• Fig. 2 Plan of Islamic
  Figure 2
  
  Plan of the ancient city of Ayla with the locations of the measured stations (numbers inside boxes) described in the text. Hatched diagonal structure is a Wadi bed. BH is a borehole near the Egyptian gate (modified after Whitcomb 1993)
  
  al-Tarazi and Khorjenkov (2007)
  Ayla from al-Tarazi and Khorjenkov (2007)

• Fig. 2 Plan of Islamic
  Figure 2
  
  Plan of the ancient city of Ayla with the locations of the measured stations (numbers inside boxes) described in the text. Hatched diagonal structure is a Wadi bed. BH is a borehole near the Egyptian gate (modified after Whitcomb 1993)
  
  al-Tarazi and Khorjenkov (2007)
  Ayla from al-Tarazi and Khorjenkov (2007)

• from Allison (2013:50-60)

Archaeological sites situated along or near seismically active fault zones provide a unique opportunity to study the paleoseismicity of a region because these sites often preserve evidence of the seismic past in the form of damaged or ruptured structures or buried artifacts. Repairs and reconstruction of damaged structures made in antiquity can help to constrain the recurrence intervals of specific faults, and the type of failure or collapse can provide information concerning the intensity of past earthquakes. Archaeological sites also usually contain a variety of datable materials such as pottery and charcoal fragments, which can provide an important means for age control and paleoseismic event correlation for multiple events.

Archaeoseismology is a subfield of seismology that focuses on the study of past earthquakes through the traces left behind in the cultural record (Sintubin and Stewart, 2008), and was essentially born in the Mediterranean region and Middle East as a result of the active tectonics and plentiful cultural ruins and deposits present in these areas (Jones and Stiros, 2000). Archaeoseismology is an interdisciplinary science because it incorporates the fields of seismology and archaeology, along with geology, geotechnics, geomorphology, and even civil engineering at times, in order to gather data about both the pre-instrumentation 50 and prehistoric seismic past (Jones and Stiros, 2000). The term prehistoric, as it is used here, is meant to delineate between historically recorded earthquakes and those without a historical record of occurrence. The use of archaeological data to investigate poorly known historical earthquakes or even unknown events actually began in the late 19th to early 20th centuries, but has only really developed into the subfield we know today in the last few decades (Hinzen, 2010). Since first used, however, the field of archaeoseismology has continued to grow as has the acceptance of this type of interdisciplinary science (Galadini et al., 2006). Sintubin et al. (2010) cite a major problem plaguing those who study ancient earthquakes (defined as pre-instrumental earthquakes that can only be identified through indirect evidence in the archaeological or geological record), which is that the instrumental record is too short to be very useful, and the historical record is often incomplete. Archaeoseismology, which can potentially delve much deeper into the historic and prehistoric past, is a subdiscipline that developed out of the need to overcome both of these issues (Noller, 2001).

With its richness of history and abundance of archaeological ruins, the Middle East is a prime location in which to utilize archaeoseismology to study the seismicity of the region. Compared to studies conducted in North America, for example, archaeoseismic studies undertaken in the Middle East and Mediterranean regions have the advantage of being able to incorporate a much longer historical record into those archaeoseismic efforts. With a record of historical earthquakes extending back over 2000 years, the seismic record of Jordan is impressively one of the longest records in the world (Thomas et al., 2007). In an attempt to better understand the behavior of the Dead Sea transform fault, the chronology of 51 historical earthquakes along with any historical accounts have been organized into earthquake catalogs for easy reference. Earthquake catalogs will be discussed in more detail in a later section. The Dead Sea seismic zone also benefits from a longer and more comprehensive record of archaeoseismic practice, which acts to add validity to the more recent and perhaps more refined studies being conducted in recent years.

Much of what we know about the seismic history of the Middle East and Mediterranean region before the Christian era has come from a collection of archaeoseismic investigations and excavations (McCalpin and Nelson, 2009). The archaeoseismic studies conducted there are more common than in North America primarily because the availability of cultural remains is exponentially more abundant. Many ancient sites are located within 100 km of the Dead Sea transform fault, for example, and therefore have a greater chance of being damaged by earthquakes in the region (e.g. Marco et al., 1997; Ellenblum et al., 1998; Meghraoui et al., 2003; Haynes et al., 2006). Ancient structures in the Middle East are also often constructed from more resistant materials such as worked and unworked stone as opposed to wood, which arid regions like Jordan generally lack in any significant quantity. The arid climate and the generally slow sedimentation rates along the Dead Sea rift also play a role in the preservation of cultural remains, and have allowed for materials like mudbrick to be used for construction purposes in antiquity, much of which is still preserved until today (Thomas et al., 2007).

