Calving season habitat selection of maternal and non-maternal female moose in Southwest Alaska, U.S.A
Data files
Oct 18, 2024 version files 29.47 GB
-
2019_NuyakukRiver_CodeKey.csv
5.74 KB
-
2019_NuyakukRiver_Cover.csv
426.95 KB
-
2019_NuyakukRiver_Heights.csv
15.06 KB
-
2019_NuyakukRiver_Sites.csv
9.29 KB
-
Covariate_alnus.zip
334.96 MB
-
Covariate_betshr.zip
575.67 MB
-
Covariate_dectre.zip
197.34 MB
-
Covariate_Elevation.zip
658.11 MB
-
Covariate_empnig.zip
598.99 MB
-
Covariate_erivag.zip
514.63 MB
-
Covariate_ForestEdge.zip
2.10 GB
-
Covariate_picea.zip
140.21 MB
-
Covariate_rhoshr.zip
565.09 MB
-
Covariate_Roughness.zip
409.67 MB
-
Covariate_salshr.zip
765 MB
-
Covariate_sphagn.zip
605.67 MB
-
Covariate_TundraEdge.zip
1.52 GB
-
Covariate_vaculi.zip
580.10 MB
-
Covariate_vacvit.zip
435.81 MB
-
Covariate_wetsed.zip
338.85 MB
-
Input_Barren.zip
9.52 MB
-
Input_dryas.zip
144.31 MB
-
Input_Lake.zip
12.19 MB
-
Input_picgla.zip
221.68 MB
-
Input_picmar.zip
42.18 MB
-
Output_Maternal_Selection.zip
4.78 GB
-
Output_Maternal_Significance.zip
71.15 MB
-
Output_Maternal_Uncertainty.zip
4.60 GB
-
Output_Nonmaternal_Selection.zip
4.68 GB
-
Output_Nonmaternal_Significance.zip
71.48 MB
-
Output_Nonmaternal_Uncertainty.zip
4.50 GB
-
README.md
22.32 KB
-
README.txt
21.24 KB
-
SouthwestAlaska_StudyArea.zip
1.66 MB
-
VegSurvey_BristolBay_HabitatExtract.csv
12.81 KB
Abstract
We predicted habitat selection of maternal and non-maternal female moose according to a use-available design with animal paths as the sampling unit. To make predictions while assessing the uncertainty from randomness in our representation of available habitat, we created a meta-model for each group from 50 independent random forest models each trained on a different set of available paths (but with the same used paths). We trained all models relative to 16 raster covariates representing topographic and plant community characteristics that we expected to be important to moose during calving season. All covariate rasters are provided in this data archive. The resulting habitat selection models successfully distinguished between used and available paths with cross-validation accuracies of 88% for non-maternal moose and 77% for maternal moose. In addition to habitat selection rasters, we provide uncertainty rasters as the per pixel 95% confidence interval width and binary significance rasters where pixels were significant (value = 1) if the 95% confidence interval did not cross 0 (i.e., from selection to avoidance or vice versa). The spatial prediction rasters enabled us to interpret selection patterns relative to landscape structure.
README: Calving season habitat selection of maternal and non-maternal female moose in Southwest Alaska, U.S.A.
This README.txt file was generated on 2022-07-22 by Timm W. Nawrocki and updated on 2024-10-18 by Timm W. Nawrocki.
I. GENERAL INFORMATION
1. Title of Dataset:
Calving season habitat selection of maternal and non-maternal female moose in Southwest Alaska, U.S.A.
2. Author Information
A. Principal Author(s) and Contact Information:
Amanda Droghini & Timm W. Nawrocki
Alaska Center for Conservation Science, University of Alaska Anchorage
3211 Providence Dr., Anchorage, AK 99508
twnawrocki@alaska.edu
B. Co-investigator(s):
Jeffrey Stetz, Alaska Department of Fish and Game
Paul Schuette, U.S. Fish and Wildlife Service
Andrew (Andy) Aderman, U.S. Fish and Wildlife Service
Kassidy Colson, Alaska Department of Fish and Game
3. Date or date range of data coverage:
2015-06-01 to 2021-10-01
4. Geographic location of data:
Bristol Bay region of Southwest Alaska, U.S.A.
