Towards large-sample data availability for applications in soil erosion and sediment transport studies in Europe

Larger samples of measurement data unlock far-reaching opportunities to improve our predictive capabilities and process understanding in erosion and sediment load modelling through: (i) streamlining model applications to compare performance and understand model generalizability across differing environments, (ii) improving upscaling capacity and predictive capabilities in unmonitored locations via data-oriented approaches, and (iii) developing new modelling approaches better suited to data ingestion. Despite the tangible benefits of applying soil erosion models over multiple spatial domains, indicative overviews of modelling efforts show efforts in 3 or more catchments remain considerably less abundant. This study synthesises the currently available measurement data available for in-stream monitoring of hillslope sediment fluxes to stream channels in Europe. By combining purpose-compiled community data from small to medium catchments in EUSEDcollab (a European Union Soil Observatory initiative) with other open-access water quality data (e.g. GEMSTAT) and other resources, we give an overview of the current state-of-the-field. Key findings show: (i) data is significantly less abundant from small catchment drainage areas, limiting the potential inferences on hillslope processes, (ii) catchment data on the long-term average annual sediment load (e.g. statistical aggregations) is significantly more abundant than time series data, reflecting limited open sharing of historical measurement data, (iii) sediment load measurements are decreasing in modern time periods, limiting our potential to capitalise on modern revolutions in domain-agnostic geospatial data (e.g. remote sensing data). Further community efforts to compile current and legacy data across Europe with FAIR (Findable, Accessible, Interoperable, Reusable) standards are vital for scientific advancements and data rescue, following similar data sharing efforts (e.g. CAMELS for hydrology). Extensions of catchment data with large-scale feature compilations of (time series) of hydrometeorological, soil and management attributes data may further strengthen efforts to provide ready-to-use data for models. To conclude, open data is pivotal for multi-scale, open, and collaborative research which requires ongoing collaboration between research groups, national agencies, and multi-national institutions.