Report to The Nature Conservancy
Central and Western New York Chapter
Rochester, New York 14604
Aerial Photography-Based GIS and Historic Map
Analysis of the
Eastern Lake Ontario Shoreline:
a follow-up study on
beach, dune, inlet and bar
changes and processes
Prepared as part of the ELOSTS
For
New York State Department of State
Under the
Environmental Protection Fund
By
Charles E. McClennen, Professor of Geology
Colgate University
Hamilton, New York 13346
March 6, 2002
Table of Contents
| Abstract | 3 |
| Introduction | 4 |
| Error Analysis | 5 |
| Comments on NYS Map Product Requirements | 6 |
| Impact of Lake Level Change | 7 |
| Beach, Dune, Inlet and Bar Changes | 8 |
| Consideration of 1862, 1878, 1960 and 1992 edition Charts and Maps | 10 |
| Thoughts on Coastal Process Questions and Sediment Budgets | 13 |
| Management Implications | 16 |
| Conclusions | 18 |
| Acknowledgements | 19 |
| Bibliography | 20 |
| Table I. Aerial Photograph Identification and Characterization | 21 |
| Table II. NOAA Lake Levels for days of Aerial Photography | 23 |
| Figures | |
| 1. Aerial Photographs of North Pond Inlet 1966-1984 | 9 |
| 2. GIS digitized land cover of North Pond Inlet 1966-1984 | 10 |
|
This research report continues the efforts to refine the understanding of the dynamics and history of coastal change along the eastern Lake Ontario shoreline. The coast is dominated by a seventeen-mile long stretch of barrier beaches and dunes with numerous ponds and small river outlets. Coastal land management issues and opportunities have motivated the search for an increased certainty about the active processes and expectations for future change along the shoreline. Analysis of potential errors and their magnitude in this GIS-based study demonstrates that available aerial photography has fundamental registration uncertainty of about 50 feet (15 meters). Of particular significance for the 1974 and 1978 images is the magnitude of the distortion caused by the lens effect in aerial photographic prints. The outer margins of all prints include location distortions of somewhat more than fifty feet at the scales used in this study. In addition, records of lake level fluctuations generated by storm conditions and seasonal climate variability, even when modulated by several decades of lake level management, demonstrate that one to five foot changes are expectable in any given year. This range of lake level fluctuations horizontally shifts the shoreline location anywhere from 50 to over 200 feet, depending on the magnitude of the lake level change and the gradient of the beach face and near-shore zone. Because of the absence of precise knowledge of lake levels at the time of acquisition of aerial survey images it is difficult to distinguish between real geological shoreline changes and the mere appearance of change caused principally by lake level fluctuations. The combined impact of image registration for GIS location and lake level variability may be additive or partially self-canceling. In either case, many of the apparent shifts in shoreline position, beach width, dune location and inlet character fall within the range of these inherent uncertainties. Thus only the most extreme changes, as at the North Pond inlet in the town of Sandy Creek, Oswego County, can be pointed to with certainty, based on the GIS analysis of aerial photographic data collected each decade between 1938 and 1994. Historic maps and charts allow the extension of the study back to 1862 in a limited way and more substantially from 1878 when detailed bathymetric values were first published for eastern Lake Ontario and the adjacent ponds. Shoaling of North Pond of one to six feet suggests an upper limit of sand removal to the pond from the barrier and near-shore bars in the order of two to five feet. Wind blown transport and inlet currents are presumed to be of major importance for the pond depositional accumulations. Reasons for caution about the validity of these calculations are reviewed and emphasized. Review of management issues and questions about coastal dynamics lead to the fundamental necessity to appreciate the unpredictable yet inevitable nature of coastal changes to the beaches, bars, dunes, inlets and associated vegetation zones. All coastal development that requires stability of shore and land should be excluded from the zones of likely natural coastal change by carefully designed management planning and regulation. Consideration for the desired human development uses and the value of natural ecosystems should also be included in future coastal zone management analyses. Introduction In reviewing the September, 2000 Report to the Nature Conservancy entitled "Aerial Photography-Based GIS Analysis of the Eastern Lake Ontario Shore: Coastal zone change and processes 1938-1994" (McClennen et al., 2000) it was determined that a follow-up GIS study would be of value. The several analysis concerns included; 1) the previous absence of any photographic coverage from the 1970s decade, 2) the lack of a quantitative consideration of the sources of error in the GIS digitization of the photographic coverage, 3) the limited quantitative analysis of lake level change as a factor and 4) the shift of priorities from the coastal zone including the extensive wetlands to the much more restricted areas of beach, dune, inlet and adjacent offshore sand bars. The sand budget focus and coastal change analysis was partially overwhelmed by the aerial extent of the coastal wetlands and their changes over the decades. This review of the beach, dune, inlet and bars with their history of change, as recorded in the aerial photographs and a few 19th century maps and charts, gives primary attention to the features of primary interest. Aerial Photographic coverage for the entire study area from the 1970s was made possible by combining images taken in 1974 and 1978. However they are less than ideal images for GIS analysis in that the registration points for precisely locating each image had to be taken from the outer limits of each photograph. That is where the optical distortion is known to be greatest. This fundamental aspect of the only available 1970s imagery for the study area makes it impossible to produce highly accurate GIS digitized comparisons between the preceding and subsequent decades. This aspect of the photographs taken in the study area during the 1970s was the reason that they were excluded from the initial study and report to The Nature Conservancy in 2000. Here they have been added as requested because of the numerous changes around inlets that occurred in the 1970s, a decade of particularly variable and high lake levels. A consideration of lake level variability and impacts on the coastal features revealed in the photographic images is presented. There are three aspects of this review that deserve attention. First, elevated and lowered lake levels directly affect both the apparent beach width and shoreline location at the time photographs are taken. Second, the impact of waves on sediment transport and geomorphic modification of beach, dune, sand bar and inlet features, particularly immediately after storms, is significantly enhanced by above average lake levels. Third, lake level data has not been systematically