Using GIS and Digital Imagery for Environmental Analysis: A Case Study of the Relicensing of a Large Hydroelectric Project

Ed Alkiewicz and John Wingfield

----------------------------------------------------------------------------------------------------

Using GIS and Digital Imagery for Environmental Analysis: A Case Study of the Relicensing of a Large Hydroelectric Power Project

Abstract

The New York Power (Authority) NYPA is undertaking the re-licensing of a large hydroelectric power project located on the U.S - Canadian border and constructed in the 1950's. In order to assess the past and present environmental impacts of the project, natural resources and processes must be identified and studied.

NYPA has developed an in-house GIS and acquired data layers needed to support environmental analysis and other resource management applications.

Due to the inadequacy of existing mapping a decision was made to create digital orthophotographic base maps of the 38,000 acre project area. After obtaining current low altitude aerial photography, Black and White and False Color Infrared imagery were acquired and digitized. The higher resolution B&W imagery was used to identify building foot prints and encroachments, while the Color Infrared imagery was used for environmental analysis, particularly the identification of vegetation signatures. The imagery was used to create data layers including wetland delineations, land use/land cover, property maps, environmentally sensitive areas, topography and cultural resources locations. Results of environmental analyses and real estate queries were also plotted for inclusion in licensing documents and presentation at public meetings and workshops. The database has also been loaded onto a laptop computer running Arcview that is linked with a GPS receiver to document structures and other geographic data in the field. This applications is being used to support a shoreline user permit system. All the map submittals required by the Federal licensing process for this project will be prepared on a GIS.

Data layers were created in ArcInfo and loaded into Arcview workstations to allow for analysis by environmental scientists and resource managers. Arcview scripts were written to allow for easier access to the entire database by casual GIS users.

-

-------------------------------------------------------------------------------------------------------------

Project Description

The international St. Lawrence Power Project is one of the largest non-federal hydroelectric facilities in North America. The project straddles the international border between the United States and Canada on the St. Lawrence River. The project is jointly operated by the New York Power Authority (NYPA) on the United States side of the border and by Ontario Hydro on the Canadian side of the border. The project was completed in 1959. The U.S side of the Project is licensed to operate by the Federal Energy Regulatory Commission. This license will expire in 2003.

The principal facilities of the St. Lawrence Power Project are the Moses-Saunders Power Dam, Long Sault Dam, associated dikes on both sides of the river, and the Iroquois Dam located approximately 25 miles upstream from Long Sault Dam. Navigation locks and the enlargement of sections of the river channel as part of the St. Lawrence Seaway permit deep water navigation between the Atlantic Ocean and the Great Lakes.

Electric generation occurs at the Moses-Saunders Dam. There are 16 turbines on the United States side and an equal number on the Canadian side. The NYPA's turbines are capable of producing greater than 900,000 kilowatts of electricity.

The impoundment formed by the Power Project is Lake St. Lawrence. This lake stores approximately 244 billion gallons of water and has a surface area of approximately 39,100 acres (61 square miles). The average flow rate at this location is approximately 243,000 cubic feet per second. The head (vertical drop) at the Power Dam is approximately 81 feet. The next impoundment below the Power Project is Lake St. Francis which is approximately 13 miles downstream.

There are approximately 100 miles of shoreline exclusive of the islands within the area. The Project boundary contains approximately 14,000 acres of land including islands.

Database Design

Base Mapping

From the start, NYPA's GIS was envisioned as a multi-purpose, multi-discipline tool. One of the keys to providing this kind of flexibility is the judicious choice of appropriate base mapping.

The most comprehensive existing map set of the St. Lawrence Project prior to the relicensing effort was Exhibit K to the Federal Energy Regulatory Commission (FERC) license. It was produced from conventional transit and plane table surveys and photogrammetry in the early to mid 1950's. While its scale (1"=500') is reasonable for its intended purpose as the official project area map it exhibits the same lack of scale consistency typical of manually drafted maps of its type. Because the survey 2

work which was the basis for Exhibit K occurred over a forty-plus year period preceding construction, it was georeferenced to three different and now obsolete coordinate systems.

