On 2 September 2026, Kajetan Chrapkiewicz of the WaterSmartLand team presented a talk, “Efficient pixel-scale upstream covariate computation for environmental machine learning”, in the General Session of FOSS4G 2026 — the global Free and Open Source Software for Geospatial conference held in Hiroshima from 1 to 3 September 2026.
The talk introduced a method developed by our team to compute a hydrologically conditioned data cube of environmental predictor variables, achieving orders-of-magnitude speed-ups over previous approaches and making a European-scale analysis possible.
Alongside the method itself, the open-source software built around it, and the European water-quality modelling case study, the presentation explained the broader objectives of the WaterSmartLand project and potential applications of our open and scalable approach. This grounded the work in a wider context, illustrating how FOSS4G tools can accelerate environmental modelling, reproducible research and applied geospatial machine learning
