{"dataset":"ebird","predictor":"longitude","description":"Longitude assigned to checklist. Locations are assigned using the most precise data available, in descending order of precision: checklists with tracks will use the track centroid, checklists assigned to hotspots will receive the mean centroid of all tracks at that hotspot, checklists assigned to hotspots with no other tracks will receive the hotspot location, and checklists assigned to personal locations will receive the location assigned by the observer."}
{"dataset":"ebird","predictor":"latitude","description":"Latitude assigned to checklist. Locations are assigned using the most precise data available, in descending order of precision: checklists with tracks will use the track centroid, checklists assigned to hotspots will receive the mean centroid of all tracks at that hotspot, checklists assigned to hotspots with no other tracks will receive the hotspot location, and checklists assigned to personal locations will receive the location assigned by the observer."}
{"dataset":"ebird","predictor":"year","description":"Integer year of observation."}
{"dataset":"ebird","predictor":"day_of_year","description":"Integer day of year of observation [1-366]."}
{"dataset":"ebird","predictor":"solar_noon_diff","description":"Time of day for checklist midpoint expressed as the difference in hours from solar noon."}
{"dataset":"ebird","predictor":"is_stationary","description":"Was this a stationary (TRUE) or traveling (FALSE) protocol?"}
{"dataset":"ebird","predictor":"effort_hrs","description":"Checklist duration in hours."}
{"dataset":"ebird","predictor":"effort_distance_km","description":"Distance traveled in kilometers."}
{"dataset":"ebird","predictor":"effort_speed_kmph","description":"Mean speed traveled kilometers per hour [effort_distance_km / effort_hrs]."}
{"dataset":"ebird","predictor":"num_observers","description":"Number of observers."}
{"dataset":"ebird","predictor":"cci","description":"Checklist calibration index, a measure of observer expertise specific to this checklist.","reference":"https://doi.org/10.1111/2041-210X.12838"}
{"dataset":"moon","predictor":"moon_fraction","description":"Illuminated fraction of the moon at the date and time of the observation; varies from 0.0 (new moon) to 1.0 (full moon).","reference":"suncalc R package, getMoonIllumination() function"}
{"dataset":"moon","predictor":"moon_altitude","description":"Moon altitude above the horizon in radians at the date, time, and location of the observer.","reference":"suncalc R package, getMoonPosition() function"}
{"dataset":"cds","predictor":"cds_{class}","description":"Hourly weather conditions at the date, time, and location of the observation from the ERA5 Copernicus product. Weather conditions are broken into 14 different variables, see the \"classes\" sheet for details. Note that these are local weather variables not climate variables and are therefore only expected to describe the observation processes rather than any ecological processes.","reference":"https://doi.org/10.24381/cds.adbb2d47"}
{"dataset":"eastness","predictor":"eastness_1km_{median/sd}","description":"East-west component of slope aggregated from the 1 km resolution topographic dataset described in Amatulli et al. 2018. Combines both slope and aspect. Steep slopes facing east will have high positive values, while steep slopes facing west will have high negative values.","reference":"https://doi.org/10.1038/sdata.2018.40\n\n"}
{"dataset":"eastness","predictor":"eastness_90m_{median/sd}","description":"East-west component of slope aggregated from the 90 m resolution. Combines both slope and aspect. Steep slopes facing east will have high positive values, while steep slopes facing west will have high negative values."}
{"dataset":"northness","predictor":"northness_1km_{median/sd}","description":"North-south component of slope aggregated from the 1 km resolution topographic dataset described in Amatulli et al. 2018. Combines both slope and aspect. Steep slopes facing north will have high positive values, while steep slopes facing south will have high negative values.","reference":"https://doi.org/10.1038/sdata.2018.40\n\n"}
{"dataset":"northness","predictor":"northness_90m_{median/sd}","description":"North-south component of slope aggregated from the 90 m resolution. Combines both slope and aspect. Steep slopes facing north will have high positive values, while steep slopes facing south will have high negative values."}
{"dataset":"elevation","predictor":"elevation_250m_{median/sd}","description":"Elevation and bathymetry in meters aggregated from the 250 m dataset described in Tozer et al. 2019.","reference":"https://doi.org/10.1029/2019EA000658"}
{"dataset":"elevation","predictor":"elevation_30m_{median/sd}","description":"Elevation in meters aggregated from the 30 m resolution ASTER Global Digital Elevation Model (v003) dataset.","reference":"https://doi.org/10.5067/ASTER/ASTGTM.003"}
{"dataset":"island","predictor":"island","description":"A unique integer ID for every distinct island or continent.","reference":"https://doi.org/10.1080/1755876X.2018.1529714"}
{"dataset":"astwbd","predictor":"astwbd_c{class}_{pland/ed}","description":"Water cover variables aggregated from the 30 m resolution ASTER Global Water Bodies Database (v001) dataset. These features are static and don't describe changes in water cover. See the \"classes\" sheet for details on the available water cover classes.","reference":"https://doi.org/10.5067/ASTER/ASTWBD.001"}
