ABSTRACT

This paper describes a data query interface developed in support of the Santa Rita Experimental Range Ecological Information System (SREREIS). The system will be used to provide access to various types of ecological information collected on the range since 1903. The system is comprised of a semantic model database, a data handler, and a unified user interface. Currently, the majority of the data is available only in hardcopy format, and is at risk to loss. Among critical issues in automating the data is method of query and access. The structure, function, and operation of the prototype is described. The interface has been developed using ArcInfo's AML and form menu functions. The prototype has provided an opportunity to focus on information access processing and associate database design issues. These issues will be critical to the further development of the SREREIS as its other components of the system are implemented. The selection of ArcInfo as the development platform for the data query interface component has been successful in serving the continuing development of the overall information system. It has enabled the research team to gain insight into the logical structure of ecological information access in a short period of time.

INTRODUCTION

The Santa Rita Experimental Range Ecological Information System (SREREIS) is an integrated system for browsing, referencing, querying, analyzing, and downloading ecological information collected over a nearly 100-year period in Arizona. The system will provided access to many types of digital environmental information, including meteorological data, biological data, physical data, and cultural data. These data range widely in format, structure, resolution, type, condition, and consistency. The scope of data collection ranges from range-wide (24,000 hectares) to small plot (< 1.0 sq. meter) areas. Data types include GIS data, tabular data, and a variety of hardcopy data such as field data forms, summaries, calculations, maps, photos, reports, articles, and study plans. The vast majority of these data are in hardcopy form, containing the field data collection from dozens of investigators over nearly a century. Many of the hardcopy records were at risk from physical deterioration, lack of archiving, and inadequate cataloging. Access to most of these data sources for researchers was by direct reference from previous investigators and physical retrieval of the, largely, hardcopy data from uninventoried storage.

The value of the information contained in the Santa Rita Experimental Range data set and the need to protect, and provide non-destructive access to it lead to the development of a funded research project. The objectives of the project are to rescue the data, develop a spatial information system, and a distributed access system. This project, funded by the USDA Forest Service and the Vice President of Research at The University of Arizona, supported the human and computing resources necessary to develop a spatial information system. The system was conceived as one that would include all available data types and provide access for browsing, query, display, and downloading. The designed system is to be comprised of modules for data management, spatial and logical query, and remote distributed access.

A conceptual framework was developed that guided the database design was based on multiple criteria, such as existing format, proposed digital format, and user query structure. Within the funding and phasing constraints of this endeavor, it was determined that the design and development of a working prototype of query interface would be a high priority. It soon became apparent that the types, structures, and formats of the information in the Santa Rita data set required a close examination of the access handling. This document discusses the structure and function of a prototype query interface developed for the SREREIS.

OVERVIEW

The Santa Rita Experimental Range (SRER) is operated by The University of Arizona for the purposes of rangeland research and education. The SRER is located approximately 50 km south of Tucson, Arizona and covers about 24,400 hectares. The site includes vegetative communities from desert grassland to oak woodland, with the majority of the area located on a large alluvial fans with incised drainages. Elevations range from about 940 to 1700 meters from the northwest to the southeast. Rainfall along those elevations ranges from about 250 to 500 millimeters.

Established in 1903 to protect native rangeland and conduct livestock production research, the SRER is the oldest range experimental station in the world. Much of the research was primarily conducted by the USDA Forest Service in cooperation with The University of Arizona. SRER was transferred by federal legislation to the Arizona State Land Department. Arizona Senate Bill 1249 provided that the area would be used by The University of Arizona for ecological and rangeland research purposes.

The Santa Rita Experimental Range (SRER) data set provides an opportunity gain access to an unsurpassed ecological record. The written records include climate, vegetation change, livestock use, land use treatments and long term photography.

The database design process was driven by a structured approach to the major identified issues. The first was to acknowledge the fact that most of the field measured data records were compiled in study-based data sets. This has serious implications for integration of information of the same kind, for example, herbage production, but with different data structures, measuring units, and field methods. A primary objective of the database design is that it be flexible and parsimonious and support the integration of future data. It was determined that all relational data would need to be compatible with the ArcInfo format of the existing GIS database. The database was to be designed to support the distributed access objectives of the SREREIS project. An additional consideration in the database design was that it would need to be capable of integrating data sets based on a funding schedule. That is, the database design would have to be modular and expandable in nature to facilitate the addition data sets still to be automated.

