Charles Van Dusen
In a cooperative effort funded by the National Oceanic and Atmospheric Administration (NOAA) specifically the Office of Coastal Resource Management (OCRM) and managed by Massachusetts Coastal Zone Management (MCZM), Applied Geographics, Inc. (AGI) performed a vector based historic shoreline change analysis using ArcInfo vector coverages, AML, and C. Linear historic shoreline data as early as 1844 and as recent as 1982 were provided and an analysis was undertaken to define and execute a procedure for deriving the historic rate of shoreline change using a vector-based methodology. Programs written in C were modified to handle the complexities of the Massachusetts historic shoreline data. The data were segmented for analysis and then appended to a single State-wide dataset comprising nearly 30,000 sampling points. Custom plots were created and delivered for distribution and a MS Access database interface was designed and delivered to permit interactive statistical query of any single sampling point or any contiguous series of sampling points.
Coastal zone managers, emergency management officials, and coastal property owners need to be aware of the potential risks to coastal property before, during, and after severe storms and hurricanes. As new sensors become available and new technologies are focused on the problems of hazard mitigation in the coastal zone, a wealth of data is being generated which will permit volumetric analyses of recent landform morphology along the coast (SAR, LIDAR). These data may be able to provide high spatial and temporal resolution surficial forms for modeling recent changes, yet to understand the longer-term fluctuations for which these data are not available, historic linear data may be exploited. Currently, historic linear data provides us with the ability to assess future changes in the shape of the shoreline by reviewing historic snapshots of the shoreline. The long-term rates of change provide managers and property owners with a clearer picture of the potential hazards confronting coastal development.
The Massachusetts coast is highly variable, characterized by rocky headlands framing sandy beaches and salt marsh. Defined in linear terms, the shorelines are convoluted, circuitous shapes. Their complexity is further complicated by time series replication showing the temporal, morphological changes in the shoreline. New methods for developing hi-resolution surficial data may supplant the need for performing these types of linear temporal shape analyses, yet historic data still provides a substantial resource archive for evaluating future coastal hazard risks from historic trends.
The study area for shoreline change analysis include the entire Massachusetts shoreline spanning approximately 1500 statute miles. Analysis was completed in all areas where the data was deemed by state coastal geologists to be sufficient for realistically estimating long-term shoreline change rates.
MCZM provided AGI with a historic shoreline dataset with a temporal span of nearly 140 years for the Massachusetts shoreline. This dataset was automated from a number of sources, including NOAA/NOS topographic map sheets, FIS (FEMA) topographic map sheets, hydrographic map sheets, USGS quadrangles, aerial photos, and orthophotos. The source data were evaluated for error and accuracy prior to conversion and then plotted for delivery to MCZM at a scale of 1:5000. These plots were subsequently digitized to create the Massachusetts historic shoreline dataset. This dataset, in ArcInfo coverage format including attributes describing the date of each shoreline, was delivered to AGI for analysis.
The shoreline data are both temporally inconsistent and spatially inconsistent. No shoreline for any year spans the entire coast, nor does any year/shoreline necessarily have a consistently earlier or later year/shoreline. Rather, the data are spatially and temporally dispersed. Further, the data as delivered had a temporal resolution of 1 year. The shorelines were thus assumed to provide the shape of the high-water line at a single date during the calendar year and were assumed to be reliable for use at 1:5000 scale. The table below outlines the temporal distribution of the shorelines and their summarized lengths across the entire Massachusetts shoreline.
|
|
|
|
|
|
The historic shoreline data were segmented for analysis. The criteria used to segment the data were developed within the analysis methods to provide consistent, accurate, and timely temporal shoreline change analysis results. The data were divided into approximately 100 analysis segments considering (in part) the following criteria:
Once the historic data were segmented into these manageable units, transecting and analysis proceeded within each of the analysis units. Baselines were constructed on the upland side of all historic shorelines to provide a starting point for the transecting operation. Baselines were digitized parallel to the general trend of the historic shorelines so that orthogonally oriented transects originating from the baseline would most closely match transects placed by manual 'best fit' methods. These baselines coverages were SPLINEd, GENERALIZEd, and DENSIFied to provide a good origin point for each of the transects cast.
With baselines and historic shoreline data coverages present for each
analysis segment, a suite of C functions were called to generate transects,
perform the analysis and deliver results in ArcInfo GENERATE format. The
original C code was developed for the USGS and is described in Open File
Report Number 92-355 (Danforth and Theiler, 1992) as the Digital Shoreline
Analysis System (DSAS). The C code was redesigned by AGI to perform accurate,
high-resolution temporal shoreline change analysis on shorelines which
are complex in shape and in orientation. These modifications included code
to perform the following tasks:
Whether long stretches of sandy beaches, migrating inlets, or salt marshes, these functions are capable of casting an orthogonal sampling transect, measure the distances, and compute the interim and overall rates of movement along the sampling transect. Additonal statistical functions provide estimations of the long-term shoreline change rate and the 30 year and 60 year projected positions of the shoreline along the sampling transect. Geographic data was output in GENERATE format and attribute data were output in a format suitable for input into INFO or any commercial database. Unique identification codes provided the link between geographic and attribute data which were subsequently joined using JOINITEM.
