Assessment of the influence of changes in land use and climate on landslide activity in a Mediterranean environment

In Mediterranean environments, changes in land use and climate affect the incidence of rainfall-induced landslides by modifying the antecedent moisture conditions and the net rainfall amount delivered during erratic storm events. This study aims to assess the influence of the likely changes in land use and climate on landslide activity on medium temporal scales (10-100 years). Because of the hypothetical nature of anticipated environmental change, a physically based model has been used to assess this influence. The model is distributed to take land use patterns and flow paths into account and couples a hydrological model to a slope stability assessment. It has been validated under present conditions and applied to static scenarios of anticipated environmental change. Landslide activity has been simulated for 30-year periods from which changes in landslide activity have been derived. A small catchment near Alcoy (Spain) has been used as a case study.
Validation of the hydrological model over the historical period has proven to be reasonably successful. Under wet conditions, favourable to landsliding, the simulated hydrology approximates the observed moisture contents adequately. Simulated groundwater levels are less accurate as they accumulate errors over space and time. The performance of the stability model has been evaluated against landslide occurrence as observed from aerial photographs. Although simulated failure does not replicate observed landslide occurrence it reflects the difference in landslide activity between land use types and can consequently be applied to assess relative differences in landslide activity as the result of environmental change.
Mediterranean land use change is mainly driven by the abandonment of marginal fields and by incentives provided by EU agricultural policy. The current trend of land use change has been projected to the likely date of CO2-doubling. This results in three scenarios: present land use, continued abandonment and a limited reorganisation of agriculture. Climate conditions are specified by two scenarios: a warmer, future climate at the date of CO2-doubling, based on downscaled GCM results, and the present one. For these scenarios a stochastic weather generator provides unbiased, synthetic timeseries of precipitation and potential evapotranspiration.
For the hypothetical scenarios, landslide activity decreases relative to the present conditions: after abandonment, the higher evapotranspiration by the more extensive semi-natural vegetation attenuates the hydrological response to rainfall. Under the present climate mainly the temporal activity of landsliding is affected by land use change. The maximum groundwater levels are hardly attenuated and, consequently, the decrease in the area experiencing failure is slight (circa 5%). For the future climate, the compounded effect of land use and climate change leads to larger decreases in the temporal activity and spatial occurrence of landsliding (between 5 and 25%). This should be attributed to the lower antecedent moisture conditions under the warmer climate. The decrease is the largest on slopes that are less susceptible to failure. Thus, the impact of environmental change on the activity of rainfall-induced landslides is large. In this case, the impact is apparently advantageous although the inherited landscape may be prone to large adaptations after disturbance