Almost every forester uses a geographic information system (GIS) in some form or fashion. Whether this may be accessing remotely hosted satellite or aerial imagery for a qualitive assessment of forested area, tree size, and density, or investigating terrain to assess wetness or slope that affect operability and site potential, using such tools as topographic map services or digital terrain models.
However, most of this data is in raster form, especially derived forest characteristics such as forest type, stand density or size, and mapped as 30-meter resolution cells at best, usually from space-based sensors such as the Landsat or Sentinel programs. LiDAR-derived products such as digital surface models, canopy height and density models do offer better precision. Being an arbitrary and often coarse-scale grid, raster data does not conform well to natural topography that may be fine-scaled, sinuous or undulating depending on geomorphology and parent material origin such as from erosion patterns (residuum), landslides (colluvium ), alluvium (water-deposited) or glacial-origin (glacial till, outwash, eskers and moraine). Topographical patterns can be mapped at the wide-scale, but are also fractal in nature (pun-intended), with complexity evident all the way down to the micro-scale, as evidenced by the vast diversity of soil profiles mapped in soil classification. Along with anthropogenic and biological interactions, the geographically fractal nature of physical site characteristics contributes to much speculation and mystery in forestry.
To better capture this fine-scale variability, vector data of forest stands is usually adopted by natural resource agencies or other land managers to better capture the fluidity of forest characteristics. Stand boundary lines can be infinitely precise depending on the spacing of points or nodes, with a usually minimum effective spacing necessary to lasso in a single tree as the fundamental biological unit of forestry (and often a folly when attempting to layout and mark on-the-ground stand boundaries). Urban forestry has long recognized individual trees as points in community tree mapping, and new technology has now enabled continuous individual tree-level mapping at the stand-level (at least of canopy trees), a move towards census-level forest inventory and a great boon for cruise planning (see LiDAR- tree segmentation!) which helps supplement or otherwise aid vector-based stand mapping.
Practically however, many managers set a minimum stand size for assessment purposes and suggest following topographically surveyed and/or observed stand-type (composition/size/density) boundaries evident in remotely sensed imagery (traditionally stereoscopically paired aerial photographs used in manual photogrammatry). Historically, manually-mapped vector data was recorded first on paper topographic maps, then later digitized or otherwise integrated into computer-based geographic information systems, often part of an institution's enterprise data system for internal use only. More modern methods such as automatic segmentation of remotely sensed data usually just transmits the same inaccuracies of the base raster data and often is less satisfactory than well-maintained legacy data products. Due to file sizes, accuracy concerns or combination thereof, these proprietary authoritative, working vector datasets are often limited to internal-use or otherwise permitted users such as with the United States Forest Service's FSVeg Spatial dataset. Given the effort to map and maintain these datasets, the extent of institutional stand-level vector datasets is usually limited only to immediate ownership of the agency. Sometimes legacy map products can be found, usually not digitized, that do cover larger areas but have inaccuracies due to interpretation errors or needed updates from disturbances such as harvest, fire, or wind events. This extent limitation of good stand vector data greatly impairs assessment at wide-ranging mid- or regional scales.
An alternative option for forestry vector data exists for the rest of us however, and comes from the world of open data and Free Open-Source Software (FOSS).
From their website:
"OpenStreetMap is a free, editable map of the whole world made by people like you. It was started in 2004 in the UK, out of frustration with the lack of availability of good map data that was free to use. OpenStreetMap includes data about roads, buildings, addresses, shops and businesses, points of interest, railways, trails, transit, land use and natural features, and much more."
Roll up your sleeves, however, as much of the world-wide map extent is incomplete or lacks precision for forestry purposes such as with distinguishing between upland and lowland forests (natural=wood, natural=wetland), this requires significant contributions to make the map better and more usable for foresters.
Historically, vetted imports of public domain vector data such as the US Census TIGER roads dataset have been used to help fill-in the map, with other possibilities from the National Wetlands Inventory as well as established workflows to vectorize public domain raster datasets such as the National Land Cover Dataset. The default iD web editor provides many built-in reference imagery datasets with the ability to add custom layers as well. Specialized OSM data editors such as the web-based Rapid or desktop JOSM can assist efforts to enrich the dataset through community-approved imports of existing data. OSM Notes and MapRoulette challenges help identify and provide a workflow to address known issues.
Though a relatively anarchic and often incomplete dataset, OpenStreetMap has wide use and adoption by institutions. For example, the US Department of Interior uses OSM base maps via MapBox for many of their website maps. Many other OSM-based services are available, with strong documentation on the OpenStreetMap Wiki. The data is published as vector tiles, and available in QGIS using the QuickOSM plug-in, based on the Overpass API.
However you use GIS in your work, whether as an institutional user or if you keep your own self-hosted geodatabase, consider including OpenStreetMap as one of your resources. If not an active contributor, consider using OSM as a reference layer, as a basemap with attribution in your published maps. Either way, you help promote open data and the Free Open Source Software initiative.