5. Abstract
We predicted habitat selection of maternal and non-maternal female moose according to a use-available design with animal paths as the sampling unit. To make predictions while assessing the uncertainty from randomness in our representation of available habitat, we created a meta-model for each group from 50 independent random forest models each trained on a different set of available paths (but with the same used paths). We trained all models relative to 16 raster covariates representing topographic and plant community characteristics that we expected to be important to moose during calving season. All covariate rasters are provided in this data archive. The resulting habitat selection models successfully distinguished between used and available paths with cross-validation accuracies of 88% for non-maternal moose and 77% for maternal moose. In addition to habitat selection rasters, we provide uncertainty rasters as the per pixel 95% confidence interval width and binary significance rasters where pixels were significant (value = 1) if the 95% confidence interval did not cross 0 (i.e., from selection to avoidance or vice versa). The spatial prediction rasters enabled us to interpret selection patterns relative to landscape structure.
6. About the data:
The continuous foliar cover datasets included in this data package are derived from the Continuous Foliar Cover Maps of Plant Species and Aggregates in North American Beringia Version 1.0 (https://doi.org/10.5281/zenodo.4770218). Alpine barrens were derived from the National Land Cover Database 2016 (https://www.mrlc.gov/data/nlcd-2016-land-cover-alaska). A lake raster was generated from lake features in the National Hydrography Dataset (version 2021; https://www.usgs.gov/national-hydrography/national-hydrography-dataset). The topographic datasets included in this data package are derived from the USGS 3DEP 5m IFSAR DEM (https://gis.data.alaska.gov/pages/elevation). We extracted all input and covariate rasters to our Southwest Alaska study area. Vegetation cover and height data were collected in the vicinity of the Nuyakuk and Nushagak rivers in 2019. For maternal and non-maternal females, we provide the following model outputs in the form of raster datasets: habitat selection, uncertainty (95% confidence interval width), and binary significance. Detailed information on the processing of all provided data is available through the code repository linked in the code archive (). Geospatial data are provided as TIF single band rasters with 10 x 10 m resolution projected into Alaska Albers Equal Area Conic (EPSG 3338). The data package contains optional .lyrx files for visualization of model output rasters in ArcGIS Pro.
7. Acknowledgements and funding sources:
Alaska Department of Fish and Game provided funding to collect vegetation observation data in Southwest Alaska in 2019, develop continuous foliar cover maps of plant species and aggregates, and conduct analysis of moose calving season habitat selection. We thank Kassidy Colson and Dr. Jeffrey Stetz for their support of this work. In kind support was provided by U.S. Fish and Wildlife Service. Kassidy Colson and Andy Aderman provided GPS and VHF collar data for female moose in the Bristol Bay region for use in these analyses.
8. Recommended citation for this dataset is from the associated manuscript:
Droghini, A., T.W. Nawrocki, J.B. Stetz, P.A. Schuette, A.R. Aderman, and K.E. Colson. 2024. Variation in habitat selection among individuals differs by maternal status for moose in a region with low calf survival. Ecosphere [volume:identifier].
Alternatively, you can cite this data archive directly as follows:
Droghini, A., T.W. Nawrocki, J.B. Stetz, P.A. Schuette, A.R. Aderman, and K.E. Colson. 2024. Calving season habitat selection of maternal and non-maternal female moose in Southwest Alaska, U.S.A. Archived data. Data Dryad.
9. Version information
A. Release date:
2024-10-02
B. Versioning:
The data package is not a versioned release. The continuous foliar cover raster vegetation maps included in this data package are version 1.0.
C. Anticipated Update Cycle:
As-needed (no updates planned).
D. Explanation:
The calving season habitat selection included in this data package will not be updated. Continuous foliar cover vegetation maps for North American Beringia will be updated regularly separately from this data package. Users interested in the continuous foliar cover maps for new analyses should acquire the most current version.