While changes in NYPA's ownership and other land rights over the years were reflected in periodic revisions, the accuracy of the revisions was inconsistent and no other changes were made to the mapping to reflect the topographic and planimetric as built condition of the project. These and other reasons rendered Exhibit K unsuitable as a GIS base map.

In 1987, NYPA contracted with the surveying and photogrammetric firm of Lafave, White, and McGivern of Theresa, New York to produce a topographic map of the majority of the project area in preparation for a surplus land study. The map scale is 1:1200 and the contour interval is 2 feet. NYPA had no digital mapping capabilities at that time so the mapping was delivered in hard copy form only, consisting of 224 individual 30" by 42" mylar sheets. This map set, while highly accurate and fairly current, does not encompass the entire project area. Since it was produced in a digital environment, however, it was possible and desirable to extract many of the map features for use in the GIS.

Balancing Scale, Accuracy, and Cost

NYPA's experience with GIS began in 1989 with a pilot program addressing transmission right-of-way management. The cost, scale, currency, and flexibility "curves" all crossed at 1:2400 scale digital orthophotography as a base map choice. Ortho imagery clearly provides the most information for the least cost of all the existing alternatives. It provides the third dimension (a digital elevation model) as a by product of its production. At appropriate scales data can be vectorized "heads up" on the screen for specific purposes, providing the intelligence that can be lacking in raster data as-needed while serving as a kind of "library" of features which can be added later. In our experience, the simple expedient of substituting human intelligence for vector/polygon symbology or attributes ("machine intelligence") is sufficient in most of our uses to obviate the need for comprehensive vector mapping. In addition, updates are far simpler and less expensive than with other base map alternatives.

In keeping with our multiple use philosophy we have developed three different GIS base maps of the St. Lawrence Project. Although relicensing has been and continues to be the primary driver of GIS development at NYPA's hydro facilities, projected uses also include land management, environmental/recreational planning and management, civil works maintenance planning and tracking, emergency action plan map updates, and transmission line right of way management.

Although the truism that a high precision base map can easily serve lower precision needs applies here as well, the cost penalties involved with small scale mapping the entire project vicinity are too great to ignore. Therefore the core project area base map, where the highest precision is required, is 1:2400 scale black and white digital orthophotography, with a pixel size of approximately 1.25' on the ground. This map is used primarily for real estate/land management and civil engineering and maintenance tasks. A second base map, covering a much larger area, is 1:12000 scale false color infrared digital orthophotography with a pixel size of approximately 1 meter. This map was developed primarily for environmental analysis and related relicensing support. Yet a third base map is composed of raster images of the New York State DOT's 7.5 minute quadrangle map series, which is a partially updated planimetric derivative of the familiar USGS's 7.5 minute quadrangle series. This map is used for all those purposes where the ortho imagery is too detailed or insufficient in area to serve a particular purpose.

All data in NYPA's GIS is registered (georeferenced) to the appropriate state plane coordinate zone. In the case of the St. Lawrence project, this is the New York Coordinate System, East Zone. The horizontal datum is the North American Datum of 1927. This map projection (coordinate system) was chosen because of the wealth of data in NYPA files which is tied to the state plane system and for maximum compatibility with existing external data such as tax maps. It has the added benefit of minimizing the distortion inherent in any map projection as the state plane zones were designed to hold such distortion to a maximum of one part in ten thousand. Data such as the above referenced DOT. quadrangles which were registered to a different projection (Universal Transverse Mercator Zone 18) have been converted to state plane in-house using Arc Info and the PROJECT command.

The Third Dimension

A number of the uses of GIS at the NYPA require elevation data. Because of the availability of the photogrammetric data noted earlier and our choice of digital orthophotography as a base map we are well situated in this respect. Development of any orthophotograph (digital or hard copy) requires elevation data sufficient to detect and compensate for terrain displacement in the raw photography. This data set is typically somewhat generalized compared to the kind of detailed contour data usually prepared for design purposes. It is, however, entirely adequate for general planning purposes and is typically significantly more detailed and current than that derived from the USGS 7.5 minute quadrangles. NYPA's digital orthophotography supplier, Photo Science of Gaithersburg, Maryland, supplies the elevation model as x,y,z point data in ASCII format which is then developed into ArcInfo compatible TIN and/or LATTICE format in-house.

