ArcGIS · Terrain modelling · Viewshed analysis · Remote sensing
A complete GIS workflow integrating terrain, Earth-observation imagery, soils, hydrology and site records to evaluate how topography, water access and visibility influenced spatial distribution.
Independent spatial analysis and technical reporting
24 recorded locations, Landsat 8, OS Terrain 5, soils, rivers and shoreline
Raster mosaicking, terrain derivatives, proximity modelling, viewshed analysis and statistical comparison
Evidence-led landscape interpretation and targeted research priorities
Source imagery, terrain derivatives and visibility models are presented as a single workflow. Use the controls or arrow keys to move through the outputs.
01 / 10Regional terrain model combining elevation, shaded relief, rivers, shoreline and the four recorded-site categories. This establishes the spatial structure used throughout the analysis.
The study tested whether terrain, access to water and visibility explain the distribution of 24 recorded locations across the Beauly, Moray and Cromarty Firths. Archaeological records functioned as the point dataset; the principal subject was the wider landscape system.
A 30 × 30 km geodatabase integrated Landsat 8 imagery, OS Terrain 5 elevation, soil classifications, hydrology, shoreline and recorded-site data. Elevation tiles were mosaicked and clipped, then standardised from 5 m to 20 m resolution to balance processing efficiency with regional analysis. Selected soil classes provided a cautious proxy for potentially cultivable land.
Terrain derivatives included hillshade, slope and aspect. Each location was evaluated against elevation, terrain form, 3 km access to rivers and coast, and individual or cumulative viewsheds. Twenty random land points provided an exploratory comparison baseline.
Most locations fell below 150 m and occupied slopes under 10 degrees. Every recorded location lay within approximately 3 km of either a river or the shore. Visibility models revealed extensive coverage across the firths, establishing water as both an access corridor and a recurrent component of the visible landscape.
The mapped patterns are robust as exploratory evidence but not predictive proof. Statistical power was constrained by a small, uneven sample; the soil layer was coarse; one source record was duplicated; and the historic shoreline was not reconstructed. All reported t-tests returned p-values above 0.05 and therefore did not establish significant differences from the random-point baseline.
Principal spatial conclusion