Probabilistic Analysis of Future Runoff Change and its Implications for International Equity

We apply probabilistic runoff projections to assess the implications of greenhouse gas-induced changes in runoff on international equity. Projections of future runoff are taken directly from a GCM ensemble and from a dynamic global vegetation and hydrological model (LPJmL 3.2) driven by climate projections from several GCMs. We investigate the relationship between various national indices of socio-economic development on the one hand and change in average runoff and agreement about the direction of runoff change on the other. Projections of change in runoff from the GCMs and from LPJmL are strongly correlated but there is no significant relationship between the confidence of these projections from the GCMs and from LPJmL. We find a weak positive relationship between baseline runoff and projections of runoff change from the GCMs, which would increase current inequities, but not from LPJmL. Future change in runoff is not significantly correlated with the level of greenhouse gas emissions per capita or with any of three national development indices (human development index, index of human wellbeing and water poverty index) but we do find a strong positive relationship between these development indices and agreement about the direction of runoff change from the GCMs. In summary, this analysis does not suggest that future changes in average runoff vary systematically between developing and industrialized countries. Combining data on current water poverty with projected changes in runoff indicates, however, that future hotspots of water poverty will concentrate in poor countries, in particular in Southern Africa, North Africa and the Middle East. Furthermore, this probabilistic analysis suggests that the direction of future runoff change is more uncertain in countries at low levels of socio-economic development, which implies that anticipatory adaptation is particularly difficult for currently disadvantaged countries. Future work will investigate the robustness of these findings for a larger ensemble of GCM projections, for analyses performed at the subnational level and for alternative hydrological indicators, such as soil moisture.