The best and most clearly defined examples of faulting happen when an earthquake fault ruptures a standing wall, aqueduct, or other cultural lines that act as piercing points where the amount and direction of offset is measureable. The Dead Sea transform not only exemplifies near-textbook left-lateral, strike-slip motion when the fault rupture makes it to the surface, but faulting of this sort allows archaeoseismologists the opportunity to measure the actual offset or distance the fault slipped during a specific event, or the cumulative offset from multiple events (e.g. Noller, 2001; McCalpin, 2009). In this way, strike-slip faults that occur within cultural deposits with substantial architecture are not only easier to study, but also provide a greater abundance of faulting information. Archaeoseismology has the potential to bridge the gap between instrumental and historical seismology on one hand, and paleoseismology and earthquake geology on the other (Galadini et al., 2006), but can only be conducted in locations where seismic activity and past human activity have intersected in a meaningful way.

Numerous archaeological sites situated along both sides of the Dead Sea transform have been the focus of archaeoseismological studies in the past. With one of the most tightly constrained historical records, the site of Vadum Jacob, also known as the Ateret Fortress, is an example of using a seismically damaged archaeological site to assess the history of faulting in an active tectonic zone (e.g. Marco et al., 1997; Ellenblum et al., 1998). Ateret, built in A.D. 1179, is a castle overlooking the Jordan River in modern-day Israel that once served as an outpost during the Crusader period. The foundation stones of the castle were actually laid down in A.D. 1178, but it was conquered and destroyed only eleven months later on August 30, 1179 by Saladin and his troops (e.g. Marco et al., 1997; Ellenblum et al., 1998). Numerous independent historical sources pinpoint this specific date of destruction, and thus the associated destruction layer exposed in cross-section in a trench located just outside of the castle wall (e.g. Abu-Shama; Lyons and Jackson, 1982; Raphael, 2010).

The Ateret Fortress is built on the active Jordan Gorge Fault, a straight segment of the Dead Sea transform located between the Sea of Galilee and the Hula Valley in the north. Archaeoseismic research conducted at Vadum Jacob by Marco et al. (1997) and Ellenblum et al. (1998) has revealed numerous offset archaeological structures. Most noticeably displaced are the four-meter-thick stone walls that surround the castle, and aqueducts situated across the DST fault are also offset. Stratigraphically, the destruction layer dated to A.D. 1179 was also offset, so the castle was clearly faulted after it was destroyed. Historical records indicate that only four sizable earthquakes occurred in the region after the late 12th C. construction date of the castle. These are the earthquakes of A.D. 1202, 1546, 1759, and 1837 (e.g. Amiran et al., 1994; Ambraseys, 2009), and thus these are the only known seismic events that could have potentially affected the region of Israel in which the castle is built. Of these earthquakes, Ellenblum et al. (1998) explain that the Ateret Fortress was likely located too far north for the A.D. 1546 earthquake to have faulted the site based on the known damage radius, although this earthquake would have been felt there (Ambraseys, 2009). Instead, excavations have shown that the fortress was most significantly deformed by the historic M>7 earthquake that occurred at dawn on May 20, 1202.

In all, both the north and south four-meter-thick stone fortress walls were offset by a total of 2.1 m due to earthquake faulting. Of this total measured displacement, 1.6 m of offset is attributed to the left-lateral motion of the Dead Sea transform during the 1202 event, and the remaining 0.5 m of measureable offset is attributed to either the 1759 and/or 1837 seismic events (Marco et al., 1997; Ellenblum et al., 1998). Post-Crusader structures built at the Vadum Jacob site were clearly not affected by the 1202 earthquake because, in addition to their construction post-dating the occurrence of this historical earthquake, they are also only offset by 0.5 m, which serves to constrain the amount of displacement resulting from each of these seismic events. In particular, a Muslim mosque at the site, originally built in the 12th C., was destroyed and rebuilt again either once or twice during the Turkish Ottoman period (1516-1918) based on archaeological excavation and study of associated pottery finds (Ellenblum et al., 1998). The latest or youngest rebuilding phase of the mosque during the Turkish Ottoman period shows 0.5 m of offset in its northern wall, but lacks the cumulative 2.1 m offset observed in the Crusader castle wall.

The thick, fortified walls that surround the Ateret Fortress are also of particular sturdy construction for the time period. Built with the intention of preventing attacks directly to the castle itself, the walls are also substantial enough to prevent potential intruders from being able to easily dismantle and break through the stone walls before being detected. Ellenblum et al. (1998) describe the walls as being composed of a parallel pair of meticulously laid ashlar supports that enclose a 3-meter-thick fill of cemented basaltic cobbles built by “highly skilled Crusader masons.” Particularly important to note is that these walls were erected with great precision along smooth, straight lines, and thus serve as quite reliable strain gauges for any deformation occurring after A.D. 1179 (Ellenblum et al., 1998). Essentially, the detailed historical records available for the site allow Vadum Jacob to be dated with unusual precision. The abundance of well-dated structures and only a small number of candidate earthquakes that could have damaged the castle walls also makes Vadum Jacob an excellent site for geoarchaeological and archaeoseismic research along the Dead Sea transform.

