Catherine S. Jarnevich Peder Engelstad Demetra A. Williams Keana S. Shadwell Cameron J. Reimer Grace C. Henderson Linnea Fraser Shelby LeClare Rich Inman Ian Pfingsten Wesley Daniel 20260123 tabular and raster digital data https://doi.org/10.5066/P13JMOQW This is a dataset containing the potential distribution of 103 invasive species found in or adjacent to freshwater environments. We developed habitat suitability models for invasive freshwater species selected by resource management agencies and other managers. We adapted the modeling workflow described in Jarnevich et al. (2024, https://doi.org/10.3897/neobiota.96.134842). We developed a national library of environmental variables known to physiologically limit freshwater species distributions (Henderson et al. 2025, https://doi.org/10.5066/P14JDTTJ) and relied on human input based on natural history knowledge to narrow the variable set for each species before developing habitat suitability models. We developed models using five algorithms (boosted regression tree, generalized linear model, multi-variate adaptive regression spline, maxent, and random forest) with VisTrails: Software for Assisted Habitat Modeling (SAHM 2.2.2, Morisette et al., 2013: https://doi.org/10.1111/j.1600-0587.2012.07815.x). For each species, we generated models for waterbody and stream environments where there were enough data available in each system. For several predictors we had future scenarios available, including up to five alternatives. We combined stream and waterbody models into a single map representing freshwater environments using a weighted ensemble across algorithms, including current and alternative future conditions. We also calculated subwatersheds (Hydrologic Unit Code [HUC] 12) summaries of information related to pathways of spread, finally integrating this information with current and alternative future habitat suitability to produce a risk score for each HUC12 east of the Mississippi River. This data bundle contains a single file of tabular summaries by management unit (including each species/ensemble type/abundance level combination), the merged data sets used to create models, tabular outputs including response curve data, variable importance information, model assessment metrics, a single file of pathways and risk scores, a spatial vector layer of the subwatersheds to which pathways and risk are summarized, a species metadata file, and R scripts to produce the model inputs and final products. The bundle documentation files are: 1) 'AquaINHABIT_V1_metadata.xml' (this file) which contains the project-level metadata. 2) 'Species_model_information.csv' contains information on specific model changes of each species from tuning algorithm parameters to ensure model quality. 3) 'Merged_dataset.csv' contains the merged data set used to create the models, including location and associated environmental data, for each species. 4) Variable_importance.csv is the tabular summaries indicating predictor importance for each of the models produced for each species. 5) Assessment_metrics.csv is the tabular summaries of assessment metrics for each model or ensemble for each species. 6) Risk_and_pathways_scores.csv contains summarized values for pathways, current suitability, future suitability, and a combined risk for each species by HUC12. 7) WBD_HUC12_EasternUS.gpkg is the spatial vector file of the HUC12 boundaries used for analyses and summarizations. 8) Rcode.zip is a zipped file of the R scripts used to pull and prep the data included in the merged dataset, code to make the derived raster outputs, and code to calculate the pathway and suitability summaries along with the final risk score. There are also three child items. 1) “response curves” child item contains files of XX_response_curves.csv, the tabular information needed to produce response curves for each predictor retained in each of the up to 10 models produced for each species, where XX represents the code for the species from 'Species_model_information.csv', all within the response curves child item. 2) "management summaries" child item contains files Management_summaries_XX.csv which are the tabular summaries by management area, where XX indicates the management area group. The management area group includes HUC12 and US Army Corps of Engineers (USACE) project areas, with 1 file for USACE and HUC12 broken out into HUC2 watersheds (01 to 18) 3) "rasters" child items contains raster files XX_YY.tif where XX is the code for the species from 'Species_model_information.csv' and YY is the raster type including seven rasters for each species: 1) Current occurrence suitability - Continuous value ensemble (XX-ens-current-mean .tif) 2) Restricted current occurrence suitability - Continuous value ensemble with restricted environmental conditions* (XX-ens-current-mean-masked.tif) 3) Future occurrence suitability - Continuous value ensemble (XX-ens-future-mean.tif) 4) Maximum future occurrence suitability - Maximum continuous value ensemble from the five alternative scenarios (XX-ens-future-max-gcm.tif) 5) Minimum future occurrence suitability - Minimum continuous value ensemble from the five alternative scenarios (XX-ens-future-min-gcm.tif) 6) Standard deviation of future occurrence suitability - Standard deviation of continuous value from each algorithm for the five alternative scenarios (XX-ens-future-stdev.tif) 7) Restricted count - Count of ensembles from each alternative scenario with restricted environmental