This is a dataset containing the input and output data used in the analysis of best practices of invasive plant species distribution modeling (Young et al. 2024). We developed habitat suitability models for 13 invasive plant species at a variety of geographic ranges and different invasion stages and modeling strategies to assess the impact of predictor quality, thinning resolution, and geographic range of occurrence points on model performance. We developed a library of environmental variables at both the global scale and at the scale of the contiguous United States known to physiologically limit plant distributions (Young et al. 2024, Table S1) 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 with VisTrails: Software for Assisted Habitat Modeling (SAHM 2.2.2, Morisette et al., 2013). We used the Continuous Boyce Index (CBI) as a metric to assess model performance. This data bundle contains the merged data sets used to create the models, including location and associated environmental data, for each species, invasion stage, and modeling strategy, grouped by predictor set. In this data bundle, we have also included a dataframe of CBI values for each species, invasion stage, and modeling strategy, used in our analyses. The species include Ailanthus altissima, Alliaria petiolata, Brassica tournefortii, Cenchrus ciliaris, Chondrilla juncea, Cirsium vulgare, Dioscorea bulbifera, Imperata cylindrica, Lonicera maackii, Lysimachia nummularia, Microstegium vimineum, Pueraria montana, and Ranunculus testiculatus.

We developed habitat suitability models for invasive plant species selected by Department of Interior land management agencies. We applied the modeling workflow developed in Young et al. 2020 to species not included in the original case studies. Our methodology balanced trade-offs between developing highly customized models for a few species versus fitting non-specific and generic models for numerous species. We developed a national library of environmental variables known to physiologically limit plant distributions and relied on human input based on natural history knowledge to further narrow the variable set for each species before developing habitat suitability models. We developed models using five algorithms with VisTrails: Software for Assisted Habitat Modeling [SAHM 2.1.2]. We accounted for uncertainty related to sampling bias by using two alternative sources of background samples, and constructed model ensembles using the 10 models for each species (five algorithms by two background methods) for four different thresholds. Each species folder contains the potential distribution of the species and all raster layers were produced using VisTrails:SAHM [SAHM 2.1.2]. Each of the 8 rasters represent the following: 1) MPP - minimum predicted presence threshold 2) 0.01 - one percentile threshold 3) 0.1 - ten percentile threshold 4) MaxSS - maximum sensitivity plus specificity threshold 5) MPP - minimum predicted presence threshold with Restricted Environmental Conditions 6) 0.01 - one percentile threshold with Restricted Environmental Conditions 7) 0.1 - ten percentile threshold with Restricted Environmental Conditions 8) MaxSS - maximum sensitivity plus specificity threshold with Restricted Environmental Conditions These rasters will be integrated into the Invasive Species Habitat Tool (INHABIT), a web application displaying visual and statistical summaries of nationwide habitat suitability models for manager identified invasive plant species. These species include: African rue (Peganum harmala), Air potato (Dioscorea bulbifera), Amur honeysuckle (Lonicera maackii), Amur peppervine (Ampelopsis brevipedunculata), Annual bluegrass (Poa annua ), Annual rye (Lolium multiflorum), Asian mustard (Brassica tournefortii), Beefsteak mint (Perilla frutescens), Bigleaf periwinkle (Vinca major), Bird vetch (Vicia cracca), Bishop's goutweed (Aegopodium podagraria), Black henbane (Hyoscyamus niger), Bohemian knotweed (Fallopia bohemica), Bradford pear (Pyrus calleryana), Buffelgrass (Cenchrus ciliaris), Bulbous bluegrass (Poa bulbosa), Bull thistle (Cirsium