A digital elevation model also developed from the aforementioned two foot contour data. This data was "spliced" into the ortho-generated digital elevation model. As a result, the St. Lawrence digital elevation model is comprised of three "concentric rings" of elevation data which is of highest accuracy in the core area derived from the two foot contour photogrammetry, surrounded by data generated in the development of the 1:2400 digital ortho base map which can be relied upon to produce accurate, if somewhat generalized, five foot contours, which is in turn surrounded by data resulting from the 1:12000 digital ortho base map production which is estimated to produce reliable ten foot contours. As with the horizontal data, a number of now obsolete vertical data have been used in since the initial design of the St. Lawrence project. The differences between these data amount to no more than a foot but can be significant when comparing inundated areas along the relatively flat shorelines that characterize much of the project area. The elevation data in the St. Lawrence database are registered to the International Great Lakes Datum of 1955.

The most recent component of the third dimension to be incorporated in our database is bathymetry. The data we are currently using was digitized from published NOAA navigational charts. While adequate as a general characterization, the source scale (typically about 1:80000 with 1:20000 detail areas) does not support the kind of detailed analysis that may be necessary for some of the planned environmental studies. We are currently researching the feasibility of obtaining and digitizing the "field sheets" from the Canadian and U.S. federal hydrographic agencies. Since these data often are sparse to non-existent in shallow near-shore areas, we are evaluating various options for filling in those areas. These options include conventional surveys, photogrammetry, and SHOALS (Scanning Hydrographic Operational Airborne Lidar Survey, an airborne remote sensing technique using a helicopter mounted laser array (LIDAR or Light Imaging Detection and Ranging) which can gather data suitable for the production of half-meter bathymetric contours to a depth that may approach 60 feet given the expected water clarity on the St. Lawrence River. We plan to append this bathymetric data to the terrestrial elevation data so as to produce a seamless, comprehensive model of the terrain above and below the water surface for the entire project area.

As evidenced by the foregoing discussion, a good deal of thought (and some trial and error!) has been put into building a base map supporting disparate applications. The false color infrared base map is particularly useful for environmental analyses due to the possibility of correlating its component spectra with vegetation types and other environmental phenomena. For this reason, and because of its greater coverage area, this is the base map used to support, and in some cases to produce, the various environmental applications and data layers developed to date.

Environmental Data

Descriptions, including location and extent, of the following environmental resources were deemed necessary in order to conduct the environmental assessment of the impact of Project operations. The resulting data layers were created, some using remote sensing techniques and stored in NYPA's GIS in order to allow spatial analysis of the impacts on specific resources:

In 1994, a land-use inventory was conducted of the 38,700 acres of lands and waters within the Project. The inventory was based upon photo interpretation of 1993 color infrared aerial photography; data were digitized and stored in NYPA's Geographic Information System (GIS). Land use and cover types were delineated using the Cornell University Land Use and Natural Resource (LUNR) classification system (Cornell University, 1970).

There are 22,637 acres of lacustrine habitats at the Project. Lacustrine habitats include wetlands and deepwater habitats associated with permanently flooded lakes and reservoirs that are at least 20 acres in size and which lack a dominance of trees, shrubs, and persistent emergent plants (Cowardin et al. 1992). Lacustrine habitats within the Project are primarily deepwater habitats. These areas are generally below the deepwater boundary of emergent wetlands but also contain beds of submerged aquatic vegetation.

Riverine habitats include wetlands and deepwater habitats that are contained within naturally or artificially created channels and which lack a dominance of trees, shrubs, and persistent emergent plants (Cowardin et al. 1992). The 1,184 acres of riverine habitats within the Project are characterized by slow moving water, mud or sand substrate, and beds of submerged aquatic vegetation.

Palustrine wetlands are non-tidal wetlands dominated by trees, shrubs or persistent emergent plants as well as non-vegetated wetlands that are less than 6.6 ft deep and less than 20 acres in size (Cowardin et al., 1992). Forested wetlands are the primary palustrine wetlands with a total of 1,081 acres (2.8% of Project lands and waters). Forested wetlands are characterized by woody vegetation that is 20 ft tall or taller (Cowardin et al., 1992). Almost all of the forested wetlands at the Project are classified as broad-leaved deciduous. Emergent wetlands, which are dominated by erect, rooted, herbaceous plants, cover 541 acres (1.4%). All emergent wetlands are classified as persistent. Scrub-shrub wetlands include areas dominated by woody vegetation less than 20 ft tall including true shrubs and young trees. Scrub-shrub wetlands cover 344 acres (0.9%). Palustrine unconsolidated bottom wetlands, which lack large stable surfaces for plant and animal attachment, account for 113 acres (0.3%).

