In a previous post, I looked at how we might model some of the modern factors that affect the distribution of PAS finds. I termed these factors “affordances”, as they can be thought of as helping to govern the opportunity for metal detectorists (and others) to be able to discover archaeological material. I then began thinking about the affordances that affect the opportunity / likelihood of sites and monuments being discovered (if they were there to begin with in the past, of course).
I would argue that the six primary routes by which sites and monuments are discovered in the present day are:
- As clusters of spot finds.
- As documentary records (or place names, etc.).
- Via excavation / other intrusive field evaluation (e.g. watching briefs, test pits, etc.).
- Via geophysical survey.
- Via aerial photography as crop marks.
- Via aerial photography as earthworks.
To these could be added other routeways, such as LiDAR prospection, but these account for quite low percentages of sites and monuments at present. Of the affordances relating to these six possible discovery pathways, number 1 has been outlined in the post referred to above, number 2 is, I think, impossible to define / quantify, and number 4 is hard to define / quantify. Therefore, I have focused on numbers 3, 5, and 6.
Excavation / other intrusive field evaluation:
The main affordance in this category is where the opportunity to undertake archaeological excavation (etc.) exists or has existed. Ideally, this variable would be based on data not explicitly linked to archaeological investigation: i.e. planning decisions on major projects and small projects in sensitive areas (albeit sensitive areas brings in an archaeological element), but this data is very hard to discover and collate. In particular, planning statistics are inconsistently archived on the UK Government websites and changes in planning authority boundaries over time make collation particularly problematic. Although I feel that putting the effort into compiling this information (since 1990 ideally, i.e. post PPG16) would be worthwhile, it would be too great an amount of work to undertake for our current project and purposes.
Therefore, I decided to base this affordance variable on data for where excavations (etc.) have taken place. The most complete source of national data on this is the Excavation Index maintained by English Heritage and hosted by the ADS. I extracted relevant types (e.g. excavation, watching brief, etc.) for post-1990 events and constructed a KDE plot of the resulting distribution. I included events resulting in discoveries of all periods in order to minimise any effects of period bias. The results were very “peaky” in the sense that the density of events in certain cities (in particular London) dwarfed that of the rest of the country. In order to create my affordance surface, I therefore capped off the variation at +3 standard deviations (c.1.45 events per sq. km.), and divided by this maximum value to produce an affordance surface that varied between 0 (low chance) and 1 (high chance):
The results look pretty convincing, with areas of well-known high amounts of fieldwork showing up in the reds (e.g. London, the Upper Thames, Peterborough, Cambridgeshire). As such, I am pretty happy with this result.
For (both types of) aerial photographic prospection, I based the affordance on two (overlapping) factors: modern land use and obscuration of the ground surface. As with the PAS affordance, the land use was taken from LCM 2007 data and the obscuration based upon my earlier EngLaID work using OS and BGS data (1)(2)(3). To keep it (relatively) simple, “cropmark” affordance was built from arable land with areas obscured by human (buildings, etc.), environmental (water, woodland, etc.) and soil factors (see previous posts) removed, and “earthwork” affordance was built from grazing land with areas obscured by human and environmental factors removed (but not the soil factor). The results can be plotted onto the same map, as the areas of affordance are mutually exclusive (green = earthwork affordance present, orange = cropmark affordance present):
Although there are some issues with this model, in particular the fact that cropmarks can appear on grazing land in dry years, the results look fairly robust.
As a final step in the process, I then combined these three affordance patterns based upon the relative percentages of records in our database recording each discovery method as an evidence type. This results in a composite affordance map for sites and monuments, albeit only based on the types of evidence that it is possible for me to map affordances for:
Although the model is clearly imperfect, the results do look intuitively correct, with the more archaeological dense areas of the country generally showing up in yellow and with clusters of high affordance around historic towns / cities. The lower affordance levels in the north west, around the Wash and across the Weald are genuinely (relatively) low areas of site / momument density in our database. Some areas of lower affordance can have higher densities of sites / monuments, which is particularly true of Bronze Age activity in upland areas (especially Dartmoor and the Peak District), but I do not think that this represents a problem. There will inevitably exist areas with low affordance but high densities of sites (and vice versa) and, in some ways, these are perhaps the areas of most interest (as they are the ones where we can say that there are definitely past high activity levels coming through rather than simply being due to intensity of investigation)?