vulgare), Bur buttercup (Ranunculus testiculatus), Burning bush (Euonymus alatus), Camelthorn (Alhagi maurorum), Canada thistle (Cirsium arvense), Cereal rye (Secale cereale), Cheatgrass (Bromus tectorum), Chinaberry (Melia azedarach), Chinese holly (Ilex cornuta), Chinese privet (Ligustrum sinense), Chinese tallowtree (Triadica sebifera), Chinese wisteria (Wisteria sinensis), Chocolate vine (Akebia quinata), Clasping pepperweed (Lepidium perfoliatum), Cogongrass (Imperata cylindrica), Common crupina (Crupina vulgaris), Common gorse (Ulex europaeus ), Common reed (Phragmites australis), Common tansy (Tanacetum vulgare), Coral ardisia (Ardisia crenata), Crape myrtle (Lagerstroemia indica), Creeping bentgrass (Agrostis stolonifera), Creeping buttercup (Ranunculus repens), Crested wheatgrass (Agropyron cristatum), Crown vetch (Securigera varia), Dalmatian toadflax (Linaria dalmatica), Diffuse knapweed (Centaurea diffusa), Dyer's woad (Isatis tinctoria), English holly (Ilex aquifolium), English ivy (Hedera helix), European beachgrass (Ammophila arenaria ), False brome (Brachypodium sylvaticum), Field brome (Bromus arvensis), Fountaingrass (Pennisetum setaceum), French broom (Genista monspessulana), Fuller's teasel (Dipsacus fullonum), Garlic mustard (Alliaria petiolata), Giant knotweed (Fallopia sachalinensis), Hairy cat's ear (Hypochaeris radicata), Halogeton (Halogeton glomeratus),

We developed habitat suitability models for invasive plant species selected by Department of Interior land management agencies. We applied the modeling workflow developed in Young et al. 2020 to species not included in the original case studies. Our methodology balanced trade-offs between developing highly customized models for a few species versus fitting non-specific and generic models for numerous species. We developed a national library of environmental variables known to physiologically limit plant distributions and relied on human input based on natural history knowledge to further narrow the variable set for each species before developing habitat suitability models. We developed models using five algorithms with VisTrails: Software for Assisted Habitat Modeling [SAHM 2.1.2]. We accounted for uncertainty related to sampling bias by using two alternative sources of background samples, and constructed model ensembles using the 10 models for each species (five algorithms by two background methods) for four different thresholds. Each species folder contains the potential distribution of the species and all raster layers were produced using VisTrails:SAHM [SAHM 2.1.2]. Each of the 8 rasters represent the following: 1) MPP - minimum predicted presence threshold 2) 0.01 - one percentile threshold 3) 0.1 - ten percentile threshold 4) MaxSS - maximum sensitivity plus specificity threshold 5) MPP - minimum predicted presence threshold with Restricted Environmental Conditions 6) 0.01 - one percentile threshold with Restricted Environmental Conditions 7) 0.1 - ten percentile threshold with Restricted Environmental Conditions 8) MaxSS - maximum sensitivity plus specificity threshold with Restricted Environmental Conditions These rasters will be integrated into the Invasive Species Habitat Tool (INHABIT), a web application displaying visual and statistical summaries of nationwide habitat suitability models for manager identified invasive plant species. These species include: African rue (Peganum harmala), Air potato (Dioscorea bulbifera), Amur honeysuckle (Lonicera maackii), Amur peppervine (Ampelopsis brevipedunculata), Annual bluegrass (Poa annua ), Annual rye (Lolium multiflorum), Asian mustard (Brassica tournefortii), Beefsteak mint (Perilla frutescens), Bigleaf periwinkle (Vinca major), Bird vetch (Vicia cracca), Bishop's goutweed (Aegopodium podagraria), Black henbane (Hyoscyamus niger), Bohemian knotweed (Fallopia bohemica), Bradford pear (Pyrus calleryana), Buffelgrass (Cenchrus ciliaris), Bulbous bluegrass (Poa bulbosa), Bull thistle (Cirsium vulgare), Bur buttercup (Ranunculus testiculatus), Burning bush (Euonymus alatus), Camelthorn (Alhagi maurorum), Canada thistle (Cirsium