gathered along the eastern shore of Lake Ontario within the study area. The nearest available hydrographic data comes from two observation stations located at Oswego and Cape Vincent, New York. Linking a particular photographic image series from a known date to these lake level data is complex since the daily, monthly and annual averages include both normal and extreme weather conditions. These kinds of lake level changes have significant impact on shoreline position since the beach faces generally have low slopes or gradients of between one in ten to one in one hundred. The error analysis of the digitized aerial photographic and GIS data set incorporates several components. The United States Geological Survey (USGS) topographic quadrangle maps of Ellisburg, Pulaski and Henderson, at a scale of 1:24,000, have an accuracy standard that requires 90% of all map points to be within 40 feet of true location. These base maps were used to pick key registration points for locating the aerial photographs in GIS coordinate space. The digitizing tablet reads to within 10 feet at the scale of 1:24,000 used for the published topographic maps. The digitizing accuracy varies somewhat from photographic image to image, depending of the scale of the prints used. They ranged from about two to six feet, with most in the range of three to four feet. Thus the initial required registration of each photographic image is the dominant source of the error, as all registration positions for each aerial photograph were taken from the topographic quadrangle maps. Thus any position shift of coastal features seen in the GIS analysis that is 50 feet or less must be seen as possibly caused by these unavoidable mapping and digitizing errors. Changes of greater magnitude are likely to be real but subject to an uncertainty of about 50 feet (15 meters) in exact location. Some time was also spent evaluating the quality and information from a series of historic maps and charts. The goal was to identify any coastal changes in location or shoreline shape as well as bathymetry. Differences in the water level datum used for each chart and map were small where it could be determined (< 2 feet between 1960 survey and 1992 chart). The 1878 chart displayed a systematic offset in latitude of approximately 800 feet to the west for all the land features, even those located to the east and fully separated from coastal processes. The significantly deeper pond bathymetry from 1878 enables the calculation of interesting sedimentation rates and depositional volumes. The 1862 property tax map for the Town of Sandy Creek has no latitude and longitude coordinates nor any bathymetry and elevations making it of limited utility for this study. A coastal inlet and channel ways are depicted for North and South Ponds. An earlier (1829) and other historic inlet locations for North Pond are noted by Weir (1977) in his illustrated Masters Thesis. Finally, an attempt to address a series of questions that have been raised about numerous aspects of the eastern Lake Ontario shore and sediment dynamics is made, utilizing the knowledge gained through the aerial photographic and map analysis. The aerial views and GIS analysis can provide useful information or perspective for only some of these questions. Others, while of considerable interest, must be addressed through different means used in the ELOSTS study. A few questions seem to extend well beyond any of the field research methods applied so far to the eastern shore of Lake Ontario. Ways of addressing those unresolved questions in the future could well require new field research of considerable magnitude. This report will be posted on the Colgate University web site along with the earlier (McClennen et al., 2000) TNC GIS Study. All the scanned aerial photographic images and digitized shape files for each decade are posted as well: see http://www.colgate.edu/academics/geology/faculty/mcclennen.html. Zip discs of the report, scanned images and back-up GIS files and materials will be provided to TNC with the final submission of this report. Error Analysis As indicated in the introduction, there are several sources of position error when conducting GIS-based aerial photographic analysis. Without restating the detailed data handling methodology included in The Nature Conservancy Report (McClennen et al. 2000), it is useful to be reminded that for any image-to-image comparisons to be effective all the aerial photographs have to be accurately positioned in space. Typically some form of latitude and longitude map projection is used, such as the Universal Trans Mercator projection of the USGS topographic quadrangle maps. We took at least four positions of known features from the topographic maps, used them to register each photograph and then proceeded to digitize the photographed boundaries of all the coastal features of interest. The USGS standard of having 90% of all positions within 40 feet of their true location on 1:24,000 quadrangle maps turns out to be the primary limiting factor because all other sources of error are considerably smaller, although important because they are potentially additive. By using at least four registration points, cross checking and internal consistency of each image registration effort can be made in the ArcView version 3.2 (ESRI, 1999) used in the research project. Registration plots with greater than 0.01 inches of error on the digitizer were rejected as unacceptable and other registration points were selected. The amount of offset depended on the aerial photograph with 0.01 inch on a standard 1:24,000 scale topographic map = 20 feet. Distortions of relative location are inherent in aerial photographs particularly around the outer perimeters of each image. This lens effect is unavoidable in standard aerial photographic prints from negatives. For this reason the central third to half of each image is preferred for precise aerial space and location analysis. Most aerial photographic series are therefore shot with extensive overlap on all sides that may be of interest. Some of the series used in this Lake Ontario study did not have the coast and registration features located in a central part of the images. Thus registration efforts were handicapped, particularly for even the best of the 1974 and 1978 photographic series used in this study. In addition the two 1970s decade photographic series only provided partial coverage of the entire study area. So any changes in shorelines that were as little as fifty feet (15 meters) were viewed as within the margin of error and thus too suspect to accept as geologically significant or real. Comments on NYS Map Product Requirements In view of this described situation with respect to error analysis it is obvious that it has not been possible to generate any map from this aerial photographic GIS study that complies with the National Map Accuracy Standards, USGS and NYSDOT General Map Product Requirements. In fact no level of mapping was attempted beyond the digitization of the aerial photographs and the superposition of such products for the assessment of coastal feature changes. All the GIS files and documentation for the TNC research effort are tabulated and delivered in their entirety on Zip Discs. This paper