II. METHODOLOGICAL INFORMATION:
1. Licenses/restrictions placed on the data:
None
2. Usage Instructions:
Raster TIFF data can be viewed, analyzed, and manipulated in any GIS, including ArcGIS, QGIS, R, and Python. We have provided lyrx files that store visualization information for rasters that can be imported into ArcGIS Pro. For use in other software, rasters must be imported with user-defined visualizations.
3. Software/instruments used to generate data:
We conducted spatial processing using ArcGIS Pro 2.8.1 with Python 3.7.10; data formatting, explanatory statistical modeling, and prediction post-processing using R 4.1.1 and RStudio 1.4.1717; and predictive statistical modeling in the Anaconda 2021.05 distribution of Python 3.8.8. We used Scikit-learn 0.24.2 (Pedregosa et al. 2011) and the following R packages: amt 0.1.4 (Signer et al. 2021), circular 0.4–93 (Lund et al. 2017), ctmm 0.6.0 (Fleming et al. 2021), data.table 1.14.0 (Dowle et al. 2021), geosphere 1.5–10 (Hijmans et al. 2019), MASS 7.3–54 (Ripley et al. 2021), move 4.0.6 (Kranstauber et al. 2020), raster 3.4–13 (Hijmans et al. 2021), rgdal 1.5–23 (Bivand et al. 2021), sf 1.0–2 (Pebesma et al. 2021a), sp 1.4–5 (Pebesma et al. 2021b), survival 3.2–11 (Therneau et al. 2021), tidyverse 1.3.1 (Wickham et al. 2019), and tlocoh 1.40.07 (Lyons et al. 2013). All scripts used to generate results for this study are available: https://github.com/accs-uaa/swak-moose-calving.
4. Description of methods used to collect or generate data:
See code archive: https://doi.org/10.5281/zenodo.13883436
5. Describe any quality-assurance procedures performed on the data:
See code archive: https://doi.org/10.5281/zenodo.13883436
6. Link to Code Repository:
https://doi.org/10.5281/zenodo.13883436
7. Links to ancillary data sets:
None
III. DATA & FILE OVERVIEW
1. Primary Data Formats:
tif -- raster geospatial data files
csv -- tabular data files
2. File List:
A. Data archive contents:
- SouthwestAlaska_StudyArea.zip -- Zip archive containing raster study area
- Input_Barren.zip --- Zip archive containing alpine barren raster
- Input_Lake.zip -- Zip archive containing lake raster
- Input_dryas.zip -- Zip archive containing percentage continuous foliar cover of Dryas shrubs
- Input_picgla.zip -- Zip archive containing percentage continuous foliar cover of white spruce (Picea glauca)
- Input_picmar.zip -- Zip archive containing percentage continuous foliar cover of black spruce (Picea mariana)
- Covariate_alnus.zip -- Zip archive containing percentage continuous foliar cover of alder (Alnus)
- Covariate_betshr.zip -- Zip archive containing percentage continuous foliar cover of birch (Betula) shrubs
- Covariate_dectre.zip -- Zip archive containing percentage continuous foliar cover of deciduous trees
- Covariate_empnig.zip -- Zip archive containing percentage continuous foliar cover of crowberry (Empetrum nigrum)
- Covariate_erivag.zip -- Zip archive containing percentage continuous foliar cover of tussock cottongrass (Eriophorum vaginatum)
- Covariate_picea.zip -- Zip archive containing percentage continuous foliar cover of spruce (Picea)
- Covariate_rhoshr.zip -- Zip archive containing percentage continuous foliar cover of Rhododendron shrubs
- Covariate_salshr.zip -- Zip archive containing percentage continuous foliar cover of non-dwarf willow (Salix) shrubs
- Covariate_sphagn.zip -- Zip archive containing percentage continuous foliar cover of Sphagnum mosses
- Covariate_vaculi.zip -- Zip archive containing percentage continuous foliar cover of bog blueberry (Vaccinium uliginosum)
- Covariate_vacvit.zip -- Zip archive containing percentage continuous foliar cover of lingonberry (Vaccinium vitis-idaea)
- Covariate_wetsed.zip -- Zip archive containing percentage continuous foliar cover of wetland sedges
- Covariate_ForestEdge.zip -- Zip archive containing distance to forest edge (m)
