The researchers developed a model called RINSPE to accurately predict surface runoff, infiltration, and pollution in a river catchment called the Ton watershed. They used a method called the Curve Number to estimate runoff during rainstorms. By testing two different scenarios with varying initial abstraction coefficients, they found that a higher initial abstraction coefficient led to slightly lower runoff values. In the forested areas of the watershed, they observed that a higher initial abstraction coefficient resulted in slightly more runoff and less infiltration. This indicates that the initial abstraction coefficient plays a crucial role in predicting surface runoff in different land cover types within the watershed.