arvense), Cereal rye (Secale cereale), Cheatgrass (Bromus tectorum), Chinaberry (Melia azedarach), Chinese holly (Ilex cornuta), Chinese privet (Ligustrum sinense), Chinese tallowtree (Triadica sebifera), Chinese wisteria (Wisteria sinensis), Chocolate vine (Akebia quinata), Clasping pepperweed (Lepidium perfoliatum), Cogongrass (Imperata cylindrica), Common crupina (Crupina vulgaris), Common gorse (Ulex europaeus ), Common reed (Phragmites australis), Common tansy (Tanacetum vulgare), Coral ardisia (Ardisia crenata), Crape myrtle (Lagerstroemia indica), Creeping bentgrass (Agrostis stolonifera), Creeping buttercup (Ranunculus repens), Crested wheatgrass (Agropyron cristatum), Crown vetch (Securigera varia), Dalmatian toadflax (Linaria dalmatica), Diffuse knapweed (Centaurea diffusa), Dyer's woad (Isatis tinctoria), English holly (Ilex aquifolium), English ivy (Hedera helix), European beachgrass (Ammophila arenaria ), False brome (Brachypodium sylvaticum), Field brome (Bromus arvensis), Fountaingrass (Pennisetum setaceum), French broom (Genista monspessulana), Fuller's teasel (Dipsacus fullonum), Garlic mustard (Alliaria petiolata), Giant knotweed (Fallopia sachalinensis), Hairy cat's ear (Hypochaeris radicata), Halogeton (Halogeton glomeratus),

We developed habitat suitability models for invasive plant species selected by Department of Interior land management agencies. We applied the modeling workflow developed in Young et al. 2020 to species not included in the original case studies. Our methodology balanced trade-offs between developing highly customized models for a few species versus fitting non-specific and generic models for numerous species. We developed a national library of environmental variables known to physiologically limit plant distributions (Engelstad et al. 2022 Table S1: https://doi.org/10.1371/journal.pone.0263056) and relied on human input based on natural history knowledge to further narrow the variable set for each species before developing habitat suitability models. We developed models using five algorithms with VisTrails: Software for Assisted Habitat Modeling [SAHM 2.1.2]. We accounted for uncertainty related to sampling bias by using two alternative sources of background samples, and constructed model ensembles using the 10 models for each species (five algorithms by two background methods) for three different thresholds (conservative to targeted). The mergedDataset_regionalization.csv file contains predictor values associated with pixels underlying each presence and background point. The testStripPoints_regionalization.csv file contains the locations of the modeled species occurring in the different geographic test strips.

We developed habitat suitability models for invasive plant species selected by Department of Interior land management agencies. We applied the modeling workflow developed in Young et al. 2020 to species not included in the original case studies. Our methodology balanced trade-offs between developing highly customized models for a few species versus fitting non-specific and generic models for numerous species. We developed a national library of environmental variables known to physiologically limit plant distributions (Engelstad et al. 2022 Table S1: https://doi.org/10.1371/journal.pone.0263056) and relied on human input based on natural history knowledge to further narrow the variable set for each species before developing habitat suitability models. We developed models using five algorithms with VisTrails: Software for Assisted Habitat Modeling [SAHM 2.1.2]. We accounted for uncertainty related to sampling bias by using two alternative sources of background samples, and constructed model ensembles using the 10 models for each species (five algorithms by two background methods) for three different thresholds (conservative to targeted). The mergedDataset_regionalization.csv file contains predictor values associated with pixels underlying each presence and background point. The testStripPoints_regionalization.csv file contains the locations of the modeled species occurring in the different geographic test strips.