includes just a couple of figures and a complete updated table of aerial photographic descriptions. See the scanned photographic images for in their entirety for 1974 and 1978 as well as the Mktc folder and sub-folders backing up this report and the associated GIS files. The horizontal datum coordinates were taken from the USGS Topographic Maps with their standard NAD27. No vertical datum was included in the GIS analysis project. Position and Map Accuracy has been discussed above. Similarly, the Additional Digital Cartographic File Requirements are not realistically applicable to this GIS-Aerial Photographic analysis because of the numerous error factors discussed above and the fact that no maps were required by the research contract agreement. No map products were generated that could be made to conform to the Edge-matching, Common Boundaries, Point Duplication, Connectivity, Line Quality, Graphic Precision and Digitizer Accuracy standard requirements. Polygon Closure was standard for each photo digitization effort. However, as no Digital-Ready Maps were generated, the related requirements provided for review by the TNC for Base Map Media, Map Scale, Map Registration, Map Title and Legend and Cartographic Quality need no further commentary in this report. Impact of Lake Level Change As indicated in the introduction, there are several lake level factors of significance in any aerial photographic analysis of coastal areas. The typically low topographic slopes seen along sandy coasts cause major horizontal shifts in shoreline location when water level is vertically displaced. The Great Lakes Information Network provides numerous useful records and tabulations of lake level data on the web. In Lake Ontario storm related lake level changes of a foot or two are documented on the Great Lakes Information Network for both Oswego and Cape Vincent (hhpm@lre02.usace.army.mil/storm/ontstrm). Storm induced rises of one to two feet have been recorded at both stations. The significant probability (20%) of exceeding a one half foot to one foot excess elevation in storms for each month of the year shows that this is a common occurrence. Storm induced higher elevations of lake level are less likely but seen at both the Cape Vincent and Oswego stations in all months. Such elevated lake levels make the beaches look narrower by 50 feet (15 m) to more than 200 feet (61m) particularly in the coastal sections with beach face slopes of less then 1:50, which occurs along the fine sandy shores of the barrier beaches and at the spits by inlets. In addition to the short term storm events the seasonal and inter annual changes in lake level has been well documented for the Great Lakes and reported on the web at (http://huron.lre.usace.army.mil/levels/maxmin.html). For Lake Ontario the difference between maximum and minimum lake levels for each month roughly span four and a half (4.5) feet to five and a third (5.33) feet. Table II, seen at the end of the paper, indicates the levels of the lake reported by NOAA for Cape Vincent and Oswego, New York on the dates that the digitized aerial photographs were taken. A one foot difference between stations on June 4, 1959 and the total range of 3.05 feet (11/10/38 vs. 05/25/84) indicate the significant magnitude of lake level changes impacting this analysis. Such vertical lake level changes produce striking shifts in the horizontal waterline location, even on steeper sloped beach faces. With the relatively steep beach face slope of 1:10, a five foot shift in lake level moves the shore line a full 50 feet (15 m). At lower beach face gradients, such as 1:50, the shoreline would be shifted horizontally by as much as 250 feet (76 m). This kind of shoreline shift is similar in magnitude to, but can obviously exceed, the uncertainties or errors inherent in this aerial photographic GIS digitizing and analysis. On the more gentle and typical beach slopes such shoreline displacements are known to occur but are not always easily identified in aerial photographic analysis because of the lack of precise knowledge of the local lake levels at the hour of photographic flight times. Published monthly mean lake level data are thus of limited value in photographic image interpretation requiring precise determinations of shoreline positions and their geological origins, if any. It is logical that changes in lake level cause a relocation of both the shoreline and the locus of wave action on the coast. Most coastal observers note the prevailing pattern of destructive or erosional impact of storm waves particularly during elevated lake levels. Mobile unconsolidated beach and dune sands are subject to rapid geomorphic change when wind driven waters flood the barriers and dunes. Because there are no tidal patterns in Lake Ontario many fail to pay sufficient attention to the impact of changing water level caused by the weather events and climate. Lake level change goes well beyond the inconvenience of narrower or widened beaches, deeper or shallower channels and occasionally flooded front lawns. In some instances storms that occur at times of raised lake level produce lasting geomorphic change, such as the creation of new inlets. This rapid change is well documented by the periodic dredging of inlets demanded by the boating public when inlets are in-filled by mobilized beach sands. The rare, but more vivid, creation of new inlets through the barrier beaches developed during particularly powerful storms that occur each century leave more lasting impacts on the ponds and coastal morphology. Beaches and boating channel ways are both clearly modified and easily recognized on any series of charts, maps and aerial photographs covering such areas. The history of inlet change for North Pond is certainly the best example on the eastern Lake Ontario shore (See; McClennen et al., 2000, Figures 1 &2 of this report, Weir, 1977 and SUNY Oswego web site; http://www.oswego.edu/Acad_Dept/a_and_s/earth.sci/geo_geochem/geol/sandy.html). Thus lake level changes are a significant part of coastline image and location mapping and they are certainly related to more fundamental coastal reconfigurations. The basic problem of aerial photograph analysis is that it is not always possible to link the exact time of photographing to the exact lake level. Changes observed in individual photographs may thus be either temporary due to short-term lake level rises or of a more permanent nature and reflect a true shift in the location of coastal sedimentary deposits such as bars, beaches and dunes. If the images of coastal changes exceed the scales of uncertainty indicated by error analysis then the change should be viewed as real. If the changes observed are seen to persist, expand in degree or continue shifting with a persistent trend then they can be taken as real. Yet if they are within the range of error and are seen to have an oscillating character they may be due to photo registration and feature location uncertainties rather than true coastal change. In order to evaluate the extent of horizontal shift in shoreline location due to any given amount of lake level change it will be necessary to know the slope of the beach face and near-shore zone along the entire length of the coastal area of interest. So