Abstract: Determining watershed response to vegetation treatment has been the subject of numerous hydrologic studies over the years. However, generalizing the information obtained from traditional paired-watershed studies to other watersheds in a region is problematic because of the empirical nature of such studies and the context dependence of hydrologic responses. This paper addresses the issue of generalizing hydrologic information through integration of process-based modeling and field observations from small-scale watershed experiments. To this end, the results from application of a process-based model were compared with the results from small-scale watershed experiments in ponderosa pine forests of Arizona. The model simulated treatment impacts reasonably well when compared to the traditional paired-watershed approach. However, the model tended to overestimate water yields during periods of low flow, and there was a significant difference between the two approaches in the estimation of treatment impacts during the first four years following treatment. The results indicate that the lumped-parameter modeling approach used here may be limited in its ability to detect small changes, and tends to overestimate changes that occur immediately following treatment. It is concluded that watershed experiments can be highly informative due to their direct examination of cause-effect relationships, while process-based models are useful for their processing power and focus on functional relationships. The integrated use of both watershed experiments and process-based models provides a way to generalize hydrologic information, illuminate the processes behind landscape treatment effects, and to generate and test hypotheses.