Robert C. Weih, Suzanne Wiley, and James B. Baker
Developing a Data Support System for Ecosystem Management Research
Abstract
Changing attitudes toward national forests have increased demands to manage forests in a socially acceptable and ecologically sustainable manner. As an alternative to clearcutting on national forests in western Arkansas and eastern Oklahoma, research was initiated to investigate the effects and trade-offs of partial cutting methods in pine-hardwood stands. An interdisciplinary group of Forest Service and University scientists organized under the Southern Forest Experiment Station developed a replicated stand level study and has begun to monitor activities in the following areas: silviculture, wildlife, biodiversity, visual quality, recreation, water quality, soils, cultural resources, insects, diseases, harvesting and management economics. Information is being collected by more than fifty researchers over a five year period.
In order to evaluate the effects of different management strategies and their interactions with forest resources, the data must be brought into a common format and made available to all researchers. A data support system was developed which incorporates Geographic Information System (GIS), Global Positioning Systems (GPS) and computer network technologies. This allows researchers to integrate both spatial and tabular data from more than two thousand plots. Initiation of this data support system involved the development of a conceptual design which allows all phases of the research to be integrated into a spatial model.
Introduction
Ecosystem Management Research Project
The Ecosystem Management Research Program was conceived during a 1990 field trip in the Ouachita National Forest attended by Arkansas Senator David Payor and USDA Forest Service Chief Dale Robertson. Discussions during this trip resulted in the elimination of clearcutting and planting as the primary silviculture method for pine and pine-hardwood stands on the Ouachita National Forest (Mersmann et al., 1994). As an alternative, research was initiated to investigate partial cutting methods that involved the retention of a mixture of tree species and natural regeneration. The research was based on the need and desire to manage national forest lands using silvicultural practices consistent with sustainable ecosystem management. The Ecosystem Management Research Program in the Ouachita and Ozark National Forests encourages increased interaction among researchers, managers, and citizens to resolve national forest issues and consists of three phases (Baker, 1994a). The first phase would establish demonstration stands; the second phase would take a statistical approach to study alternative silviculture treatments at the stand level; and the third phase would address ecosystem management practices at a watershed scale.
Phase I, an unreplicated stand-level demonstration project implemented in the summer of 1991, provided early evidence of the operational feasibility of selecting for various densities, compositions, and structures of pine/hardwood overstories. Phase I is complete.
Phase II, a replicated stand-level study, was installed in mature, shortleaf pine (Pinus echinata Mill.)-hardwood stands in the Ouachita and Ozark National Forests during the summer of 1993. A series of permanent and temporary sample plots were established to test and evaluate a range of partial cutting methods (seed-tree, shelterwood, single-tree and group selection) and vegetation management treatments (site preparation and release). The objectives of the study (Baker, 1994a) are to evaluate: (1) the biologic and economic feasibility of using partial cutting methods and long-term retention of pine-hardwood overstories to establish and maintain mixed pine-hardwood stands that reflect indigenous vegetation and historical stand structure on south-facing slopes of the Ouachita Mountains and (2) the effects and trade-offs of the partial cutting methods on various commodity and noncommodity resources and values. Thirteen treatments include both even-aged and uneven-aged reproduction cutting methods, with long-term retention of various densities, compositions, and structures of overstory pines and hardwoods. Two controls, an unmanaged control and a clearcut control, are also included as part of the 13 treatments. Four levels of vegetation management are also being investigated. The effects of harvesting/vegetation management treatments will be evaluated in terms of multiple resources and noncommodity values, including: plant and animal communities, arthropod and microbial communities, soils, water, cultural resources, scenic quality, recreational opportunities, and harvesting and management costs. Phase II data collection will continue through 1998 and is the focus of the data support system described in this paper.
Phase III, a large-scale landscape-level study, has been established on four watersheds (each drainage contains between 1,500 and 4,000 hectares) with different forest management scenarios. It will permit the evaluation of biological, physical, and social processes, functions, and linkages at various spatial and temporal scales. The Phase III project is designed to test the operational implementation of ecosystem management at a large scale.
Data Support System
A tremendous amount of data has been collected by Phase II researchers up to this point. At a 1993 symposium, researchers reported on pretreatment conditions and preliminary findings of the Ecosystem Management Research in the Ouachita Mountains (Baker, 1994b). Papers were narrowly focused on topics such as: herbaceous plant diversity, small mammal communities, breeding birds, arthropod biodiversity, water chemistry, scenic quality, and harvest management costs. Each research team has begun to accumulate a significant amount of information on the various components of the project. However, in order to evaluate treatment effects on the ecosystem as a whole, researchers must have access to information collected by other teams for each specific location and treatment. All variables must be weighed and referenced to one another before a broader focus on the ecosystem can be achieved.
