Assessment of the impact of land use and land cover change on the hydrology of Mabacan Watershed, Philippines using the Soil and Water Assessment Tool (SWAT) model

The capability of Soil and Water Assessment tool (SWAT) hydrologic model in predicting the impact of land use and land cover change without standard calibration was tested in Mabacan watershed [Philippines], an ungauged catchment with limited data. The model was complemented with the use of remote sensing and GIS (RS-GIS) technology in characterizing the biophysical aspect of the watershed which can aid in parameterization process. SWAT modelling was performed in two phases: 1) initial model development using hydrometeorological data input measured from 2005 to 2014 and 2) final model set up and evaluation using longer and more recent input data. Pseudo-calibration was performed using manually computed surface runoff values based on water balance and SCS Curve Number method. The final model performed reasonably well in simulating surface runoff with NSE of 0.76 and 0.77 and R sup 2 of 0.96 and 0.96 in comparison with the two methods. Potential land use and land over change (LULCC) scenarios were determined on the basis of 2% growth in built-up areas from 2010 to 2016, which originated mainly from agricultural land (65%) and mixed vegetation (35%). Four scenarios were formulated with a 20% increase in built-up areas in Scenario 1, 30% in Scenario 2 and 40% in Scenario 3. The 40% increase was retained in Scenario 4 but with the inclusion of agroforestry as a management strategy. SWAT simulation using the finals model showed an increasing trend in surface runoff and an overall decline in streamflow as built-up areas increase. The introduction of agroforestry as a farming practice led to a lower surface runoff and higher average monthly streamflow despite the largest increase in built-up areas. Further analysis indicated that maximum average monthly surface runoff increases as the vegetative cover decreases. On the other hand, the minimum average monthly streamflow during the dry season decreases as vegetative cover decreases. Results of this study suggest that the SWAT model is highly applicable for predicting surface runoff and streamflow in ungauged watershed in the Philippines even with limited data and for assessing the impact of LULCC on hydrology.