far no such systematic profiling survey has been conducted. Beach, Dune, Inlet and Bar Changes As previously reported (McClennen et al, 2000) there are many identifiable apparent changes in the shoreline along eastern Lake Ontario as recorded in the aerial photographs taken over the decades. When digitized with proper registration the horizontal shoreline location shifts mostly fall within the range of 33 to 66 feet (10 to 20 meters). This magnitude of horizontal relocation can easily be explained by three likely and at times interacting factors. First, the beaches actually retreated in response to erosion or advanced due to deposition. Second, the lake water levels differed sufficiently at the times each photographic image was taken thus producing an apparent gain or loss of beach sediment. Third, the part of the photographs used to locate the shoreline were too close to the edge of the image and thus subject to significant lens-effect distortion. Accordingly, based on the error analysis and associated GIS uncertainties, most of these small scale shoreline shifts should be looked upon as probably not being of geological significance. No patterns of multi-decadal trends, in shoreline shifts, were noted. As McClennen et al. (2000) previously reported, even comparisons of the earliest and latest photographic images displayed shoreline dislocations fully within the recognized range of uncertainty. The dune vegetation by contrast showed progressive development and thickening over the decades. Seasonal variations of dune vegetation were noted in some of the photographs, particularly in the early spring vs. late summer or late fall. Erosion induced retreat of the lake-ward face of the dunes was indicated locally by a narrowing of the westward extent of the dune vegetation. Re-growth of dune grasses and scrub vegetation in subsequent decades typically followed the storm erosion events. Shift in dune and beach vegetation was most extensive and noticeable around the inlets where the dune deposits can be very thin such as around the North Pond inlet. At North Pond the inlet and shoreline, as well as vegetative cover, shifts have been the most vivid in the entire study area, during the decades examined. Figures 1 and 2 show portions of the photographic images and digitized land-cover polygons from 1966, 1974 and 1984. The wave breaching of the barrier beaches during storms has created a new inlet seen only in 1974, which has migrated and been reconfigured by subsequent decades of wave driven near-shore and beach-face sediment transport. The April timing of the 1974 photograph may well account for the absence of any forested dune features in the northern spit. The preceding and subsequent photographs were taken in late May and July when trees are in full leaf and thus more readily identified in the photographs. Figure 1 (click on image for larger version) 
Figure 2 (click on image for larger version) 
Equally striking is the fact that the associated near-shore bars and inlet shoals are always different in form and abundance when comparing the photographic images from the different decades. Fully describing the changes between the decades is prevented because many of the aerial photographs suffer from poor resolution of the lakebed deposits, even in the near-shore shallow waters. Breaking waves and surf foam overwhelm some sand bar imaging. Reflected sun glare off of the waves in lake and pond surface waters is another recurring pattern that prevents a systematic, decade-by-decade, analysis of the near-shore shoals and sand bars located along the length beaches of the study area and at each inlet. However it is no surprise that the indications of greatest mobility of coastal sands occurs at the inlets where in addition to the dominant Lake Ontario wind generated waves there are also inlet associated currents. Spring melt run off and heavy rains generate some of the outflow. Changing lake water levels driven by shifting wind patterns create more frequent inlet currents that are strong enough to mobilize the sandy lakebed sediments. Consideration of 1862, 1878, 1960 and 1992 edition Charts and Maps The 1862 Sandy Creek map, by J. B. Butler, shows the surveyed real-estate property bounds. No datum, elevations or bathymetry is recorded making it of limited use in a coastal change study. The meandering channel inlet from Lake Ontario that bifurcates and simultaneously serves North Pond and South Pond is clearly depicted in detail and quite similar to that, seen with less detail, in the 1878 Corps of Engineers Coast Chart No. 2. This was surveyed in 1874 and 1875 and provides very useful bathymetry at the map scale of 1:80,000 for eastern Lake Ontario and the several adjacent ponds. Of particular interest is the comparison of the 1878 bathymetric data with the 1959-1960 Corps of Engineers' bathymetric survey of the eastern end of the lake and the 1992 edition of the National Ocean Survey chart number 14803. The 114 year period, between 1878 and 1992, provides our best long-term basis for evaluating bathymetric change and deposition or erosion. The Lake Ontario depths show no apparent or substantial changes. This assessment is limited in precision by the margin of uncertainty resulting from the depth notation units of fathoms, in 1878, and then the more precise unit of feet, in 1992. However, the obvious differences in water depths for North Pond in the Town of Sandy Creek cannot be explained away by such uncertainties because both charts have depth expressed in feet. Any possible map-datum differences cannot explain the bathymetric shoaling because it varies considerably within the area of North Pond. North of Greene Point sedimentary infilling of one to three feet is indicated by the charted water depths. West of Carl Island it is intermediate at three feet. In the main body of North Pond, south of Greene Point and Carl Island the shoaling over the same time interval is more substantial at three to six feet. The greater rate of infilling of the central and southern part of North Pond is consistent with several coastal features observed on the aerial photographs dating back to the 1930s. The barrier spit north of Carl Island has had wooded dunes for most of its length during the 20th century. The barrier spit south of Carl Island has been the locus of several inlets that have been developed and all subsequently re-closed, except for the present one just west of and adjacent to Carl Island. The trees and other vegetation of the wooded dunes to the north would have greatly inhibit the extent of wind blown transport from the beach eastwards into the pond. Just to the south the prevailing westerly winds would have had a different impact. The series of documented pond inlets and the subsequent infill sediments have provided extensive areas of exposed beach sands and low dunes. Considerable sediment seems to have been blown eastward into North Pond or carried in through the series of inlets cut through the southern portion of the barrier. These, in combination with organic deposition and land-derived sediments, have probably been the dominant causes of the recorded shoaling. While the depths of several other coastal ponds are given on the 1878 chart they are not recorded on any of