- Covariate_TundraEdge.zip -- Zip archive containing distance to tundra edge (m)
- Covariate_Elevation.zip -- Zip archive containing elevation (m)
- Covariate_Roughness.zip -- Zip archive containing roughness (m)
- 2019_NuyakukRiver_Sites.csv -- csv table of vegetation survey sites along Nuyakuk and Nushagak rivers visited in 2019
- 2019_NuyakukRiver_Cover.csv -- csv table of vegetation survey point-intercept results from surveys along Nuyakuk and Nushagak rivers conducted in 2019
- 2019_NuyakukRiver_Heights.csv -- csv table of vegetation heights recorded during surveys along Nuyakuk and Nushagak rivers conducted in 2019
- 2019_NuyakukRiver_CodeKey.csv -- csv table containing plant species and soil surface codes and corresponding values used in the table 2019_NuyakukRiver_Cover.csv
- VegSurvey_BristolBay_HabitatExtract.csv -- csv table of calving season habitat selection for maternal and non-maternal moose extracted to 2019 survey sites
- Output_Nonmaternal_Selection.zip -- Zip archive containing the calving season habitat selection raster for non-maternal moose
- Output_Nonmaternal_Uncertainty.zip -- Zip archive containing the calving season habitat selection uncertainty raster for non-maternal moose as the 95% confidence interval width across 50 model iterations
- Output_Nonmaternal_Significance.zip -- Zip archive containing the calving season habitat selection binary significance raster for non-maternal moose where significant pixels are those where the 95% confidence interval does not cross zero on the selection scale
- Output_Maternal_Selection.zip -- Zip archive containing the calving season habitat selection raster for maternal moose
- Output_Maternal_Uncertainty.zip -- Zip archive containing the calving season habitat selection uncertainty raster for maternal moose as the 95% confidence interval width across 50 model iterations
- Output_Maternal_Significance.zip -- Zip archive containing the calving season habitat selection binary significance raster for maternal moose where significant pixels are those where the 95% confidence interval does not cross zero on the selection scale
- README.txt -- Plain text readme file.
B. Zip archive contents (per archive):
Raster files are 10 x 10 m resolution TIFF format files (.tif and associated files) in Alaska Albers Equal Area Conic (EPSG 3338). One raster (.tif) is included per zip archive. All additional files are associated with the raster file and should not be interacted with independently.
IV. DATA INFORMATION:
1. Raster geospatial data:
A. Continuous foliar cover raster datasets (.tif):
i. Purpose: Provide spatial representation of foliar cover for mapped species or aggregate.
ii. Spatial Reference (EPSG): 3338
iii. Resolution: 10 x 10 m
iv. Field definitions: VALUE -- percentage foliar cover.
v. Missing data codes: -32768
B. Distance to edge raster datasets (.tif)
i. Purpose:Provide spatial representation of distance to tundra edge or forest edge.
ii. Spatial Reference (EPSG): 3338
iii. Resolution: 10 x 10 m
iv. Field definitions: VALUE -- distance in meters (m)
v. Missing data codes: -32768
C. Topographic datasets (.tif):
i. Purpose:Provide spatial representation of topographic characteristics (elevation and roughness). Roughness is the standard deviation of elevation within a 5 x 5 cell moving window.
ii. Spatial Reference (EPSG):3338
iii. Resolution: 10 x 10 m
iv. Field definitions: VALUE -- meters (m)
v. Missing data codes: -32768
D. Barren dataset (.tif)
i. Purpose: Provide spatial representation of alpine barrens. Includes the barren class from NLCD 2016 restricted to elevations over 300 m.
ii. Spatial Reference (EPSG): 3338
iii. Resolution: 10 x 10 m
iv. Field definitions: VALUE -- 50 = barren
v. Missing data codes: -32768
E. Lake dataset (.tif)
i. Purpose: Provide spatial representation of lakes. Includes lakes features from the NHD (version 2021).
