Variability in roughness measurements for vegetated rivers near base flow, in England and Scotland

The aim of this study was to reduce uncertainty in roughness information for instream vegetation. This in turn reduces the uncertainty in estimating water levels in channels where aquatic vegetation is present. Roughness plays an essential role in water level estimation as it reduces the discharge capacity through energy expenditure on boundary generated turbulence as well as through physical blockages by vegetation. Apart from the hydrology analysis for determining the inflow to the river system, the largest source of uncertainty for estimating water levels in channels is roughness (Latapie, 2003; Mc Gahey, 2006). Of the different roughness types, vegetation roughness includes the greatest natural variability (e.g. Sellin & van Beesten, 2004; Defra/EA, 2003a). The hypothesis that measures of plant growth can explain variation in roughness for vegetated channels was tested. Presented here is a snapshot of roughness measures made in vegetated channels at the height of the growing season. One source of variability in roughness values was quantified which is the regional difference in roughness values attributable to differences in aquatic plant growth. This work focused on sites dominated by either submerged Ranunculus spp. (water crowfoot) or the emergent Sparganium erectum (branched bur-reed). These 2 taxa are among the most common higher aquatic plants in UK rivers. Rivers were sampled from the Scottish border in the north to the south coast of England. Sites (36) were chosen carefully to provide a very wide range of growing conditions. Both roughness, as Manning’s n, and plant growth varied substantially between sites. At S. erectum sites Manning’s n was 0.125 sm-1/3 + / - 0.04 standard error (s.e.) and at Ranunculus spp. sites Manning’s n was 0.124 sm-1/3 + / - 0.0168 s.e. For all sites it was possible to reduce uncertainty by attributing variability in Manning’s n to standing crop. A plant index, combining 3 measures of abundance, explained 70% of the variation in Manning’s n for Ranunculus spp. sites while 28% of variation in Manning’s n at S. erectum sites was explained by the % cross sectional area occupied by the plant. The application of biometric error reduction models to conveyance estimation is discussed.