A Cultural Resource Inventory: Philadelphia's Fairmount Park Expands its Resource Preservation Strategy Through GPS/GIS Integration
Submitted By:
Amy Freitag, Fairmount Park Commission/Fairmount Park Historic Presentation Trust, Inc.
Peter Godfrey Jr., RLA/AICP, Camp Dresser & McKee
1.0 INTRODUCTION
This effort was initiated by the Fairmount Park Commission in support of the Natural Lands Restoration and Environmental Education Program (NLREEP) and was funded through a grant from the William Penn Foundation to the Fairmount Park Historic Preservation Trust. The purpose of this project was to generate geographic information system (GIS) coverages of the existing "man-made" structures in the Academy of Natural Science (ANS) designated natural areas within Fairmount Park. The park system encompasses a total of 8,900 acres with the designated natural areas accounting for 5,400 of these acres.
1.1 Cultural Resource Inventory
In mid-1998, Fairmount Park, working in partnership with the Fairmount Park Historic Preservation Trust, Inc. began initial work on establishing a functional GIS that would contain spatial and tabular data regarding all buildings ("things with roofs or were meant to have roofs") within park property. The GIS platform for this system was the Environmental Systems Research Institute, Inc. (Esri) ArcView 3.1 software.
During this effort, the Park received the following existing city-wide GIS data from the Mayors Office of Information Services, the Streets Department, and The Philadelphia City Planning Commission:
These coverages were then used to establish the first park GIS project. While there was data regarding buildings within the park, the data was spatial in nature and not attribute rich in regards to building specific information. Information regarding park buildings (dates, historic certifications, maintenance records, uses, lessees, etc.) was available from a variety of sources, none of which were digital in format. Through design and implementation of a database to manage building related data (Microsoft Access '97) and the integration of this data with the spatial GIS data, the first comprehensive building resource inventory for Fairmount Park was created.
This spatial and tabular data is now being used to assist in improving the long term preservation of park structures through day-to-day maintenance operations as well as prioritizing cyclical maintenance activities. The GIS is also being used to improve stewardship, increase collaboration between the park and other city departments, and improve strategies for historic preservation and overall cultural resource planning. This project was known as the Cultural Resource Inventory project.
1.2 Cultural Resource Inventory: Man-Made Features
The project described in the following report is an extension of the Cultural Resource Inventory project and the intent is to gather and integrate both spatial and tabular data regarding "man-made" features within the ANS defined natural areas.
The natural areas have been designated through vegetation codes that have been field verified and will be explained within this document. Unlike the initial Cultural Resource Inventory project which already had the buildings identified in the GIS, this project had minimal initial spatial data regarding "man-made" features (roads, fences, trails, walls, etc.). As a result, the use of Global Positioning System (GPS) technology was employed to locate man-made features within these natural areas.
Project Equipment and Methodology
The GPS hardware used for this project was a Trimble AgGPS 132 DGPS receiver. The type of GPS software deployed for this effort (DynaMo GIS) allowed for the use of existing park and city-wide GIS data to be loaded into a Fujitsu Stylistic 2300 pen-based computer as a basemap to be used for locational verification. The software and GPS would then display a real-time cursor of where the field data gathering staff was located using the basemap as a reference. The software also provided the functionality of populating feature attributes such as feature type, material, and condition while in the field.
The GPS field gathered data was then imported into the park's existing GIS. The field gathered feature attributes were supplemented with data such as name aliases, owner, builder, maintenance responsibility, etc. This data was found in either existing park hardcopy records or through institutional knowledge.
As a result of this effort, the park has now created another "module" of GIS data regarding the park system and its facilities. As with the previous GIS effort, this data will be used for tracking and locating features within the various parks, prioritizing maintenance, and providing needed information for park managers and users alike.
1.3 Report Organization
This report is intended to be used by park staff as a reference document for this project and will detail the following:
Section Two: Spatial data needs, uses, and accuracy
Section Three: Existing spatial data, accuracy and integration
Section Four: Global Positioning System and field data gathering methodology
2.0 INTRODUCTION
This section of the report will identify the spatial data needs for the second phase of the Cultural Resource Inventory project. It is important to document this information to provide an understanding of the overall project's intent as well as for future reference. Prior to this project effort, existing park GIS data has been documented as part of the initial Cultural Resource Inventory project and this report, dated March 1999, should be used for reference as well.
Both these project phases have had a tremendous head start on initiating park GIS efforts resulting from both the GIS and tabular data available on a city-wide scale. There has been both technical and data gathering assistance in the project efforts from the following city departments and other entities:
Without this true enterprise-wide collaboration and data sharing, both of these efforts would still be in the initial GIS data conversion process.