ii. Spatial Reference (EPSG): 3338
iii. Resolution: 10 x 10 m
iv. Field definitions: VALUE -- 0 = non-lake; 1 = lake
v. Missing data codes: -128
F. Habitat selection datasets (.tif)
i. Purpose: Provide spatial representation of calving season habitat selection for maternal and non-maternal female moose.
ii. Spatial Reference (EPSG): 3338
iii. Resolution: 10 x 10 m
iv. Field definitions: VALUE -- relative scale of habitat selection where values greater than 0 indicate selection and values less than 0 indicate avoidance
v. Missing data codes: -3.4e38
G. Uncertainty datasets (.tif)
i. Purpose: Provide spatial representation of model uncertainty caused by random representation of available habitat.
ii. Spatial Reference (EPSG): 3338
iii. Resolution: 10 x 10 m
iv. Field definitions: VALUE -- 95% confidence interval width along relative scale of habitat selection
v. Missing data codes: -3.4e38
H. Binary significance datasets (.tif)
i. Purpose: Provide spatial representation of model significance.
ii. Spatial Reference (EPSG): 3338
iii. Resolution: 10 x 10 m
iv. Field definitions: VALUE -- 0 = non-significant, 1 = significant
v. Missing data codes: 3
2. Tabular data:
A. 2019_NuyakukRiver_Sites.csv
:
i. Purpose:
Provide site metadata and location for 2019 survey sites along the Nuyakuk and Nushagak rivers used to assess willow heights and diversity relative to moose habitat selection.
ii. Field definitions:
site -- Alphanumeric code that uniquely identifies site
observer -- First and last name of the vegetation observer
recorder -- First and last name of the vegetation recorder
date -- Date of observation in m/d/yyyy format
latitude -- site midpoint latitude in decimal degrees
longitude -- site midpoint longitude in decimal degrees
datum -- Horizontal datum of site coordinates
physiography -- Descriptor of site physiography
strata -- Descriptor of site vegetation strata
geomorphology -- Descriptor of site geomorphology
macrotopography -- Descriptor of site macrotopography
microtopography -- Descriptor of site microtopography
microrelief_cm -- Microrelief distance in cm
drainage -- Descriptor of site drainage regime
moisture -- Descriptor of site moisture regime
waterdepth_cm -- Depth of water in cm with positive values indicating water above soil surface and negative values indicating water below soil surface
iii. Missing data codes:
Blank
B. 2019_NuyakukRiver_Cover.csv
:
i. Purpose:
Provide plant community composition data for 2019 survey sites along the Nuyakuk and Nushagak rivers used to assess willow heights and diversity relative to moose habitat selection.
ii. Field definitions:
site -- Alphanumeric code that uniquely identifies site
line -- Line number within the plot
point -- Point number within the line
layer1 -- First plant/soil intercepted (codes correspond to values in 2019_NuyakukRiver_CodeKey.csv)
layer2-- Second plant/soil intercepted (codes correspond to values in 2019_NuyakukRiver_CodeKey.csv)
layer3-- Third plant/soil intercepted (codes correspond to values in 2019_NuyakukRiver_CodeKey.csv)
layer4-- Fourth plant/soil intercepted (codes correspond to values in 2019_NuyakukRiver_CodeKey.csv)
layer5-- Fifth plant/soil intercepted (codes correspond to values in 2019_NuyakukRiver_CodeKey.csv)
layer6-- Sixth plant/soil intercepted (codes correspond to values in 2019_NuyakukRiver_CodeKey.csv)
layer7-- Seventh plant/soil intercepted (codes correspond to values in 2019_NuyakukRiver_CodeKey.csv)
layer8-- Eighth plant/soil intercepted (codes correspond to values in 2019_NuyakukRiver_CodeKey.csv)
layer9-- Ninth plant/soil intercepted (codes correspond to values in 2019_NuyakukRiver_CodeKey.csv)
layer10-- Tenth plant/soil intercepted (codes correspond to values in 2019_NuyakukRiver_CodeKey.csv)
iii. Missing data codes:
Blank (table cells that are blank indicate that the layer indicated by the corresponding field name did not occur at the row line & point)
C. 2019_NuyakukRiver_Heights.csv
:
i. Purpose:
Provide plant height data for 2019 survey sites along the Nuyakuk and Nushagak rivers used to assess willow heights and diversity relative to moose habitat selection.