On the opposite side of the rift valley, the site of Qasr Tilah located in Jordan approximately 30 km south of the Dead Sea in the Wadi ‘Arabah is another well-documented example of an archaeological site that has been clearly faulted and offset by the Dead Sea transform, and is the subject of archaeoseismic study. The site of Qasr Tilah consists of a Roman-period fort (caravanserai), a Byzantine - Early Umayyad water reservoir (birkeh), an aqueduct irrigation system, and agricultural fields. Detailed archaeological excavation of the site, which was first conducted by Niemi (2000, 2002) as a part of the Wadi ‘Arabah Earthquake Project, has revealed that structures situated across the DST fault were offset laterally by earthquakes in the past, including the north wall of the 30 m x 30 m stone birkeh and a portion of the aqueduct system. The aqueduct system was built so that two aqueducts first funneled water into the birkeh from the adjacent Wadi Tilah where particulates could settle out of suspension in a small settling pool, then a short NW-trending aqueduct connects the settling pool to a W-trending aqueduct that carried water away from the reservoir and, presumably, into a nearby agricultural field during periods of overflow (Haynes et al., 2006).

The northwest corner of the birkeh has a large crack caused by left-lateral offset of the DST by 1.5-2 m which is indicative of earthquake rupture (e.g. Niemi, 2000; Niemi and Atallah, 2000; Klinger et al., 2000a; Galli and Galadini, 2001). Charcoal collected from within the mortar of the original birkeh construction was radiocarbon dated to A.D. 641-687 (Niemi, 2000), a date that agrees with radiocarbon published by Klinger et al. (2000b). Investigation of the structure revealed that a repair made to the northwest corner of the birkeh wall suggests that an earthquake damaged the birkeh not long after it was constructed 56 (Niemi and Atallah, 2000). Due to subsidence of the damaged corner of the birkeh, another course of stones was laid on top of the wall and covered in a layer of plaster in an effort to repair the structure. A radiocarbon date from the repair made to the corner of the birkeh agrees very closely with a radiocarbon date collected from the lower portion of the original birkeh wall, which supports the hypothesis that the birkeh was initially damaged shortly after construction, or perhaps that the birkeh was rebuilt over an earlier structure after being damaged (Haynes et al., 2006). Archaeological excavation by Haynes et al. (2006) also revealed the NW-trending portion of the aqueduct system, which was determined to have been constructed below ground using mortar and stones laid in a trench, to be laterally offset 1.6 +/- 0.4 m.

The site of Qasr Tilah is faulted by the Wadi ‘Arabah Fault, which is a segment of the southern Dead Sea transform fault. Radiocarbon dating of charcoal samples from significant subsurface horizons provided one means of age control, and ceramic analysis of pottery sherds and artifacts collected during excavation of the site also provided age control by relative typological dating (Haynes et al., 2006). Using offset stratigraphic horizons, upward terminations of faults, fissure fills, and outward splaying faults (flower structures) visible in the trench walls, Haynes et al. (2006) identified evidence of four ground-rupturing earthquakes at the site ranging from the sixth to the nineteenth centuries. The archaeoseismic data was combined and examined with historical records to determine possible calendar dates for each event. The data suggests that these four ruptures likely represent the historical earthquakes of A.D. 634 or 659/660, 873, 1068, and 1458 or 1546 (Haynes et al., 2006). In all, the site of Qasr Tilah provides an excellent example of how studying seismic damage and deformation of an archaeological site can provide insight into the seismic past. The ability to date the structural repairs made to the damaged birkeh act to further constrain these earthquakes based on radiocarbon dating, and stratigraphic analysis of a trench excavated across the fault trace acts to corroborate this data as well.

North of both Qasr Tilah and the Ateret Fortress, the site of Al Harif in Syria is another archaeoseismically relevant site that has been directly faulted and offset by the Dead Sea transform. The Al Harif site is located in northwestern Syria and consists of a Roman-period aqueduct that has been repeatedly left-laterally displaced by the approximately 90 km-long Missyaf fault, a northern segment of the DST (Meghraoui et al., 2003; Sbeinati et al., 2010). The site was first studied by Meghraoui et al. (2003) who used microtopographic surveys and trenching techniques to determine that the Roman aqueduct was faulted three times in the last 2000 years and displaced a total of 13.6 +/- 0.2 m (Meghraoui et al., 2003). Meghraoui et al. (2003) documented the occurrence of earthquakes between A.D. 100 and 750, between A.D. 700 and 1030, and between A.D. 990 and 1210 based on the radiocarbon dating of faulted alluvial deposits.

