The excavation took place c. 150 m north of Kh. Qumran (325 m below sea level), on a broad marl plain cut by deep narrow wadis as much as 20 m below the plain. The rock in the area consists of soft horizontal strata (laminae) of marl and small pebbles of the Lissan Formation (lake and beach sediment dating 70,000–14,000 years BP). The weathering processes that formed the landscape, particularly erosion and undercutting, were very rapid in this area. Depressions, craters and anomalies were identified on the surface in the survey that preceded the excavation. Tunnels and underground caves in and around this surface had been formed by an erosion process known as piping, in which cavities are created by surface runoff that penetrates the ground through a local crack. Erosion, combined with slow dissolution, disintegrated the soft rock and created tunnels, most of which are shallow and broad in section (Frumkin 2015:17, 262). Excavations conducted in the nearby caves yielded numerous sherds in unstratified contexts, apparently the result of looting (Broshi and Eshel 1999). The results of Broshi and Eshel’s excavation and the findings from the survey that preceded the current excavation indicate that a good many of the craters were formed by the collapse of caves and due to artificial activity surfaces.
Two excavation areas were conducted (A3-006, A3-013; Fig. 2), in search of any evidence of artificial rock cuttings and of dense piping craters lacking a drainage system, which do not fit with any known geomorphological phenomena.
 
Cave A3-006 (map ref. 243638/627807; Broshi and Eshel 1999: Cave D; Fig. 3) is the only one of the marl caves in which clear signs of rock cuttings were identified. The cave had partially collapsed but it seems that an archaeological or looting excavation had taken place in it. A natural depression (c. 7 × 8 m) was discovered outside the cave, with a surface runoff channel running through its center. No signs of piping were noted. The western side of the depression is bounded by a cliff (length c. 5 m; height c. 3 m). Lacking drainage, the depression seems to have been created by the collapse of an artificial cave, and, therefore, four excavation squares were opened east of the cave opening. The excavation results showed that any sign of human activity, if there had ever been such, had eroded and washed away long ago.
 
Cave A3-013 (map ref. 243507/628864; Figs. 4, 5) is located on the northern side of the ridge, 10 m away from another cave that was previously excavated (Broshi and Eshel 1999: Cave C). Four squares (1–4) were excavated in a crater-like depression, enclosed on three sides and open to the north (c. 5 × 6 m; height c. 5 m); a piping system flowing eastward (length c. 5 m) formed in southeastern corner. An ancient artificial activity area that had filled with alluvium and changed its shape following rapid weathering of the marl could be discerned. At the bottom of Square 1, excavated in the center of the space, was a compact surface (L106) upon which were sherds dating from the first century BCE to the first century CE. To the north, near the slope descending to the wadi bed, Square 4 was excavated, revealing leveled bedrock surface on which was a bowl dated to the same time period (L113; Fig. 6). A layer of fill over Surface 113 consisted of soft soil and small to medium-sized pebbles (L108; Fig. 4: Section C–C). Crushed chunks of rock, partially stratified (L104), were incorporated into this layer. These chunks of rock may indicate that this activity area was once a closed space. In a section opened in the western part of Sq 4 (L109; depth c. 1 m), the leveled bedrock on which the occupation layer was discovered was seen sloping upward and meeting the natural bedrock; this appears to have been the western boundary of the complex. Square 2 was excavated on the western side of the central part of the depression with the goal of finding its western wall. Bedrock was identified mainly on the western side of the square, where the southwestern boundary of the complex was detected. Another square (3) was excavated near the back of the crater to identify the southern boundary of the complex and its relation to the piping system. The southern wall of the depression was uncovered after the removal of alluvium, revealing its original size at c. 4 × 4 m. At the base of the southern wall signs of a collapsed roof were detected as well as numerous sherds along with a fragment of a stone vessel—an ossuary?—which was found in secondary use (Fig. 7).
 
At this stage, sufficient data is lacking to reconstruct the plan of Complex A3–006. Geomorphological evidence suggests that there was either another activity surface or an additional cave, which did not survive, in addition to the hewn cave. In contrast, Complex A3-013 has a clearer plan. It is c. 16 sq m in size and it apparently included an open activity surface outside the cave. The meager finds impede our understanding of the nature and use of the complex. The survey and excavation yielded evidence of hewn or partly hewn activity areas, which were abandoned in an organized fashion and fell into ruins over time. The ceiling collapsed and buried under it the meager finds from the first century BCE–first century CE—the main period of activity at adjacent Kh. Qumran. The spaces later filled with alluvium and eroded naturally. It seems that the site featured a variety of activity complexes whose typological classification requires further research.