Data and code for: Rapid shallow-water saturation and deep-water expansion of an invasive freshwater ecosystem engineer in a deep European lake

Dataset Overview These datasets contain five years of quantitative sediment sample data on quagga mussel population dynamics in Lake Constance. Sampled mussels were measured to determine their shell length and ash-free dry weight (AFDW). Hofstetter_et_al_dataset1.csv Contains the target depth and coordinates of each sampled station, as well as their actual depth, coordinates and sampling date for every sampled year. Hofstetter_et_al_dataset2.csv Contains individual shell length measurements and AFDW for every measured mussel. Mussels < 5 mm shell length were counted in all years but were measured inconsistently in 2021 and 2022. For subsamples where some smaller mussels were not measured, the “count” column contains the number of unmeasured mussels in the sample. Hofstetter_et_al_dataset3.csv Contains mussel density and AFDW per subsample. For sites where no Ponar sediment samples were taken in 2021, Ponar density equivalents were estimated from mussel counts collected using a Benthic Imaging System (BIS). The column “pred_density_ponar_adult" contains these predictions, the column “density_ponar_adult_filled” the combination of measured and predicted density values used in all density analysis. Hofstetter_et_al_dataset4.csv Contains mussel density and AFDW, averaged per station and year from dataset 3. R script overview File paths in [square brackets] must be adjusted before running Hofstetter_et_al_Rscript1.R Generates a map of all sampled stations in Lake Constance and of its location in Europe (Fig. 1). Hofstetter_et_al_Rscript2.R Fits a regression between Ponar and BIS mussel count data and predicts quagga mussel densities for all stations where no Ponar sample was taken (Fig. S3). Hofstetter_et_al_Rscript3.R Tests how the skewness of shell length distribution per year and depth category changes between 2023, 2024 and 2025, and plots shell lengths and skewness by year and depth (Fig. 2; Fig. S4). Hofstetter_et_al_Rscript4.R Plots density and biomass time series in Lake Constance for each depth category and lake-wide (weighted by the depth distribution of three-dimensional lakebed area; Fig. 3). Hofstetter_et_al_Rscript5.R Uses nonparametric Mann-Kendall Tests to check for significance and direction of biomass and density change over the years by depth category and for the whole lake. Hofstetter_et_al_Rscript6.R Fits per-station linear and log-linear biomass trends across all survey years and generates forest plots (Fig. 4), showing slopes at each station, plus pie charts showing the proportion of stations with positive/negative slopes by depth (Fig. S5) Hofstetter_et_al_Rscript7.R Generates quagga mussel biomass (AFDW) time series since first detection, for Lake Constance and the Laurentian Great Lakes by depth category and lake-wide (weighted by the depth distribution of three-dimensional lakebed area; Fig. 5).