This dataset accompanies the manuscript entitled "Temporal and Granulometric Variability of Fluvial Sand Composition: An Annual Time Series from Four Rivers in SW Germany" by Laura Stutenbecker et al., submitted to JGR Earth Surface in March 2023. The study aims at analyzing the temporal variability of fluvial sediment composition. For this purpose, sediment of four rivers was sampled monthly over the course of one year between April 2019 and December 2020 at the same locations, using the same operator, tools, and sampling strategy, resulting in a total of 46 grab samples. The sampled rivers were the Gersprenz, Modau, Mümling and Upper Neckar rivers in southwestern Germany.The summary of sample names, coordinates, and sampling dates can be found in the data table "Sampling dates". The sediment was wet-sieved for grain-size analysis using six sieve sizes (63, 125, 250, 500, 1000, and 2000 micrometers). The percentage of each fraction after drying and weighing is provided in the table "Granulometry". The individual fractions (172 sub-samples in total) were analyzed for chemical composition. Geochemical analysis was performed using wavelength-dispersive X-ray spectrometry on pressed powder pellets using a S8 Tiger 4 kW (Bruker) spectrometer at Technical University of Darmstadt. Geochemical data of major element oxides and selected trace elements (in wt% and ppm, respectively) is provided in data table "XRF". Mineralogical analysis was performed on selected samples using X-ray diffraction on powdered samples using a Panalytical X'Pert Pro diffractometer at Goethe University Frankfurt am Main. The percentages of the minerals quartz, alkali feldspar, plagioclase, calcite, dolomite, illite/muscovite, kaolinite, chlorite, amphibole, goethite, and hematite are provided in data table "XRD". The data was input into an unmixing model following Lizaga et al. (2020, https://doi.org/10.1007/s11269-020-02650-0). The data table "Mixing modeling results" sums up the model performance (goodness of fit) as well as the modeled sediment source contributions (in percent, mean and standard deviation) of three considered sediment sources from the literature.