A logical step in the design of the SRER database was a preliminary inventory and cataloging of the entire, assembled written data set. The hardcopy information, of all types, was numbered and categorized by type of information and subject. The information was input to a Microsoft Access database, which was partially developed at the site where the hardcopy records are stored. Data types include reports, field data sheets, photos, study plans, and summary sheets and the subject areas include information on the physical environment, vegetation, management treatments, and range infrastructure.

Prior to the inventory and cataloging, selected data sets had been identified as priority data sets. The prioritization was based on the type of data set, its uniformity, and length of time. Each data set was ranked on a scale of 1 to 3, with 1 being the highest priority class. Specific data sets are funded for automation from the ranked list. At the present time, five data sets have been funded for automation. These data sets and related tabular information, such as synonym and name histories, comprise the study-based themes in the SREREIS database. In addition to this, the database includes over 300 scanned photographs taken from repeat photo points for over 70 years, [put this in a table].

The Santa Rita Experimental Range Ecological Information System (SREREIS) has been designed as an integrated system for browsing, referencing, querying, analyzing, and downloading ecological information. The system will provided access to the full range of data and information developed during field studies as well as additional information created during the development of the system. It will serve user based views of data in the form of dynamic selection queries. For example, a user would be able to view rain gage, forage production, and fire history data for the same period on the range. The user would be able to analyze the data in a combinatorial manner, such as selecting all stock tanks over 0.1 hectares in area, within 500 meters of a wash. A major component of the completed SREREIS will be a fully functional user interface that supports user-based spatial and logical queries of the data. A spatial data information query interface is one of the three main components of the complete system, along with a data handler and a semantic model database to handle relational mapping.

DATA QUERY INTERFACE

A data query interface will be a major component of the spatial data information system. Its role will be to provide a structured approach to accessing user-defined views of data. These data could include collections of vegetation records, associated study plans, published reports, site photos, imagery, and GIS data. In addition, the system will provide the user with dynamically generated metadata, specific to the selected data set.

The data query interface will provide such functions as categorical metadata, selection interface, and data download preparation. Categorical metadata provides general information on the selected data type category, such as vegetation data or infrastructure. This provides users with the ability to learn more about the data represented by these categories and verify the utility of the data for a particular use. The interface provides the user with critical selection criteria for the currently available data. For example, the user may select the category "precipitation" and the proceed to select subsets by location, date, name, or any other recorded parameter. Specific user interfaces, based on ArcInfo AML, menus, and display canvases are used to help the user define data subsets.

Prior to downloading a dynamic data subset, a couple of key procedures will be executed. First, the user will be allowed to view and verify the contents of the current data subset. Once verified, the spatial data information system will generate content-specific metadata for the user data set. This includes format and structure information about both the source data but also about the user's data subset. This information is currently produced by the system as an ASCII report from INFO. The current implementation of this interface simply produces .dbf format database files and ASCII text files. Future implementations will deliver these files to the user via a file transfer utility.

The data query interface was developed and implemented in an ArcInfo environment for several reasons. It was decided early in the interface planning process that among the key issues in the interface were those procedural sequences affecting information content, integration, and generation of synonymy and name history relational mapping tables. Inasmuch as the final distributed data access system had not been funded, it was determined to use existing software applications. The Advanced Resource Technology (ART) Group of the School of Renewable Natural Resources provides GIS support to The University of Arizona. This group maintained the existing SRER GIS database, and some relational data, in an ArcInfo format. The students, staff, and faculty were familiar with the AML programming language and have a bit of facility with it. The AML and INFO programming functions were reviewed to provide assurance that the conceptual functions of the data query interface could be developed.

The SREREIS query interface application is run inside WorkStation ArcInfo version 7.0.3. The opening menu is called from the "Arc:" prompt in its current form. After viewing metadata about the available information categories, more detailed selections of sub sets can be made. These selection choices establish subset selection by logical or spatial criteria. Logical criteria vary with the data category selected. For example, choices available for precipitation are not all the same as those for vegetation cover. The data selection menu includes choices for spatial selection or logical selection. Logical selection choices will vary by data category. Spatial selection can currently be made by single point or box method and includes a map display canvas run in ARCPLOT. Once a selection has been made, the user is prompted to view the current set of data records from the INFO database. Once verified by the user, the data records can be formatted for output in the common .dbf file format and downloaded by the user. At the same time, an INFO program generates all relevant metadata into an ASCII report form. This includes information about the files delivered , format definition, and any sampling information that is specific to the current selected set. Sequence control functions such as CANCEL and option specific instructions are part of the application interface. The user may return to the opening menu to view information about a different data category. The interface currently contains nominal levels of automated query validation control. suited to the prototype.