Submerged aquatic vegetation includes both algae and vascular plants that grow below the surface of the water. The plants are either attached to the substrate or float freely above the bottom (Cowardin et al., 1992). The submerged aquatic vegetation community occurs in Lake St. Lawrence to water depths of 23.0 ft. The distribution of this community may overlap in shallow waters with emergent wetlands, which are typically restricted to waters less than 6.6 ft deep (Cowardin et al., 1992). Submerged aquatic vegetation communities provide habitat for benthic invertebrate production, fish feeding and spawning, and water bird foraging.

The occurrence and development of aquatic plant communities in the St. Lawrence River and Lake St. Lawrence results from the interaction of a variety of environmental influences: water depth, drawdown, chemistry and turbidity; substrate depth and texture; and the physical action of waves, current and ice scouring.

Twenty-two types of soil are found within the Project Area ranging from very well drained fine sands to frequently flooded muck. These soils were formed from glacial till, glacial outwash, or lake sedimentation .

Upland plant communities, at the Project were generally classified as open or forested systems. The open-uplands system includes five communities: 1) agricultural cropland, 2) agricultural pasture, 3) successional old field, 4) successional shrubland, and 5) developed or residential . Agricultural cropland and agricultural pasture occur commonly throughout the Project . The successional old field and successional shrubland communities occur commonly throughout the Project where farms were abandoned. Due to the past and ongoing agricultural uses on Project lands, mature upland forests are relatively uncommon. Notable exceptions include relatively small stands that occur on the larger islands.

Significant Ecological Communities

In New York State, significant communities include all occurrences of a rare community, regardless of element quality, and all high quality examples of a community, regardless of rarity.

A 1995 study identified eight occurrences of four significant ecological communities within the Project Area: 1) cobble shore wet meadow, 2) maple-basswood rich mesic forest, 3) red maple-hardwood swamp, and 4) silver maple-ash swamp.

A 1995 study compiled existing wildlife occurrence records in Project lands and waters and completed field surveys within representative areas for each major ecological community (hereafter referred to as "habitat"). Wildlife species were categorized by the habitats in which they are expected to occur. These habitat delineations were subsequently digitized and stored as a GIS layer. A total of 17 species of amphibians, 14 reptiles, 215 birds and 40 mammals may occur in Project lands and waters. Of the birds, 168 species may breed within the area of the Project . The fairly high species diversity results from the diversity of habitats as well as the abundance of water resources and wetlands. Specific wildlife communities are associated with each habitat, while some more widely distributed species such as the white-tailed deer and blue jay use a mixture of habitats.

The predominant habitat types at the Project are the lacustrine and riverine wetland and deepwater habitats which cover 23,821 acres (61.4% of Project lands and waters). These habitats support a fairly large number of amphibian, reptile and bird species, but relatively few mammals.. These habitats, particularly areas supporting submerged aquatic vegetation beds, provide an important resting and feeding habitat for spring and fall migrating waterfowl, gulls, wading birds, and shorebirds. They are also used by breeding waterfowl, gulls and wading birds.

Palustrine wetlands support a diverse assemblage of amphibians, reptiles, birds and mammals with the greatest concentration of reptiles and amphibians in Project lands and waters.

Open upland habitats at the Project include agricultural and successional communities. Field habitats provide important year-round habitat for raptors, such as the red-tailed hawk and American kestrel.

Most of the Project Area is in a natural state of maturing forests, wetlands and successional old fields, benefiting many wildlife species. These lands are available for public wildlife uses including hunting, viewing, bird-watching, photography, artistry and environmental education.

A 1995 inventory of rare species in Project lands and waters documented the presence of five threatened and endangered wildlife species and two species of special concern. The locations of these species within the Project were digitized. Access to this layer is restricted due to the sensitivity of the data.

