DEVELOPMENT AND IMPLEMENTATION OF FEATURE LEVEL METADATA FOR EPA REGIONS

With the expanded efforts to gather spatial data in the U.S. Environmental Protection Agency (EPA) and the increased use of GIS as a tool for environmental programs, the EPA Regions concluded that there is a need to manage the locational accuracy of feature data. Up to this point, EPA's Regional GIS programs have had no consistent way to store or retrieve locational data obtained from sources such as GPS or geo-coding, or from most of the Agency's own databases.

Although the GIS programs do not own the Agency data, and, therefore, have little control over the locational data stored in these databases, they do have an inherent responsibility of using the most accurate data available. In order to meet this need and to provide a means of improving efficiency, and maintaining and improving product quality, the EPA Regions and the National GIS Program worked jointly to develop a standard Oracle-based feature level metadata database.

INTRODUCTION

A brief assessment of EPA's existing databases revealed that many have fields for locational data, but in most cases associated fields containing information on the accuracy of the data and the method used to collect the data (i.e., metadata) are ether lacking or not populated. Also, no EPA databases contained adequate fields to store a high accuracy latitude and longitude and its locational metadata. While there are ongoing initiatives within the Agency to modify the databases to include appropriate metadata and locational information, the need for accurate locational data is already upon us.

The EPA GIS Technical Sub-Workgroup, comprised of technical GIS representatives from the National GIS program in EPA Headquarters and from each of the 10 EPA Regions, formed as a mechanism to share information and advance GIS technology throughout the Agency. Through this workgroup's activities, a growing concern was expressed over the lack of adequate standards and a uniform methodology for storing collected locational data and their attributes. Because spatial data is expensive to gather and highly critical for GIS work being performed within the Agency, the workgroup wanted to design a database that all EPA Regions could use for storing their locational data.

A team was formed to perform an analysis of requirements, formulate design specifications and develop a system that could be used to maintain accurate locational data until the time came when the Agency databases contained the capacity to meet this need. This system was named the Spatial Preferred Locational Attribute Table, or SPLAT for short.

SPLAT database diagram

DESIGN CONSIDERATIONS

• Store, retrieve, update and manage locational data of different origin and quality through a standardized system
• Preserve integrity of all feature data
• Standardize feature level metadata consistent with EPA's Method Accuracy Description (MAD) code guidelines
• Provide easy access to locational entities
• Use a consistent, documented criteria for selecting preferred locations for use in GIS products
• Support a standard method of GIS data distribution
• Provide regulated entity cross-referencing through logical and maintainable linkages
• Work in conjunction with EPA's Envirofacts Warehouse
• Be a distributed locational data warehouse for the Agency

WHAT IS AN ENTITY?

Records defining entities use a GIS_ID as a unique identifier. GIS_ID's are unique to the method of data collection, location or shape, and type of feature. In SPLAT, location means all the information necessary to describe the shape and location of a feature. Type of features are lines, points, areas, routes, or regions. Methods of data collection include all of the elements from the EPA's MAD codes.

An example of using more than one GIS_ID for one entity would be having addressed-matched coordinates and map interpolated coordinates for the same entity. Because there are two different methods used to obtain locational coordinates, there would be two feature types.

The matrix below displays various GIS_ID/ENTITY combinations:

Method        Location       Feature Type   GIS ID

Same          Same           Same           Same
Same          Different      Same           Different
Same          Same           Different      Different
Same          Different      Different      Different
Different     Same           Same           Different
Different     Different      Same           Different
Different     Same           Different      Different
Different     Different      Different      Different

PREFERRED LOCATION SELECTION METHODOLOGY

The following sequence describes the logical flow for selection of a preferred coordinate pair from a group of coordinate pairs for the same feature.

1. Check accuracy_value in MAD table. All null records will be populated with a valid value from the accuracy lookup table.
2. Convert accuracy_value to raw_score:

   if accuracy_value le 20 then raw_score = accuracy_value
   else raw_score = 15 + (accuracy_value ** 0.61)

3. Apply modifier weights: verify_weight - scores of 0%, 5%, 10% and 15% depending on the precision of the test or user_def_weight - 0%, 5%, 10% and 15% for those locations known to be highly valid.

   modified_score = raw_score - (raw_score * verify_weight) - (raw_score * user_def_weight)

4. For duplicate epa_id's, lowest modified_score record, calculate preferred = y

CONCLUSION

The system will continue to be evaluated, tested and documented. Upon completion of this process, the final product will be made available for distribution to other Regions and interested state agencies.

ACKNOWLEDGEMENTS