Paper 125: Esri User Conference - 1997
DEFINING ISSUE: California Assembly Bill 1039 requires the Department of Fish and Game to meet five goals relating to natural diversity:
1) Develop and maintain a data management system for natural resources;
2) Identify the most significant natural areas in California;
3) Ensure the recognition of these areas;
4) Seek the long-term perpetuation of these areas;
5) Provide coordinating services for other public agencies and
private organizations interested in protecting natural areas.
Determination of Significant Natural Areas (SNA) within California
needed to be automated so that current information would be available
to agencies for use in planning and protection.
GIS SOLUTION: Using the California Natural Diversity Database
(CNDDB) and ArcInfo, an AML was developed to determine which
element occurrences from the CNDDB meet the criteria for a Significant
Natural Area. An SNA region coverage is created, identifying
the element occurrences that comprise each SNA. This coverage
can then be updated as new information is added to the CNDDB.
APPLICATION OR METHODOLOGY: Oracle SQL scripts were written to extract any potential element occurrence records from the CNDDB that meet the criteria for a SNA. SNAs are identified using biological and spatial criteria. The following are the criteria used to select SNAs:
- Areas supporting extremely rare species or natural communities
- Areas supporting associations or concentrations of rare species or communities
- Areas exhibiting representative examples of common or rare communities
- Areas of high species-richness or habitat-richness.
ArcInfo is then used to reselect element occurrences that meet
proximity and area criteria by using buffers, reselects, unions,
frequencies and statistics. Additional coverages are unioned
with the SNA region coverage to identify SNAs that are in developed
or disturbed areas. Cursors are used for automating a unique
site number for each Significant Natural Area.
SOFTWARE: The application uses SQL scripts written for
Oracle, and AML, using the ARC and ARCPLOT modules. The purpose
of this paper is to demonstrate how GIS solved the problem and
met the need for automating and updating the selection of Significant
Natural Areas.
Natural resource managers and land use planners need a tool that
gives them accurate information about the location of priority
natural areas. Identification of significant natural areas is
the necessary element to being able to ensure recognition and
protection of priority areas. A method of automating the selection
of significant natural areas from digital data needed to be created
so that this information would be current and readily available
for use.
Use and purpose of the SNA data:
One of the goals of SNA data is to raise awareness among developers, planners, conservationists and others about the presence of SNAs in their local area. The SNA sites can be assured increased protection when others become aware of their location and significance. Experience has demonstrated that agencies and developers will adjust their plans to avoid environmental impacts if:
Early awareness of potential conflicts can avoid hearings and litigation, saving all parties concerned potentially millions of dollars and years of time.
A second goal is to improve communication between public and private organizations regarding identification and protections of significant natural areas. This cooperation is needed to continually identify, refine and update information about these areas. SNA information has been used for:
California is one of the most biologically diverse areas in North America. Unfortunately, much of this diversity has already been lost due to human activity, specifically agriculture and urban expansion. By the late 1970's, organizations concerned with protecting natural areas began to recognize that no comprehensive inventory of California's important biological resources existed to guide the protection of natural areas. It was also becoming clear that poor communication between conservation organizations was resulting in an inefficient allocation of funding and staff resources by duplicating efforts, not tapping potential assistance from other organizations and focusing on high-profile species while other more-critically endangered species were being ignored. In an effort to remedy this situation, a consortium of sixteen conservation organizations formed in 1978, proposing the development of a statewide natural areas program and in 1979 this program was administratively established in the California Department of Fish and Game. In 1981, this program received legislative basis with the passage of Significant Natural Areas legislation (Assembly Bill 1039).
Assembly Bill 1039 requires California Department of Fish and Game to meet the following goals:
In response to Assembly Bill 1039, The California Natural Diversity
Database (CNDDB) and the Significant Natural Areas Program (SNAP)
were established. The CNDDB continues to serve as California's
most comprehensive inventory of rare species and natural communities,
currently containing over 31,000 records. SNAP is responsible
for the identification of significant natural areas.
No specific criteria was specified for establishing significant natural areas (SNAs). Different approaches exist to identify priority natural areas. As one example, Margules proposes the use of iterative procedures which incorporate the principle of complimentarity and offer advantages of flexibility, efficiency and explicitness. Due to limited staff, limited knowledge of individual sites, and a need to keep results simple and easy to understand, we have developed a simple model. A statewide analysis model requires a statewide database of biological data. Until the recent development of a statewide vegetation map (GAP vegetation data, UCSB) there has been no reasonable way to identify the distribution of common habitats and species (collectively known as "elements" of natural diversity). CNDDB has been the only statewide database. Originally the CNDDB was intended to inventory all natural resources, but limited staff has restricted data collection to rare elements. Due to these limitations, our analysis model focuses on priority natural areas for rare elements.
SNA criteria are strictly biological, and were determined by Marc Hoshovsky, Biodiversity Protection Planner for California Department of Fish and Game, Natural Heritage Division. SNA sites must meet at least one of the following criteria:
The extremely rare and best example criteria can be determined directly from the Oracle database of the CNDDB. CNDDB occurrences meeting the concentration or species/habitat-richness criteria need to be determined using a GIS to analyze proximity. The following is additional criteria that has been established to determine SNAs of concentration or species/habitat richness:
The maximum 500 meter distance between occurrences was determined
by use of professional judgment.
The CNDDB resides in an Oracle database and as an ArcInfo region coverage. Oracle SQL scripts have been developed to determine which records in the CNDDB would be eligible to be evaluated for the automated SNA selection process. An Oracle script selects records:
Approximately 2/3 of the occurrences in the CNDDB meet eligibility at this point.