conditions* (XX-gcm-mask-count.tif) *Restricted environmental conditions = only display areas where environmental characteristics are inside the range of the values used to develop the model. For example, a location with a minimum winter temperature of 12 C would be outside the range of -10 to 10 C used in model development. These data will be integrated into the first version of AquaINHABIT, a web application displaying visual and statistical summaries of freshwater habitat suitability models for manager identified invasive species. These species include: Aldrovanda vesiculosa, Alisma plantago-aquatica, Alosa pseudoharengus, Alternanthera philoxeroides, Ambloplites rupestris, Ameiurus catus, Ameiurus melas, Ameiurus natalis, Arundo donax, Astronotus ocellatus, Azolla cristata, Azolla pinnata, Bithynia tentaculata, Butomus umbellatus, Bythotrephes longimanus, Cabomba caroliniana, Callitriche stagnalis, Canna glauca, Carassius auratus, Carassius gibelio, Channa argus, Chrosomus oreas, Cipangopaludina chinensis, Cipangopaludina japonica, Clarias batrachus, Colocasia esculenta, Corbicula fluminea, Crassula helmsii, Ctenopharyngodon idella, Cyperus blepharoleptos, Cyperus papyrus, Cyprinella lutrensis, Cyprinus carpio, Daphnia lumholtzi, Didymosphenia geminata, Dorosoma cepedianum, Dorosoma petenense, Dreissena bugensis, Dreissena polymorpha, Echinogammarus ischnus, Egeria densa, Egeria najas, Eichhornia azurea, Eichhornia crassipes, Eichhornia paniculata, Eubosmina coregoni , Fallopia japonica, Faxonius rusticus, Faxonius virilis, Gambusia affinis, Gambusia holbrooki, Glyceria maxima, Gymnocephalus cernua, Hemimysis anomala, Hottonia palustris, Hydrilla verticillata, Hydrilla verticillata peregrina, Hydrilla verticillata verticillata, Hydrocharis morsus-ranae, Hydrocotyle ranunculoides, Hygrophila polysperma, Hypophthalmichthys molitrix, Hypophthalmichthys nobilis, Ictalurus furcatus, Ipomoea aquatica, Iris pseudacorus, Lagarosiphon major, Landoltia punctata, Lepomis gulosus, Limnobium laevigatum, Limnobium spongia, Limnocharis flava, Limnophila indica, Limnophila sessiliflora, Ludwigia decurrens, Ludwigia grandiflora, Ludwigia hexapetala, Ludwigia octovalvis, Ludwigia peploides, Ludwigia peruviana, Lythrum hyssopifolia, Lythrum portula, Lythrum salicaria, Marsilea quadrifolia, Melaleuca quinquenervia, Melanoides tuberculata, Mentha aquatica, Micropterus henshalli, Misgurnus anguillicaudatus, Monochoria vaginalis, Monopterus albus, Murdannia keisak, Mylopharyngodon piceus, Myosotis scorpioides, Myriophyllum aquaticum, Myriophyllum heterophyllum, Myriophyllum sibiricum, Myriophyllum spicatum, Najas marina, Najas minor, Nasturtium microphyllum, Nasturtium officinale, Nelumbo lutea, Nelumbo nucifera, Neogobius melanostomus, Nitellopsis obtusa, Nymphaea alba, Nymphaea lotus, Nymphaea mexicana, Nymphoides cristata, Nymphoides grayana, Nymphoides indica, Nymphoides peltata, Oreochromis aureus, Oreochromis niloticus, Oryza sativa, Panicum repens, Persicaria hydropiper, Phalaris arundinacea, Phragmites australis australis, Pistia stratiotes, Pomacea canaliculata, Pomacea canaliculata, Pomacea maculata, Potamogeton crispus, Potamopyrgus antipodarum, Procambarus clarkii, Proterorhinus semilunaris, Pterygoplichthys anisitsi, Pterygoplichthys disjunctivus, Pterygoplichthys multiradiatus, Pterygoplichthys pardalis, Pylodictis olivaris, Rotala indica, Rotala rotundifolia, Sagittaria guayanensis, Sagittaria montevidensis, Salmo trutta, Salvinia auriculata, Salvinia minima, Salvinia molesta, Salvinia natans, Salvinia oblongifolia, Sander lucioperca, Scardinius erythrophthalmus, Sporobolus anglicus, Stratiotes aloides, Tamarix chinensis and ramosissima, Tamarix sp, Tinca tinca, Trapa bispinosa, Trapa natans, Typha domingensis, Utricularia inflata, and Veronica anagallis-aquatica. To provide information and documentation on the potential distribution of invasive aquatic species in the contiguous United States. 1980 2024 date of source data Complete Irregular Contiguous United States -128.3863 -65.0916 51.2443 23.0860 ISO 19115 Topic Category biota None Species Distribution Modeling VisTrails SAHM habitat suitability USGS Thesaurus invasive species biogeography modeling habitat suitability indices USGS Metadata Identifier USGS:685039bcd4be025b9e8c6368 None Contiguous United States Common geographic areas United States None. Please see 'Distribution Info' for details. None. Users are advised to read the dataset's metadata thoroughly to understand appropriate use and data limitations. Catherine S Jarnevich U.S. Geological Survey, Rocky Mountain Region Ecologist mailing

2150 Centre Avenue Bldg C

Fort Collins CO 80526 970-226-9439 jarnevichc@usgs.gov Funding to support AquaINHABIT has come from USGS Northeast Climate Adaptation Science Center, U.S. Army Corps of Engineers, and U.S. Geological Survey Invasive Species Program. The U.S. Geological Survey Fort Collins Science Center participated in the project. Created using version 2.2.2 of the VisTrails/SAHM package (Morisette et al., 2013), and version 4.3.1 of R, including "taxize", "CoordinateCleaner", "enmSdmX" packages. Citations: Scott Chamberlain, Eduard Szoecs, Zachary Foster, Zebulun Arendsee, Carl Boettiger, Karthik Ram, Ignasi Bartomeus, John Baumgartner, James O'Donnell, Jari Oksanen, Bastian Greshake Tzovaras, Philippe Marchand, Vinh Tran, Maëlle Salmon, Gaopeng Li, and Matthias Grenié. (2020) taxize: Taxonomic information from around the web. R package version 