We developed habitat suitability models for three invasive plant species: stiltgrass (Microstegium vimineum), sericea lespedeza (Lespedeza cuneata), and privet (Ligustrum sinense). We applied the modeling workflow developed in Young et al. 2020, developing similar models for occurrence data, but also models trained using species locations with percent cover ≥10%, ≥25%, and ≥50%. We chose predictors from a national library of environmental variables known to physiologically limit plant distributions (Engelstad et al. 2022 Table S1) and relied on human input based on natural history knowledge to further narrow the variable set for each species before developing habitat suitability models. We developed models using five algorithms with VisTrails: Software for Assisted Habitat Modeling [SAHM 2.1.2]. We selected background samples using the target background approach, and took an alternative approach to construct model ensembles by combining first percentile and ten percentile threshold rules (suitability values associated with the lowest one percent and lowest ten percent of the training data) to categorize the continuous output from each algorithm into low (below the one percentile), moderate (between the one and ten percentile), and high (above the ten percentile) suitability. Finally, we summed these to create an ensemble. This data bundle contains the merged data sets used to create the models, the composite raster files for each abundance threshold associated with each species, tabular summaries by management unit (including each species/ composite type combination), and the occurrence points with their associated cover. The spatial data are organized in a separate folder for each species, each containing 5 rasters describing potential habitat suitability for the species at the different abundance thresholds. Each of the rasters represent the composite map (composite_abundX.tif) for each abundance threshold. The bundle documentation files are: 1) 'thresholded_abundance_project_metdata.xml' (this file) which contains the project-level metadata 2) 'mergedDataset.csv' contains the merged data set used to create the models, including location and associated environmental data, for all three species for each thresholded abundance. 3) XX.tif where XX is the raster type explained above (abundance threshold). 4) managementSummary.csv is the tabular summaries by management unit.

We developed habitat suitability models for three invasive plant species: stiltgrass (Microstegium vimineum), sericea lespedeza (Lespedeza cuneata), and privet (Ligustrum sinense). We applied the modeling workflow developed in Young et al. 2020, developing similar models for occurrence data, but also models trained using species locations with percent cover ≥10%, ≥25%, and ≥50%. We chose predictors from a national library of environmental variables known to physiologically limit plant distributions (Engelstad et al. 2022 Table S1) and relied on human input based on natural history knowledge to further narrow the variable set for each species before developing habitat suitability models. We developed models using five algorithms with VisTrails: Software for Assisted Habitat Modeling [SAHM 2.1.2]. We selected background samples using the target background approach, and took an alternative approach to construct model ensembles by combining first percentile and ten percentile threshold rules (suitability values associated with the lowest one percent and lowest ten percent of the training data) to categorize the continuous output from each algorithm into low (below the one percentile), moderate (between the one and ten percentile), and high (above the ten percentile) suitability. Finally, we summed these to create an ensemble. This data bundle contains the merged data sets used to create the models, the composite raster files for each abundance threshold associated with each species, tabular summaries by management unit (including each species/ composite type combination), and the occurrence points with their associated cover. The spatial data are organized in a separate folder for each species, each containing 5 rasters describing potential habitat suitability for the species at the different abundance thresholds. Each of the rasters represent the composite map (composite_abundX.tif) for each abundance threshold. The bundle documentation files are: 1) 'thresholded_abundance_project_metdata.xml' (this file) which contains the project-level metadata 2) 'mergedDataset.csv' contains the merged data set used to create the models, including location and associated environmental data, for all three species for each thresholded abundance. 3) XX.tif where XX is the raster type explained above (abundance threshold). 4) managementSummary.csv is the tabular summaries by management unit.

