Rattenborg and Blanke (2017:11) report that structural collapse in Area GO, adjacent to the Mosque, destroyed the subsurface compartment around the courtyard cistern shaft and forced the abandonment of two rooms in the northern wing of Building B. In the destruction layer of the qiblat Hall—described but not dated by Barnes et al. (2006:295)—excavators recovered more than five tonnes of roof-tile fragments from a thick, uniform collapse layer. The hall’s layout and surviving architectural elements point to a triple-gabled roof that once spanned the space. Looters removed whole tiles and much of the paving after the collapse. Outside the mosque, massive stone tumble contrasts sharply with the interior deposits, showing that the weight of the roof drove the walls outward when they failed. Together, the tile concentrations, the uniform collapse layer, and the outward fall of masonry point to a single destructive event.

Rattenborg and Blanke (2017:11) argue that this collapse occurred no earlier than the second half of the 9th century CE and may match damage recorded in residential units on the Southwest Hill. Pottery from Building A’s paved floors (Phase 3/II) supports a late 8th–9th century date. The assemblage includes painted red terracotta, cream, and pale orange wares. A nearly complete Cream Ware jug lay beneath wall collapse over a paved floor. Comparable wares appear in Abbasid houses farther north (Gawlikowski 1986, 1995) and in Area Z at Umm Qais (el-Khouri and Omoush 2015:17–20). Even so, Rattenborg and Blanke (2017:15) warn that archaeological evidence for the 9th–11th centuries remains sparse and chronologically imprecise.

Rattenborg and Blanke (2017:13) also describe how a late 9th- or early 10th-century collapse triggered a partial rebuilding of the mosque. Only the western third of the prayer hall returned to use, while the courtyard stayed in disrepair. Collapse debris blocked the laneway along the mosque’s northwest perimeter and cut off access from the main street. However, the passage leading from the west entrance to the southwest corner shows no evidence of collapse, suggesting that residents in the adjacent dwellings continued to use this route after the disaster.

Deformation Map - Click on image top open in a new tab - Modified by JW from Fig. 12 of Rattenborg and Blanke (2017)