Mersa Gawasis (Red Sea - Egypt): UNO/IsIAO and BU 2003-2004 Field Season under direction of Rodolfo Fattovich and Kathryn Bard

Geoarchaeology/Micromorphology

In order to better understand the site formation processes and the nature and extent of the anthropogenic contributions at the Middle Kingdom site of Mersa Gawasis, field observations were made by Trina Arpin. Several test pits (usually 50 x 50 cm) were excavated and a number of samples for future micromorphological analysis were collected. The field observations below are described by the excavation unit. Since the site was occupied intermittently for short periods over the course of some 300 years, most of the observations made here can only apply to the general span of occupation rather than to a particular episode of use.
The site is located on the slope of a Pleistocene marine overlooking a large wadi and near the modern coastline. Numerous features are found on the terrace itself. Playa lakes have formed in the wadi in historical times. The site contains a complicated mixture of eolian, colluvial and alluvial deposits and there is a strong possibility that lacustrine (playa) deposits will also be found. The location of the ancient shoreline in relation to the site is also unclear.

WG 18: The deposits visible in both WG 18 and the geological trench excavated to the west of it (WG 15) contain a sequence of eolian and alluvial deposits (Fig. 7). The basal layer in WG 18 is an eolian deposit, over 1.5 m thick, containing occasional well preserved hearths. Excavations in the final days of the season appeared to expose an earlier alluvial unit below this eolian deposit, but this identification is not certain.
Above the eolian deposit was a channel filled with sub-rounded granule to pebble sized stones. This channel deposit is visible in the western wall of WG 18 and the eastern wall of WG 15. Both large (over 1 cm) charcoal fragments and potsherds are found throughout these layers, indicating that the wadi was actively depositing sediments during the use of the site. Wadi deposition is typically associated with wetter periods. Visible in WG 15, but less apparent in WG 18, is a 10-15 cm zone at the base of the channel deposit composed of well sorted sands with bedded organic material distributed throughout. This layer may represent organic material deposited in standing water and may reflect ponding at the site.
Also visible in WG 18 is a salt pan which occurs at the transition from the eolian to the channel deposit. This same pan also appears in WG 15 where it is extremely well developed. A second stratigraphically lower salt pan also appears in WG 15. These salt pans are a post-depositional feature that form at or very near the surface and are associated with increased aridity.

WG 16 and WG 17: These trenches were excavated into the slope northwest of WG 18. The surface of the slope is covered by loose sand and abundant large, usually rounded cobbles of basalt. Other volcanic rocks and chert are visible on the surface. Both units reached bedrock, a loosely consolidated conglomerate. The rounded cobbles from the conglomerate are the source of the cobbles visible on the surface. Some of the cobbles may have been used as pounding or grinding tools and the presence of these readily available tools may have been another advantage of the site.
The bedrock conglomerate contains rounded cobbles throughout, but the matrix into which these cobbles are set grades upward from a coarse matrix dominated by granule sized material, through coarse sand, to medium sand. This transition is visible in WG 17. The uppermost unit of a medium sand matrix is also marked by inclusions of gray to yellow sandstone. The surface of this unit has weathered, creating a moderately compact sand layer visible in both WG 16 and 17. Where weathered, the sandstone inclusions have created a mottled appearance, which is distinctive of this surface. In WG 16 a salt crust has formed on the surface of the weathered bedrock and numerous post-holes cut through this crust.
Occupation of the site occurred on the weathered surface of this bedrock. The existence of a stable surface (unlike the loose and generally steeper colluvial and eolian slopes found elsewhere below the terrace) may have been another desirable quality of this location.
The deposits in WG 16 that overlie the salt encrusted bedrock are a series of eolian and colluvial deposits. The colluvial deposits are a mixture of weathering of the conglomerate exposed at the base of the coral and erosion of surface material. A colluvial unit overlies the bedrock, but this unit contains common inclusions of preserved vegetal material (I observed no wood fragments, but the material may include many leaves). This unit may also represent a period of frequent use at the site and a relatively rapid build up of this deposit. At the top of this unit is a thin layer with abundant wood fragments, which is one of the most anthropogenic layers at the site. Most likely it is also one of the most in situ layers. In the colluvial deposit below and the eolian layer above are 1–5 cm aggregates of a clay-silt material, which may be the remains of construction material (mudbricks, mortar, etc.).
Overlying this anthropogenic layer is an eolian layer which contains only occasional pottery sherds. These may be eroded from the surface; however, the presence of two in situ hearths, an intact mudbrick, and a 20+ cm deposit of leafy vegetation indicate that occupation did continue throughout the deposition of this layer. This was likely much more infrequent than the earlier occupation. This eolian deposit may correspond to the eolian deposit in WG 18 since both are relatively thick and contain well-preserved hearths.
The eolian deposit is capped by the colluvial deposit which forms the modern surface.

