On the Application of the Hasselmann’s Stochastic Theory to Assess the Uncertainties of Climate and River Runoff Forecasts

A new approach is proposed to assess the uncertainties of long-term river runoff forecasts—for both globally averaged runoff and for the runoff of rivers whose annual discharges are most sensitive to global warming. Analysis of the series of global temperature obtained in the framework of “historical” experiments on 63 Earth System Models (ESM) (project CMIP-6) with the help of the system of statistical and stochastic estimates proposed by the authors, and in accordance with the K. Hasselmann’s theory, gives estimations of uncertainty ranges for global surface temperature projections, which are much larger than those published by the IPCC. Using river runoff projections from 39 CMIP-6 models, continuous changes in the mathematical expectation of the total runoff were calculated, qualitatively coinciding with existing estimates. At the same time, the width of the obtained ranges of uncertainty of forecasts (by recalculation through the ranges of uncertainty of global temperature) significantly exceeds those appearing in the literature. Examples of such calculations are given for the Lena basin, which traditionally serves as a “testing ground” for “greenhouse” forecasts of runoff, as well as for other large rivers—for the north and south of Eurasia, the north of North America.