Dataset and code to accompany the manuscript 'Consistency of clustering analysis of complex 3D ocean datasets'

Data and code needed to recreate all analyses and figures presented in the manuscript 'Consistency of clustering analysis of complex 3D ocean datasets'. Data 'all_data_for_paper.nc': model data, 2000-2004 mean of all variables used, provided at all depth levels. 'mesh_mask.nc': domain and depth data file to be used alongside model data. ModelViz (code) Tool to classify marine biogeochemical output from numerical models Written by rmi, dapa & dmof Preprocessing preprocess_amm7_functions.pyFunctions needed to run different preprocessing scripts.preprocess_all_depths.pyFirst script to run. Extracts relevant variables and takes temporal mean for physical,biogeochemical and ecological variables. For physical - calculates PAR from qsr.preprocess_amm7_mean.pyUse for surface biogeochemical and ecological sets (faster)preprocess_DI_DA.pyUse for depth integrated, depth averaged and bottom biogeochemical and ecological sets. Can use for surface but slower.preprocess_amm7_mean_one_depth.pyExtracts data at specified depth (numeric). Works for biogeochemical and ecological variables.preprocess_physics.pyTakes all_depths_physics and calculates physics data at different depths. Metrics silhouette_nvars.pyCalculates silhouette score for inputs with different numbers of variables and clustersrand_index.pyrand_index_depth.pyremove_one_var.pyCalculates rand index between cluster sets with one variable removed and original set Clustering Modelviz.pyContains functions for applying clustering to data Plotting kmeans-paper-plots.ipynbProduces figure 4kmeans-paper-plots-illustrate-normalisation.ipynbProduces figure 2kmeans-paper-plots-depths.ipynbProduces figures 5-7plot_silhouette.ipynbProduces figure 3