Scholarly article on topic 'The assimilation of historic photography and cartography into longterm coastal geomorphological analysis'

The assimilation of historic photography and cartography into longterm coastal geomorphological analysis Academic research paper on "History and archaeology"

CC BY-NC-ND
0
0
Share paper
Academic journal
Procedia Environmental Sciences
OECD Field of science
Keywords
{Rephotography / "Historic Imagery" / "Coastal Geomorphology"}

Abstract of research paper on History and archaeology, author of scientific article — Peter G. Bannon, Ian W. Marshall, Mike James, Suzana Ilic

Abstract Long-term coastal changes have traditionally been analyzed using cartographic sources supplemented by anecdotal descriptions of the coastal configuration where available. These studies have been susceptible to many problems as the cartographic sources differ in terms of the surveying techniques, the accuracy, and the representation of features on the final map, amongst other factors. Modern research on coastal areas has focused on using data collected over recent time-scales and often ignores the historical information provided through maps, because of the problems that are faced with defining the true accuracy and reliability of the information that they provide. In this study, historic photographs used on popular picture postcards are integrated as a data source for the purpose of assessing and supplementing information contained in historic maps. Using the coastline of Heysham, Lancashire, UK as a case study, it is shown that historic photography (postcards) allows the researcher to quantify the accuracy of historic Ordnance Survey (OS) maps at a fine spatial scale, and identify the timing of coastal change that occurs between mapping surveys. The improved information is useful for a variety of purposes, such as land-use change or urban development history.

Academic research paper on topic "The assimilation of historic photography and cartography into longterm coastal geomorphological analysis"

Available online at www.sciencedirect.com

W ScienceDirect

Procedía Environmental Sciences 2 (2010) 527-534

International Society for Environmental Information Sciences 2010 Annual Conference (ISEIS)

The assimilation of historic photography and cartography into long-term coastal geomorphological analysis

Peter G. Bannon*, Ian W. Marshall, Mike James, Suzana Ilic

Lancaster Environment Centre, Lancaster University, Lancaster, LA1 4YQ, United Kingdom

Abstract

Long-term coastal changes have traditionally been analyzed using cartographic sources supplemented by anecdotal descriptions of the coastal configuration where available. These studies have been susceptible to many problems as the cartographic sources differ in terms of the surveying techniques, the accuracy, and the representation of features on the final map, amongst other factors. Modern research on coastal areas has focused on using data collected over recent time-scales and often ignores the historical information provided through maps, because of the problems that are faced with defining the true accuracy and reliability of the information that they provide. In this study, historic photographs used on popular picture postcards are integrated as a data source for the purpose of assessing and supplementing information contained in historic maps. Using the coastline of Heysham, Lancashire, UK as a case study, it is shown that historic photography (postcards) allows the researcher to quantify the accuracy of historic Ordnance Survey (OS) maps at a fine spatial scale, and identify the timing of coastal change that occurs between mapping surveys. The improved information is useful for a variety of purposes, such as land-use change or urban development history.

© 2010 Published by Elsevier Ltd.

Keywords: Rephotography; Historic Imagery; Coastal Geomorphology

1. Introduction

Historical Maps, Charts and Literature have traditionally been used as the only potential source of information for analysing long-term morphodynamics of Coastal Systems. Naturally, the sources of information can contain errors which should not be overlooked when deriving quantitative estimates of coastal indicators. The reliability of the information given on maps changes through time as survey techniques, sampling strategies and mapping standards improve [1-3]. It is therefore important that supplementary sources of information are used within coastal assessments [1]. Photography has become an essential utility in the scientific community in a broad range of circumstances since its advent during the 19th Century. Developments of better camera models served to create a

* Corresponding author.

