Multi-variable calibration of SWAT in a small experimental catchment area for afforestation scenario analyses

Simulation modelling is essential for scenario testing but requires careful calibration. Multi-variable calibration improves model realism but implies decisions about variable weighting in the objective function. An alternative is the sequential introduction of calibration variables. An example of this is related to water policy is the impact assessment of forest plantations on streamflow in South America. This requires a realistic implementation of forestry systems, where evapotranspiration and plantation growth are simulated realistically. This study used La Corona paired catchment experiment to demonstrate how a sequential multi-variable calibration improves Pinus taeda and grasslands representation in a landscape model. The model was based on long-term flow observations and satellite-based ET and LAI data. Results showed that flow prediction performance increased, with a shift in the parameter distributions and decreased uncertainty compared to calibration on streamflow alone. This ultimately resulted in more robust simulation results to assess afforestation and deforestation impacts on streamflow.