The researchers wanted to see if data from small areas could help predict runoff in larger watersheds. They tested this using different types of land cover like grass and wheat. They found that using data from small areas to calibrate the model for larger areas can lead to inaccurate predictions of runoff. This is because factors like soil type and land management practices can affect how water runs off the land. So, using data from small areas may be better for adjusting how water soaks into the ground or checking values for predicting runoff in larger areas.