Some of the shoreline along Lake St. Lawrence (the Project impoundment) consists of steep bluffs and soils which are more prone to erosion than other portions of the Project. Shoreline stabilization has been accomplished in selected areas where the shoreline has eroded inland and has approached the Project Boundary. The shoreline of the Lake is inspected each year for erosion, and areas of specific concern documented. Areas of visible erosion and sedimentation are visible in various digital and conventional orthophotography.

A variety of habitats is available for fish spawning in Lake St. Lawrence and adjoining waters. Spawning habitats include tributary streams and rivers, marshes, shallow embayments, shoreline areas, shoals, and large embayments. Some of these features can be identified from the orthophoto base maps and other data layers.

Environmental Analysis

Introduction

The issuance of a new license for the Project constitutes a major federal action that could

significantly affect the quality of the existing human and natural environment. Consequently, an Environmental Impact Statement (EIS) will be prepared to address the impacts of the operation of the Project on fish, wildlife, botanical resources as well as water, recreation, land use, cultural and socioeconomic resources. In support of the EIS and in order to address issues brought up by interested parties, approximately thirty environmental studies will be undertaken during the relicensing process. The data collected from these studies will need to be analyzed in order to characterize the natural resource and evaluate the impact on those resources.

The following is a partial list of issues that will be addressed during the relicensing process and for which GIS analysis would be used:

Since water level fluctuations created by natural phenomena or Project operations may contribute to shoreline erosion, the potential effects of these actions must be evaluated. The studies for this issue will identify causes of erosion and will characterize the rate of recession and location of the erosion prone areas. Additionally, areas of sediment deposition located downstream of the erosion will also be identified. Existing natural color aerial photography and the digital orthophoto base map will be used at a reconnaissance level to delineate eroded sections of the shoreline and corresponding sedimentation areas. Digital soils and vegetation information will also be investigated to assist in determining causal factors. The proximity and aspect of the nearby St. Lawrence Seaway shipping channel will also be located in order to assess the effects of wave induced action.

Water level fluctuations related to natural phenomena or Project operations may affect aquatic habitats used by fish and other organisms. In order to undertake this analysis bathymetric data will be collected and converted into a model of the lake bottom. This model in conjunction with water level data collected from gauges will be used to identify the level changes and extent attributable to Project operations. The location of submerged aquatic vegetation beds and bottom substrate characteristics delineated from aerial photography will be compared to water level fluctuations. The magnitude and extent of the water level fluctuations will be characterized in order to determine any impacts on aquatic habitats.

Water level fluctuations may also affect terrestrial habitats. Emergent wetland areas and associated species that may occur near the shoreline in shallow water areas may be influenced by slight changes in water level. Wetland areas delineated from color infrared aerial photography will be overlayed with water level change extents in order to identify affected areas.

In addition to identifying the location and occurrence of threatened and endangered species, habitats used by these species will be analyzed to determine if any impacts are occurring due to water level fluctuations or land management practices. For example, this may entail an analysis of the extent of water level fluctuation superimposed on know nesting site locations. Additionally, areas where land management practices, such as agricultural use, are allowed but may be affecting threatened and endangered species habitats would be identified.

The existence of the Project and its operation affect land uses within the 14,000 acres of land within the Project boundary. The land is necessary as a buffer zone to protect from potential flooding along the shoreline due to water level regulation. In order to prepare a Project wide land management plan a characterization of all lands within the boundary is required. All lands will be categorized according to suitability criteria (e.g. needed for power generation, recreation or open space) and evaluated to determine whether the lands need to be retained or possibly reallocated as surplus to Project operations. This analysis will seek to identify resource constraints on the lands, such as, significant ecological communities, threatened and endangered species habitat or open space and recreation. Additionally, land use outside the Project boundary will be categorized to determine effects on land use designations inside the boundary.

Development of an Analytical Tool

Although the Unix workstation version of ArcInfo is a very robust software package, it is difficult to impossible for casual users to access. NYPA's GIS technical staff is hard pressed to handle all data requests and special projects in a timely fashion, and the ever more competitive electric utility market makes staff expansion unlikely in the foreseeable future. The solution to this problem at NYPA was originally envisioned as a custom graphical user interface (GUI) developed in house. However, the functionality built into Arcview by Esri has rendered the continued development of our in-house GUI unnecessary. Staff resources previously devoted to the in-house GUI have been reallocated to database management, data creation and integration, special projects, and the design and promulgation of customized Arcview scripts.