Methods
Database design
The first step in developing a data support system for the Ouachita/Ozark National Forest Ecosystem Management Research effort was to determine common denominators which applied to the data of every research team. Research topics include silviculture, biodiversity, wildlife, water quality, soils, cultural resources, visual quality, recreation, arthropod and microbial communities, as well as logging and management economics. It was also necessary to determine what types of data were being collected and devise methods to store, view, integrate and evaluate many different data formats. Data formats included tabular records, tables, charts, analyses, color slides, color and black and white photographs, stand maps, and plot diagrams. Although each research team devised data collecting methods appropriate to its needs, all data could be tied to specific real world locations where treatments had been installed at the stand and plot level.
A conceptual design was developed to integrate both spatial and tabular data from 52 stands and more than two thousand plots. Because all data can be tied to specific treatment locations, Geographic Information System (GIS) software provides a logical means to integrate research efforts. Global Positioning Systems (GPS) data serve as a coordinate framework to bring the real world data into a computer world analysis. This spatial model has evolved into a data support system which allows integration of all phases of the research.
Developing a spatial reference system
Providing coordinate references for each treatment location was a critical step in the development of this data support system. Stands were located on USGS quadrangle maps based on hand-drawn stand maps provided by Ouachita/Ozark National Forest Service personnel. The roads and streams on the quad maps were digitized in the area of each stand using PC ARC/INFO*. Quad maps were then scanned into tiff images which were later rectified and registered using the road and stream cover in Workstation ArcInfo*. UTM coordinates were determined for each stand and their locations in northwestern Arkansas and eastern Oklahoma are shown in Figure 1.
Figure 1. Location of the 52 Phase II stands are shown in red. The Arkansas/Oklahoma border is in yellow and county names and boundaries are in black.
Aerial photographs were also used as base references for stand data. Color stereo aerial photography was flown at a scale of 1:7,200 over all 52 stands in the study (Figure 2). Aerial photographs were scanned into tiff images and converted to ER Mapper* files using Image Alchemy*. Image Alchemy and customized programs were used to convert registered raster quad map files to ER Mapper registered images. The quad maps were then used to register and rectify the aerial photos in ER Mapper. This procedure was necessitated by the paucity of roads and streams in the area of many stands making it impossible to register the aerial photographs using the vector road and stream coverages. In addition, GPS coordinates were collected and differentially corrected to assist with the rectification of photos and to pinpoint various wildlife and silviculture subplots.
Figure 2. Aerial photo of Stand 458-10, pine seed tree treatment, on the Fourche Ranger District in the Ouachita National Forest, Yell County, Arkansas.
Developing a GIS database
Data submission, metadata, and data correction forms were designed and distributed to all researchers involved in the project. Compartment and stand numbers were selected to serve as primary database keys in the relational database model and link records to a central treatment database which contains coordinate information for each stand. Data distribution policies and procedures were established.
Because of the diversity of data types, from digital records to visual images, it was necessary to choose a software package that could easily display a wide range of data sources. ArcView 2.1* served this purpose. Many types of data could be integrated and manipulated as various themes and projects. The ArcView interface could also be modified using Avenue* scripts so that researchers who have little experience with GIS software could easily interact with many layers of data. D-BASE 5.0* was chosen as the relational database package because of its compatibility with ArcView and with most of the database and spreadsheet formats in which data would be submitted. Digital records from each research team member were brought into d-BASE as separate files and linked by treatment, compartment, and stand number. Color slides, black and white photographs, and plot diagrams were converted to tif images and hot-linked to the appropriate stands and/or plots using ArcView. Stand maps and aerial photos were imported as image data and displayed as themes in ArcView projects. Tables, charts and analyses were also linked to stand data in ArcView.
Avenue scripts are being written to modify the ArcView interface for easier manipulation of research topics within the study. Icons will be incorporated into the ArcView button bar which correspond to specific data sets, such as wildlife, silviculture, visual quality, management economics, etc.
Data have been organized in an easily accessible manner. Both images and tabular data are stored on optical CD-ROM disks in a format that is accessible to many computer platforms. Aerial photographs and printed ArcView layouts of stand and quad maps have been organized in notebooks for quick reference during the development of the project. Data and documentation submitted by researchers is organized in a separate notebook.
