Data from: Trajectories and drivers of compositional and functional variation in sagebrush steppe restoration
Data files
Jun 23, 2026 version files 127.42 KB
-
allvariables_ESO-25-09-229.csv
47.15 KB
-
functionaltraits_ESO-25-09-229.csv
8.03 KB
-
README.md
15.69 KB
-
speciesabund_ESO-25-09-229.csv
56.55 KB
Abstract
Restoration outcomes are often variable and can fall short of targets due to poorly understood factors, especially in arid ecosystems. Identifying the drivers of variation during community assembly is critical for achieving restoration outcomes. Community trajectories and levels of recovery could be contingent on conditions at the time of establishment, a point at which practitioners may have the most control, though we currently lack a predictive understanding of these factors. Furthermore, it has been proposed that functional traits will respond more predictably than compositional metrics and could improve forecasts of restoration outcomes. Here, we use NMDS ordination and mixed effects models to evaluate sagebrush steppe restoration trajectories across 13 sites and 11 years. We compare the relative importance of time since restoration (age) and establishment conditions (soil, seed mix, establishment-year climate) to variation in restoration outcomes. Most compositional and trait metrics responded predictably through time, with sites converging with each other towards reference conditions despite differences at establishment. However, communities fell short of or overshot reference targets after 11 years, reaching a maximum 15 % compositional similarity and 69 % functional similarity compared to reference sites. Restored communities were dominated by native and exotic graminoid species, corresponding with a conservative and drought-tolerant composition of species with high leaf dry matter content and low turgor loss point. Restoration outcomes and convergence toward reference sites were contingent upon establishment year climate (year effects), soil sand content (site effects), and, to a lesser degree, seed mix design. These establishment effects on restoration trajectories often persisted or increased in strength over 11 years. Practical implication: Full restoration may require many more decades, but practitioners could speed success and increase control over outcomes by designing soil-specific seed mixes and scheduling seed sowing or post-seeding management to match short and long-term climate forecasts. When restoration efforts fall short of targets or lead to undesirable stable states, practitioners may improve recovery through actions such as supplemental targeted planting of desired shrub and forb species to redirect successional trajectories. Seed mixes with a higher proportion of forbs and shrubs to graminoids should be further tested.
Dataset DOI: 10.5061/dryad.cjsxksnnt
Description of the data and file structure
This dataset contains community-level compositional and functional trait variables from nine reference and 13 restored sagebrush steppe plant communities in Grand Teton National Park measured at multiple time intervals since restoration (restoration ages) from 2011 to 2020. Restoration sites were planted in eight different years. Restored sites were monitored using the line-point intercept method (LPI) at 1 m intervals along permanently marked 50 m transects. Transect data were converted to percent cover by summing species hits across canopy layers, dividing by the total intervals (50), and multiplying by 100 (Stefano et al., 2021). The nearby reference sites were monitored annually using a nested random 1 m2 quadrat/0.1 km2 permanent frame design, following pre-existing National Park Service Inventory and Monitoring (I&M) protocols (Jean et al., 2014, Yeo et al., 2009). Daubenmire cover classes (1 = 0-5 %, 2 = 5-25 %, 3 = 25-50 %, 4 = 50-75 %, 5 = 75-95 %, 6 = 95-100 %) were converted to percent cover estimates by taking the midpoint of each class per species (Hulvey et al., 2018). To ensure that methodologies were comparable, we analyzed LPI and quadrat data using nine I&M frames on which both methods were used and found that differences should underestimate overall restoration success (i.e. a conservative estimate). Percent cover estimates were then used to calculate compositional and functional metrics of management interest at the transect level. We measured five functional traits on the 70 most common species representing > 99 % of total species abundances across all sites following standardized protocols on mature and healthy individuals which were rehydrated to full turgor. We then calculated community-weighted means (CWMs), or the mean trait value weighted by species relative abundances, and multivariate functional dispersion, or the mean distance of all species to the weighted community centroid in trait space, to represent functional responses at each site.