Since the development of initial data sets and associated access protocols would establish database development procedures, logical structure issues and implementation procedures were carefully detailed. Decisions were made by the research team to go to the data format closest to the actual field data sheets, for example. This issue came up because the SRER data set contained, in some cases, multiple versions of the same study data. It was considered that uniform policies like this be documented and followed to assure consistency in the sources of the automated data sets. In other cases, there were inconsistencies in values for specific records in different forms of the same data. Some of these issues were resolved by direct interview of the individual researchers. In such situations, the fact of the resolution by researcher interview is also documentation. The philosophy is that we are trying to protect and archive automated data sets in as close to original content as possible, while providing the end user with seamless access to selected information.

The data query interface is a key component of what will be a larger information system. The proposed form of that system will include components for the semantic model database and the data handler. At the present time, the data query interface is implemented in ArcInfo using AML, menu, and INFO programs.

While developing the prototype, many issues in information access were identified and addressed conceptually. However, not all coded routines were necessarily implemented in the prototype. There are two reasons for this. The first is that information access processes were identified and categorized and programmed to the point that a sample procedure could be demonstrated Once this type of query was demonstrated, the programming for additional queries of that type was no longer needed to understand the issues. Another reason is that the final platform for the relational database management portion of the system has not been established. Since low level data handling will be specific to this platform, it was not feasible to develop large numbers of procedures on a temporary platform.

As developed, the current query interface is being iteratively evaluated, refined, and tested by faculty, staff, and student researchers at The University of Arizona. The current automated data from specific studies has been provided to the USDA Forest Service along with complete system output. This output includes the data, metadata, and system documentation. As new data sets are automated, access issues unique to these new sets will be addressed within the development environment.

CONCLUSIONS

This paper provides a description of a data query interface prototype developed to gain an understanding of ecological information access, its processing requirements, and its affect on database design. The selection of ArcInfo as the development platform has been very effective for several reasons. First, it provided all required functional capabilities for the development of the prototype. These included spatial data management, relational data management, and user interface development functions. We found it easy to develop and maintain the relational data using INFO and to develop and refine the interface using AML and form menus. All issues relevant to the investigation of ecological information access for the SREREIS could be explored using these tools.

The second reason is because the existing SRER GIS database was developed and maintained at ART in a WorkStation ArcInfo format. This is consistent with the majority of data sets maintained by ART and other public sector organizations. In addition, the ready availability of GIS data in this format made it easy to integrate other external GIS data themes into the database. These themes are used to enhance the thematic richness of the GIS portion of the SRER database.

Another reason that the development platform was successful was the availability of human resources familiar with it. Since The University of Arizona is a participant in the Esri University Site License Program, there was an adequate supply of expertise in data processing using the ArcInfo software products. The software and expertise with it were used to strategically shorten the prototype development process. In this way, research focus could turn to the investigation of information access and database design without the burden of learning new software products. This will allow us to transfer the conceptual access procedures to the next development platform.

In this implementation of the prototype query interface in AML was good choice as it facilitated the exploration of research and design questions applicable to the final data query interface platform. The software provided a suitable environment in which operationalized responses to query problems could be developed. These procedures provide the conceptual basis for the next implementation of the system.

In conclusion, the selection of ArcInfo as the development platform for the data query interface has been successful in serving the continuing development of the SREREIS. It has enabled the research team to gain insight into the logical structure of ecological information access in a short period of time.

ACKNOWLEDGEMENTS

The authors of this paper would like to thank the faculty, staff, and students of the Advanced Resouce Technology Group of The University of Arizona who have supplied the human resources for the project. We also wish to thank the United States Department of Agriculture Forest Service and the Office of the Vice President of Research at The University of Arizona for funding the database and applications development. All AML and INFO programming by Karen Borstad.

AUTHOR INFORMATION

Craig Wissler: Craig Wissler is the GIS Coordinator at the Advanced Resource Technology Group. He is the database administrator for the project. (520) 621-9588 craig@sisyphus.srnr.arizona.edu

Karen Borstad: Karen Borstad is a Graduate Research Assistant. borstad@nexus.srnr.arizona.edu

Deborah Angell: Deborah Angell is a Graduate Research Assistant. angell@aruba.ccit.arizona.edu

Mitchel McClaran: Dr. McClaran is Associate Professor of Range Management. He is a principle investigator on the SRER data project. grama@ccit.arizona.edu

School of Renewable Natural Resources
College of Agriculture
The University of Arizona
BSE 325
Tucson, Arizona 85721