We developed habitat suitability models for invasive plant species selected by Department of Interior land management agencies. We applied the modeling workflow developed in Young et al. 2020 to species not included in the original case studies. Our methodology balanced trade-offs between developing highly customized models for a few species versus fitting non-specific and generic models for numerous species. We developed a national library of environmental variables known to physiologically limit plant distributions (Engelstad et al. 2022 Table S1: https://doi.org/10.1371/journal.pone.0263056) and relied on human input based on natural history knowledge to further narrow the variable set for each species before developing habitat suitability models. We developed models using five algorithms with VisTrails: Software for Assisted Habitat Modeling [SAHM 2.1.2]. We accounted for uncertainty related to sampling bias by using two alternative sources of background samples, and constructed model ensembles using the 10 models for each species (five algorithms by two background methods) for three different thresholds (conservative to targeted). This data bundle contains a single file of tabular summaries by management unit (including each species/ ensemble type combination) and a subfolder for each species that contains the merged data sets used to create models, the six raster files associated with the species, and tabular outputs including response curve data, variable importance information, and model assessment metrics. Each of the six rasters represent the following: 1) 0.01 - one percentile threshold 2) 0.1 - ten percentile threshold 3) MaxSS - maximum sensitivity plus specificity threshold 4) 0.01 - one percentile threshold with Restricted Environmental Conditions 5) 0.1 - ten percentile threshold with Restricted Environmental Conditions 6) MaxSS - maximum sensitivity plus specificity threshold with Restricted Environmental Conditions The bundle documentation files are: 1) 'INHABIT_V3_metdata.xml' (this file) which contains the project-level metadata 2) managementSummaries.csv is the tabular summaries by management unit. 3) 'mergedDataset.csv' contains the merged data set used to create the models, including location and associated environmental data, for each species. 4) XX.tif where XX is the raster type explained above (threshold; masked or not). 5) responseCurves.csv is the tabular information need to produce response curves for each predictor retained in each of the 10 models produced for each species. 6) variableImportance.csv is the tabular summaries indicating predictor importance for each of the 10 models produced for each species. 7) assessmentMetrics.csv is the tabular summaries of assessment metrics for each model or ensemble for each species. These data will be integrated into the third version of the Invasive Species Habitat Tool (INHABIT), a web application displaying visual and statistical summaries of nationwide habitat suitability models for manager identified invasive plant species. These species include: African rue (Peganum harmala), Air potato (Dioscorea bulbifera), Alkali swainsonpea (Sphaerophysa salsula), Amur honeysuckle (Lonicera maackii), Amur maple (Acer ginnala), Amur peppervine (Ampelopsis brevipedunculata), Annual bluegrass (Poa annua), Annual rye (Lolium multiflorum), Asian mustard (Brassica tournefortii), Autumn olive (Elaeagnus umbellata), Balloon vine (Cardiospermum halicacabum), Beefsteak mint (Perilla frutescens), Bermudagrass (Cynodon dactylon), Bigleaf periwinkle (Vinca major), Bird vetch (Vicia cracca), Bishop's goutweed (Aegopodium podagraria), Black henbane (Hyoscyamus niger), Bohemian knotweed (Fallopia bohemica), Bradford pear (Pyrus calleryana), Brazilian peppertree (Schinus terebinthifolius), Briton's wild petunia (Ruellia simplex), Broad leaved helleborine (Epipactis helleborine), Buffelgrass (Cenchrus ciliaris), Bulbous bluegrass (Poa bulbosa), Bull thistle (Cirsium vulgare), Bur

We developed habitat suitability models for invasive plant species selected by Department of Interior land management agencies. We applied the modeling workflow developed in Young et al. 2020 to species not included in the original case studies. Our methodology balanced trade-offs between developing highly customized models for a few species versus fitting non-specific and generic models for numerous species. We developed a national library of environmental variables known to physiologically limit plant distributions (Engelstad et al. 2022 Table S1: https://doi.org/10.1371/journal.pone.0263056) and relied on human input based on natural history knowledge to further narrow the variable set for each species before developing habitat suitability models. We developed models using five algorithms with VisTrails: Software for Assisted Habitat Modeling [SAHM 2.1.2]. We accounted for uncertainty related to sampling bias by using two alternative sources of