Within each analysis segment, the output transect coverages were GENERATEd, attributed, and APPENDed to a single statewide line coverage. This line coverage consists of nearly 30,000 transects with attributes describing the interval rates of change, overall long-term rates of change, miscellaneous statistical measures, and the 30 year and 60 year projected coordinate positions of the shoreline.
A map index for the Massachusetts coastal zone was built and plots conforming to the index were created and plotted at 1:10000 scale on paper and on mylar for overlay with orthophotography. These plots are being distributed to coastal zone managers, town planners, and the public to support coastal hazard assessment, coastal economic impact analysis, and property assessment applications. In addition to plots, an MS ACCESS database and query interface was designed and delivered to allow MCZM geologists to analyze the shoreline change results for any individual transect or series of transects.
Shoreline change data plots were converted to gif format and embedded within an HTML application for ease of browsing and display. This application is available for browsing for a limited time at www.appgeo.com beginning in May, 1997
A limited section of the shoreline change data and analysis results are presented below. Note the shift from net overall loss (erosion) to net overall gain (accretion) as the analysis moves from left to right. Uplands are at the top of the image, offshore areas at the bottom of the image. Transects are spaced at 50 meter intervals. Scale is 1:4500.
| Historic Shoreline Change Analysis |
|
|
|
|
As the above images illustrate, some areas of the Massachusetts coast are eroding and some are accreting. For the entire Massachusetts shoreline, sampling transects which indicate erosion outnumber those which indicate accretion roughly 2 to 1. Additional temporal summary statistics are provided only for areas where temporal and spatial consistency of the data allows them to be generated. A multitude of additional statistics are available for transect analysis within the ACCESS interface, including interim temporal rates of change with their Variance and Standard Deviation for all interim rates, End-Point Rate, Average of Rates, and Linear Regression Rate (used as long-term rate for this project).
The figure below presents a summary of the temporal change for a stretch of coastline spanning 100 transects (approximately 5 kilometers) of Nantasket Beach in the town of Hull. Each bar in the chart represents the overall summarized linear change in the shoreline for the time period. In this sample area, summary temporal statistics suggest a trend in the data showing overall erosion for the selected area from 1847 to 1895, accretion from 1895 to 1938, and erosion from 1938 to 1978.
| 1847 to 1895 | 1895 to 1938 | 1938 to 1978 |
|
Summary Change in Linear Feet for 100 transects at 50 meter intervals. Nantasket Beach, Hull Massachusetts. |
Vector based shoreline change analysis provides a model of temporal erosion and accretion for any set of linear historic shoreline data. The vector approach to analyzing historic shoreline change data contrasts with a raster approach in its sampling flexibility and temporal scaleability. The vector approach developed above can accept any number of temporal linear representations of the shoreline and can flexibly sample those shorelines to calculate past variability and project future changes.
Historic rates of shoreline change provide valuable data on erosion trands and permit limited forecasting of shoreline movement. Automated GIS shoreline change analysis provides rapid, high-resolution evaluation of multiple temporal shoreline delimitations. There is room for improvement on the methodology described above, which will be implemented in future analyses as new shoreline data becomes available. ArcInfo embeds the functionality of the C programs used here in COGO. Nonetheless, this methodology demonstrates that the plotting strengths of ArcInfo can be combined with C functionality using standard ASCII files for data transfer and communication.
Other applications include linear shape change analysis for any spatial phenomenon which can be defined and delimited for at least 2 time periods. Spatial migration of distinct geographic features is a documented phenomenon whose morphology is of substantial interest to resource managers and geographers.
Anders, Fred J. and Byrnes, Mark R., 1991, Accuracy of Shoreline Change Rates as Determined from Maps and Aerial Photographs, Shore and Beach, January 1991, pp. 17-26.
Danforth, W. W., and Theiler, E. R., 1992, Digital Shoreline Analysis System Users Guide, Reston, Virginia, U.S. Geological Survey Open-File Report Number 92-355, 42 p.
Dolan, Robert, Fenster, Michael S., and Holme, Stuart J., Journal of Coastal Research, Volume 7, Number 3, 1991, pp. 723-744.
Crowell, Mark, Leatherman, Stephen P., and Buckley, Michael K., 1993, Shoreline Change Rate Analysis: Long Term Versus Short Term Data, Shore and Beach, April 1993, pp. 13-20.
Crowell, Mark, Leatherman, Stephen P., and Buckley, Michael K., 1991, Journal of Coastal Research, Volume 7, Number 3, pp. 839-852.
Leatherman, Stephen P., 1983, Shoreline Mapping: A Comparison of Techniques, Shore and Beach, July 1983, pp. 28-33.
Massachusetts Coastal Zone Management, Massachusetts Shoreline Change
Project, 1989
Smith, George L. and Zarillo, Gary A., 1990, Calculating Long-Term Shoreline Recession Rates Using Aerial Photographic and Beach Profiling Techniques, Journal of Coastal Research, Volume 6, Number 1, pp. 111-120.
AGI would like to acknowledge Jim O'Connell, MCZM coastal geologist, for providing guidance and management throughout the analysis process.
Charles Van Dusen
Senior GIS Analyst
Applied Geographics, Inc.
100 Franklin Street
Boston, Massachusetts 02110
Telephone: (617) 292-7125
Fax: (617) 292-7911
email: dusen@appgeo.com