0.9.98. https://github.com/ropensci/taxize Zizka A, Silvestro D, Andermann T, Azevedo J, Duarte Ritter C, Edler D, Farooq H, Herdean A, Ariza M, Scharn R, Svanteson S, Wengstrom N, Zizka V, Antonelli A (2019). “CoordinateCleaner: standardized cleaning of occurrence records from biological collection databases.” _Methods in Ecology and Evolution_, -7. doi:10.1111/2041-210X.13152 <https://doi.org/10.1111/2041-210X.13152>, R package version 3.0.1, <https://github.com/ropensci/CoordinateCleaner>. Smith A, Murphy S, Henderson D, Erickson K (2023). “Including imprecisely georeferenced specimens improves accuracy of species distribution models and estimates of niche breadth.” _Global Ecology & Biogeography_, *32*, -13. doi:10.1111/geb.13628 <https://doi.org/10.1111/geb.13628>. Catherine S. Jarnevich Peder Engelstad Demetra Williams Keana Shadwell Cameron Reimer Grace Henderson Janet S. Prevey Ian S. Pearse 20241125 publication NeoBiota vol. 96 n/a Pensoft Publishers ppg. 261-278 https://doi.org/10.3897/neobiota.96.134842 Henderson, Grace Engelstad, Peder Reimer, Cameron J LeClare, Shelby K. Fraser, Linnea S. Williams, Demetra A. Shadwell, Keana S. Daniel, Wesley M. Pfingsten, Ian A. Jarnevich, Catherine S. 2025 publication https://doi.org/10.5066/P14JDTTJ Jeffrey T. Morisette Catherine S. Jarnevich Tracy R. Holcombe Colin B. Talbert Drew Ignizio Marian K. Talbert Claudio Silva David Koop Alan Swanson Nicholas E. Young 20130125 publication Ecography vol. 36, issue 2 n/a Wiley ppg. 129-135 https://doi.org/10.1111/j.1600-0587.2012.07815.x R Core Team 2025 application/service Vienna, Austria R Foundation for Statistical Computing https://www.R-project.org Before ensembling layers, we examined model outputs for quality. We ensured that the training and test Area Under the Curve (AUC) were greater than 0.7. We also visually assessed the complexity of response curves to identify overfitting. With evidence of overfitting based on those criteria, we explored alternative model-specific tuning parameters to decrease overfitting while maintaining good performance. We also ensured that the train and test Continuous Boyce Index (CBI) values were greater than 0.5. Model algorithms for each species that did not meet this threshold were assessed individually and when deemed appropriate, subsequently dropped. After layers were ensembled, maps were visualized for consistency and ecological plausibility. We used the hashtag method to split our data into non-random, spatially-sampled cross-validation and independent test data to evaluate model performance. This methodology uses a hashtag shape (#) overlaid on the 99% binary kernel density estimate (KDE) of the occurrence points. Observation points falling within the buffered hashtag are assigned to the test split (and withheld for model evaluation) with data desired ratio of 70% train/ 30% test. We evaluated location data for accuracy including taxonomic and spatial coordinates. All known synonyms were collected (excluding subspecies, variants, and hybrids) using the Global Biodiversity Information System backbone (ITIS; www.itis.gov) as an authoritative taxonomy in the R library ‘taxize’. We filtered observations by observation type (observation or specimen only), observation date (1980 to present), and coordinate uncertainty (≤ 30 m). Furthermore, we cleaned the species data and corresponding geographic coordinates by flagging potentially erroneous coordinates that are known common issues in biological collection databases using the R package “CoordinateCleaner”. This included removing records from within country capitals, within country centroids and within a 100m radius of known biodiversity institutions (which include research centers, universities, herbaria, museums, zoos, and botanical gardens). We also removed records with coordinates that fall into the oceans, coordinates with latitude and longitudes equal to zero, and species with equal latitude and longitude coordinates. We filtered to species that fell within the stream template for stream models and within the waterbody template for waterbody models. We checked the entire dataset for duplicate records. Locations compared with reported distributions. Data sets are considered complete for the information presented, as described in the abstract. Users are advised to read the rest of the metadata record carefully for additional details. We filtered our dataset to coordinates with ≤ 30 m accuracy and visually screened our data for positional accuracy. We filtered our dataset to coordinates with ≤ 30 m accuracy and visually screened our data for positional accuracy. U.S. Environmental Protection Agency (USEPA) U.S. Geological Survey 2012 vector digital data n/a US Geological Survey https://www.sciencebase.gov/catalog/item/56c38ad8e4b0946c6520aa52 https://www.epa.gov/waterdata/get-nhdplus-national-hydrography-dataset-plus-data Digital and/or Hardcopy 2012 date of source data NHDPlus HR geospatial extent of freshwater habitats and mean annual flow predictor The University of Georgia - Center for Invasive Species and Ecosystem Health 20240319 tabular digital data EDDMapS. 