Separate attached datasets contain the underlying species composition data as percent cover by species and the underlying functional trait data as the average value for each trait by species. Also included are variables characterizing establishment conditions at each restored site, including seed mix design and soil conditions. Seed mix design data were provided by Grand Teton National Park and soil variables were derived from cores taken at each site. Data on climatic variation between establishment years (average planting year snow depth and first growing season temperature) were sourced from the NOAA National Climatic Data Center (https://www.ncdc.noaa.gov/cdo-web/) station in Moose, WY (43°39'12.9" N 110°43'00.5" W, GHCND:USC00486428) and are not included here for copyright reasons. Similarities between each restored site and reference site and between all restored sites were calculated on untransformed species abundances using the Bray-Curtis dissimilarity (0-1 scale) and on scaled and centered CWM traits using Euclidean distance and then converted to similarities by subtracting each dissimilarity value from 1. Data were used to assess restoration success, restoration trajectories, and the influence of establishment conditions on compositional and functional outcomes using non-metric multidimensional scaling (NMDS) ordination and generalized linear mixed effects models (GLMMs).
Files and variables
File: allvariables_ESO-25-09-229.csv
Description: All final compositional and functional variables calculated from species abundances and species-level trait values recorded in restored and reference sagebrush steppe communities in Grand Teton National Park. Missing or inapplicable values are indicated by NA.
Variables
- site_type: Indicates whether the site is a restored unit or a reference unit.
- subunit: Site name as indicated by Grand Teton National Park.
- sample_year: Year that the community data was recorded.
- rest_age: Number of years since the site was planted.
- tranID: Unique identifier composed of the unit code-plot-transect. Only applicable to restored sites which were monitored with the line-point-intercept method.
- tran: Unique identifier composed of the plot-transect.
- unitcode: Unique identifier corresponding to the subunit.
- seeding_year: Year that the restoration site was planted.
- Plot: Unique identifier corresponding to a specific section of a subunit within which transects were grouped.
- Transect: Unique identifier corresponding to the individual transect. Only applicable to restored sites which were monitored with the line-point-intercept method.
- Frame: Unique identifier corresponding to each 0.1 km2 permanent frame. Only applicable to reference sites which were monitored using a nested random 1 m2 quadrat/0.1 km2 permanent frame design, following pre-existing National Park Service Inventory and Monitoring (I&M) protocols.
- NMDS1: Values extracted from Axis 1 of a non-metric multidimensional scaling (NMDS) ordination, as a univariate metric representing one dimension of plant community composition.
- NMDS2: Values extracted from Axis 2 of a non-metric multidimensional scaling (NMDS) ordination, as a univariate metric representing one dimension of plant community composition.
- seededrich: The number of species present in the plant community that were included in each restored site’s unique seed mix.
- volrich: The number of species present in the plant community that were not included in each restored site’s unique seed mix.
- gram_prop: The proportion (0-1) of graminoid to shrub and forb cover in the community.
- HII: The exotic proportion of only graminoid and forb cover (Wainwright et al., 2020).
- CWM.SLA_log: Community-weighted mean (CWM), or the mean trait value weighted by species relative abundances, for specific leaf area (SLA, mm2 one-sided area mg -1 dry mass), log-transformed to normalize model residuals.
- CWM.LDMC: Community-weighted mean (CWM), or the mean trait value weighted by species relative abundances, for leaf dry matter content (LDMC, mg dry mass mg -1 fresh mass).
- CWM.TLP: Community-weighted mean (CWM), or the mean trait value weighted by species relative abundances, for leaf turgor loss point (TLP, MPa).
- CWM.Height: Community-weighted mean (CWM), or the mean trait value weighted by species relative abundances, for mature vegetative height (m).
- CWM.Seed_Mass: Community-weighted mean (CWM), or the mean trait value weighted by species relative abundances, for oven dry seed mass (seed mass, mg).
- FDis: Multivariate functional dispersion, or the mean distance of all species to the weighted community centroid in trait space, as a metric of functional diversity.
- sand: Percent sand content from soil texture analysis of 6 cm deep soil cores collected along each transect or within each reference site frame, representing a soil texture gradient.
- OM: Percent organic matter from analysis of 6 cm deep soil cores collected along each transect or within each reference site frame, to represent a pH and fertility gradient.