The aqueduct, originally dated to between A.D. 30 and A.D. 70 by bracketing the age of the sedimentary units below its foundation and from early travertine accumulation on its wall, was used to bring fresh water from the wet, western high mountains to wheat and olive oil fields (mills), and local cities and villages located in the eastern plains (Meghraoui et al., 2003). A bridge even allows the aqueduct to cross the Al Harif River without disrupting the flow of water. A later study conducted by Sbeinati et al. (2010) consisting of four archaeological excavations, three paleoseismic trenches, and coring of accumulated tufa deposits (previously identified as travertine by Meghraoui et al., 2003) on the aqueduct at the Al Harif site, further refined the construction date of the Roman aqueduct to between A.D. 63 and A.D. 70.

Like the abundant historical records in Jordan, the historical record of seismicity of Syria and surrounding regions is also very rich because of the numerous archaeological and historical sources that date back at least to 1365 B.C. (e.g. Ben-Menahem, 1979; Ambraseys and Jackson, 1998; Guidoboni, 1994). Combining historical sources with data collected from trenching, excavation, and sediment coring, Sbeinati et al. (2010) identified as many as four earthquakes that have affected the site in the last 3400 years. The fourth seismic event back (Event W) occurred sometime between 2300-510 B.C., and a cluster of three other seismic events occurred in A.D. 160-510 (Event X), A.D. 625-690 (Event Y), and A.D. 1010-1210 (Event Z) (Sbeinati et al., 2010). The most recent seismic event, Event Z, is likely correlated with the large historical earthquake that occurred on June 29, 1170 (Meghraoui et al., 2003; Sbeinati et al., 2010). Events X, Y, and Z each ruptured and progressively offset the Roman aqueduct, while Event W occurred prior to the construction of the aqueduct. Further, two reconstruction phases at the Al Harif site were identified through the analysis of building stone type, related cement dating, and tufa deposits on the aqueduct. These reconstruction episodes most likely occurred in A.D. 340 +/- 20 and A.D. 720 +/- 20 (Sbeinati et al., 2010).

The large earthquakes of Events X and Z have been estimated as Mw = 7.3-7.5 based on detailed historical reports of ground shaking and damage (Meghraoui et al., 2003), and incurred successive faulting events, fault segment length, and related amount of coseismic 59 slip (Sbeinati et al., 2010). The study by Sbeinati et al. (2010) also helped to constrain the slip rate of the Dead Sea transform in northwestern Syria along the Missyaf fault segment to between 4.9-6.3 mm/yr. Further, with the goal of producing a refined assessment of the seismic hazards of the region, the findings of Sbeinati et al. (2010) suggest that the recent seismic quiescence along the Missyaf fault segment of the DST implies that a large earthquake is overdue and may result in a major catastrophe to the population centers of Syria and Lebanon in the near future. This archaeoseismic approach to seismic hazard analysis is particularly applicable in this region and especially important along the Dead Sea transform since it corresponds to a major tectonic plate boundary (Meghraoui et al., 2003; Sbeinati et al., 2010).

Sites like Ateret in Israel, Qasr Tilah in Jordan, and Al Harif in Syria are unique in that information about past earthquakes can be obtained from measurements collected from directly offset structures and from careful recording of the stratigraphy across the fault trace. While many archaeological sites have been damaged from earthquakes in the historical past, few sites are able to demonstrate such strong evidence of faulting and rupture of the site so clearly. The vast majority of archaeological sites damaged by earthquakes, like the site of Early Islamic Ayla (7th-12th C.) to be discussed herein, were damaged as a result of ground-shaking, and may show signs of earthquake damage in the form of subsidence, liquefaction at the site, and/or failure of walls, columns, or other structures that were present at the time of the event.

• from Allison (2013:61-63)

Historian and archaeologist Dr. Donald Whitcomb of the Oriental Institute at the University of Chicago excavated the archaeological site of Early Islamic Ayla from 1986 to 1995, and unearthed a large portion of the city near the modern shoreline in Aqaba. His excavations revealed the original rectangular city wall, a mosque, and numerous commercial and residential structures in and around the ancient city. Archaeological evidence shows that the city of Ayla was a fishing port likely occupied from the mid-seventh century to the twelfth or thirteenth centuries (Khouri and Whitcomb, 1988; Whitcomb, 1986, 1987, 1988a, 1988b, 1989a, 1989b, 1992, 1993a, 1993b, 1994a, 1994b, 1995a, 1995b). Donald Whitcomb’s extensive excavations at Ayla over the course of nearly a decade ultimately illuminated the organization of an early Islamic city.