2021. Early Detection & Distribution Mapping System. The University of Georgia - Center for Invasive Species and Ecosystem Health. Available online at http://www.eddmaps.org/; last accessed March 19, 2024. http://www.eddmaps.org/ Digital and/or Hardcopy 19800101 20201201 observed EDDMAPS Species location data Global Biodiversity Information Facility 20240225 tabular digital data Derived dataset GBIF.org, accessed from R via rgbif (https://github.com/ropensci/rgbif) on 2023-02-07 and 2024-03-19. Filtered export of GBIF occurrence data. https://doi.org/10.15468/dl.ckybbt https://doi.org/10.15468/dl.p2c8k2 Digital and/or Hardcopy 19800101 20240225 observed GBIF Species location data USDI-BLM 20210101 tabular digital data https://gbp-blm-egis.hub.arcgis.com/pages/aim Digital and/or Hardcopy 19800101 20230314 observed AIM Lotic Species location data Laís Petri Evelyn M. Beaury Jeffrey Corbin Kristen Peach Helen Sofaer Ian S. Pearse Regan Early David T. Barnett Inés Ibáñez Robert K. Peet Michael Schafale Thomas R. Wentworth James P. Vanderhorst David N. Zaya Greg Spyreas Bethany A. Bradley 20230112 publication Ecology vol. 104, issue 3 n/a Wiley https://doi.org/10.1002/ecy.3947 Digital and/or Hardcopy 19800101 20230112 observed SPCIS Species location data NatureServe 20230419 tabular digital data http://www.imapinvasives.org Digital and/or Hardcopy 19800101 20230419 observed iMapInvasives Species location data LANDFIRE Earth Resources Observation and Science Center (EROS) U.S. Geological Survey 20210805 LF Remap raster digital data Sioux Falls, SD Earth Resources Observation and Science Center (EROS), U.S. Geological Survey https://www.landfire.gov Digital and/or Hardcopy 2016 ground condition Landfire Species location data Amy J Benson Holly J Beck Matthew E Neilson Wesley M Daniel Aidan L Devantier Amanda Cousins Anastashia M King Mason Schermerhorn Shanamon Tangkulwarodom 2023 dataset https://www.sciencebase.gov U.S. Geological Survey https://doi.org/10.5066/p9zirvf0 Digital and/or Hardcopy 20220531 ground condition Boatramps boatramp density pathway Grace Henderson Demetra A. Williams Keane S. Shadwell Cameron J. Reimer Shelby K. LeClare Linnea S. Fraser Peder S. Engelstad Catherine S. Jarnevich 2025 raster digital data ScienceBase U.S. Geological Survey https://doi.org/10.5066/P14JDTTJ Digital and/or Hardcopy 1971 2059 observed Freshwater predictors Raster predictor layers U.S. Geological Survey 20240408 tabular digital data Gainesville, FL U.S. Geological Survey http://nas.er.usgs.gov Digital and/or Hardcopy 19800101 20240408 observed NAS Occurrence data and pathways information Janet S Prevey Cameron J Reimer Peder Engelstad Pairsa N Belamaric Terri Hogan Jillian M. Laroe Colter J Mumford Jennifer L Sieracki Catherine S Jarnevich 20250609 Dataset https://www.sciencebase.gov U.S. Geological Survey https://doi.org/10.5066/p13mg455 Digital and/or Hardcopy 1981 2023 observed human transport risk human_transport_risk raster layer for pathway summary Bethany Bradley Annette Evans Catherine S Jarnevich Evelyn Beaury Peder S Engelstad Nathan B Teich Jillian M Laroe 20240426 Dataset https://www.sciencebase.gov U.S. Geological Survey https://doi.org/10.5066/p14vvres Digital and/or Hardcopy 1980 2023 observed Bradley et al. Occurrence data Species location data were acquired from online occurrence data repositories using data_pull R scripts. Sources included Nonindigenous Aquatic Species (NAS) database, Global Biodiversity Information Facility (GBIF), the Bureau of Land Management’s (BLM) Assessment, Inventory and Management (AIM) Lotic database, the Early Detection and Distribution Mapping System (EDDMapS), the Standardized Plant Community with Introduced Status database (SPCIS), a plant abundance dataset from Bradley et al. 2024, iMapInvasives database, and the LANDFIRE reference database. AIM data were obtained directly from the BLM by request. See the Logical Consistency Report section for how species location data were further filtered and cleaned to ensure data quality using the data_prep R scripts. Occurrence data were spatially thinned to 1 location per 0.9 kilometer (km) to avoid issues of autocorrelation using the R package “enmSdmX” (Smith et al., 2023). These thinned data were then inputs to VisTrails Software for Assisted Habitat Modeling (SAHM, Morisette et al., 2013, https://doi.org/10.1111/j.1600-0587.2012.07815.x) workflows to create each habitat suitability layer. Predictor layers were from Henderson et al. 2025 (Freshwater predictors) and include predictors capturing climate, topography, land use, hydrologic characteristics, and soils. EDDMAPS GBIF AIM Lotic SPCIS iMapInvasives Landfire NAS Bradley et al. Freshwater predictors 20240713 Each observation record was assigned to the stream template or waterbody template from Henderson et al. 2025 (Freshwater predictors) based on spatial location. Where enough data were available, we created a model for each habitat. We used the Target Background approach to create background points to use in modeling. Species distribution models (SDMs) were fit following modified methods from Jarnevich et al. (2024, https://doi.org/10.3897/neobiota.96.134842) using the SAHM (version 2.2.2, Morisette et al., 2013) in the VisTrails software. For species with >143 spatially thinned occurrences, ~70% of occurrences were used to train the model and the other ~30% were withheld to evaluate the model. All models were fit using n-fold spatial cross validation (where n ranges from 3-10). The model algorithms included Random Forest (RF), Multiple Adaptive Regression Splines (MARS), Maxent (v 3.4.1), Generalized Linear Model (GLM), and Boosted Regression Trees (BRT). See Jarnevich et al. (2024) for additional details. SAHM produces several standard outputs for each model including assessment metrics for training and test