the later charts used, so sedimentation rates elsewhere along the coast could not be determined by this method of chart analysis. In order to get a volumetric estimate of the amount of sediment lost from the beach and barrier dunes into North Pond some simple reasoning and calculations were utilized. First, it was assumed that perhaps only half of the thickness of recorded shoaling was attributable to beach and dune sources. Second, for convenience of analysis it was determined that the mean shoaling was approximately two feet north of Carl Island and four feet to the south. The area of the two sections is taken to be 26 million and 76 million square feet respectively. Using these two assumptions and the square area estimates the total volume of beach and dune sand transported eastward from the barrier into North pond is approximately 178 million cubic feet or just about 6.5 million cubic yards. A reasonable question then is, what impact would this volume of sediment have had if it had been in fact removed from the beach, barrier dunes and near-shore bar deposits. Spreading it evenly over the three and a half mile length of the North Pond barrier with an approximated average width of 2,000 feet would produce nearly a five foot thick deposit over the barrier and near-shore bars. However, a more reasonable model would be to assume that the loss of barrier deposits fed into North Pond also came in part from the beaches to the north and south. Long shore and beach transport, driven by northwest and southwest winds, is certainly important along the seventeen mile stretch of the eastern shore of Lake Ontario. One could thus reasonably consider the impact of spreading the calculated volume over the full seventeen mile length, but on a narrower width band of only 1000 feet to account for the lack of barrier dunes in several sections. This calculation indicates that one would have had to remove a thickness of just under two feet of sand from the beach and near-shore bars for the entire seventeen mile length of the study area in order to obtain half the calculated volume of sediment needed to fill North Pond since the 1878 bathymetric survey. Before we take these calculations too seriously the rough nature of the assumptions must be recalled and there are other considerations. First, the depth measurements of the nineteenth century were most likely determined by the leadline technique or possibly by using a pole, since all depths were less than twenty feet. When the lead weight or pole rests on the bottom it has penetrated most of the soft organic rich sediments resting on the bottom of this type of pond. Secondly, the twentieth century surveys are done with sonar, which gives a different depth values. The reason is that at the 10 to 24 kHz frequency range used in echo sounding equipment the sound waves reflect off of the top of the soft organic rich sediments as well and the more solid underlying sandy deposits that would stop a leadline weight or be felt by someone using a pole to determine the water depth. These survey factors would tend to cause an over estimate of the sediment volumes used in this analysis. Other factors should be mentioned. It probably is safe to assume that the other back barrier ponds have received some wind blown and wash-over sediments from the beach and dune areas. The aerial photographs and field observations indicate that these processes are and have been active but not the full extent of this sediment redistribution within the coastal system. The above calculations are thus viewed with considerable caution. They are never the less valuable as providing an order-of-magnitude estimate and possibly an upper limit guide to the volumes of beach, dune and bar sediment transported by selective processes to the adjacent ponds in this area over the last century or more. The 1878 chart has another set of notations that is of interest when considering the Lake Ontario bottom sediments in the offshore zone. The bottom is characterized mostly as sand or more rarely as hard and this pattern extends five to ten or more miles offshore. So for more than a century sandy sediments seem to have prevailed from the shoreline out several miles into the lake where water depths reach 150 to 300 feet. This is not a surprise given the knowledge that the lake water surface and thus the beaches have been as much as eighty feet lower in the last eight thousand years. The presence of sand indicated periodic mobilization, possibly from some combination of storm waves, internal waves and currents. The lack of indications of the presence of finer silt and clay accumulations precludes any thoughts about this being a stagnant lakebed setting, even out in these depths. The magnitude of any exchange of sand through transport between the beach and near-shore bars and these greater depths has not been possible to analyze based on the available data and field sampling. Thoughts on Coastal Process Questions and Sediment Budgets During the winter of 2000-01 two sets of questions were provided to The Nature Conservancy with respect to the ELOSTS project. The aerial photographic analysis described above seemingly has relevance to some of them as addressed below. Other of the questions are apparently well beyond the scope of this GIS study. First, are the points raised by Barry Pendergrass, of the NY DOS, then those of Geoffrey Steadman, originally formulated back in April of 1997 in a report prepared for The Nature Conservancy. Pendergrass: 1) This second GIS report for TNC does focus strictly on the barrier beach changes including the offshore bars (when detectable), dune features and inlets, with careful consideration of lake water levels during the surrounding period in which the images were taken. As discussed above, it is impossible to know the precise lake level at the time of taking each photograph along the eastern edge of Lake Ontario. Table II located at the end of this report provides the daily heights of the lake on the dates of the digitized aerial photographs for the two nearest water level gauge locations; Cape Vincent and Oswego, New York. Note at these two stations there is as much as a one foot of difference in elevation, as recorded on 06/04/59. On the dates of the photographs a range of 3.05 feet has been recorded. Typical lake level changes of several feet do significantly impact the digitized location of the shoreline in the photographs because the beach face slopes are generally quite gentle (< 1:10 gradient). It is also important to recognize that storm events associated with higher than usual lake levels frequently coincide with beach and dune erosion. Inlet reconfiguration and relocation or even breaching of the barriers to form new inlets are all seen in the photographic record following some of these periods of major lake level fluctuation. This knowledge does not however provide predictive foresight. Only the management concept that during periods of high lake level storms can have greater coastal erosion impact than when the levels are lower than average. Wind direction, intensity and duration of each storm are also known to be significant. Seasonal variations of both lake level and the weather events should thus play a significant role in future management analyses. 