In 2000, S.T. Parker excavated a trench (R1) across the alignment of the city wall of Early Islamic Ayla as a part of the Roman Aqaba Project. The specific purpose of this excavation was to determine once and for all whether the site’s origin was the legionary fort of the Legio X Fretensis. The trench was placed underneath the bridge of Corniche Road (alIstiqlāl) in the wadi that now runs through Ayla in a section of the site referred to as Area R (Parker, 2002) because it was an area of lower elevation that would allow the deeply buried layers to be more easily exposed. Parker’s R1 trench exposed a portion of the city wall that was found to be in good condition with no apparent earthquake damage. Pottery collected from at and beneath the wall foundation confirmed Whitcomb’s hypothesis of an A.D. seventh century construction date for the Islamic city wall (Parker, 2002).

In an attempt to better understand the history of the Islamic city and the seismicity of the region, the Wadi ‘Arabah Earthquake Project (WAEP), led by Dr. Tina Niemi, has excavated several locations along the outer walls of the site, looking for evidence that the Dead Sea transform (DST) fault offset the city walls of Ayla. Whitcomb (1993a) hypothesized that the DST fault trends directly through the city, along the trend of the wadi that crosses the site. Excavations conducted by the WAEP in 2001 and 2003 focused on articulating several portions of the city wall leading to the east corner tower, as well as the corner tower itself, which is situated in the wadi under the same bridge where Parker (2002) excavated. Based on these excavations, Rucker and Niemi (2005) proved conclusively that there is no evidence of faulting offset through this section of Islamic Ayla. They do not rule out the hypothesis presented by Galli and Galadini (2001), however, that the DST fault may trend inland from the Gulf of Aqaba NNE-SSW from the Sea Gate to the reconstructed wall NW of the Egyptian Gate, although further excavation is required to examine this hypothesis (Rucker and Niemi, 2005).

Kristoffer Damgaard also excavated portions of Islamic Ayla starting in 2008 as a part of the Islamic Aqaba Project, directed by Dr. Johnny De Meulemeester from the University of Ghent, in an effort to investigate further the town’s occupational phases and morphology. A trench was laid out out specifically to investigate the possible marketplace expansion of Ayla outside of the city walls near the Egyptian Gate, as well as a trench positioned within a previously unexcavated portion of the southwestern quadrant of the city (Damgaard, 2008). This preliminary excavation revealed that at least the southwestern quadrant of Ayla remained occupied for the entire lifespan of the city (Damgaard, 2008). Further, the expansions made along either the inner or outer walls of the city were found to have been constructed in-line with earlier structures despite those earlier structures being buried by a fill layer and presumably hidden from view during each new phase of construction. As such, Damgaard (2008) suggests that the inhabitants of Ayla understood the previous construction plan of the city in many respects and were able to carry on this plan through the centuries. In 2010, Damgaard returned to excavate at Ayla as director of the Aylah Archaeological Project (AAP) after Dr. De Meulemeester’s untimely passing in 2009. This season focused again on determining the extent and location of mercantile facilities related to the Red Sea trade, as well as further defining the morphology of occupation and urban planning from the seventh through the eleventh centuries (Damgaard and Jennings, 2010; Damgaard et al., 2010; Damgaard, 2011).

• from Allison (2013:63-81)

Much of what we know about the history of the region comes from the archaeological excavations in and around the city of Aqaba. Buried by sand until the mid-1980s when Donald Whitcomb first rediscovered the Islamic settlement through well-placed archaeological probes (Whitcomb, 1994a), Ayla has become one of the best examples of an early Islamic city anywhere, and is unique in the country of Jordan (Whitcomb, 1986). In terms of conducting archaeological research, the Aqaba region is easily accessed because of its position at the head of the Gulf, although until rather recently it was archaeologically unexplored. While there are accounts of western travelers passing through the Aqaba coastal zone in the early 19th - 20th centuries and noting the strewn ancient ruins, it was not until the 1930s that archaeological research on the region was conducted and published.

Ayla is a walled city with a rectangular plan approximately 170 m x 145 m (Figure 3.1), now situated between the Mövenpick Hotel and Aqaba Yacht Club along the Gulf of Aqaba. As is the case with the majority of historical events that took place at Islamic Ayla, there is no documentary evidence for dating the foundation of the city. Whitcomb (1995a), however, places the establishment of Ayla within the caliphate of ‘Uthman ibn ‘Affan in A.D. 650, the third of the Rashidun Caliphs after the death of Muhammad in A.D. 632 (Table 3.1), based on data collected from archaeological excavations.

Today, Aqaba functions as both a growing tourist resort and vital commercial port, providing invaluable access to the Red Sea, as it has since antiquity. The port of Ayla was a very important route of trade and cultural connections because it lies at the critical junction of both land and sea routes uniting Asia, Africa, and Europe (Khouri and Whitcomb, 1988). During the medieval period this region also served as an important commercial center based on the pottery assemblages, which are the primary dating tool used at this site, as well as other archaeological materials and artifacts recovered from medieval Aqaba (Khouri and Whitcomb, 1988).

