The role of lateral pipe flow in hillslope runoff response: an intercomparison of non-linear hillslope response

The importance of lateral pipe flow on runoff generation in wet steep hillslopes has been described in many case studies from around the world. However, most hillslope studies are done in isolation and usually focus on the idiosyncrasies of individual hillslope characteristics and pipe flow response at a single site. Consequently, the first order controls on pipe flow responses remain poorly understood. We present a new intercomparison of pipe flow response to storm rainfall at four well instrumented sites: Panola (Georgia, USA), Toinotani (Kyoto, Japan), Jozankei (Hokkaido, Japan) and Hakyuchi (Tokyo, Japan). Our objective was to minimize the complexities of lateral pipe flow on hillslopes by looking for commonality across different hillslope types. Despite the large differences between the study sites in topography, climate, soil type and soil matrix hydraulic conductivity, we found several common pipe flow responses to storm rainfall: (1) the relationship between total rainfall amount and total pipe flow volume was highly non-linear at each site, (2) initiation of measurable pipe flow was threshold-dependent, controlled by the total rainfall amount and the pre-storm wetness, (3) once significant pipe flow response occurred, the maximum pipe flow rate was sensitive to the measured rainfall intensity, and (4) the ratio of total pipe flow to total hillslope discharge for each site was constant, regardless of total rainfall amount once the precipitation threshold for significant pipe flow response was reached. We used these results to develop a decision tree to determine the general conditions necessary for significant pipe flow to occur. The controls derived from our comparisons agree largely with the controls reported by the previous individual hillslope pipe flow studies in the literature. The commonality of response between our four very different forested hillslope settings indicates that site intercomparison may be effective for extracting common first order controls on complex hillslope processes. We argue that a decision tree may be a useful organizational framework to summarize and organize comparative analyses and may provide a structure for defining the hierarchy of process controls necessary for model development. (c) 2005 Elsevier B.V. All rights reserved.