- seed_gf_rat: The ratio of graminoid pure live seed (PLS) weights (lbs/lb) to the combined weights of shrubs and forbs (PLS).
- trint_sim: Similarity of functional trait compositions to intact reference sites, based on scaled and centered CWM traits using Euclidean distance which we then scaled to 0-1 by dividing by the maximum distance value. Dissimilarities were then converted to similarities to improve interpretation by subtracting each dissimilarity value from 1.
- trrest_sim: Similarity of functional trait compositions between restored sites, based on scaled and centered CWM traits using Euclidean distance which we then scaled to 0-1 by dividing by the maximum distance value. Dissimilarities were then converted to similarities to improve interpretation by subtracting each dissimilarity value from 1.
- sprest_sim: Similarity of species compositions between restored sites, based on the Bray-Curtis dissimilarity (0-1 scale). Dissimilarities were then converted to similarities to improve interpretation by subtracting each dissimilarity value from 1.
- spint_sim: Similarity of species compositions to intact reference sites, based on the Bray-Curtis dissimilarity (0-1 scale). Dissimilarities were then converted to similarities to improve interpretation by subtracting each dissimilarity value from 1.
File: speciesabund_ESO-25-09-229.csv
Description: Raw species composition data as species percent cover within transects or frames for restored and reference sagebrush steppe sites in Grand Teton National Park. Missing or inapplicable values are indicated by NA.
Variables
- site_type: Indicates whether the site is a restored unit or a reference unit.
- subunit: Site name as indicated by Grand Teton National Park.
- sample_year: Year that the community data was recorded.
- rest_age: Number of years since the site was planted.
- tranID: Unique identifier composed of the unit code-plot-transect. Only applicable to restored sites which were monitored with the line-point-intercept method.
- tran: Unique identifier composed of the plot-transect.
- unitcode: Unique identifier corresponding to the subunit.
- seeding_year: Year that the restoration site was planted.
- Plot: Unique identifier corresponding to a specific section of a subunit within which transects were grouped.
- Transect: Unique identifier corresponding to the individual transect. Only applicable to restored sites which were monitored with the line-point-intercept method.
- Frame: Unique identifier corresponding to each 0.1 km2 permanent frame. Only applicable to reference sites which were monitored using a nested random 1 m2 quadrat/0.1 km2 permanent frame design, following pre-existing National Park Service Inventory and Monitoring (I&M) protocols.
- sample_type: Monitoring protocol use to collect species composition data.
- ACHLET: Achnatherum lettermanii percent cover
- ACHMIL: Achillea millefolium percent cover
- ACHNEL: Achnatherum nelsonii percent cover
- AGOSPP: Agoseris spp. percent cover
- ALYSPP: Alyssum spp. percent cover
- ANTSPP: Antennaria spp. percent cover
- ARASPP: Arabis spp. percent cover
- ARESPP: Arenaria congesta percent cover
- ARTARB: Artemisia arbuscula percent cover
- ARTTRI: Artemisia tridentata percent cover
- ARTTRIP: Artemisia tripartita percent cover
- ASTRSPP: Astragalus spp. percent cover
- BALSAG: Balsamorhiza sagittata percent cover
- BROINE: Bromus inermis percent cover
- BROMAR: Bromus marginatus percent cover
- BROTEC: Bromus tectorum percent cover
- CAMMIC: Camelina microcarpa percent cover
- CAPBUR: Capsella bursa-pastoris percent cover
- CARNUT: Carduus nutans percent cover
- CARSPP: Carex spp. percent cover
- CASSPP: Castilleja spp. percent cover
- CHEALB: Chenopodium album percent cover
- CHRVIS: Chrysothamnus viscidiflorus percent cover
- COLLIN: Collomia linearis percent cover
- COLPAR: Collinsia parviflora percent cover
- COMUMB: Comandra umbellata percent cover
- CREACU: Crepis acuminata percent cover
- DESINC: Descurainia incana percent cover
- ELYELY: Elymus elymoides percent cover
