Evidence of long-term improvements data quality in a national hydrometric dataset through a quantitative, indicator-based quality assurance programme

Globally, access to hydrometric data of adequate quality, consistency and coverage to answer pressing research questions and manage operational freshwater systems use remain a major issue. The UK National River Flow Archive (NRFA) provides stewardship of daily, monthly and flood peak river flow data from over 1,500 gauging stations in the UK. These data are disseminated to a very broad user community and provide crucial underpinning for water management (e.g. water resource assessments, flood risk estimation, environmental flows). Large-scale datasets are essential for detecting hydrological change in space and time, and to be a valuable resource they must have good spatial coverage and long, substantially complete and regularly updated records. To this end, river flow records are subject to both automated data screening and manual methods of quality control by a team of trained hydrologists which ensure the data disseminated by the Archive are of the highest quality and fit-for-purpose for a range of applications. Examples of the methods and resultant improvements in data quality and completeness are presented here. Following increased gaps in river flow records, a potential decline in data quality and changes to the gauging station network coverage in the 1990s, the UK introduced a Service Level Agreement (SLA) in 2002. Central to the Agreement is a set of quantifiable indicators of data quality, completeness and provision. Here, we present the results from 15 years application of the SLA system and discuss UK experiences of applying it to a diverse gauging station network maintained by multiple data providers. The improvements shown demonstrate the benefits of such a system to the overall utility of data, but to our knowledge there are few comparable initiatives, globally, to the SLA system. The quality control and performance indicator systems outlined here could be applied to other monitoring networks around the world and demonstrate one method of helping to ensure hydrological databases provide information of adequate quality to meet pressing water management needs both today and into the future.