How did East Sussex Really Appear in 1066? The Cartographic Evidence

Military history has provided significant insight into the factors determining the outcome of armed conflict through time. At the same time, it often fails to adequately assess variables unrelated to historical accounts per se that may contribute to military outcomes. For example, in 1066, English and Norman forces engaged in a decisive battle near Hastings, U.K. Numerous historical accounts have chronicled this event, using a combination of eyewitness and participant testimony, as well as written records, and art forms. Few, however, have paid significant attention to the role of the local landscape in shaping events. In the case of Hastings, the battlefield itself provides an example of the way in which geography can contribute to our understanding of historical events. By applying environmental sources and a regressive cartographic analysis, this study demonstrates that there is, in fact, considerable evidence to suggest how the landscape appeared back to the time of the battle. This finding is significant, insofar as it opens the door to new research on the Battle of Hastings which may shed additional light on the events that occurred there and the factors that influenced the outcome of this crucial conflict in British history. It also reveals the importance of applying new methodological approaches to traditional disciplines such as history, to deepen and expand existing analysis.


Introduction
For the most part, our understanding of military history relies upon a scholarly assessment of qualitative accounts of participants, eyewitnesses, stories, chronicles and other forms of evidence directly related to the event in question. However, other than broad discussions associated with terrain, academic reconstructions rarely consider the full environmental context of the site in question. By completing a comprehensive environmental context of the battle in question, military researchers would be able to better assess factors such as the choice of battle sites, the decisions taken by military commanders, and most importantly, military outcomes (Morillo 2013, 3-5).
The 1066 Battle of Hastings-the most famous in British history-provides an interesting case in point. To many historians of the Battle of Hastings, the reconstruction by Williamson is sufficient for our understanding of the region in 1066 (Tetlow 1974, 134-35, J. A. Williamson 1959, 77, Lawson 2016. However, contrary to Tetlow's claims of a "painstaking reconstruction" (1974,, Williamson, in fact, provides little in the way of environmental sources or methods to back his reconstruction. Instead, he only provides a narrative history of the region (Williamson 1959). This perspective is problematic because we do not know which sources he used or the reliability of those sources. In fact, this seems to represent yet another case of where, in the words of Rackham, historians "confine themselves to the written word or, worse still, to the literary word; they are reluctant … to see … the land itself … At best this shortens perspectives … At worst it manufactures false conclusions" (Rackham 1986, 6). In the specific case of the Norman Conquest, Williamson's approach could lead to errors in the understanding of William's campaign and hence could render many historians' claims about the battlefield environment to be false. 1 In order to 'set the record straight,' it would seem more than appropriate to revisit and as necessary revise current interpretations of the local landscape in 1066.
This study offers just such a corrective. Using the temporal categories in the East Sussex Historic Environmental Record (ESHER) as a reference, it undertakes an evidence-based comprehensive reconstruction of the Hastings landscape, stepping it back to how it likely appeared in 1066 (East Sussex County Council 2013). Temporal categories examined include Present-1946Present- , 1945Present- -1914Present- , 1913Present- -1800Present- , 1799Present- -1600Present- , 1599Present- -1500Present- , and 1499Present- -1066. As will be revealed, only by reconstructing the Hastings area back to the eleventh century in this way, can work subsequently proceed to effectively test historians' interpretations of the choices and decisions made by the English and Norman armies of the time. In pursuing this task, this study begins with a brief discussion on the context of the battle followed by a survey of the literature on the British environment back to the medieval period. This is followed by a discussion of the sources to be used in the reconstruction, and of the methodologies that are employed in its execution. Finally, in conclusion, the study will consider the implications of its findings for historical accounts of the Battle of Hastings and more generally the importance of geographic analysis to historical studies.

Historical and Environmental Context
In the year 1066, there was a succession crisis to the English throne. In early January, King Edward of England died and his great nephew Edgar, by then in his early teens, was considered still too young to rule (Lawson 2016, 26 Agricultural land use, farming, yields and weather patterns are also examined by Brandon (1971aBrandon ( , 1971bBrandon ( , 1972

Cartographic Sources
To begin with, there are a series of historical maps from the twentieth to the fourteenth centuries which can prove useful in interpreting the local landscape. There is also a wide range of Ordnance Survey maps from the early twentieth and nineteenth centuries which can be accessed. In this study, the maps from the third and first edition will be reviewed. The third edition was surveyed during the First World War and represents the start of the interwar period between the First and Second World Wars (Brighton & Eastbourne, 1920;Hastings, 1921).

Analytical Methods
For this study, analysis of these cartographic sources was undertaken using "map regression". According to Rippon (2004, 79), map regression is an analytical technique, where the researcher begins with a modern map of a location and then proceeds back in time to progressively analyze older maps of the same location.
This process typically involves a sequence where one begins with "Ordnance Survey  Kayan andErol 1980, Kraft, Rapp, et al. 2003).