- ELYTRA: Elymus trachycaulus percent cover
- EPISPP: Epilobium spp. percent cover
- ERISPP: Erigeron spp percent cover
- ERIUMB: Eriogonum umbellatum percent cover
- FESIDA: Festuca idahoensis percent cover
- HACSPP: Hackelia spp. percent cover
- HELMUL: Heliomeris multiflora percent cover
- HELUNI: Helianthella uniflora percent cover
- HESCOM: Hesperostipa comata percent cover
- KOEMAC: Koeleria macrantha percent cover
- LACSER: Lactuca serriola percent cover
- LACTAT: Lactuca tatarica percent cover
- LEYCIN: Leymus cinereus percent cover
- LITRUD: Lithospermum ruderale percent coveer
- LOMSPP: Lomatium spp. percent cover
- LUPSPP: Lupinus spp. percent cover
- MAHREP: Mahonia repens percent cover
- MELSPP: Melica spp. percent cover
- MICGRA: Microsteris gracilis percent cover
- MICSPP: Microseris spp. percent cover
- NEMBRE: Nemophila breviflora percent cover
- PENPRO: Penstemon procerus percent cover
- PERMON: Perideridia montanum percent cover
- PHLHOO: Phlox hoodii percent cover
- PHLLON: Phlox longifolia percent cover
- PHLPRA: Phleum pratense percent cover
- POABUL: Poa bulbosa percent cover
- POASEC: Poa secunda percent cover
- POASPP: Poa spp. percent cover
- POLDOU: Polygonum douglasii percent cover
- PSESPI: Pseudoroegneria spicata percent cover
- PURTRI: Purshia tridentata percent cover
- SISALT: Sisymbrium altissimum percent cover
- SOLSPP: Solidago spp. percent cover
- SYMASC: Symphyotrichum ascendens percent cover
- SYMORE: Symphoricarpos oreophilus percent cover
- TAROFF: Taraxacum officinale percent cover
- THLARV: Thlaspi arvense percent cover
- TRADUB: Tragopogon dubius percent cover
- VERSPP: Veronica spp. percent cover
- VIOSPP: Viola spp. percent cover
File: functionaltraits_ESO-25-09-229.csv
Description: Average values of five functional traits measured on the 70 most common species representing > 99 % of total species abundances across all sites. Together, these traits correspond to drought tolerance and major axes of plant survival strategies. Traits were measured following standardized protocols on mature and healthy individuals which were rehydrated to full turgor. Measurements were taken on 3-5 individuals per species as replicates and then averaged for final values. Missing trait data (< 2 %) were sourced from the TRY database or imputed using the ‘mice’ function in R (Kattge et al., 2011; van Buuren & Groothuis-Oudshoorn, 2011). Averaged TRY values are not included here due to copyright reasons but were used to fill the following gaps in the final version of this data (indicated here as NA): seed mass for Artemisia tripartita, seed mass for Phlox hoodii, and seed mass for Viola spp. represented by the average TRY value for Viola purpurea.
Variables
- Species_Code: Six letter code corresponding to the first three letters of the genus and first three letters of species.
- Family: Plant family that the species belongs to.
- Functional_Group: Functional group that the species belongs to.
- Species: Latin binomial for the species.
- SLA: Average specific leaf area (SLA, mm2 one-sided area mg -1 dry mass) for the species.
- LDMC: Average leaf dry matter content (LDMC, mg dry mass mg -1 fresh mass) for the species.
- TLP: Average leaf turgor loss point (TLP, MPa) for the species, representing drought tolerance based on the water potential at which a leaf loses turgor and closes its stomates. TLP was calculated by repeatedly measuring the osmolarity of four 1 mm2 leaf sections stored at -20 °C using a vapor pressure osmometer (Wescor VAPRO 5600, Logan, UT, USA) until equilibrium was reached. TLP was then estimated from osmolarity values using separate linear models developed for forbs, shrubs, and graminoids (Bartlett et al., 2012; Griffin-Nolan et al., 2019).
- Height: Average mature vegetative height (m) for the species, measured as the tallest photosynthetic tissue without any manual manipulation.
- Seed Mass: Average oven dry seed mass (seed mass, mg) for the species.
Code/software
All csv files were created in R version 4.2.0 for use in Microsoft Excel 16.101.3.