A data dictionary accompanied the development of the database. The dictionary contains brief descriptions of projects and themes developed in ArcView (Appendix A) and item definitions and coding descriptions of for each variable represented in a theme (Appendix B). It also includes variable definitions supplied by the researcher for each data file submitted (Appendix C).
Developing researcher access
Several data access options are proposed. Access approval procedures were formulated and access request forms were designed. Access on site in the Spatial Analysis Laboratory (SAL) in the School of Forest Resources will be the most direct route. Electronic access may be obtained over a high-speed modem, an internet connection, or through the World Wide Web (WWW). Access over a modem or internet connection using remote PC clients and a Unix server in the SAL to manage data would require installation of ArcView 2.1 and PC-NFS* on the remote PCs. Data may also be accessed using remote PC clients or an X terminal emulator with ArcView 2.1 and the database installed on the Unix server in the SAL. Another option would set up WWW links to the appropriate software and databases with password protection making them accessible only to researchers involved in the project.
Summary
Initiation of this data support system involved the development of a conceptual design which allows all phases of the research to be integrated into a spatial model. ArcInfo and ArcView served this need. Individual observations, statistics, tables, graphics, photographs, aerial photos, maps, etc. are linked based on common spatial coordinates. When this database support system is fully developed, researchers will be able to investigate interactions between their data and that of others by overlaying themes and preforming different spatial analysis procedures.
An ArcView/Avenue interface is being developed to allow easy access to the various components of each database topic. Electronic access will provide support and integration of all research activities. Researchers can not only share information resources, but can graphically visualize many facets of the forest ecosystem and its changes under various management strategies. Researchers can also produce map products that clearly illustrate the effects and trade-offs of partial cutting methods on various commodity and noncommodity resources and values. The methodology used to develop and configure this very large, relational database into an easily accessible form usable in an interactive GIS environment should be transferable to many other areas of natural resource management.
_________________
* Use of registered trade names is solely for the reader's information and does not imply endorsement of the product.
Appendix A. Example of project layers and themes described in the data dictionary.
Appendix B. Example of item definitions and coding descriptions for the soils coverage contained in the data dictionary.
Appendix C. Example of file and variable definitions included in the data dictionary.
Literature Cited
Baker, James, B. 1994a.
- An overview of Stand-Level Ecosystem Management
Research in the Ouachita/Ozark National Forests.
IN: Baker, J. comp. Proceedings of the Symposium on Ecosystem
Management Research in the Ouachita Mountains: Pretreatment
Conditions and Preliminary Findings. October 26-27, 1993; Hot
Springs, AR. Gen. Tech. Rep. SO-112. New Orleans, LA: USDA,
Forest Service, Southern Forest Experiment Station. p 18-28.
Baker, James. comp. 1994b.
- Proceedings of the Symposium on Ecosystem Management Research in the Ouachita Mountains:
Pretreatment Conditions and Preliminary Findings. October
26-27, 1993; Hot Springs, AR. Gen. Tech. Rep. SO-112. New
Orleans, LA: US Dept. of Agriculture, Forest Service, Southern
Forest Experiment Station. 259 pp.
Mersmann, Timothy, J., James B. Baker, James M. Guldin, and William F. Pell. 1994.
- Implementing Ecosystem Management Research:
Bringing Researchers, Managers, and Citizens Together. IN:
Baker, J. comp. Proceedings of the Symposium on Ecosystem
Management Research in the Ouachita Mountains: Pretreatment
Conditions and Preliminary Findings. October 26-27, 1993; Hot
Springs, AR. Gen. Tech. Rep. SO-112. New Orleans, LA: USDA,
Forest Service, Southern Forest Experiment Station. p 10-17.
Author Information
Dr. Robert C. Weih, Director of the Spatial Analysis Laboratory
School of Forest Resources, Arkansas Forest Resources Center
University of Arkansas, P.O. Box 3468
Monticello, Arkansas 71656
Telephone: (501) 460-1248
Fax: (501) 460-1092
E-mail: weih@uamont.edu
Suzanne Wiley, Technical Analyst, Spatial Analysis Laboratory
School of Forest Resources, Arkansas Forest Resources Center
University of Arkansas, P.O. Box 3468
Monticello, Arkansas 71656
Telephone: (501) 460-1548
Fax: (501) 460-1092
E-mail: wiley@uamont.edu
Dr. James Baker, Team Leader, Ecosystem Management Research
USDA Forest Service
Southern Research Station, P.O. Box 3516
Monticello, Arkansas 71656
Telephone: (501) 367-3464
Fax: (501) 367-1164
E-mail: /s=j.baker/ou1=s30a@mhs-fswa.attmail.com