John, Chelsea MariaChelsea MariaJohnLunet, ThibautThibautLunetGötschel, SebastianSebastianGötschelHerten, AndreasAndreasHertenKesselheim, StefanStefanKesselheimRuprecht, DanielDanielRuprecht2026-07-032026-07-032026-0626th International Conference on Computational Science, ICCS 2026https://hdl.handle.net/11420/63808We propose an improved Fourier Neural Operator (FNO) for modeling two-dimensional Rayleigh–Bénard convection by predict-ing time increments instead of full solutions, achieving higher accuracy than a standard FNO baseline. The resulting model is compact (314k parameters, 1.26 MB) and fast (7 ms inference), while maintainingsim-ilaraccuracyasdemonstratedinpreviousbenchmarks.WeshowthatalthoughFNOsgeneralizetofinermeshes,accuracyremainslimitedbytheresolutionofthetrainingdata.enFourier Neural OperatorRa yleigh–BénardconvectionNatural Sciences and Mathematics::510: MathematicsConference Paper10.1007/978-3-032-29924-6_40