As with any information system (IS) planning and development, the key initial factor to consider is what data are needed, how will they be used, and what are the accuracy requirements. For this project, representatives of the following stakeholders were consulted:
As an initial resource for identifying the project "needs" and the potential spatial features to include in this project, the Fairmount Park NLREEP grant "narrative" was consulted.
The relevant section of the narrative (page 10) reads as follows:
Existing Conditions Survey: Structures Survey (Within the Natural Lands)
The past uses of the large natural parks can still be seen in the remnants of mills, mill dams and walls. But the most outstanding evidence of human presence are the trails, bridges, shelters, retaining walls, fences, and even golf courses constructed by the thousands of men employed on relief by the Commission under the Local Works Division and Works Progress Administration projects. Less evident as the works of man, are the thousands of trees planted through out our "natural lands". The Commission, in managing the use and maintenance of the natural landscape, must consider the historical significance and utility of these structures, as well as whether they are themselves intrusions or assets to the preservation of the natural landscape. A shelter may have to be removed, if located in an area slated for woodland or meadow restoration. Conversely, a new structure, such as a fence, may be needed to reduce off trail use. We will:
This document initially identifies the needs and purpose of the project. During further discussions, the following "needs" were also identified:
After defining the overall uses of the information to be gathered through this project, the specific features and their relevant attributes had to be defined.
2.2 Cultural Resource Inventory : Man-Made Spatial Features
This component of the project was focused on identifying the "man-made" features to be field gathered and the specific information regarding these features that would be necessary for system applications. This process involved four (4) meetings between project participants and resulted in a list of 46 individual features presented below:
Each designated feature was defined by its GIS class (point, line, or polygon), original data source (field gathered, PWD planimetrics, digitized), and specific attributes. The attributes were divided into categories of data that could be gathered while in the field (i.e. feature type, materials, condition, width, etc.) and data that would need to be entered through a PC (i.e. age, aliases, maintenance responsibility, historical certification, etc.). The final data dictionary with full definitions of all the above features can be found in a companion document titled: Data Dictionary Phase Two - Man Made Features.
Once the needed features and their respective attributes were defined, existing city-wide GIS data was imported in support of this project effort. The following section details this effort as well as discusses spatial accuracy requirements.
3.0 INTRODUCTION
This section of the project report will detail the existing GIS spatial data that were imported into the Cultural Resource Inventory GIS as part of this effort. The spatial data provided to the park was used for locational verification during the data collection efforts, as well as a reference for the overall extent of project efforts. As mentioned earlier, only the natural areas, which encompassed 5,400 acres, were to be inventoried during this project. The GIS data that was provided was either supplied in Esri ArcView format as shapefiles (.shp) or Esri Arc/Info export files (.e00) which had to be imported into ArcView and then converted into shapefile format.
3.1 Existing GIS Spatial Features and Accuracy
Previously established spatial data were imported into the park's GIS environment from the following sources:
The hardcopy maps were used for field reference and locational verification while gathering data. The latter two sources of digital spatial data were used for basemap generation and/or field verification of existing features as shown on the PWD planimetrics. The ANS data was used as a guide to determine the areas where field data was to be gathered during this project. Descriptions of the two digital data sources are provided below.
The planimetric data supplied by PWD was referenced to the following coordinate datum:
The planimetric data relevant to this project was "clipped" out of citywide GIS coverages (using Arc/Info functionality) by MOIS and new coverages were generated containing only data falling within the six watershed and one estuary parks. This data was then delivered to the park for use as base data while collecting field data.
The positional accuracy of the planimetric feature data is "sub-meter" (+/- 39" from its digital representation). This positional accuracy was determined to be appropriate for the project purposes and was used to determine specifications for the GPS equipment that would be used for gathering field data.
The following GIS coverages were delivered to the park in Arc/Info export format (.e00), imported into the Cultural Resources GIS and then converted into ArcView shapefile format (.shp).
3.1.2 Academy of Natural Sciences GIS Data
The data supplied by ANS was referenced to the following coordinate datum:
The ANS vegetation class spatial data brought into the park's GIS was established initially by Munro Ecological Services, Inc. from 1" - 800' scale aerial photographs. The identification of vegetative polygons grouped by disturbance class was performed by Munro and then these hardcopy polygons were digitized and ortho corrected ("rubber sheeted") by the ANS Patrick Center for Environmental Research staff to graphically fit within existing roads and hydrology base layers.
