The Khabur River (arabic, also transliterated as Khabour, Habour, or Habur River) is a river in northern Mesopotamia that begins in the Tauros Mountains in southeastern Turkey and flows south to Syria, where it empties into the Euphrates River.

The "Harvard-Heidelberg Atlas of Settlement Patterns at the Upper Khabur River" is a repository of – mostly ancient – settlement structures in the upper Khabur basin, mainly recorded from the analysis of sources such as satellite imagery (CORONA, SPOT, ASTER) and digital elevation models (SRTM).

Mudbrick is the predominant traditional building material in the Near East since early times. Places of long term occupation are characterized by anthropogenic soils mainly resulting from decomposed mudbrick, with specific signatures in the infrared spectrum, and – from the accumulation of debris over millennia – by artificial mounds, "tells", often large enough to be spotted in global elevation models.

This information can be used to identify possible locations of ancient human occupation at large scale. Examples on Tell-like mounds in Northern Mesopotamia, the general mapping of anthropogenic soils, and the measurements of patterns of long-term occupation are given below.

Average pattern of 133 tells in the western Khabur basin in multi-spectral imagery and digital elevation model.

Tell-like mounds in northern Mesopotamia

The wider picture – a view onto northern Mesopotamia. Mounds with a tell-like elevation pattern in the SRTM model are shown as black dots (click to enlarge).

Mounds with a characteristic, tell-like elevation profile were identified in a "virtual survey" in the SRTM model [Menze, Muehl et Sherratt 07]. A semi-automated detection strategy, based on a classifier adapted to the spot-like elevation pattern of known tells in the western Khabur basin [Menze, Kelm et Hamprecht 07], provided the technical means for this region-wide tell spotting in the elevation model [Menze, Ur et Sherratt 05, 06].

The data sets available as overlay for Google Earth:

Mapping anthropogenic soils

Khabur basin – Results of the ASTER classification. Dark spots indicate a high probability for anthropogenic soils (click to enlarge).

Probability maps show soil marks resulting from ancient and modern occupation. The maps result from a classification of approx. 160 ASTER images from the time between 2003 and 2007, with a spatial resolution of 15m*15m [Menze et Ur 07]. A total of 14000 - 15000 sites could be identified in a visual inspection of the results and are now part of the HAPUR data base. Optimal thresholds on the probability map were calculated locally for each sites, geo-referenced outlines were determined in a postprocessing.

Khabur basin – Coverage by ASTER imagery and possible settlement sites identified from the probability maps (click to enlarge).

Up to 30 images were available for the east-central basin, while less than 3 ASTER images could be used at the margin. The optimal classification result were obtained in the center, after pooling probability maps of 15-30 observations.

Screen-shot of data set loaded to Google Earth. – View onto a 15km*10km large area in the north-central part of the basin, close to the Syrian-Turkish border (indicated yellow). Extensions of areas with a distinct spectral signature are outlined red.

Area outlined red in the Google Earth overlay correspond to the outlines shown in the image patch of the probability map. Areas shown dark in the probability map have a spectral signature which is highly similar to the spectral signature of verified settlement sites in the same region (i.e., with a distance up to 100km radius around the site).

Settlement locations around 'Tell Brak' (modern Nagar), northeastern Syria, as detected from ASTER images.

The data sets available as overlay for Google Earth from the Harvard open data repository:

Long-term occupation

Khabur basin – low and high mounded sites. Places with mounds visible in high resolution SPOT imagery are indicated red. Other, low-mounded sites visible in ASTER are indicated black (click to enlarge).

Height and volume of each of the 14000+ sites in the HAPUR data base was determined using the 90m SRTM digital elevation model (DEM). A topographical background (i.e., the natural elevation profile without the mound) was estimated from the elevation model. The residual between DEM and background – within the spatial limits defined by the ASTER classification – served as an estimate for the height of the site and the volume of the mound (similar to [Menze, Ur et Sherratt 06]). Fit and model error were determined from the margin pixels; the typical standard deviation was 1m – the digitation error of the SRTM model. Pixels which were more elevated than two times the standard deviation of pixels at the margin of the site were defined as "high mounded" areas. High mounded sites may be uses as proxy to human long term occupation.

Different quantities encoded by symbols, as overlay in Google Earth – Sites in the central part of the Basin, close to the Syrian-Turkish border (indicated yellow). Here the white circles encode the volume of a site.

White circles in the center of the sites encode area, volume, or elevation. They are linked to image patches of the SRTM (after correction for the background topography), which also show the outlines of the spectrally distinct area (green in the image patch, red in the overlay) and the high mounded pixels in the DEM (outlined blue in the image patch). The image patch in the popup also shows extensions of the site (bounding box BB), model and fit error for the background removal (standard deviations of the pixels outside the green contour line Std), the maximal elevation of the site in the SRTM model (height), area – in hectares – and volume – in cubic meters – of the pixels outlined green and blue, i.e. the low-mounded and high-mounded area of the site.

The data sets available as overlay for Google Earth from the Harvard open data repository:

Related publications

Menze BH and Ur JA. Mapping patterns of long-term settlement in Northern Mesopotamia at a large scale. PNAS 2012. 109:E778 (publisher, open access)

Menze BH, Ur JA, Sherratt AG. Detection of ancient settlement mounds – Archaeological survey based on the SRTM terrain model. Photogrammetric Engineering and Remote Sensing 2006. 72:321-327 (publisher, open access)

Menze BH, Ur JA. Classification of multi-spectral ASTER imagery in the archaeological survey for settlement sites of the Near East. In: ME Schaepman, S Liang, N Groot, M Kneubuhler (eds.), Proc 10th International Symposium on Physical Measurements and Signature in Remote Sensing (ISPMSRS 07), International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol. 36, Davos, Switzerland. 2007 (publisher, open access)

Menze BH, Muehl S, Sherratt AG. Virtual survey on north Mesopotamian tell sites by means of satellite remote sensing. In: B Ooghe, G Verhoeven (eds.), Broadening horizons: Multidisciplinary Approaches to Landscape Study, 5–29. Cambridge Scholars Publishing, Newcastle upon Tyne, UK, 2007 (publisher) (preprint)

Menze BH, Kelm BM, Hamprecht FA. From Eigenspots to Fisherspots – latent spaces in the nonlinear detection of spot patterns in a highly variable background. In: HJ Lenz, R Decker (eds.), Advances in Data Analysis., vol. 33 of Studies in Classification, Data Analysis, and Knowledge Organization. Springer, Heidelberg and Berlin, 2007 255–262 (publisher) (preprint)

Menze BH, Ur JA, Sherratt AG. Tell Spotting – surveying Near Eastern settlement mounds from space. In: S Dequal (ed.), Proc 20th CIPA International Symposium (CIPA 05), Torino, Italy. 2005 (publisher, open access)

Related links

For further information and publications see the hompages of Bjoern Menze and Jason Ur.