Lower slopes: Deposits visible on the surface of the slope, below WG 17, show very noticeable color contrasts. Immediately below WG 17, there is a zone of reddened sand. This sand is similar in appearance to that which was found immediately adjacent to the "kiln" identified in WG 17. This red sand could have been produced by smelting processes, which caused oxidation of the quartz sand. A test pit showed that this red sand deposit was at least 30 cm deep. In a 1  1 m unit excavated into this deposit, the sand had been strongly consolidated by halite. This unit was too far upslope for the halite to have come from a playa, and may be associated with a subterranean spring. This deposit contains no charcoal and very few artifacts.
Below the reddened sediments (reaching to the foot of the slope) was a zone of very dark, charcoal rich deposits. A test pit excavated into these deposits indicated that these deposits extended to a depth of 45 cm before a layer of reddened sand was reached. These charcoal rich deposits contained more bedding than the reddened sand (which tended to be more massive) and contained varying percentages of reddened sand. Some layers had a reddish cast to them. These charcoal rich layers contained very few artifacts.
The presence of such a large volume of sediments that are the product of heating is surprising given that only one possible smelting furnace has been found thus far. The processes that produced the distinct zonation (reddened sand higher on the slope, charcoal rich sediment on the the slope) are unclear.

WG 19: This unit contained a complex sequence of anthropogenic layers, many of them containing abundant charcoal and ash. This unit also contained abundant precipitated gypsum and halite. These minerals may be associated with the presence of a playa in the wadi. Observations in this unit were limited due to time constraints.

Terrace: The terrace associated with the site is comprised of limestone (derived from ancient corral) overlain by a relatively thin layer of coarse alluvium. A desert pavement (including many pottery sherds) has formed on the surface, but has been removed by construction equipment near the modern road. In areas unaffected by construction, there are numerous (at least 15) circular features visible on the surface. These circles are 1.0 1.5 m in diameter and the edges are marked by very low mounds of stone. They contain a pavement that is much less dense and finer grained than that found exterior to the features. Test pits and excavation units within these features have not been able to trace them to any depth, nor have substantial numbers of artifacts been found. As there is no natural process to account for these features, they appear to be anthropogenic. Given the lack of stratigraphy and artifactual content, it is difficult to interpret them. They may have been tent circles, but the lack of a clear pattern of post holes is perplexing.
Excavations into the terrace surface revealed a layer of gypsum 10 cm below the present surface. Gypsum precipitates below a stable surface, but normally at depths greater than 10 cm (often 30–40 cm below the surface). Its presence at 10 cm below the surface on the terrace indicates a period of landscape stability at some point in the past, followed by a period of erosion (most likely during a dry phase). This sequence of gypsum precipitation and erosion of the terrace surface predate the construction of the circular features on the terrace surface. If these features date to the Middle Kingdom, as seems likely, then the formation of the gypsum and the erosion of the terrace surface would pre-date this period.

Conclusion: Field observations of the geology of the site show that a variety of processes were at work. These include depositional and erosional processes as well as post-depositional changes to the deposits. The types of deposits include eolian, alluvial, colluvial and possibly lacustrine. The observations have identified the basic processes at work in the units excavated thus far. Outstanding questions include the existence and timing of a playa lake, the location of the sea level, and the processes responsible for the formation of halite and gypsum in some of the deposits.