Yosemite National Park - Vegetation Map 1937 - Open Data

This map of the vegetation of Yosemite National Park was developed in the 1930s as part of an effort to map the vegetation of California. It is often referred to as the "Weislander map", because A. E. Weislander was the United States Forest Service employee who oversaw mapping in the Sierra Nevada (and perhaps other portions of the state as well). The work was carried out by park staff under the direction of the park forester, Emil Ernst. The data development and field work were overseen by Yosemite NP Forester Emil Ernst from 1933-1935. The polygons were originally delineated on 1937 map drawn by Wieslander, et al (see below for reference publication.) This map was digitized by the USGS Biological Resources Division in the early 1990s (date currently unknown.) The field work occurred over multiple years, with different areas of the park mapped in different years. There are two historic hardcopy maps that indicate which areas were mapped in which years, but these two maps do not always agree. The general areas mapped between 1932 and 1936 are listed below: 1932: North side of Tuolumne Canyon below Hetch Hetchy 1933: Western boundary of park from South Entrance to Hetch Hetchy entrance, Yosemite Valley, White Wolf, Porcupine Flat, Illilouette Creek, Glacier Point Road area, Wawona, Lyell Canyon, Gaylor Lakes 1934: north rim of Yosemite Valley, Clark Range (northern portion), Little Yosemite Valley, Ostrander Lake area, upper Chilnualna Creak area 1935: South Fork Merced River above Buck Camp, Upper Merced watershed (above Little Yosemite Valley), Tenaya Canyon, watershed above Lake Eleanor, Tiltill Canyon, Stubblefield Canyon, Kerrick Canyon, Thompson Canyon, Slide Canyon, Matterhorn Canyon, Virginia Canyon, Young Lakes area 1936: Tuolumne Meadows, Glen Aulin, Tuolumne Canyon above Hetch Hetchy, Smith Peak It is not known precisely when the final map was compiled and completed, although it was likely done in the winter of 1936-1937. Hence, it is not possible to put a precise date on the map, a date that represents the vegetation of the park on a precise date (or even year). It is probably most useful to consider this map to be somewhat fluid in its birth date: the map represents the vegetation of the park that existed in the 1930s, notwithstanding year-to-year changes in vegetation due to fire, disease, forest succession, etc. This time period also represents the beginning of effective exclusion of fire throughout the park, although fuel loads at the time were probably light and frequent low-intensity fires were the norm. The exclusion of fire led to unnatural fuel accumulations and type conversions through various ecological mechanisms. Field work included primarily establishment of vegetation plots and field mapping (i.e. delineations). The vegetation plots collected data on dominant plant species (by no means a complete species list), distribution, leaf litter, soils, etc. The field maps were assembled and transferred to the final map. The final map is in "panels", presumably so that they could be folded and put into files or taken into the field. The panels were attached to linen and a protective coating applied to the map surface. The mapped area does not correspond exactly to the park's boundary. For instance, areas within the park near the east boundary were not mapped. These unmapped areas are generally defined by high mountains and may have been too difficult to map. Also, some areas outside the park were mapped by park staff. Minimum Mapping Unit (MMU): The concept of an MMU was apparently not used for this project; certainly, no documentation has been found. There are polygons as small as .1 acre, and polygons as large as many hundred acres, indicating that the "MMU" size varied depending on the accessibility of the area. Given the coarseness of the mapping in some of the less accessible areas of the park, a reasonable guess for MMU is 15-25 acres. Accuracy: The accuracy varies depending on the accessibility of the area, with less accessible areas being mapped very quickly, or from a distance (i.e. binocular botany). The small field crews (2 to 5 people) covered vast areas, presumably by foot or horseback. In August 1935, about 180,000 acres were mapping by a 4-person team. The most accurate information about the vegetation in an area is probably found in the plots that the field crews established. The maps were found in a filing cabinet in the Forestry office, apparently little used and in danger of being lost or thrown away. The maps were stuck together, and some damage occurred while separating them. The maps were seized by the park's Research Forester, Jan van Wagtendonk, who began an effort to maximize the utility of the data by converting it to a GIS format. The maps were transferred to mylar (62,500 scale?) and digitized by Chicago Aerial Graphics around 1986. The polygons and a single attribute (text, representing the species code or codes found in each polygon) were put into SAGIS format. SAGIS allowed a single attribute for each polygon, limiting the utility of the multiple attributes found in the original data. Sometime around 1994, the polygons were converted to GRASS format, the text labels were converted to RIM format (RIM was a database management system with a link to GRASS, allowing users to take advantage of the multiple attributes contained in the original data), and a link between the RIM database and the GRASS data was created. Like SAGIS, GRASS allowed for only a single attribute for each polygon, but the ability to link to an RDBMS using this single attribute allowed for powerful manipulations of the multiple attributes contained in the original data. During the SAGIS to GRASS conversion, the codes used on the original map were replaced with 6-character codes that were used in other vegetation-related datasets. For example, the code "J" represnting Jeffrey pine was replaced with the code "PIJE10" representing Pinus jeffreyi. In the early 1990s, Rich Walker improved the rectification by using 7000 control points and Arc/Info software. The improved data was delivered in GRASS format, as the park was still using GRASS. Around 1995, the data was converted from GRASS to Arc/Info format by Joe Meyer. The polygons were exported from GRASS using the v.out.arc command imported into Arc/Info using the generate command (the RIM-CODE attribute was still attached to the polygon labels). The attributes in the RIM database were exported from RIM, converted to INFO format, and linked to the polygon labels using the JOINITEM command. The resulting coverage was