data, rasters that map predicted habitat suitability across geographic space, metrics of predictor importance (change in Area Under the Curve (AUC) values when predictor values are permuted between presence and background locations), and response curves describing how relative habitat suitability varies across the range of values for each predictor. Finally, we assessed model performance using the continuous Boyce index (CBI) calculated for each of the habitats per species to determine overall performance using the data_prep/boyce_functions.R script. We used the calculated CBI metric for each algorithm plus habitat to determine inclusion in the ensemble step (Step 3). Individual models with CBI values of less than 0.50 were assessed and dropped if deemed appropriate. See Species_model_information.csv for which algorithms were excluded from a species' ensemble map. 20250115 We created maps of continuous habitat suitability for each species, including applying to the five future scenarios (raster layers 1, 3). These maps represent a weighted ensemble of the algorithms used to generate the models (up to 5 occurrence maps), weighted according to their CBI values and calculated from model output maps using data_postprocessing R scripts. They represent habitat suitability on a 0-100 scale, where lower values represent less predicted suitability for that species and higher values represent greater predicted suitability. These maps were also restricted based on environmental conditions, meaning areas where environmental conditions are outside the environmental range of location points used to build the model are masked from the map (raster layers 2) calculated with data_postprocessing/mess-calc-function.R. For future scenarios we counted the number of future scenarios with environmental conditions outside the range (raster layer 7). We also calculated the maximum and minimum values calculated for each pixel by the five future scenarios and the standard deviation across future scenario predictions. 1) Current occurrence suitability - Continuous value ensemble (XX-ens-current-mean .tif) 2) Restricted current occurrence suitability - Continuous value ensemble with restricted environmental conditions* (XX-ens-current-mean-masked.tif) 3) Future occurrence suitability - Continuous value ensemble (XX-ens-future-mean.tif) 4) Maximum future occurrence suitability - Maximum continuous value ensemble from the five alternative scenarios (XX-ens-future-max-gcm.tif) 5) Minimum future occurrence suitability - Minimum continuous value ensemble from the five alternative scenarios (XX-ens-future-min-gcm.tif) 6) Standard deviation of future occurrence suitability - Standard deviation of continuous value from each algorithm for the five alternative scenarios (XX-ens-future-stdev.tif) 7) Restricted count - Count of ensembles from each alternative scenario with restricted environmental conditions* (XX-gcm-mask-count.tif) *Restricted environmental conditions = only display areas where environmental characteristics are inside the range of the values used to develop the model. For example, a location with a minimum winter temperature of 12 C would be outside the range of -10 to 10 C used in model development. 20250430 We created spatial layers using data_pathways R scripts representing potential pathways of movement using the included subwatershed vector layer (WBD_HUC12_EasternUS.gpkg) for a study area in the Northeast, including boat ramp density per subwatershed (pathway-boatramps-by-huc12.R), water area per subwatersheds (from waterbody template and river template in Freshwater predictors; pathway-perm-water-by-huc12.R), average subwatershed human transport risk (pathway-remoteness-by-huc12.R), anthropogenic connectivity across (HUC4) boundaries from features in the NHDPlus HR (pathway-anthropogenic-connectivity-by-huc-4.R), and water connectivity per species per subwatershed based on occurrences and direction of flow from NHDPlus HR (pathway-water-connectivity-by-huc12.R). The final scores were generated using composite-risk-score-species-by-huc12.R. The values for each pathway were rescaled between zero and one across subwatersheds. We then summed the pathway scores for each species, using pathways reported in Species_model_information.csv. We also calculated the average suitability score for current and future conditions for each species and subwatershed. The final risk score was the sum of the combined pathway score, the average current suitability score, and the average future suitability score for each HUC12. NHDPlus HR Boatramps human transport risk 20250430 Vector G-polygon Albers Conical Equal Area 20.0 60.0 -96.0 40.0 0.0 0.0 row and column 98.4693338923368 98.4693338923368 meters North_American_Datum_1983 GRS 1980 6378137.0 298.2572221010042 Species_model_information.csv Comma Separated Value (CSV) file containing data. Producer Defined scientific_name Species identity flag in the format genus_species Producer Defined Scientific name for 159 species in the format genus_species common_name Species identity flag as common name associated with the scientific name in scientific_name Producer Defined Common name for species Taxa Taxonomic group the species (scientific_name) belongs to Producer Defined plant Scientific name in Plantae Kingdom Producer defined fish Scientific name classified as fish Producer defined invertebrate Scientific name classified as invertebrate Producer defined area_thresh Area threshold parameter