The changing spatial variability of subsurface flow across a hillside

Subsurface flow has been measured using thirty troughs across the base of a steep forested hillslope at Maimai, near Reefton, New Zealand. Measurements are reported for a 110 day period, including extended dry periods and periods of intense rainfall. Flows vary with surface topography; the flow per unit width of trough is greater at the base of convergent regions, simply because the catchment area is larger. However, the spatial patterns also vary with the total flow being discharged by the hillside. There is a consistent relationship between the proportion of the total flow carried by each trough and the total measured flow from the hillside. That relationship depends on the local topography and upstream catchment area for each trough. The flow-related changes in spatial variability suggest that 1-2 metre long troughs are inadequate for accurate estimates of hillside or small catchment total subsurface flow. Large variations in flow rates between adjacent troughs indicate that catchment-wide subsurface flow cannot be estimated simply by multiplying the length of a trough by the estimated length of stream. Estimates of runoff are improved if adjacent troughs are grouped using common topographic features. The spatial variability between groups of troughs changes with storm size and antecedent wetness, with less variability during larger storms after wetter antecedent conditions. Some understanding of the relationship between catchment area, topography and flow is needed to interpret such data. Implications for experimentalists, users, and developers of hydrological models are discussed, and a new form of topographic wetness index is proposed for predicting patterns of subsurface flow.