Simulating and assessing future land-use scenarios using land-use and hydrological models with landscape metrics in the Wu-Tu watershed, Northern Taiwan

摘要

Development of an approach for simulating and assessing watershed land-use changes, land-use patterns and their effects on hydrological processes is an essential task in watershed land use, planning and management. This study combines an empirical land-use change model with landscape metrics, a combined distributed/lumped parameter watershed hydrological model and two-way ANOVA. These are used to simulate and assess the impacts of land use scenarios on land use patterns and hydrological processes under various spatial and non-spatial policies in the Wu-Tu watershed in northern Taiwan. The analyses of these land-use pattern scenarios reveal that the patterns of further land use differed among spatial policies, but did not significantly differ among conversion policies under low demand for land use change. The further pattern of the target land use type differed among conversion policies and interaction of conversion and spatial policies. The policy priorities were a conservation spatial policy withno agricultural land release policy, a conservation spatial policy with an agricultural land release policy, a regular spatial policy with no agricultural land release policy and regular spatial policy with an agricultural land release policy. The predicted streamflow, runoff and groundwater discharge were successfully simulated using a lumped hydrological model based on various land use demands that can assess the impact of land-use change in watersheds. The models revealed that both variability and magnitude of further hydrological components are significantly and cumulatively impacted by land-use change scenarios during the simulation period, especially runoff and groundwater discharge. This novel integrated approach is an effective tool for modelling a landscape eco-hydrological decision-support system for watershed land-use planning, management and policy in Taiwan.