used to define the hashtag shape that defined the test data area Producer Defined NA Not applicable, indicated a hashtag shape was not used. Producer defined 0.03 18.0 index habitat_type Type of freshwater habitat modeled Producer Defined waterbody Waterbody template used Producer defined stream Stream template used Producer defined Notes Any species habitat type model notes related to model_changes or removal. Producer Defined << empty cell >> No notes were needed Producer defined Freeform text where needed to explain model_changes or removed_models. train_pts Number of training points used to fit the model described in habitat_type Producer Defined Not enough points Not enough training points to fit a model Producer defined 50 12048 Count test_pts Number of test points used to assess the model described in habitat_type Producer Defined NA Not applicable (no test point split) Producer defined 61 4438 n_splits Number of training data splits for the model Producer Defined NA No Data Producer defined 5 10 bg_pts Number of background points for the model Producer Defined NA No Data Producer defined 857 10000 model_changes Freeform text to describe any model algorithm changes for the species model group. Producer Defined << empty cell >> No model changes were recorded Producer defined Freeform text to describe any model algorithm specific changes for the species model group where brt is boosted regression tree, glm is generalized linear model, mars is multivariate adaptive regression spline, rf is random forest, and kde is kernel density estimate. removed_algorithms Freeform text list of model algorithms removed from the habitat_type Producer Defined NA No Data Producer defined List of model algorithms dropped from consideration in the final model ensemble for the specified species' habitat_type, where brt is boosted regression tree, glm is generalized linear model, mars is multivariate adaptive regression spline, rf is random forest, and kde is kernel density estimate. min_cbi Minimum continuous boyce index for any algorithm included in the ensemble Producer Defined NA No Data Producer defined 0.22 0.97 index max_cbi Maximum continuous boyce index for any algorithm included in the ensemble Producer Defined NA No Data Producer defined 0.63 1.0 index ensemble_cbi Continuous boyce index for the ensemble for the species model habitat type including CBI based on training data and with test data CBI in parenthesis. Producer Defined NA No Data Producer defined Continuous boyce index for the ensemble for the species model habitat type. pathways Freeform text highlighting the pathways applicable to the species Producer Defined NA No pathway information collected Producer defined Freeform text highlighting the pathways applicable to the species species_code Code generate to represent the species Producer Defined NA No Data Producer defined Code generated based on the scientific_name to represent the species in file names threshold Calculated threshold values to create binary maps Producer Defined NA Not applicable; no threshold values calculated for the species for the habitat type Producer defined Freeform text of calculated first, fifth, and tenth threshold values Merged_dataset.csv Comma Separated Value (CSV) file containing data. Producer Defined scientific_name Species identity flag in the format genus_species Producer Defined Scientific name for modeled species in the format genus_species model_type Type of freshwater habitat modeled Producer Defined stream Data to fit model using stream template Producer defined waterbody Data to fit model using waterbody template Producer defined set Identified if the data were used to fit the model (train) or assess the model (test) Producer Defined train Locations used to train the model Producer defined test Locations used to assess (test) the model Producer defined X X coordinate in Alber's Equal Area (ESRI:102008) Producer Defined -2236332 2124678 meter Y Y coordinate in Alber's Equal Area (ESRI:102008) Producer Defined -1686461 1328473 meter Split Spatial cross-validation split value assigned to the point Producer Defined test No split because set = test Producer defined 1 10 index responseBinary Flag indicating if presence or background location Producer Defined 1 Occurrence location Producer defined -9998 Background location Producer defined Annual_Mean_Temp mean annual temperature [1981 to 2010] Producer Defined 2666 2997 degree C *100 Mean_Temp_of_Warmest_Quarter mean temperature of warmest quarter [1981 to 2010] Producer Defined 2783 3116 degree C * 100 Mean_Temp_of_Coldest_Quarter mean temperature of coldest quarter [1981 to 2010] Producer Defined 2571 2949 degree C * 100 Annual_Precip annual precipitation [1981 to 2010] Producer Defined 674 47993 kg m-2 year-1 Precipitation_of_Wettest_Month precipitation of wettest month [1981 to 2010] Producer Defined 99 7483 kg m-2 year-1 Precipitation_of_Driest_Month precipitation of driest month [1981 to 2010] Producer Defined 1 1601 kg m-2 year-1 Precipitation_Seasonality precipitation seasonality [1981 to 2010] Producer Defined 62 1042 kg m-2 Precipitation_of_Wettest_Quarter Mean precipitation of wettest quarter [1981 to 2010] Producer Defined 272 22265 kg m-2 month-1 Precipitation_of_Driest_Quarter mean precipitation of driest quarter [1981 to 2010] Producer Defined 22 5438 kg m-2 month-1 Precipitation_of_Warmest_Quarter mean precipitation of warmest quarter [1981 to 2010] Producer Defined 22 7876 kg m-2 month-1 