 5) a climatic process exhibits a scaling behavior, also known as long-range dependence or the Hurst phenomenon; and 6) because of this dependence, the uncertainty limits of the future are affected by the available observations of the past. The last two lines differ from classical statistical considerations and produce uncertainty limits that eventually are much wider than those of classical statistics. A combination of analytical and Monte Carlo methods is developed to determine uncertainty limits for the nontrivial scaling case. The framework developed is applied with temperature, rainfall, and runoff data from a catchment in Greece, for which data exist for about a century. The uncertainty limits are then superimposed onto deterministic projections up to 2050, obtained for several scenarios and climatic models combined with a hydrologic model. These projections indicate a significant increase of temperature in the future, beyond uncertainty bands, and no significant change of rainfall and runoff as they lie well within uncertainty limits.