E-mail address: p.bannon@lancs.ac.uk

1878-0296 © 2010 Published by Elsevier doi:10.1016/j.proenv.2010.10.057

boom in the production of picture postcards in the early 20th Century and quickly established a modest trade around the world. There is therefore a potential wealth of ground based photography available that could be used in conjunction with the analysis of historic information to identify problems with maps in analysis of long term morphodynamic changes on coastlines. A commonly used technique for analysing changes in landscape morphology is Repeat Photography (Rephotography). This involves a researcher re-locating the position of an old photograph as accurately as possible, either via an iterative process or through photogrammetric application using known coordinates for identifiable points within the original photograph. Although it has classically been used as a process for collecting qualitative information there have been studies that have derived quantified results [4,5]. Recent advances in computer vision technology has brought about the use of historical photographs and online collections to model cityscapes in 3 and 4 dimensions [6,7]. The purpose of this study is to highlight the potential for the integration of historical photographs (postcards) as an additional data source for diagnosing inaccuracies in historical maps as well as potentially deriving quantitative measurements of coastal change. This is shown through select examples where rephotography and simple observational analysis are integrated with observations derived from Cartographic analysis.

2. Study Area

Heysham (Figure 1) is a small village situated on the eastern coastline of the largest single intertidal zones in the UK (~343.39 km2)[8]; Morecambe Bay. This vast space hosts dynamic channels that have been known to migrate large distances over short periods of time [8-10]. Heysham is one of a few coastal areas of Morecambe Bay that has both hard and soft rock cliffs in close proximity to dynamic intertidal channels as well as historic tourist areas. It is expected that this proximity would have a positive influence on the accessibility of photographs leading to a greater understanding of morphodynamics within the region.

Fig .1. Study Area, Heysham, Lancashire. It is divided into 5 distinct zones (A-E) for analysis. (© Crown Copyright Ordnance Survey. An EDINA Digimap/JISC supplied service.)

Table 1. Determination of map quality from Historical information.

Map Edition

Year(s) of Publication

Quality

OS County 1st Revision 1:2,500

OS County 2nd Revision 1:2,500

OS County 3rd 1:2,500

Revision(a)

1:2,500

OL: 1891; ED: 1889-1891 OL: 1911; ED: 1913 OL: 1930 - 1931; ED: 1933

OL: 1937; ED: 1937

Ambiguous / Unreliable Good / Reliable Mixed / Partly Reliable

Mixed / Partly Unreliable

OS County 3rd Revision(b)

National Grid Survey 1:2,500 OL: 1961; ED: 1963 Good / Reliable

Abbreviations. OL: Derived from Oliver (2005); ED: Derived from Edina Digimap Metadata.

3. Background

The Ordnance Survey did not begin surveys of Lancashire until 1841, when the survey of Ireland was nearing completion. A lot of details regarding the surveyors and methods used for map production have been lost due to bombings in the 2nd World War, but where information is available, speculation as to the reliability of the maps may be permitted (Table 1). For more detailed information on the History of the Ordnance Survey, the reader is referred to Seymour (1980) [11] and Oliver (2005)[12].

With speculation regarding the quality of the Ordnance Survey maps drawn from scarce information, it can be hypothesized that the information available alone is insufficient to assess the quality of historical OS maps. Old photographs and postcards may however be utilised for assessing these reliability speculations therefore presenting some hypotheses with which to base the study.

Hypothesis A: From historical analysis of the Ordnance Survey, published maps can be given a simplified quality rating (Table 1).

Hypothesis B: Historical Information on the Ordnance Survey is insufficient alone to assess the quality of information provided on historical maps.

Hypothesis C: Old Images and Photographs can be used as a source of information for assessing the reliability of the use of Historical Information in long-term Coastal geomorphological analysis..

3. Method

Figure 1 shows the Heysham area separated into distinct zones of interest. The maps used in this study are of scale 1:2,500, with tiles (1km x 1km) analysed individually in order to account for any distortions that may have arisen in the scanning of the maps into a digital format. For each tile, 5+ Ground Control Points (GCP's) are selected for geo-rectification. These points are chosen to be as close to the coastline as possible in order to deduce the best fit between subsequent tiles. The cliff-top edge is then traced to allow comparison of change throughout the temporal extent of the Ordnance Survey maps as it is the most ideal shoreline indicator for high bluffs and cliffed shorelines [13].