Precipitation_of_Coldest_Quarter mean precipitation of coldest quarter [1981 to 2010] Producer Defined 181 18487 Mean_Monthly_Temp_Range mean diurnal range [1981 to 2010] Producer Defined 11 173 degree C * 10 Isothermality isothermality [1981 to 2010] Producer Defined 104 590 degree C * 10 Temp_Seasonality temperature seasonality [1981 to 2010] Producer Defined 1201 12599 degree C/100 Max_Temp_of_Warmest_Month maximum temperature of warmest month [1981 to 2010] Producer Defined 2875 3172 degree C * 100 Min_Temp_of_Coldest_Month minimum temperature of coldest month [1981 to 2010] Producer Defined 2490 2933 degree C/100 Annual_Temp_Range temperature annual range [1981 to 2010] Producer Defined 46 456 degree C * 10 Agriculture_Land_Cover Percent cropland in subwatershed [HUC12] in 2023 Producer Defined NA No Data Producer defined 0 100 percent Developed_Land_Cover Percent developed in subwatershed [HUC12] in 2023 Producer Defined NA No Data Producer defined 0 100 Forest_Land_Cover Percent forest in subwatershed [HUC12] in 2023 Producer Defined NA No Data Producer defined 0 100 Grassland_Land_Cover Percent grassland in subwatershed [HUC12] in 2023 Producer Defined NA No Data Producer defined 0 100 Mean_Annual_Stream_Flow streamflow Producer Defined NA No Data Producer defined 0.0 685478.88 cubic feet per second Elevation Elevation Producer Defined NA No Data Producer defined -84.0 4163.58 meter Monthly_Growing_Season_ET Mean monthly evapotranspiration of the growing season (April through October) Producer Defined NA No Data Producer defined 5 184200 millimeter Monthly_Summer_ET Mean monthly evapotranspiration of the summer (June through August) Producer Defined NA No Data Producer defined 5 96157 millimeter Monthly_Spring_ET Mean monthly evapotranspiration of spring months (March through June) Producer Defined NA No Data Producer defined 5 101000 millimeter Monthly_Early_Spring_ET Mean monthly evapotranspiration of spring months (March through May) Producer Defined NA No Data Producer defined 5 70500 millimeter Fall_to_Spring_ET Mean monthly evapotranspiration of fall to spring (October through June) Producer Defined NA No Data Producer defined 5 170163 millimeter Human_Modification_Index global human modification index Producer Defined 0 65108 index mTPI multi-scale topographic position index Producer Defined -316 269 index Median_NDMI median normalized difference moisture index Producer Defined NA No Data Producer defined -647 934 index SD_NDMI normalized difference moisture index standard deviation Producer Defined NA No Data Producer defined 0 426 standard deviation Potential_Water_Deficit potential evapotranspiration Producer Defined 45.25 197.66 millimeter Shore_Density shore length / waterbody area Producer Defined NA No Data Producer defined 0 98.38 m/m3 Slope slope Producer Defined NA No Data Producer defined 0 90 degree Soil_Bulk_Density predicted soil bulk density at a 5cm depth (oven dry bulk density) Producer Defined NA No Data Producer defined 11.43 198 grams per cubic centimeters *100 Soil_Calcium_Carbonate_Content percent soil calcium carbonate at a 5cm depth Producer Defined NA No Data Producer defined 0.0 78.0 percent Soil_Clay_Content predicted soil clay content at a 5cm depth Producer Defined NA No Data Producer defined 0.0 80.77 percent Soil_Potassium_Content predicted soil k content at a 5cm depth Producer Defined NA No Data Producer defined 0 294.43 centimoles per kilogram Soil_Nitrogen_Content predicted soil n content at a 5cm depth Producer Defined NA No Data Producer defined 0 228.78 percent * 100 Soil_pH predicted soil pH measured using 1:1 method (-log10([H+])) at a 5cm depth Producer Defined NA No Data Producer defined 340.9 942.3 ph * 100 Soil_Depth soil depth to any restrictive layer right censored at 201cm Producer Defined NA No Data Producer defined 3.0 201.0 centimeter Soil_Sand_Content predicted soil sand content at a 5cm depth (total sand content) Producer Defined NA No Data Producer defined 0.0 99.88 percent Soil_Organic_Carbon_Content predicted soil organic carbon content at a 5cm depth Producer Defined NA No Data Producer defined 7.0 52617 percent * 1000 Assessment_metrics.csv Comma Separated Value (CSV) file containing data. Producer Defined scientific_name Species identity flag in the format genus_species Producer Defined Scientific name for species in the format genus_species model_type Type of freshwater habitat modeled Producer Defined waterbody model using waterbody template Producer defined stream model using stream template Producer defined algorithm Algorithm used to generate the model being assessed Producer Defined BRT Boosted regression tree Producer defined GLM Generalized linear model Producer defined MARS Multivariate adaptive regression splines Producer defined RF Random forest Producer defined Maxent Maxent Producer defined masked ensemble Continuous value ensemble (of all algorithms) with restricted environmental conditions Producer defined unmasked ensemble Continuous value ensemble (of all algorithms) with unrestricted environmental conditions Producer defined metric Model evaluation metrics Producer Defined AUC Area Under the Curve - Receiver Operator Curve Producer defined AUC-PR AUC - Precision Recall Producer defined correlation coefficient correlation coefficient Producer defined percent correctly classified percent correctly classified Producer defined sensitivity sensitivity Producer defined kappa Kappa Index Producer defined true skill statistic