With a settlement pattern of southern migration down the coast (Parker, 1997), in the municipality of Aqaba proper three individual settlements dating from the 1st century and later have been identified from ancient ruins and detailed archaeological excavation (Figure 3.2). Located approximately 250 m to the northwest of Islamic Ayla is the Roman/Byzantine town of Aila (2nd - 7th C.), which is largely built upon 1st century Nabataean ruins, and located some 850 m to the south of Islamic Ayla is the Aqaba Castle dated to the Mamluk and Ottoman periods (e.g. Khouri and Whitcomb, 1988; Parker, 1997).

The ancient town of what is now referred to as Roman/Byzantine Aila is thought to have been founded by the Nabataeans, and was a center of trade and industry starting in the late 1st century B.C. (Parker, 2007). The most important early evidence about Aila is in Strabo’s 1st century Geography, according to Parker (1997). Strabo confirms that the area was inhabited by Nabataean Arabs, and mentions that frankincense and myrrh as well as other aromatics from south Arabia were sold to merchants (Jones, 1930), which is quite significant because items like these, which would have been expensive commodities, often do not survive in the archaeological record or in the historical record as being bought or sold (Parker, 1996). This evidence supports the idea that Aila was an important town along the spice trade route during antiquity (Parker, 1996). Although there is little direct evidence of such, Aila would have made a likely and logical location for a Nabataean port because of its close proximity to the Nabataean capital of Petra, but the port has yet to be discovered in Aqaba (Khouri and Whitcomb, 1988). After the annexation of the Nabataean Kingdom in A.D. 106 by Roman Emperor Trajan, Aila fell under Roman control and became the southernmost extent of the Roman province along the via nova Traiana (“Trajan’s new road”), which was extended from Syria in the north all the way to Aila between A.D. 111-116 (e.g. Khouri and Whitcomb, 1988; Parker, 1996). Around the start of the 4th century, the Legio X Fretensis (10th Legion Fretensis) from Jerusalem was stationed in Aqaba to secure the region militarily, which they then called Ailana (Khouri and Whitcomb, 1988). The transition from this to the Byzantine period seems smooth in the archaeological record, but relatively little about this shift is known. The legionary presence in the city remained until at least the 5th century, and Aila is believed to have prospered during the Byzantine period, likely because of this military presence (Parker, 2007).

The reason for the collapse of Roman/Byzantine Aila and the move to Islamic Ayla is a matter of debate, but based on recent excavations the city is believed to have been damaged and/or ruptured by no fewer than seven earthquakes since the second century A.D. (Thomas et al., 2007). Multiple repairs have been documented at what is referred to as the Monumental Building at Aila, and after the A.D. 363 earthquake during the Byzantine period, there was a notable reduction in occupation and a change in function of the structure to a more domestic capacity based on artifacts recovered (Thomas et al., 2007). It is clear that the seismic activity of the region played an ongoing role in functional and social changes at Aila, likely resulting in its decline and eventual collapse somewhere between the mid-7th to mid-8th centuries (Thomas et al., 2007). It is also likely that when the earthquake damage became too severe to make repairs any longer worthwhile, most people started moving away from the dilapidated town, and Islamic Ayla was newly constructed. A small population of people, however, must have remained at Aila to some extent or re-inhabited the site because there is evidence of construction and reuse of earlier Umayyad structures as determined from archaeological deposits dating to the mid- to late 8th century at the site (Thomas et al., 2007).

Muslim rule, the start of the Islamic period, began in A.D. 630 when Yuhanna ibn Ru’ba, a representative or possible bishop of Roman/Byzantine Aila, personally negotiated the city’s surrender to Muhammad at Tabuk in present-day northwestern Saudi Arabia (e.g. Guillaume, 1955; Zayadine, 1994; Parker, 1996). As a part of this arrangement, Muhammad guaranteed the protection of Roman/Byzantine Aila’s ships and caravans on both land and sea, implying that trade and commerce were of great importance to the city of Aila right up until the Muslim conquest (e.g. Guillaume, 1955; Zayadine, 1994). Very soon after the conquest, which was not accompanied by widespread destruction (Avner and Magness, 1998), the new fortified city of Ayla was founded by the Muslims, beginning a new of chapter of development in Aqaba (Whitcomb, 1994a).