The base layers used for this ortho correction (roads and hydrology) were obtained from the Philadelphia City Planning Commission (PCPC). These coverages are maintained by the PCPC and are defined as follows:
The following coverages were delivered to the Park from ANS:
These coverages consisted of vegetation polygons encompassing both areas within the specific park area, as well as some areas that extended outside the park area. The polygon attribute table (PAT) for these coverages contained codes for the following polygon attribute fields:
The key polygon attribute field (PAT) for this project was the natural areas that were coded with the following values:
These values were used to determine the extent of the project data gathering effort. The ANS natural areas were queried and those areas that were defined by the above values were used to generate new shapefiles displaying the project delineation areas. The ArcView shapefiles maintained the same naming convention (cobbs, eastwest, fdr, penny, poquess, tacony, wissa) as the Arc/Info export files.
The use of Global Positioning System (GPS) technology was integrated into the data gathering methodology for this project. Many of the features that were to be inventoried within the natural areas did not have any available spatial data. While there was some data contained on hardcopy maps (1983 Park Maps produced by Aerial Data Reduction), these were used strictly for field reference purposes. All spatial features represented within this project had their spatial location and certain attributes entered into the system through the use of GPS equipment and field data gathering software.
This section of the report will discuss basic GPS concepts and then describe the methodology implemented by the park staff for field data acquisition.
4.1 GPS Concepts
The use of GPS technology relies on a constellation of satellites transmitting radio signals around the world that can be used to calculate coordinates any time of day and in any type of weather. Currently, there are 27 satellites orbiting the earth at an altitude of approximately 11,000 miles.
In general detail, some important GPS concepts are discussed below.
4.1.1 Triangulation
The GPS measures the distance from the satellite to the antennae of the receiving unit that is usually mounted in a pack on the field crew's back. The accuracy of the physical location of the GPS unit is dependent upon satellite triangulation. Triangulation refers to the number of measurements that were taken from satellites regarding a certain point on the earth's surface. The best measurement reading would be a "3D" which means that four satellites were used for geographic location and the X (latitude), Y (longitude), and Z (height) coordinates were all gathered for that point. Often, the only available measurement is a "2D" which indicates that there were only three satellites available and in this type of measurement, the Z coordinate is not measured. This will impact the accuracy, but not sufficiently enough to mandate only the capture of3D readings.
4.1.2 Differential Correction
Because of the nature of the satellites and their potential use (they are "owned" by the U.S. Department of Defense), the federal government distorts the transmitted signals so that without proper "correction" the horizontal measurement (X, Y) is only accurate to within 100 meters (300 feet) of the actual location. In order to get a true horizontal reading, there are two types of "differential" correction methods: real-time and post-processing.
Real-time differential correction means that the signal is corrected at the time of data gathering by either another satellite or a coast guard beacon service. Post-processing correction means that the field gathered data is taken back to the office, downloaded onto a computer and processed.
The best method is real-time, especially in this project because the data being gathered can be validated in the field and another reading can be taken if necessary. In post-processing, we do not know if our data is positionally accurate until it is processed. Another trip to the field may be required because the data is inaccurate, but one would not know this until having returned to the office and processed it.
Real-time differential correction can come from either a commercial source, that requires a subscription to either the Landstar or Omnistar service, or from the US Coast Guard. In this project, we are using both the Landstar service and the US Coast Guard beacon (Sandy Hook, New Jersey), depending on geographical position and the terrain. In the case of the Landstar service it is necessary to have a clear 15 degree southern sight line in order to receive the differential signal. This 15 degree sight line equates to a 15 degree angle from the ground plane into the sky.
Because GPS relies on the signals and measurements of numerous satellites, the accuracy of the measurement can be impacted by the satellites position in the sky. While we may be able to get readings from 4 satellites (3D), if they are in close geometric position, as opposed to spread out, the accuracy will suffer. In general, and contrary to intuition, a low DOP reading (3.2) indicates that there is a high probability of accuracy, and a high DOP reading (8.4) indicates a low probability of accuracy.
The best overall indication of satellite geometric accuracy is called the position dilution of precision (PDOP). While the horizontal dilution of precision (HDOP) and the vertical dilution of precision (VDOP) are also important, for our purposes, the PDOP was the best indication of accuracy. As with the DOP, a low PDOP indicates a higher probability of accuracy.
4.1.4 Other GPS Issues and Factors
Listed below are some other issues and factors that need to be considered when using GPS equipment:
The next section of this report will discuss the project specific equipment that was acquired for GPS data gathering.