background samples, and constructed model ensembles using the 10 models for each species (five algorithms by two background methods) for three different thresholds (conservative to targeted). This data bundle contains a single file of tabular summaries by management unit (including each species/ ensemble type combination) and a subfolder for each species that contains the merged data sets used to create models, the six raster files associated with the species, and tabular outputs including response curve data, variable importance information, and model assessment metrics. Each of the six rasters represent the following: 1) 0.01 - one percentile threshold 2) 0.1 - ten percentile threshold 3) MaxSS - maximum sensitivity plus specificity threshold 4) 0.01 - one percentile threshold with Restricted Environmental Conditions 5) 0.1 - ten percentile threshold with Restricted Environmental Conditions 6) MaxSS - maximum sensitivity plus specificity threshold with Restricted Environmental Conditions The bundle documentation files are: 1) 'INHABIT_V3_metdata.xml' (this file) which contains the project-level metadata 2) managementSummaries.csv is the tabular summaries by management unit. 3) 'mergedDataset.csv' contains the merged data set used to create the models, including location and associated environmental data, for each species. 4) XX.tif where XX is the raster type explained above (threshold; masked or not). 5) responseCurves.csv is the tabular information need to produce response curves for each predictor retained in each of the 10 models produced for each species. 6) variableImportance.csv is the tabular summaries indicating predictor importance for each of the 10 models produced for each species. 7) assessmentMetrics.csv is the tabular summaries of assessment metrics for each model or ensemble for each species. These data will be integrated into the third version of the Invasive Species Habitat Tool (INHABIT), a web application displaying visual and statistical summaries of nationwide habitat suitability models for manager identified invasive plant species. These species include: African rue (Peganum harmala), Air potato (Dioscorea bulbifera), Alkali swainsonpea (Sphaerophysa salsula), Amur honeysuckle (Lonicera maackii), Amur maple (Acer ginnala), Amur peppervine (Ampelopsis brevipedunculata), Annual bluegrass (Poa annua), Annual rye (Lolium multiflorum), Asian mustard (Brassica tournefortii), Autumn olive (Elaeagnus umbellata), Balloon vine (Cardiospermum halicacabum), Beefsteak mint (Perilla frutescens), Bermudagrass (Cynodon dactylon), Bigleaf periwinkle (Vinca major), Bird vetch (Vicia cracca), Bishop's goutweed (Aegopodium podagraria), Black henbane (Hyoscyamus niger), Bohemian knotweed (Fallopia bohemica), Bradford pear (Pyrus calleryana), Brazilian peppertree (Schinus terebinthifolius), Briton's wild petunia (Ruellia simplex), Broad leaved helleborine (Epipactis helleborine), Buffelgrass (Cenchrus ciliaris), Bulbous bluegrass (Poa bulbosa), Bull thistle (Cirsium vulgare), Bur

We developed habitat suitability models for four invasive plant species of concern to Department of Interior land management agencies. We generally followed the modeling workflow developed in Young et al. 2020, but developed models both for two data types, where species were present and where they were abundant. We developed models using five algorithms with VisTrails: Software for Assisted Habitat Modeling [SAHM 2.1.2]. We accounted for uncertainty related to sampling bias by using two alternative sources of background samples, and constructed model ensembles using the 10 models for each species (five algorithms by two background methods) for four different thresholds. This data bundle contains the presence and abundance merged data sets to create models for medusahead rye, red brome, venanata and bur buttercup, the eight raster files associated with each species/ data type (presence or abundance), and tabular summaries by management unit (including each species/ data type combination). The spatial data are organized in a separate folder for each species, each containing four rasters. Each of the rasters represent the following, with an occurrence (occ) and abundance (abund) version: 1) 1st - one percentile threshold 2) 1st_masked - one percentile threshold with Restricted Environmental Conditions This file specifically, 2) 'mergedDataset.csv', contains the merged data set used to create the models, including location coordinates and associated environmental covariate data values. The bundle documentation files are: 1) 'AbundOccur.xml' contains FGDC project-level metadata 2) 'mergedDataset.csv', which this metadata file specifically describes, contains the merged data set used to create the models, including location and environmental data. 3) XX.tif where XX is the raster type explained above (occ or abund; masked or not). 4) managementSummaries.csv is the tabular summaries by management unit.