Methods


Maps

The data mapped in the Biodiversity Atlas of Los Angeles come from a variety of sources, which are listed under each map. Some layers are based on the spatial analyses of the Atlas authors or collaborators. This work is typically performed using GIS software.

The map layers may be based on two different data formats: vector or raster. Vector data is represented by points, lines, or polygons. The shape of these vector features is informed by x, y coordinates. Examples include the geology or wildfire maps. Rasters store data in a grid of cells, which are like pixels in a digital photograph. The precipitation and black bear distribution maps are examples of raster-based data layers, which have been converted to a polygon format for use online.

Species Distribution Models

Many of the species maps in the Biodiversity Atlas of Los Angeles are based on species distribution models, which combine species observations (i.e. locations where a species is known to occur) with influential environmental variables (like those in the “Environment” and “People” maps) to predict the likelihood that a species or its habitat will occur in an area.

The model data and parameters described below apply only to the species distribution models that were developed by the Atlas authors. If a species map or model was contributed by another scientist, it is likely based on a different modeling approach.

Species data: Distribution models require species records in the form of point data (i.e. longitude and latitude coordinates). The source of the species observations used in the Biodiversity Atlas of Los Angeles are listed under each map. Many of the occurrence records that we use come from citizen science projects like iNaturalist or eBird. Others have been contributed by Biodiversity Atlas authors or collaborators.

Environmental variables: Several predictor layers are used to model the species in the Biodiversity Atlas of Los Angeles, including: topography (e.g. slope, elevation, aspect), climate (e.g. precipitation and temperature variables), geologic substrate, and land cover type (e.g. tree canopy cover, impervious surfaces). Each dataset, with the exception of geologic substrate, is raster-based. The geology layer is based on polygon vector data, which was converted to a raster format in ArcMap in order to match the other layers.

Modeling method: The species distribution maps have been developed using the Maxent model, a well-studied, open-source correlative modeling method.

Model resolution: The species distribution models have all been developed at a 1-kilometer resolution, which means that each pixel represents a 1 square kilometer area.