2) The decade of the 1970s is a period that includes episodes of particularly high lake levels and not surprisingly particularly vivid changes in the barrier and inlets west of North Pond in the township of Sandy Creek, Oswego County. As described above the GIS analysis of the 1970s photos is somewhat problematic in terms of precise east-west shoreline location because the areas of interest in the available images are confined to the outer limits and thus most distorted portions of aerial photographs. There is certainly no question about major northward relocation and even creation of a new primary inlet location during this decade. 3) The issue of error analysis including sources and magnitude of each type is discussed above. All the GIS conclusions of this paper have been evaluated with the known uncertainties clearly in mind. It would have been better to know the precise lake level at the site of each photograph at the time it was taken and to also be able to identify the location of photo registration control points with greater than a 40 foot (12 meter) uncertainty. However, that is not possible with the historic photographic and lake level records and base maps available for the data collected over the last seven decades. Steadman's (1997) sand transport and management implication issues are worth discussion, as seen in light of this GIS aerial photograph analysis report. 1) The natural forces affecting littoral processes and shoreline modification indicated in the aerial photograph analysis are clearly wind and water transport, particularly associated with storm waves, rip currents, inlet currents and barrier wash-over events. Meteorologically controlled lake levels are another natural influence on the just mentioned forces. Vegetation, or lack of it, on the beach and dune deposits has significant influences as well. 2) The principal human activities and man made features affecting beach and dune portion of the eastern Lake Ontario shoreline are associated with the increased development, resident and visiting population, and associated land uses. Dwellings, access roads, parking lots and recreation roads as well as boat launch facilities, docks, navigation channel creation with improvements and periodic maintenance are all quite visible on the series of aerial photographs covering the decades. Erosion control structures are only sometimes extensive enough to be recognized on the photographic images. All of these human activities and features are designed in fact to modify the natural coastal features and/or processes. 3) The photographic analysis does not, by design, extend to comparative ocean coasts. However, the lack of tides and saline waters each have obvious impacts on the sediment dynamics and biological population factors. The fetch limited wave spectrum reduces the prevalence of swell waves and the associated beach berm development that is so typical of sandy ocean front shorelines. Once sands are moved off shore by storm waves in Lake Ontario it is thus less likely that subsequent onshore sediment transport during swell wave conditions will occur. Seasonality of lake level fluctuations is not typically a dominant factor in coastal ocean settings. 4) The sediment sources along the eastern shore of Lake Ontario are not indicated in the aerial photographs. However, the local post-glacial reworking of late-Wisconsinan deposits by wind and lake waters seems the most likely primary source of the barrier beach and dune deposits. During this multi-thousand year period of shoreline evolution and beach development lake levels are believed to have fluctuated by many tens of feet above and below present levels, not just the few feet recognized in the historic record. 5) The photographs provide no clear indication of littoral transport direction along most of the shoreline. At the inlets there are indications of reversals in transport direction, possibly with a northward bias at the North Pond inlet during the time interval between the nineteen sixties and seventies images. 6) No quantification of sediment movement is possible in this limited GIS study. 7) Relative importance of onshore-offshore vs. longshore transport is also not determinable in this type of GIS aerial photographic study. 8) If by the term "offshore boundary" the outer limit of beach sand transport is intended, then the photographic data is not sufficient to make a valid determination. Even detailed field studies with this goal have typically failed to establish such boundaries, if they even do exist. Fine sediment fractions are extensively dispersed when part of a suspended transport load. 9) Following from 8) above it is not possible to use photographs to answer this question about transport of sand in or out of the littoral zone. The uncertain definition of the "littoral zone" and unspecified time period of consideration make this a very complex question to address. 10) The extent of offshore sand deposits is best determined from a combination of the lakebed sampling, coring and sub-bottom profiling. This is addressed in the larger ELOSTS report. How extensively coastal sand deposits are affected by littoral processes has not been sufficiently studied along the eastern Lake Ontario shoreline. 11) The cobbled sections of the shoreline along the barriers are not detectable in the aerial photographs because of their small size and limits of scale. Objects of less than ten feet in diameter are hard to identify. The shades of gray in black and white photos are sufficiently variable along pure sand beach segments that the darker tones expected with some cobble areas are not distinctly recognizable. 12) Not enough is known from the photographic images, or any other source, to identify or theorize about trends in erosion and accretion along the eastern shore of Lake Ontario. Too many of the recorded changes fall within the range of known GIS analysis and other inherent photographic error sources. Accordingly, there is no justification to further quantify these changes over time. 13) No prominent examples of efforts to modify the sand transport in this sector of Lake Ontario are known or indicated in the aerial photographs. Periodic maintenance dredging has been reported for a number of the inlets with no systematic monitoring or measurement of the sedimentary dynamics, volumes or consequent patterns of modifications to even local erosion, transport and deposition. Thus it is hard to judge in terms of success and failure or even compare any management efforts on the ocean shores or those of the other Great Lakes. Most dredged inlets are reported to fill in rapidly even in the next major storm or two. 14) This GIS study was not designed or funded with sufficient scope to include comparisons with the other Great Lakes sand dune areas. 15) This GIS study was not designed or funded with sufficient scope to include any physical or mathematical modeling for the study site or the Great Lakes shorelines more inclusively. 16) Littoral transport of the beach sediments will continue to modify the eastern Lake Ontario shoreline and inlet locations as well as the extent and location of dune deposits along the barrier beach segments. Lake level fluctuations and management, particularly elevated periods with coincident storm events will most likely continue to be of greatest significance for any changes to the coast. 