true skill statistic Producer defined CBI Continuous Boyce Index Producer defined type Whether a metric was evaluated on test, training, or cross-validation data Producer Defined train training data Producer defined cross-validation average across cross-validation splits Producer defined test test data Producer defined value Value of model evaluation metric Producer Defined Possible value ranges differ based on the metric defined in the 'metric' column Variable_importance.csv Comma Separated Value (CSV) file containing data. Producer Defined scientific_name Species identity flag in the format genus_species Producer Defined Scientific name for species in the format genus_species model_type Type of freshwater habitat modeled Producer Defined waterbody model using waterbody template Producer defined stream model using stream template Producer defined algorithm Algorithm used to generate the model the variable importance values are for Producer Defined BRT Boosted regression tree Producer defined GLM Generalized linear model Producer defined MARS Multivariate adaptive regression splines Producer defined RF Random forest Producer defined Maxent Maxent Producer defined predictor Name of predictor that variable importance value is associated with for the model specified by the model_type and algorithm columns Producer Defined Predictor variables defined as columns in Merged_dataset AUC_difference Calculated change in AUC for the model when predictor values are permutated between presence and background Producer Defined -0.944 0.789 area under the curve difference split cross-validation split number or train for all training data with which the variable importance values is associated Producer Defined Split number from 1 to 10 or train Risk_and_pathway_scores.csv Comma Separated Value (CSV) file containing data. Producer Defined huc12 Hydrologic Unit Code (huc) 12 Producer Defined 12 digit Hydrologic Unit Code (HUC) Watershed Boundary Dataset, USGS scientific_name Species identity flag in the format genus_species Producer Defined NA Data is applicable to the huc12 and not species specific Producer defined Scientific name for species in the format genus species pathway Pathway that the value is for Producer Defined combined pathway score Sum of the normalized pathway scores (each scaled 0 to 1) by species (scientific_name) applicable to the specified species (scientific_name), renormalized summed scores (0 to 1) Producer defined current habitat Mean current suitability for each species by huc12 from the raster 'SPECIES_ens-current-mean-masked', normalized by species across the study area (0 to 1) Producer defined future habitat Mean future suitability for each species by huc12 from the raster 'SPECIES_ens-future-mean', normalized by species across the study area (0 to 1) Producer defined combined risk score Sum of combined pathway score, current habitat, and future habitat with values ranging from 0 to 3. Producer defined water connectivity Distance (meter) from closest known occurrence of the species to this huc12 along waterways within the NHDPlus HR to capture natural spread Producer defined boat ramp density Number of boat ramps/ (area of waterbody template + area of stream template) to capture hitchhiking on boats/ trailer/ equipment Producer defined water area Ratio of waterbody area (defined by waterbody template) to HUC12area to capture hitchhike on/ in waterfowl Producer defined human transport index Sum relative human transport risk scores across the HUC12 weighted by area of the HUC12 to capture release or escape from captivity/water gardens/aquaculture/live bait Producer defined anthropogenic connectivity Count of human-built water infrastructure (e.g., canals, waterpipes) crossing HUC4 boundaries to capture spread via artificial connections Producer defined value Value for the metric described by pathway Producer Defined Range of values depends on the pathway, either 0 to 1 or 0 to 3 WBD_HUC12_EasternUS.gpkg GeoPackage file with polygons representing 12-digit hydrologic units used in analyses Producer Defined huc12 Hydrologic Unit Code (huc) 12 Producer Defined 12 digit Hydrologic Unit Code (HUC) Watershed Boundary Dataset, USGS necasc Flag identifying HUC12s within risk score study area Producer Defined 0 Not within risk score study area Producer defined 1 Within risk score study area Producer defined GS ScienceBase U.S. Geological Survey mailing address

Denver Federal Center, Building 810, Mail Stop 302

Denver CO 80225 United States 1-888-275-8747 sciencebase@usgs.gov Unless otherwise stated, all data, metadata and related materials are considered to satisfy the quality standards relative to the purpose for which the data were collected. Although these data and associated metadata have been reviewed for accuracy and completeness and approved for release by the U.S. Geological Survey (USGS), no warranty expressed or implied is made regarding the display or utility of the data for other purposes, nor on all computer systems, nor shall the act of distribution constitute any such warranty. Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government. Digital Data https://doi.org/10.5066/P13JMOQW None 20260123 FORT Data Management USGS Fort Collins Science Center FORT Data Management mailing address

2150 Centre Avenue Bldg C

Fort Collins CO 80526 US 970-226-9100 fortdatamanagement@usgs.gov FGDC Biological Data Profile of the Content Standard for Digital Geospatial Metadata FGDC-STD-001.1-1999