Can we give up gauging? A comparison of statistical certainty of gauged and modelled flows.

In the absence of a local gauging station, the flow regime of an ungauged catchment can be determined by the collection of short term flow data. This can be transposed to long term flow duration statistics by linking the site to an analogous gauging station with a long term data set. An alternative approach is to deploy desk-based software such as the ‘LowFlows’ model, which uses a regionalised approach to generate a flow duration curve for the ungauged catchment in question. Both approaches are common in Scotland and are used by both SEPA and those abstracting from the water environment to determine available resource for new and existing abstractions. It has generally been thought that collecting short term flow data will produce a flow duration curve which is more robust than desk based methods. However, until this study, a rigorous analysis of the performance of analogue transposition methods commonly used in Scotland has not been available. This paper presents estimates of the statistical uncertainty associated with two methods of analogue transposition, together with the equivalent estimates of uncertainty for the LowFlows software. This allows the performance of the transposition methods to be evaluated against the length of short term flow data and the ‘density’ of the surrounding analogue network, as well as drawing direct comparisons between each transposition method and the LowFlows software.