17) While there are always theoretical opportunities to modify littoral sediment processes, the economic climate and limited anticipatable benefits under any current and reasonably predictable uses appear to preclude any substantial large-scale private or public investment. Accepting the natural changes recorded and observed in the photographic series over the decades appears to make a much better management strategy. 18) The Nature Conservancy can use the understanding of littoral processes and resulting dynamics, revealed in the aerial photographic, and other ELOSTS studies, to educate the users and managers of Conservancy properties. Rather than expecting and working toward managed stability of coastal beach, inlet and dune features the policy of accepting natural change should be embraced. If inlets are to be periodically dredged in the future, care should be given to the selection of dredge spoil disposal sites. Lake level management should be encouraged with the fullest possible understanding that elevated water levels coincident with severe storms are the conditions most likely to engender rapid beach, dune and inlet changes along the barrier and cliff shorelines of eastern Lake Ontario. The best practices management includes keeping permanent man-made structures well back from the zone of natural migration of any dynamic shoreline, whether it is eroding or accreting. Management Implications With these several considerations of the aerial photographs, lake level and the historic charts and maps described above we now have a better handle on the processes and rates of change that can be determined. We also know that many interesting questions cannot be answered with the desired quantitative precision because of the fundamental and recognized limits to the available databases and methodologies. Never the less, there are obvious management implications worthy of consideration. They can be clustered into several related categories for convenience of presentation. However the interconnections and sometimes-contradictory aspects should not be ignored. Overall, one must appreciate the fundamentally dynamic nature of coasts when thinking realistically about the many different aspects of coastal management. The bars, beaches, dunes, inlets, ponds, and wetlands are subject to intermittent relocations and other modifications. The timing of any such changes is as unpredictable as the stormy weather that causes much of the change. On the other hand, it is certain that eventually, and perhaps only rarely, there will be major changes along any stretch of barrier shoreline. Because of the simultaneous certainty of change with uncertain timing, it is best to prevent any development that depends on coastal stability. Homes and other structures, such as access roads and parking lots, should be kept well back or inland from the zone of dynamic coastal modifications. Forcing stability for human convenience will eventually require excessive expense and confrontations with tragic losses. The mobile sands of bars, beaches, dunes and inlets are central to this dynamic aspect of unconsolidated coasts. Wise management planning excludes the installation of stable or fixed development structures in the dynamic portions of the coastal zone. Calculations of coastal retreat and erosion as well as deposition rates are hard to establish because of several variables and lack of sufficient databases. Beach erosion often alternates with deposition as sand is exchanged between the beach-face and near-shore bars. Similarly, inlet migration and relocation rates as well as the timing and intensity of storms are essentially unpredictable. Any management plan that depends on such rates and predictions is subject to probable failure. Without quantitative knowledge of sediment transport by waves, wind and currents it is virtually impossible to anticipate the consequences and successes of inlet maintenance dredging, dredge spoil deposition or other beach nourishment programs. Similarly, dune building and stabilization programs must be viewed as essentially uncontrolled experiments with little or no predictive certainty. Implications of lake level management are complex but some trends can be anticipated with confidence. High lake levels enable greater navigation access into shoal waters. While attractive to many, particularly in the ice-free boating season, it has the obvious associated risk of increased flooding and the likelihood of beach or dune erosion during windstorms. The impact of elevated water levels on coastal vegetation in wetlands and usually dry land is indeed complex. With elevated water levels the seasonality, duration and depth are worthy of consideration when trying to anticipate the impact on plant survivability and longevity as well as vigor of growth and competitive advantage for individual species. Lowered lake levels relocate the focus of wave and current action while expanding the coastal portions exposed to direct wind erosion and transport. As the off-shore lake bed slopes are typically less than those in the breaker zone, any lowering of the lake surface will stimulate a reconfiguration of the near-shore sands so as to conform more nearly to the equilibrium profile. Current flow, through the pond inlets, is more restricted when lake levels are low. So one can anticipate reduced pond level fluctuations and perhaps greater rates of inlet scour. Also dune growth and reworking should theoretically be enhanced by the wider and more exposed dry beaches during low lake levels, unless the wind blown surface is ice or snow covered. Lowered lake levels should reduce the probability of storm wash-over events across barriers and the related formation of new inlets into the ponds. However, the wider beach faces exposed during periods of reduced lake level are an attraction to those who like to drive off-road vehicles along the shore. Such traffic moves sand down the beach face toward the lake and can also lead to crushing and killing of dune grasses and other vegetation. This in turn provides greater access by the wind to any un-protected or un-stabilized barrier sediments. Finally, the variability in weather and climate patterns makes coastal zone prediction and management problematic. Since the weather (storm wind) provides the primary source of energy for many forms and processes of coastal change, this is of fundamental importance for coastal managers. Day to day changes, month-to-month variability and inter-annual fluctuations all play their part. As one decade is different from the preceding and following, so centuries are somewhat contrasting. These realities have important consequences for precipitation, runoff, evaporation, lake levels, waves, wind and presumably currents. Thus the ambition for reliable predictive models and associated specific management strategies remain an unresolved challenge. We can be certain of the expected change direction in some circumstances but not in others. Furthermore, we are not sure of how to predict when particular circumstances will combine to bring about the kind of significant changes that have occasionally