YOSE_VEG37. In 1996, during the development of the "Vegetation Management Plan: Yosemite National Park" (developed by Jeri Hall, finalized in June 1997), a crosswalk was developed between the 1930s vegetation data and a modified Holland classification. Moore (1993) developed a classification for the Yosemite region that was based on Holland; this classification used by the vegetation management plan with a few deviations. See descriptions of the ALLIANCE and FORMATION codes below. The plot data has been entered into a digital (i.e. database management system) format and the plot locations have been digitized into a GIS format. Source Data: The 1898 1:125000 scale base of Yosemite National Park and Vicinity was used as the base map. This map was enlarged to an approximate scale of 1:50,500 (this was determined by measuring landmarks on the 1898 base map against those same landmarks on modern-day 7.5 minute topo maps). The vegetation "type" delineations were plotted on this enlarged map. Technically, the accuracy of this map is the same as the accuracy of the 1:125000 scale base map. However, it is probably more accurate than this, given the effort by Walker to register to known locations derived from the modern 7.5 minute quads. Geographic extent: The data has been clipped to the park boundary. As noted above, not all of the park was mapped, probably due to inaccessibility of high-elevation areas near the Sierra Nevada crest. Unmapped areas should be considered to be NODATA. It is reasonable to assume that unmapped areas are "barren", if necessary (this is probably why they were not mapped, that plus their inaccessibility). RIM-CODE: RIM was the database management system used by GRASS and was used to track multiple attributes (GIS software of the time was typically not capable of tracking more than one attribute). The attribute (emphasis on singular) originally entered for each polygon was a text string derived from the hand-written labels (emphasis on plural) on the original map. The attributed derived from the original map are DOMINANT, W1, W2, W3, W4, W5, W6, W7, W8. RIM-CODE is/was the code that links each individual polygon to an individual record in the RIM database. DOMINANT: This field contains the common name of the species that is "dominant" in the polygon. It is usually, but not always, a tree species. It was derived from the contents of W1. W1, W2, ... W8: These 8 fields contain a code that represent species found in a polygon. The 6-character codes have been standardized for Yosemite NP and are derived from the scientific name for a species (four characters), a differentiation between species with similar scientific names (one character), and a differentiation among varieties of a species (one character). For a species to be listed in W1 thru W5 (usually tree species), it had to comprise at least 20% of the polygon. W1 contains code of the species that is most abundant in that polygon, W2 contains the code of the species that is second most abundant, etc. For a polygon that contains only an entry for W1, that species must comprise at least 80% of the polygon, and no other species comprises 20% of the polygon. For a polygon that contains an entry for W1 and W2, the W1 species could be between 50% and 80% of the polygon, and the W2 species could be between 20% and 50% of the polygon. ALLIANCE: The original data was reclassed into the modified Holland types used for the Vegetation Management Plan (VMP). This field contains a code for the types used in the VMP. These types are not true alliances, as defined by the Federal Geographic Data Committee. The codes and full names for this field are listed below: Subalpine Coniferous Forests A1 whitebark pine forest A2 lodgepole pine forest A3 whitebark pine - lodgepole pine forest A4 whitebark pine - mountain hemlock forest A5 mountain hemlock forest Upper Montane Coniferous Forests B1 red fir forest B2 western white pine forest B3 Jeffrey pine forest B4 Jeffrey pine - fir forest Lower Montane Coniferous Forests C1 westside ponderosa pine forest C2 Sierra mixed coniferous forest C3 ponderosa pine mixed coniferous forest C4 white fire mixed coniferous forest C5 Douglas-fir mixed coniferous forest C6 Sierra white fir forest C7 giant sequoia mixed coniferous forest Broadleaved Upland Forests and Woodlands D1 California black oak woodland and forest D2 interior live oak woodland and forest D3 interior live oak forest D4 interior live oak woodland D5 white alder riparian woodland and forest Broadleaved Upland Forests E1 canyon live oak forest E2 mixed north slope forest E3 aspen forest Broadleaved Upland Woodlands F1 blue oak woodland F2 foothill pine - live oak - chaparral woodland F3 cismontane juniper woodland Scrub and Chaparral Communities G1 montane chaparral G2 northern mixed chaparral G3 interior live oak chaparral G4 montane and alpine riparian scrub G5 montane and alpine riparian scrub Grassland, Meadow and Other Herb Communities H1 subalpine and alpine meadows H2 montane meadow H3 California annual grassland H4 sphagnum bog FORMATION: This is a lumping of the ALLIANCE field to facilitate broad-brush analysis and development of GIS-generated maps. ACRES: This is a conversion from the internally-generated AREA in square meters to acres, to facilitate analysis. FMP_TYPE: This is a reclass of the ALLIANCE field into Fire Management Plan types, based on fire behavior. A separate data layer was created based on this field, and the metadata for the data layer explains the reclass rules. The attribute table contains both dominant species and alliance information for each polygon. The item used for mapping purposes, "Formation", is an aggregate of the dominance and alliance categories. The definitions for "Formation" are as follows: a - Subalpine Coniferous Forest Communities b - Upper Montane Coniferous Forest Communities c - Lower Montane Coniferous Forest Communities d - Broadleaved Upland Forest and Woodland Communities e - Broadleaved Upland Forest Communities f - Woodland Communities g - Scrub and Chapparal Communities h - Grassland, Meadow and Herb Communities i - Cultivated z - Barren Reference Literature: Wieslander, Jensen, Yates (1933). "Instructions for the Preparation of the Vegetative Type Map of California." Coffman (1934). "Suggestions for the Mapping and Study of Vegetative Cover Types in Areas Administered by the National Park Service." Metadata above compiled by J. Meyer and A Baird. For information on vegetation alliance names and CWHR classes, join the DBF table "Alliances" to the attribute table using the Alliance field in both tables. IRMA Data Store Reference