It is important to clarify that the cities of Islamic Ayla and Roman/Byzantine Aila, are both pronounced the same way, and because a variety of spellings have been used for both sites, this can be confusing in the historical record and even in modern publications at times. Roman/Byzantine Aila, excavated from 1994-2002 by S. Thomas Parker, has also been referred to as Ailana, Aelana, Elana, Haila, Ailath, and even Ayla (e.g. Parker, 1996; Whitcomb, 1997). Islamic Ayla, because it was founded centuries after classical Aila in the 7th century, was also called “Wayla” by Shams ad-din Muqaddasi, the first Arab geographer to study the area as he traveled through the region during the 10th century (de Goeje, 1906; Whitcomb, 1997). “Wayla” is thought to mean “diminutive little Aila,” and was presumably introduced by Muqaddasi to reduce confusion about the two similar city names, although the name seems to have remained idiosyncratic to him and his writings (Whitcomb, 1997). Referring to this perceived naming problem, Muqaddasi wrote, “The city is usually called Ayla, but the true Ayla is in ruins nearby…ask them concerning the town by the sea” (e.g. de Goeje, 1906; Khouri and Whitcomb, 1988). Interestingly, until 1986 this quote could not be understood by modern historians and archaeologists because the ruins of Islamic Ayla and most of Roman/Byzantine Aila were completely buried by sand by the time Whitcomb first visited Aqaba in the early 1980s. According to Khouri and Whitcomb (1988), however, Muqaddasi’s 10th century observations, coupled with a few scattered sherds found on the surface above where Islamic Ayla was buried, are the reasons that Whitcomb dug a few small excavation probes, eventually finding the Islamic city once again.

In A.D. 661, Damascus became the new capital of the Arab-Muslim Umayyad caliphate. Jordan prospered during the Umayyad period (A.D. 661-750), likely because of its proximity to Damascus. With walls preserved up to 4.5 meters tall, Early Islamic Ayla, an example of this Umayyad prosperity, contains four gates or entrances leading into the city: the Egyptian Gate, the Syrian gate, the Hejaz (Hijaz) Gate, and the Sea Gate (Whitcomb, 1995a). Once established, the city flourished from A.D. 650 until A.D. 746 or 749 when a large earthquake struck Ayla, requiring that substantial repairs be made to its walls and towers (Whitcomb, 1995b). The extensive damage documented by excavation records and the likely weakened infrastructure of Ayla may have contributed to the defeat of the Umayyads by the Abbasids in A.D. 750 (Whitcomb, 1995b). This earthquake event, and all others that may have affected the city of Ayla and the greater Aqaba coastal zone, will be discussed in the next chapter in a section that will summarize the known seismic activity of the region.

Following this mid-8th C. earthquake, the city experienced a serious investment of capital in the form of rebuilding, expansion, and reorganization of architectural space (Cobb, 1995). This post-Umayyad re-build can be seen, most readily, in the restructured Abbasid mosque, and in the extramural expansion associated with this period, according to Cobb (1995). Also, the city of Ayla is marked by axial streets dividing the town into four quadrants, and at this central meeting place stood a tetrapylon, referred to as the “Central Pavilion” (Whitcomb, 1994a). At this central crossing, another change in the administration of the city can be seen through the modification of the function of the Central Pavilion from that of an administrative center to what appears to be a domestic residence in the Abbasid period. Khouri and Whitcomb (1988) suggest this may have been turned into the home of a governor or other city official. The city was possibly damaged yet again by another earthquake only a few years later in A.D. 757, but because these earthquakes happened in such a close period of time, they are not individually distinguishable in the archaeological record.

Islamic Ayla, originally built as a “misr,” or fortified camp town, seems to have been an economic center located along the trade route, as well as an oasis of sorts to weary pilgrims making their way to the city of Mecca in the south (Whitcomb, 1995a). After recovering from at least one, if not two, mid-8th century earthquakes and the change to Abbasid rule, it seems that the town of Ayla enjoyed relative prosperity based on the varied pottery collected through Whitcomb’s numerous excavations, which continues to show extensive trade throughout the region, albeit under increasing political destabilization (Cobb, 1995).

Egyptian Fatimid rule in the mid-10th century began a period of adversity as hinted at by geographer Muqaddasi when he writes that, “… in Wayla there is disagreement among the people of Syria, the Hejaz, and Egypt…but I join it to Syria because its customs and measures are Syrian. It is the port of Palestine, from which come its imported goods” (e.g. de Goeje, 1906; Khouri and Whitcomb, 1988). Starting in A.D. 961, a garrison was stationed at Ayla in order to contend with the local tribesmen and other unruly outsiders who were becoming increasingly prevalent, and who would eventually succeeded in sacking the city in A.D. 1024 (Khouri and Whitcomb, 1988).