{"dataset":"gsw","predictor":"gsw_c{class}_{pland/ed}","description":"Water cover variables derived from the Global Surface Water dataset, which describes the location and persistence of surface water. These features have annual temporal resolution from 2006-2021. See the \"classes\" sheet for details on the available water cover classes.","reference":"https://doi.org/10.1038/nature20584"}
{"dataset":"intertidal","predictor":"intertidal_c1_{pland/ed}","description":"Tidal wetland (defined as either saltmarsh, mangrove, or tidal flats) aggregated from the 30 m Global Intertidal Change dataset. This dataset is provided for 3 year periods: 2005-2007, 2008-2010, 2011-2013, and 2014-2016. Data are only available below 60 degrees north and above 60 degrees south; outside this range there will be NA values.","reference":"https://doi.org/10.1126/science.abm9583"}
{"dataset":"ntl","predictor":"ntl_{mean/sd}","description":"Nighttime lights aggregated from 500 m resolution mean annual reflectance values collected by VIIRS. A proxy for urban development. Annual temporal resolution from 2015-2021.","reference":"https://eogdata.mines.edu/products/vnl/"}
{"dataset":"road_density","predictor":"road_density_c{class}","description":"Density of roads (km of roads per square km of area) derived from the GRIP global road dataset. Roads are classified into 5 categories, see the \"classes\" sheet for details.","reference":"https://www.globio.info/download-grip-dataset"}
{"dataset":"mcd12q1_lccs1","predictor":"mcd12q1_lccs1_c{class}_{pland/ed}","description":"Land cover aggregated from the 500 m MODIS MCD12Q1 (v061) dataset using the FAO-Land Cover Classification System 1 (LCCS1). See the \"classes\" sheet for details on the available land cover classes.","reference":"https://doi.org/10.5067/MODIS/MCD12Q1.061"}
{"dataset":"mcd12q1_lccs2","predictor":"mcd12q1_lccs2_c{class}_{pland/ed}","description":"Land cover aggregated from the 500 m MODIS MCD12Q1 (v061) dataset using the FAO-Land Cover Classification System 2 (LCCS2). See the \"classes\" sheet for details on the available land cover classes.","reference":"https://doi.org/10.5067/MODIS/MCD12Q1.061"}
{"dataset":"mcd12q1_lccs3","predictor":"mcd12q1_lccs3_c{class}_{pland/ed}","description":"Land cover aggregated from the 500 m MODIS MCD12Q1 (v061) dataset using the FAO-Land Cover Classification System 3 (LCCS3). See the \"classes\" sheet for details on the available land cover classes.","reference":"https://doi.org/10.5067/MODIS/MCD12Q1.061"}
{"dataset":"shoreline","predictor":"has_shoreline","description":"The shoreline variables can have missing values if there are no shoreline segments to summarize within a given neighborhood. In these cases all shoreline variables will be set to 0 and has_shoreline will be FALSE, otherwise has_shoreline will be TRUE."}
{"dataset":"shoreline","predictor":"shoreline_{class}_{mean/sd}","description":"Characterization of shoreline derived from the global coastal segment units dataset described by Sayre et al. 2021. This dataset breaks shoreline into 1 km or shorter segments and segments are assigned continuous values for 7 variables describing the coast line itself and the adjacent water. This set of features summarizes these five variables. See the \"classes\" sheet for a description of the variables.","reference":"https://doi.org/10.5670/oceanog.2021.219"}
{"dataset":"shoreline_erodibility","predictor":"shoreline_erodibility_c{class}_density","description":"Characterization of shoreline derived from the global coastal segment units dataset described by Sayre et al. 2021. This dataset breaks shoreline into 1 km or shorter segments and segments are assigned to one of four erodibility classes describing the adjacent land. This set of features calculates the density (km of coast per square km of area) of each erodibility class. See the \"classes\" sheet for a description of these erodibility classes.","reference":"https://doi.org/10.5670/oceanog.2021.219"}
{"dataset":"shoreline_erodibility","predictor":"shoreline_erodibility_n","description":"The number of distinct erodibility classes."}
{"dataset":"shoreline_emu_physical","predictor":"shoreline_emu_physical_c{class}_density","description":"Characterization of shoreline derived from the global coastal segment units dataset described by Sayre et al. 2021. This dataset breaks shoreline into 1 km or shorter segments and segments are assigned to one of 23 Ecological Marine Unit Classes (EMUs) that describe the sea surface temperature, salinity, and dissolved oxygen. This set of features calculates the density (km of coast per square km of area) of each EMU class. See the \"classes\" sheet for a description of these EMU classes.","reference":"https://doi.org/10.5670/oceanog.2021.219"}
{"dataset":"shoreline_emu_physical","predictor":"shoreline_emu_physical_n","description":"The number of distinct EMU classes."}
{"dataset":"evi","predictor":"has_evi","description":"The EVI variables can have missing values due to cloud cover. If EVI is missing it is replaced with 0 and has_evi will be FALSE, otherwise has_evi will be TRUE."}
{"dataset":"evi","predictor":"evi_{median/sd}","description":"Enhanced Vegetation Index (EVI) at bi-weekly temporal resolution. Aggregated from the 250 m resolution MODIS MOD13Q1 (v061) dataset.","reference":"https://doi.org/10.5067/MODIS/MOD13Q1.061"}