CAMELS-GB : a large sample, open-source, hydro-meteorological dataset for Great Britain

Data underpins our knowledge and understanding of the hydrological system; they are used to drive, test and evaluate hydrological models and advance our understanding of hydrological processes and dynamics. With the increasing availability of observational datasets, the integration of information from many catchments for data and modelling analyses is becoming increasingly common. The production of new, open source, datasets for large samples of catchments is vital to advance knowledge on hydrological processes and to ensure hydrological research is reusable and reproducible through the use of common datasets and code. However, the availability of open source, large-sample catchment datasets is notably sparse. In this study, we present CAMELS-GB, the first large sample, open-source, hydro-meteorological catchment dataset for Great Britain (GB). CAMELS-GB integrates a wealth of different datasets derived from national, continental and global products based on observational, satellite and modelled data. The dataset consists of hydro-meteorological timeseries, catchment attributes and catchment boundaries for >800 catchments that cover a wide range of climatic, hydrological, landscape and human management characteristics across GB. Long daily timeseries is provided for a range of hydro-meteorological data (including rainfall, potential-evapotranspiration, temperature, radiation, humidity and flow) from 1970-2015 covering several major hydrological events. A comprehensive set of catchment attributes are provided describing a range of catchment characteristics including topography, climate, hydrology, land cover, soils and (hydro)-geology. Importantly, we also derive human impact attributes (including abstraction returns, percentage urban and gauge distance from reservoir), as well as attributes describing the quality of the flow data (including discharge uncertainty estimates and out of bank flow). The dataset and code used to derive the data will be made open source and provided with comprehensive metadata to allow its use in a wide range of hydro-meteorological data and environmental modelling analyses.