occurred in the past. Both the sedimentary and biotic changes can have impacts on human activities and thus are appropriate subjects of management consideration. Ecological factors have their own innate value for many people. The relative importance of human desires and impact vs. natural coastal variability is thus important to evaluate in any worthy management analysis. Such analysis is going to be most informed if there is intentional inclusion and consideration of the full range of factors and variability that has been experienced and recorded over the last few centuries. Using shorter periods of time reduces the likelihood that a valid understanding can be developed. Short-term fluctuations must be clearly distinguished from long-term trends as discussed earlier in this report. Conclusions The analyses in this report provide a valuable follow-up to the McClennen et al. (2000) Report to The Nature Conservancy and address the several requested issues raised for consideration. The 1974 and 1978 aerial photographic coverage was registered and digitized for inclusion in the GIS database. Due to the fact that the 1970s coastal sections of interest were primarily from the perimeter sections of the photographic images the lens-effect distortion was significant; perhaps over 50 feet (15 meters). Changes in the location of the shoreline caused by erosion and deposition were often masked or overwhelmed by the simultaneous uncertainties grounded in the changes of lake level and restricted GIS registration precision. However, the changes in North Pond inlet location and associated spits and shoal location were recognizable and greater than the sum of all the other uncertainties. The attempts at quantitative assessment of shoreline migration observed through the six decades of the aerial photographic coverage in the prior study was put in perspective by the error analysis. Shoreline location changes of up to 60 feet observed in the aerial photographic images can be due to differences in lake level, photo registration limitations, or true changes of the coast. Subtle discrimination between the relative importance of these simultaneously active factors is usually not possible. Larger changes are limited to inlet locations, along the eastern shore of Lake Ontario. Calculations of the sediment volume based on charted shoaling of North Pond (178 million cubic feet = 6.6 million cubic yards) and coming in part from the barrier beach and dunes to the west enabled an estimate as to the amount of sand possibly removed from the barrier since 1878. The two to five foot sediment thickness is seen as a likely upper limit when the total length and width of the barrier and the bathymetric measurement techniques of the 19th and 20th centuries are carefully considered. Terrestrial and organic sediment sources entering from the east are also presumed to be significant in the coastal ponds. Quantitative estimates of sediment transport and volumes of greater precision are not seen as possible given the present available field observation data and limits of photographic resolution. Management planning implications are numerous and clearly interrelated based on the set of considerations and observations of coastal processes addressed in this study. The fundamentally dynamic nature of coasts, when driven by weather energized forces of waves, wind and currents, as well as changing lake levels, leads to complex possibilities and low accuracy for predicted rates and timing of critical events. The desire to develop permanent and rigid or stable structures on unstable and mobile coastal deposits should be minimized or prevented by informed coastal zone management programs. Flooding, erosion rates and deposition are all influenced by lake levels in combination with the continuously changing variables of weather and climate. The weighted values of human preferences for development and usage over natural processes should be carefully balanced by the inclusion of as many biological and sedimentary processes and factors as possible. Because of the interrelationships between so many variables and coastal factors it is virtually impossible to anticipate the development of any quantitatively precise predictive models for coastal evolution. However, long-term trends and the expectation of change as a certainty is valid and of significance for any coastal management planning. Acknowledgements Thanks are expressed to Myongsun Kong, Environmental Studies Research and Teaching Support Technician, of Colgate University, for her detailed attention to the digitizing of the 1974 and 1978 aerial photographs, the thoughtful review of GIS error sources and magnitudes, preparation of figures as well as the preparation of the associated digital record backing up this study and report. The Geography Department of Colgate University kindly provided use of the GIS laboratory, software and equipment. Don Woodrow and Sandy Bonanno have provided the considerable guidance and wisdom used to frame this follow-up study of coastal change along the eastern shore of Lake Ontario. John DeHollander of the Oswego County Soil and Water Conservation District generously provided the loan of the 1970s and other aerial photographs reviewed in this study. Bibliography Environmental Systems Research Institute, Inc. 1999, Arcview, version 3.2, Redlands, CA. Great Lakes Information Network lake level data at websites addresses; http://huron.lre.usace.army.mil/levels/text/ontario_hydrographs_, http://huron.lre.usace.army.mil/levels/maxmin.html, http://huron.lre.usace.army.mil/storm/strmini.html, hhpm@lre02.usace.army.mil/storm/ontstrm, http://huron.lre.usace.army.mil/storm/ontstrm/capevinc.html, and http://huron.lre.usace.army.mil/storm/ontstrm/oswego.html. McClennen, Charles E., McCay, Deanna H. and Pearson, Marcus E., 2000, Aerial Photography-Based GIS Analysis of the Eastern lake Ontario Shore: Coastal zone change and processes 1938-1994, Report to The Nature Conservancy, Central and Western New York Chapter, Rochester, New York and for the New York State Department of State, produced at Colgate University, Hamilton, New York, 43 p. McClennen et al., 2000 may also be accessed on the web with scanned copies of the aerial photograph images used in the GIS analysis at: http://wwwcolgate.edu/academics/geology/faculty/mcclennen.html. National Ocean Survey Website: http://co-ops.nos.noaa.gov/data_res.html. Select the Great Lakes Data Inventory or Great Lakes Station Water Level Plots for Lake Ontario. Parillo, C. T. and McClennen, C. E., 1999, Historical Air-photo and Map Analysis of the Eastern Lake Ontario Coastal Zone, Progress Memo to The Nature Conservancy, July 30, 1999, produced at Colgate University, Hamilton, New York, 18 p. Steadman, G., 1997, Eastern Lake Ontario Littoral Processes: Review of Information and Management Implications, prepared for The Nature Conservancy, Central and Western New York Chapter, Rochester, New York. SUNY Oswego illustrated study of "Changes in Sandy Pond Inlet since 1898", http://www.oswego.edu/Acad_Dept/a_and_s/earth.sci/geo_geochem/geol/sandy.html. Weir, Gary Marton, 1977, Inlet Formation and Washover Processes at North Pond, Eastern Lake Ontario, Master of Arts Thesis, SUNY Buffalo, illustrated. |