Results
Based on these historical maps, historical land uses stepping back to 1066 were developed. The first step in developing and graphically presenting the historical land uses was to georeference the large-scale historical maps. These maps were geo-referenced through the affine transformation to the 2012 Ordnance Survey data.
These geo-rectifications typically went through several iterations with points being added or removed until an acceptable Root Mean Square Error (RMSE) was achieved.
For most of this study, the RMSE was between 10 and 20, but some were much lower 2 Accuracy is indicated by an R2 value of at least 0.75 for a geographically weighted regression model as depicted in Figure 12 in Lloyd and Lilley's paper (2009,(41)(42)(43)(44).
Hewitt: How did East Sussex Really Appear in 1066? The Cartographic Evidence Art. 9, page 9 of 25 (the geo-rectifications where the RMSE was under 10 were maps with only a small area to be referenced). These values were considered acceptable because the maps were historical (accuracy may vary) and in some areas, there were significant changes to the landscape since the maps were developed, such as in the case of urban Hastings.
There were also potential errors in the surveying when the maps were initially drawn.
This type of geo-referencing was applied to the maps from 1920 -1921, 1884 and 1778-1783. For the older maps, their evidence was considered in conjunction with the detailed land use maps and the data in the ESHER database.
Once the historical maps were geo-referenced, the areas in the ESHER database were classified into the various land uses by a period as defined in the ESHER. Table 1 defines the land uses depicted in the figures. This classification was based upon the maps or database respectively and was conducted through several methods. The first was to associate each period with the land uses which were classified as being from that period. For example, woodland classified as early medieval (410-1065) was assumed to be woodland through the entire study period. This assumption provided a rough starting point for the land uses in each period. From here the unclassified areas in between the classified ones had to be interpreted. The classification was estimated by overlaying the ESHER map over the geo-referenced historical maps, as discussed above, and setting the ESHER as transparent. This process enabled the historical maps to be seen beneath the ESHER map. Then the land use for each area was re-determined based upon the interpretation of the historical maps. For example, if the historical map indicated woodland, then the corresponding ESHER area was classified as woodland.
The interpretation text in the ESHER database was used for the periods before the eighteenth century insofar as no detailed map existed before that point in time.
Once all of the land use maps were complete, the Hastings area data was generalized to a resolution of 250 m by 250 m. This resolution was selected because a finer resolution in particular with the historical reconstructions "could lead one to the conclusion that we know more than we really do" (Heinen 1998, 189). The generalization was calculated as a vector to raster transformation where the vectorbased data were converted into a raster surface. A raster surface is "[a] regular "grid cell" approach to defining space" (Bolstad 2005, 511). The technique assigned a land

Coastal
Land that is along the coast and could be flooded at high tide.

Communications
Forms of communication typically from the nineteenth century onward.

Cropland
Areas which are used for crops.

Designed Landscapes
Areas which were planned such as urban and rural parks.

Fieldscapes
Land which is employed for farming or pasture but are unsure which is most correct.

Horticulture
Areas designated for growing such as orchards or greenhouses.

Industry
Sites currently or formerly employed to extract or process minerals.

Military
Defensive structures such as castles.
Pasture Areas for pasture.

Reclaimed Marshland
Areas which were formerly marshland -either tidal or freshwater.

Recreation
Locations of sporting events such as race tracks or marinas.

Settlement
Areas employed for human habitation originating with historical cores.
Unimproved/Unenclosed Open areas such as common lands.

Water
Areas with open water.

Woodland
Areas covered by trees.