ii. Field definitions:
site -- Alphanumeric code that uniquely identifies site
taxon -- Scientific name of the recorded species or infraspecies
height_cm -- 98th percentile height within the plot of the recorded species or infraspecies in cm
iii. Missing data codes:
N/A
D. 2019_NuyakukRiver_CodeKey.csv
:
i. Purpose:
Provide plant and lichen species and soil surface codes and corresponding values used in fields layer1, layer2, layer3, layer4, layer5, layer6, layer7, layer8, layer9, and layer10 in the table 2019_NuyakukRiver_Cover.csv.
ii. Field definitions:
code -- Code value that identifies the plant or lichen species or soil surface observed.
value -- Scientific name of the recorded plant or lichen species or infraspecies or soil surface description.
iii. Missing data codes:
N/A
E. VegSurvey_BristolBay_HabitatExtract.csv
i. Purpose:
Provide extracted habitat selection predictions for sites along the Nuyakuk and Nushagak rivers used to assess willow heights and diversity relative to moose habitat selection
ii. Field definitions:
OBJECTID -- Sequential number that uniquely identifies the site within this csv table
site_id -- Sequential number that uniquely identifies the site within all sites used to develop the continuous foliar cover maps version 1.0
site_code -- Alphanumeric code that uniquely identifies site
initial_project -- Name of the project during which the first visit to the site was made
perspective -- Records whether the data was recorded from a ground or aerial perspective
cover_method -- Records the method used to determine plant community composition data
scope_vascular -- Records the scope of vascular plant identifications
scope_bryophyte -- Records the scope of bryophyte identifications
scope_lichen -- Records the scope of lichen identifications
plot_dimensions -- Records the dimensions of the plots in meters (m)
datum -- Horizontal datum of site coordinates
latitude -- site midpoint latitude in decimal degrees
longitude -- site midpoint longitude in decimal degrees
error -- GPS positional error in meters (m)
preference_maternal -- Predicted habitat selection of site for maternal females (unitless)
preference_nonmaternal -- Predicted habitat selection of site for non-maternal females (unitless)
uncertainty_maternal -- Uncertainty in model prediction at site for maternal females (unitless)
uncertainty_nonmaternal -- Uncertainty in model prediction at site for non-maternal females (unitless)
significance_maternal -- Binary significance (based on 95% confidence interval) of model prediction at site for maternal females
significance_nonmaternal -- Binary significance (based on 95% confidence interval) of model prediction at site for non-maternal females
iii. Missing data codes:
Blank
Methods
The continuous foliar cover datasets included in this data package are derived from the Continuous Foliar Cover Maps of Plant Species and Aggregates in North American Beringia Version 1.0 (https://doi.org/10.5281/zenodo.4770218). Alpine barrens were derived from the National Land Cover Database 2016 (https://www.mrlc.gov/data/nlcd-2016-land-cover-alaska). A lake raster was generated from lake features in the National Hydrography Dataset (version 2021; https://www.usgs.gov/national-hydrography/national-hydrography-dataset). The topographic datasets included in this data package are derived from the USGS 3DEP 5m IFSAR DEM (https://gis.data.alaska.gov/pages/elevation). We extracted all input and covariate rasters to our Southwest Alaska study area. Vegetation cover and height data were collected in the vicinity of the Nuyakuk and Nushagak rivers in 2019. For maternal and non-maternal females, we provide the following model outputs in the form of raster datasets: habitat selection, uncertainty (95% confidence interval width), and binary significance. Detailed information on the processing of all provided data is available through the code repository linked in the code archive (). Geospatial data are provided as TIF single band rasters with 10 x 10 m resolution projected into Alaska Albers Equal Area Conic (EPSG 3338). The data package contains optional .lyrx files for visualization of model output rasters in ArcGIS Pro.
Usage notes
All provided raster datasets can be opened in a GIS or in R or Python. Files ending in .xlsx or .csv can be opened in Microsoft Excel, R, Python, OpenOffice, or other programs.