Comparison of a historic photograph and a photo taken at the same location today can allow the researcher to visualise and potentially quantify change over the time-frame. There are however some obstacles that need to be overcome; date of the photograph, original image quality and constraint of the location/perspective of the photograph[4]. These are mitigated through the use of features that are inherent with postcards, such as the postage date, publication record number, publisher and image contents such as women's fashion, automobiles, buildings etc. Where it was possible to replicate the position of an historic image, a number of exposures were taken of the scene in and around the generalised predicted area using a Canon EOS 300D Digital SLR with an 18-55mm lens and a compact Samsung S1070. A simple comparison, by overlaying the most recent image over the postcard is then carried out in order to identify missing blocks from hard rock cliff faces and/or visible changes in the cliff line.

4. Results

Figure 2 shows the top-of-cliff position for OS maps in all zones over the last 120 years. Of the 5 zones depicted the most significant changes are visible in Zone A with greater than 20m cliff recession between 1889 and 1937. The maps also record changes of the coastline with construction of sea defences in zones A and B and of minor cliff recession throughout the other zones. There are also anomalous changes of the coastline with some cliff-top traces showing progradation, particularly of the promontory at Lions Head in Zone C. Table 2 shows the coverage provided by the image collection so far. Presented here are the interesting cases in the areas that show change unreported within the OS maps.

Figure 3 presents an example of re-photography within Zone C where the highlighted area shows a significant change which is not recorded in the maps after 1937 (maximum date of photograph ~1936). Images of the soft sediment cove to the east of Lions Head (Zone C - Figure 4) show visible change over time, which is only represented on the OS maps by 1963. This highlights that the cliff lines of 1933 and 1937 are unreliable and that the 1891/1911 revisions need more scrutiny. Within Zone E, all the images are focussed on one particular area given the name "Fairy Chapel" (Figure 2). The images in Figure 5 record the degradation and eventual disappearance of a sea stack and arch some time after the early 1930's. The OS maps only show significant change in the area with the publication of the 1963 National Grid map. There is currently insufficient coverage of photographs available after 1933 to narrow the time period of its disappearance.

5. Discussion

Results show that in cases where there is significant recorded change (e.g. Zone A and B), OS maps are quite reliable. For the coastal sections that are relatively distant from the main population centre of Heysham (Zones C -E) postcards have elucidated inconsistencies with predicted map reliabilities derived from historical information sources. The use of historical imagery in identifying these inconsistencies serves to support hypothesis B and C, allowing the researcher to diagnose reliable tiles for use in coastal change analyses. Where re-photography was used to identify block removal events (Zone C and E) the biased temporal skew of collected postcards does not allow more accurate inferences of the time of block removal within the latter half of the 20th century.

Fig. 2. Changes in the Cliff top line over the last 120 years. (© Crown Copyright Ordnance Survey. An EDINA Digimap/JISC supplied service.)

Peter G. Bannon et al. /Procedia Environmental Sciences 2 (2010) 527-534 Table 2. Temporal coverage of images within the defined zones.(Note: Images can span multiple zones)

Zone 1891 - 1911 1911 - 1933 1933 - 1937 1937 - 1962 1962 - Total

Present

A 13 9 2 7 2 33

B 13 11 7 7 3 41

C 3 7 4 5 5 24

D 1 1 1 0 1 4

E 7 6 2 0 1 16

Fig. 3. A significant event occurs from c.1937 (A) - present; (B) that is not represented on OS maps.

Fig. 4. Visible change behind Throbshaw Point / Lions Head. A) c. 1900; B) c. 1970;C) 2009.

Fig. 5. The disappearance of "Fairy Chapel". A) c. 1900;B) RY - 1929; C) No post date or publisher;D) 2010.