Hewitt: How did East Sussex Really Appear in 1066? The Cartographic Evidence
Art. 9, page 11 of 25 World War. There are also notable sites of horticulture, recreation and water bodies such as reservoirs. 3 Figure 3A and B represent the land uses in the period between the two world wars of the twentieth century. Compared to the post-war period (Figures 1 and 2), the main change is in the urban coverage. Development of these maps was based on the evidence in the ESHER as well as the Ordnance Survey (OS) maps from the 1920's (Brighton & Eastbourne 1920, Hastings 1921, ESHER 2013.    The next two maps as presented as Figure 3I and Sussex was consulted for further information as well (1999,(36)(37)(38)(39)(40)(41)(42)(43)(44)(45)(46)(47)(48)(49)(50)(51). Given the available data for this period, these figures represent an approximation of the land use in the late medieval period. Figure 3J represents the land uses for the Battle area for the same time period. As it is difficult to assess where the agricultural land, meadow, and pasture were specifically, the land use "fieldscapes" was applied across the entire area. Figure 3I indicates that a large portion of the area was woodland in this period with intermixed agriculture (P. F. Brandon 1969Brandon , 1972. Small settlements had developed on the landscape as well as a small number of castles such as Hastings and Pevensey (Bradbury 1998, 87 and 122-23). Figure 3I also reveals the expanse of the arable land in the middle ages, reflecting the population increase in the early part of this period. Some of this land represented woodland clearances known as assarts which were transformed into arable or pasture (Hoskins 1977, 86-7, Rackham 1980. However, following droughts and disease, especially the Black Death in To estimate this landscape, Figure 3I was revised based on the estimated changes to the landscape between 1499 and 1066 (Harvey, 1985).
Specifically, it can generally be assumed that approximately 80 percent of the 1086 acreage was still considered arable land in 1914 (Cantor 1982, 17). 6 This knowledge provides us with a starting point from which the 1066 land use can be 6 It was also indicated that this value varied throughout England for example; the proportions of arable were higher in the Sussex plains but lower in the Weald (Cantor 1982, 17). To estimate crop land coverage, two estimates were deduced from a selection of local Domesday sites. These methods, utilized by Lennard and Maitland produced estimates based upon the number of plough-teams (100 acres) and lands (120 acres) per manor (Lennard 1959, 393, Maitland 1987. As the study area included 374.5 plough-lands or 474.875 plough-teams then the rough extent of arable was between 44940 and 47487.5 acres (Campbell 2000, 386-87, Darby 1977.  In terms of woodland, the largest change in land use presented in Figure 3I and  (2013). An assart is a Middle English word, defined as forest cleared for agriculture (Field 1972, 267). In the ESHER database, around 86 percent of the assarts were dated to the medieval period (1066 to 1499) with none dated before 1066 (P. F. Brandon 1969). Therefore, they were begun after With regard to characterization of the Battle area (Figure 5), fortunately much has been written about Battle Abbey in the medieval period to assist in the regression from Figure 3J. For example, Searle has characterized the Battle environment as without "strips because there were no common fields. There was only the vill, and beyond it, the leuga, divided into the wists characteristic of the weald: single arable farmsteads enclosed by the forest" (1974, p. 82).
Another author suggests that the woodland tended to be on the marginal sections of the landscape such as the small valleys or on the steeper slopes and the remainder of the landscape was fieldscapes (Baker 1973, 423-24). A historian of the battle has indicated that the original hill would have been covered in local grasses (Bradbury 1998, 142). Thus, based upon these characterizations, Figure 5 presents what is in all probability, the landscape the English and Norman combatants would have seen when they arrived in 1066.

Discussion and Conclusion
This study has improved on the Williamson limited interpretation (1959)  undertaken here provides for a detailed analysis of nature and extent to which the terrain has been affected by erosion, perhaps providing clues as to where valuable physical evidence of the conflict may be buried. To date, in fact, no artifacts from the battle have been located (Porter 2012, 15-16).
Thirdly, the analysis would facilitate a much more comprehensive assessment of the local resources available to the English and Norman armies at the time, and how this, in turn, may have affected their battle-readiness. As another often-neglected area in the study of military history, this would force a great focus on the resources required to maintain an army in a battle-ready state, not to mention how the local economy and society may be affected by the presence of large fighting forces.
In terms of methodology, this research would have been more labor and time intensive without the benefit of digital resources. Specifically, had the ESHER and the extensive research behind it not been available, a significant portion of this project would have been devoted to data input and management. This requirement would have increased project costs and required the expertise of a data analyst to manage the data. With respect to the historical maps, before their digital appearance online, researchers would have had to search through lists of manuscripts in archives and visited their locations to see if the map was relevant to the project. Now a simple internet search was all that was required to locate many of these maps. Thus, these digital resources have increased the efficiency of the research process and increased access to sources while reducing time spent searching and project costs.
More broadly, this study clearly reveals the importance of maps and cartographic analysis in both broadening and deepening understanding of military history per se.
As a research framework, maps bring an environmental focus to a topic and center it on the environment. While accounts and chronicles are extremely important in scholarly analysis, a fuller understanding of the environmental context through environmental sources such as maps can provide a much more comprehensive reconstruction of a historical environment. With the reconstruction historians are thus able to provide an interpretation of a historical military event which is grounded in the environment it occurred in or as close as we can get to that environment. Thus, Hewitt: How did East Sussex Really Appear in 1066? The Cartographic Evidence Art. 9, page 19 of 25 this perspective can provide a pathway to a much more sophisticated understanding of the factors affecting the context and consequences of armed conflict.
Overall, this study underscores the importance of applying up-to-date methodological techniques as a means to test, further explore, and expand upon existing disciplinary methods. This is particularly relevant to history, which through time has relied almost exclusively on written accounts and qualitative analysis.
This study opens the door much wider to the application of novel approaches that not only push the boundaries of traditional thinking but provide a much more sophisticated-and potentially more accurate-window on past events.