Additional scripts then determine and isolate the extremely rare occurrences and best example occurrences. Extremely rare occurrences are selected from the eligible occurrences by their CNDDB state rank of S1. Best example occurrences are determined by checking for the highest CNDDB occurrence rank-id for the species (A, B, C, D). The remaining eligible occurrences are evaluated using ArcInfo to determine their eligibility for becoming an 'ensemble candidate' (meeting the concentration or species/habitat-richness criteria). Occurrences that survive all the selection processes are buffered 500 meters to create 'blobs' or polygons surrounding the actual occurrence. These polygons become the actual SNA.
Three separate files are written as a result of the SQL selection process, each containing the list of a primary key identifier for each occurrence. One file contains the list for extremely rare occurrences, one for best example occurrences and one for the ensemble occurrences. These files are used in ArcPlot to reselect from the CNDDB, using the keyfile option. Separate region coverages are made for the extremely rare (S1), best example (BX) and ensemble candidate (EC) data. Acreage for each region is calculated and regions with acreage greater than 200,000 are eliminated. Occurrences that are larger than 200,000 acres were eliminated because these areas would be harder to protect and these large polygons may make the data appear less viable. An item is added to each region coverage and calculated the name of the coverage. This identifies why the occurrence is eligible for evaluation as an SNA in the final SNA coverage. Elements in the S1 and BX coverages automatically become SNAs (having met the above listed criteria). These coverages are unioned and buffered 500 meters. Elements in the EC coverage that are within 500 meters of the unioned S1-BX coverage are reselected in ArcPlot using the overlap option and unioned with the elements in the S1-BX coverage.
To determine which ensemble occurrences meet the species/habitat
richness criteria, the EC coverage is buffered 500 meters and
polyregion used to convert to regions. The EC coverage and the
newly created region coverage are unioned and regionquery is then
used to identity the coverages. This allows for the use of frequencies
and statistics to determine which of these regions contains at
least three distinct species. The ensemble occurrences that
are identified as meeting SNA criteria through this process are
then unioned with the other selected occurrences to produce the
final occurrence coverage. The SNA region coverage is created
by buffering the final occurrence coverage 500 meters, using polyregion
to convert to regions, and then unioning with the final occurrence
coverage. Regionquery is then used to identity the coverages
so that the user will have the ability know which elements belong
to which SNA, and why they are eligible as an SNA.
Automating the selection of SNAs with the above methodology is based entirely on biology, and does not take into account the quality of the site. To help identify SNAs that are located in disturbed areas, a coverage was developed from the GAP data created by University of California, Santa Barbara, selecting for polygons in urban areas and for specific agriculture polygons. The selected GAP coverage is unioned with the SNA coverage, and regionquery used to identify which SNAs are located in disturbed areas.
Users of SNA data need some way of identifying and distinguishing each SNA from each other. It was decided to label each SNA with a county abbreviation and incremental numeric code. The labeling method was automated by unioning the SNA coverage with a county coverage, using regionquery and then cursors to increment the numeric count. In addition to this label, each SNA will have a site name. The site names have yet to be determined and will be decided by the DFG Regional offices.
Reports generated from the Oracle database accompany the SNA spatial
data. For the purpose of report writing, the SNA coverage is
unioned with a land ownership and a USGS 1:24,000 quad coverage,
allowing for the identification of the quads and ownership for
the location of each SNA. Additional tables created through ArcInfo
processes are brought back into Oracle using DBMSINFO.
Field visits were conducted on SNA sites generated for pilot
areas. Field studies gave us the opportunity to check the validity
and accuracy of the data the model generated, as well as presenting
additional methods, such as using watersheds to connect or divide
sites, to further improve the model. SNA site maps and reports
were sent to the 5 DFG regional offices for staff review. The
regional staff is much more familiar with these areas and can
offer valuable input in the evaluation of these sites.
The goal of automating the selection process is to make it easy
and convenient to update, so that the data remains current. Each
time the SNA layer is updated the entire CNDDB database needs
to be re-evaluated. The CNDDB science staff is continually updating
information on occurrences and adding new occurrences, consequently
each occurrence needs to be evaluated or re-evaluated to check
if it meets the criteria. Labels and site names that have been
assigned to SNA sites in previous coverages need to be carried
to the new coverage. This is done by relating the tables of the
old and new coverage by the primary key identifier and calculating
the label and site names in the new SNA coverage to match the
label and site names of the old SNA coverage. New labels and
site names are then added to the new SNAs that didn't exist in
the old coverage.
Now that the initial data and methodology has been developed,
it is important to continue to improve the model. Additional
data layers, such as watersheds and remotely sensed data, will
be evaluated in future analysis. Input from users of SNA data
will also be valuable and help guide the direction and development
of future analysis.
Hoshovsky, Marc. Significant Natural Areas of California, 1992 Annual Report, Volumes I and II. California Department of Fish and Game, Natural Heritage Program, 1992.
Jones and Stokes Associates. 1987. Sliding Toward Extinction: The State of California's Natural Heritage, 1987. Prepared at the request of the California Senate Committee on Natural Resources and Wildlife. Commissioned by the California Nature Conservancy, San Francisco.
Margules, C.R., Cresswell, I.D., Nicholls, A.O. Systematics and Conservation Evaluation. Systematics Association Special Volume No. 50, pp 327-50, Clarendon Press, Oxford, 1994.
Lora Konde GIS Analyst California Department of Fish and Game Natural Heritage Division 1220 'S' St. Sacramento, CA 95814 Telephone (916) 445-5758 Fax (916) 324-0475 email: lkonde@dfg.ca.gov