Modeling run-off and sediment yield in Calumpang Watershed, Batangas, Philippines and its response to land cover change

This study aims to determine the impact of land cover change and establish the relationship between landscape pattern and hydrologic processes within Calumpang watershed. Land cover change was measured using supervised classification of SPOT images in 2003 and 2008. Results showed an increase of built-up areas (69%) and a reduction of mixed vegetation including riparian vegetation (-10%), which have significant effects on hydrology. The SWAT model was used to quantify the impacts on hydrology. Manual adjustment of parameters based on literature and sensitivity analysis was done. Comparing outputs with field-measured data showed the model is acceptable. Results showed a significant increase in surface runoff (5%) and sediment yield (6%) and a reduction in baseflow (-11%) for the whole basin. It also showed the spatial and temporal variation of the impacts of land cover change. Sub-basins that underwent land cover degradation experienced increased steamflow (up to 31%) during stormy months and reduced baseflow (up to -26%) during dry months. In contrast, sub-basins that exhibited improved land cover although minimally, have decreased stormflow (up to -4%), and increased baseflow (up to 5%) during dry months. At the class level, nine landscape metrics known to affect hydrology were quantified using Fragstats and was correlated with surface runoff, baseflow and sediment yield as response variables. Partial least square regression (PLSR) technique was used to address multicollinearity of the metrics. The final regression model suggested surface runoff and sediment yield tend to decrease with higher patch density (agricultural) and largest patch index (forest), and tend to increase with higher cohesion (agricultural) and aggregation index (agricultural). On the other hand, baseflow tend to increase with higher patch density (agricultural) and largest patch index (forest), and tend to decrease with higher cohesion (agricultural) and aggregation index (agricultural). To some extent, all other landscape metrics are related with the response variables; however, their contributions are not as significant. It is recommended that the results of this study be used for watershed planning and management.