6. Conclusions and Further Work

Supplementary records to Historical OS maps do not provide sufficient information to assess the reliability of the cartographic representation. This study has shown that with sufficient spatio-temporal coverage, historical imagery can be used in conjunction with historical maps to diagnose the quality of the information preserved more reliably. Further work is required in order to derive quantitative estimates of coastal change; however it is possible with extra information derived from modern 3D scanning equipment. The process of re-photography on cliffed shorelines is hindered in areas of significant land loss and requires development of specialised equipment and software to access positions that are no longer possible on foot. Where it is possible to reproduce images from postcards, the temporal information provided along with simple comparisons can allow for identification of events at finer spatial and temporal resolutions than is possible through historical maps alone. It is also shown that there is a greater need for a more equally distributed density of images both spatially and temporally. In order to aid this collection a web portal is soon to be made available, allowing the uploading, online storage and visualisation of postcards and photographs.

Acknowledgements

This project is funded by the Engineering and Physical Science Research Council (ESPRC). Postcards were donated to the project by Rex Wilmhurst, Dr. Julia Gillen (Lancaster University), the North West Film Archive (Manchester), the Lancaster Maritime Museum, John Holding and Jean Kirkbride of the Heysham Heritage Association and Lucy Belanger of the Heysham Library.

References

[1]Carr AP. The significance of cartographic sources in determining coastal change. Timescales in Geomorphology, Edited by Cullingford RA, Davidson DA and Lewin J1980: p. 69-78.

[2]Baily B, Nowell D. Techniques for monitoring coastal change: a review and case study. Ocean & Coastal Management 1996; 32 (2): 8595.

[3]Van der Wal D , Pye K . The use of bathymetric charts in a GIS to assess morphological change in estuaries. The Geographical Journal 2003; 169(1): p. 21-31.

[4]Munroe JS. Estimates of Little Ice Age Climate Inferred through Historical Rephotography, Northern Uinta Mountains, U.S.A.. Arctic, Antarctic and Alpine Research 2003; 35(4): p. 489-498.

[5]Bierman PR, Howe J, Stanley-Mann E, Peabody M, Hilke J , Massey CA. Old images record landscape change through time. Geological Society of America Today 2005, 15(4): 4-10.

[6]Snavely N, Seitz SM and Szeliski R . Photo tourism: Exploring photo collections in 3D. SIGGRAPH 2006; p. 835-846.

[7]Schindler G, Dellaert F, Kang SB, Inferring Temporal Order of Images From 3D Structure. IEEE Conference on Computer Vision and Pattern Recognition; 2007.

[8]Mason DC, Amin M, Davenport IJ, Flather RA, Robinson GJ , Smith JA. Measurement of Recent Intertidal Sediment Transport in Morecambe Bay using the Waterline Method. Estuarine, Coastal and Shelf Science 1999, 49(3): p. 427-456.

[9]Kestner FJT. The Effects of Water Conservation Works on the Regime of Morecambe Bay. The Geographical Journal1972; 138(2): p.178-196.

[10]Pringle AW. Erosion of a cyclic salt marsh in Morecambe Bay, north-west England. Earth Surface Processes and Landforms 1995; 20: p.387-405.

[11]Seymour WA (ed.). A History of the Ordnance Survey, Folkestone: Dawson Publishing; 1980.

[12]Oliver R. Ordnance Survey Maps: A Concise Guide for Historians, Second Revised Edition, London: Charles Close Society; 2005.

[13]Crowell M, Leatherman SP, and Buckley MK, Historical Shoreline Change: Error Analysis and Mapping Accuracy. Journal of Coastal Research 1991; 7 (3): p. 839-852.

Peter Bannon is presently a second year PhD student with the Lancaster Environment Centre at Lancaster University. He received his BSc (Hons) in Geology from University College Dublin in 2005 and a Diploma in

Computing from the Open University in 2008. Before taking his current role in the Lancaster Environment Centre he received his MSc in Environmental Informatics from the same institution in 2008. Mr. Bannon's main research areas include, assimilation of historical data sources, long-term morphodynamics, image processing and environmental informatics.