The following criteria was used to determine the best GPS equipment for the project:
Based upon these criteria, the following equipment and software was acquired from Geo Informational Service, Inc located in Lebanon, New Jersey:
The GPS receiver acquired for the project was the Trimble Ag132 sub-meter receiver. This receiver allows for the use of either Coast Guard beacon or satellite differential correction. There is a 12 channel GPS receiver with a built-in display and keyboard. Peripheral equipment including antennae and magnetic antennae mount (for mounting onto vehicles) was included in the purchase. A subscription (one year) to the Landstar North American satellite differential correction service was also obtained.
In order to visually validate data gathering, use existing city and park GIS data, and facilitate data entry, a pen-based computer was acquired. This piece of hardware would save time on data entry and quality control.
A Fujitsu Personal Systems, Inc. Stylistic 2300 model was acquired for this project. The tablet came loaded with Windows 95, 233 MHz Pentium MMX processor with 512 KB external cache, 4.0 GB hard-drive, and an internal 56k V.90 modem. Also acquired was a Fujitsu keyboard and external floppy disk drive. One of the key aspects of this piece of equipment was it's 8.4" display screen and color display. A one year display damage service plan was also purchased. Also acquired as part of the Fujitsu configuration was a network card for future data transfer over the park's Local Area Network (LAN).
One of the critical aspects of any field data gathering effort involving GPS is the type and functionality of the software used to store and manipulate the data. For this project, DynaMo GIS Version 1.2.1 modular field data collection software was acquired.
This software provides a live GPS position cursor that is displayed on the computer screen to verify field data locations. The functionality also includes data entry screens for real-time feature attribute gathering, displays existing GIS shapefiles for location verifications, and all data is collected in native Esri ArcView shapefile format. This format allows the field collected data to be immediately downloaded into the park's GIS environment.
Once the equipment was acquired and training on the use of the hardware and software was completed, data collection began in earnest. In order to facilitate efficient data acquisition, the following steps needed to be completed before gathering field data.
The software provided the opportunity to gather and populate feature attributes in the field. In order for this to function properly, "data dictionaries" needed to be created for each unique feature. These data dictionaries contained fields and value options for data. These fields were a result of data modeling sessions with FPC and NLREEP. An example of a data dictionary is shown to the right. This data is then loaded onto the Fujitsu pen-based computer. It is important that when naming all shapefiles, they do not exceed 8 characters for this will have an impact on the Microsoft Access database.
The Park has acquired numerous GIS ArcInfo coverages (which were converted to ArcView shapefiles) or ArcView shapefiles to be used as reference base maps. The key shapefiles to have loaded into the Fujitsu are the following:
As mentioned earlier, it is very important to know both the terrain that will be visited, as well as the actual satellite constellation to assure accurate readings. This accuracy is related to the position dilution of precision (PDOP). The lower the actual PDOP number, the more accurate the readings.
The example below is taken from the Trimble website using the actual values that need to be entered for mission planning:
In the example, the areas that are circled indicate times when the PDOP rises to a high level indicating a lower probability of accuracy.
The Trimble website is an easy and free source for mission planning, provided the Park has internet access. The website address is www.trimble.com/cgi/satview.cgi.
Before going into the field, all batteries (GPS and Fujitsu) should be charged and functional. Charged back-up batteries should also be taken into the field for replacement.
Once the data dictionaries have been built and all equipment has been checked, field data gathering can begin. The following steps must be taken upon arriving at the data gathering site after turning on and connecting the GPS unit to the Fujitsu notebook:
Once the data has been gathered, it is time to transfer it to the PC and into the ArcView environment. This is accomplished by taking the following steps:
Note: The initial steps in this process involving discs will not be needed once the data can be transferred via the parks network through use of the network card.
When going back into the field for further data collection:
4.4 Conclusion
This project has allowed the Fairmount Park Commission to further integrate technology into the management of its multitude of resources. As time passes, the creation of data as a result of this project will serve not only the park in the management of its cultural resources, but also other city agencies in gaining an understanding of the natural and man-made features that are encompassed by the Fairmount Park system. This will include the Philadelphia Water Department, The Public Art Office, the Streets Department, and a variety of other city entities.
This project will eventually capture all man-made features within the entire Park system, not just the ANS delineated natural areas. This will serve to enhance the GIS data available encompassing Fairmount Park, its resources, and its heritage.
Amy Freitag, New York City Department of Parks and Recreation, (212) 360-8203, Friday@parklan.cn.ci.nyc.ny.us
Peter Godfrey Jr., Camp Dresser & McKee, Philadelphia, (215) 636-0600, godfreyp@cdm.com