The CalFlora Species Database

This database contains summary geographic and ecological distribution information for 7660 California vascular plant taxa (8363 records including species level taxa where there is more than one var. or ssp.), as well as additional habitat information for rare taxa and species of the Sierra Nevada. It was originally compiled from two major literature sources, an electronic transcription of distribution and lifeform data from A California Flora and Supplement (Munz 1959 and 1968) and the CNPS Inventory of Rare and Endangered Vascular Plants of California ( Skinner and Pavlick 1994, electronic version) These data were supplemented with information from The Jepson Manual (Hickman 1993). Species summaries now include county distributions compiled from current contents of the CalFlora Occurrence Database, with links to the specific observations. Species summaries are illustrated in the online version with images from the UC Berkeley Digital Library Project - CalPhotos California Plants and Habitats collection.

Information Sources

The transcription of distribution information from Munz (1968) was created by Kwei-Lin Lum as part of her thesis work with Peter J. Richerson at U.C. Davis. (Lum 1975, Richerson and Lum 1980). Lum's database was subsequently modified by Richard Walker in the course of his thesis work at U.C. Santa Barbara (Walker 1992). For Sierran species, Walker added elevation limits and habitat descriptors, and also replaced Lum's 'many plant communities' entries with specific community lists. The Lum/Walker/Munz68 data covered all but about 24 of the species-level taxa included in A California Flora and Supplement. In compiling species summaries, we translated the abbreviated names and coded data of the Lum/Walker/Munz68 database. We then used data from the CNPS Inventory to update distribution and habitat data for 893 species in the Lum/Walker/Munz68 database and to add records for 849 taxa not already included in that database (that is, infraspecific and new taxa). Using the CNPS Inventory data, we also added fields for Rarity and Listing Status as well as a field for matching records in this database with records in the CNPS Inventory. Listing Status data has been updated to match the 1997 CNPS electronic inventory.

The database was subsequently expanded to include all infraspecific taxa recognized as occurring in California and updated to reflect current nomenclatural usage. Adjustments to distributions for taxa affected by changes in taxonomic delineation (lumps and splits) were made using information from the Jepson Manual (Hickman 1993) to modify existing records. Literature-based distribution data for taxa not previously included in CalFlora (mostly infraspecific taxa that are not rare) represent our interpretation of distribution descriptions in the Jepson Manual.

Distribution Information Errors

Error characteristics of the occurrence-based distribution data are unknown. Users should make their own decisions on the reliability of data of different types and from different sources. Documentation and source contact information are provided for each occurrence observation.

For literature-based distribution data, is important to note that the type and rate of errors in geographic distribution are very different for the three major literature sources. Lum was transcribing generalized and sometimes vague range descriptions into concrete presence/absence data for specific county and subcounty regions. For most species, data were not verified with any other source of information and, in addition to coding errors, represent only a best estimate of the range of the species in question. Lum (1975) showed that while her data are far from error-free, errors of assigning taxa to regions where they do not occur are about equal in number to errors of not assigning taxa to regions where they actually do occur. Consequently, her data produce unbiased and quite accurate estimates of species numbers but somewhat inaccurate species lists for given regions. The CNPS Inventory data, on the other hand, represent only documented natural occurrences of the taxon in question, are not intended to imply that the taxon does not occur elsewhere, and hence represent a minimum estimate of the full range. CNPS data underestimate number of species present in a given region but have a very low rate of error in assigning taxa to counties where they do not occur (see notes for Regioncode in Database Field Descriptions). Regions used to describe plant distributions in The Jepson Manual span multiple counties. Consequently interpretation of Jepson Manual descriptions in terms of presence and absence in particular counties and subcounties is indefinite. We have identified all items of distribution data that are based on the Jepson Manual to alert the user to this fact. ('*' in regionlist field, '6' in regioncode field)

For species level taxa, the composite literature-based distribution summaries presented in CalFlora can be expected to have error characteristics much like those Lum described. It can be expected to perform well for analyzing broad patterns and general relationships, as demonstrated by both Lum and Walker. It clearly performs less well in generating accurate species lists for particular locations, but not so badly that it is not useful for producing preliminary checklists for a variety of applications. For infraspecific taxa, the composite literature-based distribution summaries combine data that tend to underestimate range (CNPS) with data that overestimate range (Jepson Manual). Error characteristics of the combined data are unknown.

Downloading the CalFlora Species Database

We encourage serious users to download CalFlora data in order to make full use of their own query and analysis tools, and to make use of fields not included in the online display and query forms. The CalFlora Species Database is available as an ASCII file in quote-delimited, comma separated format with maximum field length of 254 characters. A second ASCII file in tab-separated format with some fields exceeding 254 characters (county distributions in a single field rather than separated into three fields) is also available. Please see database Field Descriptions.

Ann Dennis, CONTACT
Director, The CalFlora Database
updated August 1999