Nonlinear responses of soil erosion to climate change: a modelling study on the UK South Downs

Abstract

A modelling approach is used to estimate some effects of changed climate upon rates of soil erosion on agricultural land on the UK South Downs. Previous studies have concentrated only on estimating shifts in long-term mean erosion rate: these were found to be approximately linear. However such simple shifts mask changes in the underlying distributions of annual erosion. A first series of simulations indicated that, under a wetter climate, erosion rates in wet years will generally increase more than rates in dry years. Under a “best guess” rainfall scenario with a 10% increase in winter rainfall, annual erosion increased by up to 150%. Erosion rates for individual years were shown to change in more complex nonlinear ways however, with decreases as well as increases occurring. These could be explained by the interaction of timing of rainfall with changes in the rate of crop growth. Most earlier work also assumed an equilibrium climate for the simulations, with climatic parameters such as mean monthly rainfall having stabilised at some new value, usually for a 2 × CO 2 atmosphere. This however leads to an “initial conditions” problem: how will soil characteristics have changed by the time of CO 2 doubling? A decrease in erodibility of about 20% by the time of CO 2 doubling was indicated, resulting from changed soil profile properties. However, a second series of runs employed “transient” weather sequences (i.e. with a trend imposed). For these, present-day soil profiles could legitimately be used.