Arcview scripts have been developed by our GIS technical staff to support environmental and real estate functions.

The Environmental APR includes the following data layers : shoreline, tiger roads, project boundary, roads, contours, town borders, basemap (vector), National Estuarine Research Reserve, islands, D.O.S. Significant Habitats, wildlife habitats, RTE habitats, land use, wetlands, rivers, and NYSDEC quadrangles. The Real Estate APR adds zoning, tax maps, and docks to the preceding list while deleting the environmental data layers such as wetlands, landuse, and habitats. Because the number of layers involved in this APR makes finding a particular layer cumbersome, the layers most often used in each discipline were separated into the task specific APRs described above.

The 1:2400 and 1:12000 orthophotographic base maps have been written to CD. Access to these data is through the "add theme" function of ArcView. A .dbf file has been written to each CD which, when chosen at the "add theme" prompt, retrieves and displays the imagery corresponding to the current map extent. Resampled image files are also stored as a layer in the APR's for quick access to the images, where resolution is not critical.

Arc View is most often used for relatively straightforward map view and data query. Queries such as the number of acres of a particular class of wetland in a certain town, the approximate elevation of a habitat site, or the feasibility of alternate drainage schemes for a wildlife management area are easily accomplished by staff professionals with relatively minimal GIS training, allowing the Unix workstations to be dedicated to 3D modeling, large file image processing, and data creation tasks. All relicensing studies with a spatial component (so far nearly all of them) are specified for delivery in ArcView compatible formats, but we often have occasion to modify or merge various delivered coverages using ArcInfo. As an example, the shoreline coverage was developed in house by extracting and combining the polygons defining open water in the wetland coverage developed under contract.

Summary

In order to support the preparation of an Environmental Impact Statement for the relicensing of the St. Lawrence-FDR Power Project a series of studies that will describe the resources contained within the Project boundary must be carried out. Additionally, the effects of the existence of the Project and power generation operations on the environment must be assessed.

While developing strategies for the relicensing of the Project, it became apparent that the volume of studies, reports, and analyses that would be required to successfully complete the process could not be effectively undertaken without the use of a tool such as GIS. An additional goal was not build a narrowly focused GIS solely for the purpose of supporting relicensing of the Project but which would be useful for other management purposes. The commitment to a multi-purpose GIS has the effect of spreading the cost of database creation and conversion over a larger number of potential users, building support across departments and helping to convince senior management of the cost effectiveness of this technology in an increasingly competitive market.

As part of the preparation for relicensing, NYPA constructed a GIS database consisting of a series of current digital orthophotographic base maps that would be used to depict conditions at the Project. A number of environmental studies were undertaken to gather information regarding sensitive resources and potential impacts from the Project. Whenever a study was commissioned it was mandatory that any data gathered would be provided in digital format referenced to NYPA's coordinate system. This action allowed for subsequent analysis of environmental resources and impacts by overlaying them in an in-house GIS. The environmental layers consisted of all resource data that could be spatially referenced. Most of the resources present within the Project could be delineated this way and stored for later analysis.

Since some of the analyses would have to be carried out interactively by non-GIS proficient users, an Arcview interface was chosen as the tool to make the data available. This interface allowed easy access to the data layers and comparative analysis of environmental changes over a large study area.

References Cited

Cowardin, L.M., V. Carter, F.C. Golet, and E.T. LaRoe. 1992. Classification of Wetlands and Deepwater Habitats of the United States. U.S. Fish and Wildlife Service, Wahsington D.C.

About the Authors

Edward Alkiewicz is a Senior Environmental Scientist for the New York Power Authority's Environmental Division. Correspondence may be sent to: New York Power Authority, 123 Main St., White Plains, NY 10601; Phone (914) 287-3247; Fax (914) 287-3284; Internet email: walkiee@Ip3GATE.USA.COM

John Wingfield is Survey Coordinator for the New York Power Authority's Real Estate Division. Correspondence may be sent to New York Power Authority, Blenheim-Gilboa Power Project, P.O. Box 200, Gilboa, NY 12076; Phone (607) 588-6061, Fax (607) 588-7778