Detailed Methodology

Sources and Base Date of Assessment

The Hampshire historic landscape assessment is based essentially on interpretation of present day OS maps backed up by several other published works, historic maps and data held by the County Council. The mapping was carried out using the latest (1996/7) "Explorer" and New Forest "Outdoor Leisure" series of OS 1:25,000 maps, which covered almost the whole county.

Definition of Historic Landscape Types

In defining the range of Historic Landscape Types to be mapped, no particular limit was set on the number of variants that might be defined within these broad principles. The basic approach to the final decision of what range of variants should be defined was pragmatic based on the following criteria:

The morphological, spatial, functional or chronological, distinctions within the broad types should be useful for analysis
The types should be identified and mapped reasonably easily using the sources available
There should be a sufficient range of types to avoid having to make difficult choices about "lumping" areas into inappropriately general categories, losing useful distinctions
There should not be so many types that impossibly fine distinctions would be required which could not readily be discerned from the sources

In effect this meant that the final definition of Historic Landscape Types to be mapped was based on what would both achieve the objectives of the study and was comfortable to work with in terms of the constant choices and decisions needing to be made in the mapping process.

A total of 85 Historic Landscape Types were defined and these were grouped into 14 broad categories and used in the mapping. These are listed and explained in more detail in the Historic Landscape Types Listing.

Pilot Study

The first stage of the study consisted of the identification of sources, the drafting of the list and character of Historic Landscape Types, and the piloting of the mapping technique, together with checking air photographs and some field observation.

Six 10 km squares were mapped for the pilot, covering different representative parts of the county. This successfully showed that the typology was useable and led to various refinements in the types defined.

A rapid visual assessment concluded that generally the Historic Landscape Types do not conflict with landscape character, but that they often represent finer distinctions that may be relatively difficult to see. This was particularly the case with some field patterns where the shape of fields may be very difficult to discern on strongly undulating ground. This reflects a difference between conventional landscape assessment, which includes the "horizontal" viewpoint of field assessment, compared with the essentially "vertical" viewpoint of the almost entirely map-based historic landscape assessment.

Mapping

At the beginning of the project it was an objective to develop a full GIS that would provide the County Council with considerable flexibility and power of interactive analysis (it is not possible to represent this fully within this site).

Two researchers worked together on the mapping. Using two people had the obvious potential disadvantage of introducing systematic inconsistencies reflecting their particular perceptions and interpretations, but it also had several distinct advantages. It meant that:

there was automatic checking for consistency when edge matching was carried out for different overlays mapping
there was continuous, mutually supportive discussion of instances of difficult interpretation with the back-up of a third opinion
there was flexibility to share more tedious tasks helping to maintain morale
the execution of the mapping was faster than it would otherwise have been.

A checking and amendment process provided a third internal checking mechanism, which was carried out on the basis of agreement with the two researchers, so that the final map is in effect a three-way consensus of interpretation.

Mapping of the Historic Landscape Types was carried out manually in pencil on drafting film overlays pre-printed with 10 km square referencing points and a series of check boxes to record the checking and revision process. Up to two adjoining 10km squares were mapped on each overlay. The overlays consist of a continuous mosaic of polygons representing areas assigned to the Historic Landscape Types, each identified by the appropriate type code number.

The mapping was carried out at 1:25,000, with the film overlays directly positioned over the 1:25,000 OS maps. A major advantage of carrying out the mapping using original 1:25,000 maps rather than reduced copies (as used for the an earlier study of Cornwall) or large scale digital black and white digital mapping, is that the colour greatly helps interpretation. This was especially the case with the Woodland and Valley Floor and Water Management types described above. The use of drafting film over original maps also minimised possible distortion, though some slight movement of overlays did occur. Direct digitising over 1:10,000 digital mapping would probably now be more appropriate.

Commentaries on the mapping of each 10 km square were completed as the work proceeded in order to make a record of points of interpretation or to explain the rationale for particular category choices.

Checking

A good deal of cross checking with particular sources was carried out as the mapping proceeded, as part of the interpretative process, and a constant process of checking also formed aspects of the day to day teamwork and supervision of the mapping as it proceeded. Once substantial parts of the mapping were complete all the overlays were systematically checked and amended. Amendments were marked in red and the 10 km square descriptions were up-dated. When complete these checks were noted in the check-boxes on the overlays.

Digitising, Topology Building and Polygon Colour Coding

After the checking was complete, the film overlays were scanned and cropped to create raster map tiles which were then joined together by geo-referencing two 10 km square intersections for each tile. With hindsight, the use of four reference points would have been better as there was some minimal distortion from this process in a few cases.

The polygons were then digitised "head-up" drawing vector lines over the raster scans. The polygon topology was then created, cleaning up overshoots and undershoots at node points. The resultant polygons were then labeled.

In order to create a visually effective map the polygons were then colour-coded according to their Historic Landscape Types, each type being assigned a separate layer within the drawing so that any combination could be switched on or off. The maps included in this site illustrate this facility to some extent.

Other map data sets were added to the digital map, including a post-code classification of settlement pattern. For the purposes of the study Hampshire County Council made available their digital mapping of Landscape Character Areas and Landscape Types, modern Civil Parish and District boundaries, and the 1:50,000 OS raster base map.

Data structure and GIS analysis

The data attached to each polygon are its number, the Historic Landscape Type code, an alternative HLT code for small defence sites; centroid grid reference, and area.

The GIS capability of the map has been successfully tested by combining polygon area data for the Historic Landscape Types and the Hampshire Landscape Character Areas and Parishes respectively in order to compare the proportion of different Historic Landscape Types that occur in each Character Area or Parish. The resultant breakdown of historic landscape attributes of these areas was exported to Excel spreadsheets, which form part of the digital archive of the project.

The digital version of the map also contains several layers generated during the creation, correction and analysis of the map, including for example the raster scans of the film overlays.

Settlement Pattern Data

Additional settlement pattern data derived from postcode information is represented by a layer within the digital map, but is not covered within these Hantsweb pages. The original data was supplied by the South East Regional Research Laboratory, Birkbeck College, London. The data analysis was originally carried out by SERRL with some advice from OAU for part of the English Heritage input to the National Countryside Character Mapping programme. The analysis is based on a statistical calculation of the density and distribution of individual postal delivery points (letterboxes) per kilometre square. The problem of settlements straddling the boundary of more than one square was overcome by applying an automatic "smoothing" formula which would scan the characteristics of the surrounding squares to adjust that of the square being plotted.

The data used relates only to domestic addresses (some commercial town centres therefore appear remarkably blank). The data was converted to a kilometre grid square basis and maps can be generated indicating the degree of nucleation based on statistical formulae applied to the postcode data. A map was generated to characterise the settlement pattern of every kilometre square in terms of the number of postal delivery points and the degree of dispersal or nucleation. It thus provides a measure of density of settlement, and also gives an indication of the dispersal of the addresses. This facilitates comparisons of the settlement pattern of different Parishes, Landscape Character Areas, Landscape Types or Historic Landscape Types.

Back to Previous Page Historic Landscape Homepage Hantsweb Homepage