The city of Islamic Ayla took another major hit when, according to historical sources, and as cited by the earthquake catalogs of Guidobini and Comastri (2005) and Ambraseys (2009), the city of Ayla was severely damaged in the large March 18, 1068 earthquake with an epicenter believed to be in or very near the Gulf of Aqaba. This catastrophic earthquake resulted in the buildings at Ayla being destroyed and all its inhabitants killed, except for twelve residents who were out fishing on the Gulf at the time of the earthquake (e.g. Makdisi, 1956; Guidobini and Comastri, 2005; Ambraseys, 2009). The historical record also shows another earthquake occurred just two months later on May 29, 1068 that is said to have destroyed the city of Ramla in Palestine along with at least 15,000 inhabitants. The effects of these two events are not, however, altogether separable because some locations may have felt both earthquakes. Unfortunately, the close occurrence of these seismic events within the same year caused confusion in the historical record, and sources often combine these two major seismic events into one event (e.g. Guidobini and Comastri, 2005; Ambraseys, 2009).

Destroyed by this large seismic event, the city of Islamic Ayla was eventually captured in A.D. 1116 by King Baldwin I of Jerusalem, and he was not thought to have been met with much resistance as most of the city’s residents had either died or relocated as a result of the A.D. 1068 earthquakes (Whitcomb, 1994a). While a few people had come from other cities like Ramla (Guidoboni and Comastri, 2005) to try and carve out a living in the damaged city, the archaeological record indicates that Islamic Ayla was never fully inhabited nor fully functional again (Whitcomb, 1994a). After the attack by Baldwin I, the Crusaders built a small fort for their garrison, which is thought to be the Aqaba Castle located 850 m south of the collapsed city of Islamic Ayla. Whitcomb (1997) suggests that this move by the Crusaders to a new location down the coastline, rather than repairing or building on top of the old city, can be compared to that of rundown modern inner cities -- that at some point it is just easier and less expensive to construct new buildings at a new location than to restore and reconstruct damaged or dilapidated buildings. The city of Ayla changed hands again when Saladin took the city for Muslim rule from the Crusaders in A.D. 1170, only to have Crusader Renaud de Chatillon briefly occupy the city about a decade later (e.g. Khouri and Whitcomb, 1988).

Whitcomb believes that it is best to turn to the Arab geographers of the 9th and 10th centuries for first-hand descriptions of the development of the town of Ayla, including Ya’qubi who in the late 9th century wrote, “The city of Ayla is a great city on the shores of the Salt Sea, and in it gather the pilgrims of Syria, Egypt and the Maghreb. There are numerous merchants and common people…” (e.g. de Goeje, 1892; Khouri and Whitcomb, 1988). This personal account from geographer Ya’qubi (de Goeje, 1892) attests to the commercial prosperity of the town at that time and, combined with Muqaddasi’s (de Goeje, 1906) insight a century later, it also reveals that multiple sets of ruins were present and visible. Despite the collapse of Roman/Byzantine Aila, the younger Islamic Ayla was prospering (Khouri and Whitcomb, 1988).

Several westerners are known to have visited the Islamic city of Ayla throughout the 19th century. Travelers who visited the area in the first half of this century include John Lewis Burckhardt in 1816, German geographer and naturalist Eduard Rüppell and surgeon Michael Hay in 1822, French explorers Leon de Laborde and Linant de Bellefonds in 1826, and British Navy lieutenant J. Raymond Wellsted in 1833 (e.g. Burckhardt, 1822; Rüppell, 1829, 1840; de Laborde, 1838; Wellsted, 1837, 1978). The Aqaba area was also visited by a number of Western travelers in the second half of the 19th century, including English theologian Arthur Stanley in 1852-1853, English scholar Charles Doughty in 1876, English diplomat and orientalist Richard Burton in 1877, and Czech theologian and orientalist Alois Musil visited the area in 1898 and again in 1902 and 1910 (e.g. Musil, 1907, 1908, 1926; Stanley, 1910; Burton, 1913a, 1913b; Doughty, 1921).

One of the earliest 20th century visitors to the city was T.E. Lawrence in 1913 while on his trek through the region with C. Leonard Woolley. In describing what seemed to be the ruins of Islamic Ayla, Lawrence suggested the ruins and scattered artifacts represent “an Arab settlement of some luxury in the early Middle Ages” (Woolley and Lawrence, 1936). He then mentions that “there are ruins a little farther inland, and these represent probably a small village outside the gates of the larger place,” which, based on geographic placement, must have been the even older ruins of classical Aila (Woolley and Lawrence, 1936). These early accounts of Islamic Ayla and Roman/Byzantine Aila are particularly interesting and important because by the time archaeologist Donald Whitcomb visited the region in the 1980s, these ruins were either buried in the sand and out of sight, or partially built on by the expanding modern city of Aqaba. As mentioned, Whitcomb based the location of his first small trenches at Islamic Ayla on a very small surface scatter of sherds he found, as well as the primary source accounts of early geographer Muqaddasi (Khouri and Whitcomb, 1988). But the historical accounts throughout the centuries, such as Woolley’s and Lawrence’s, also reaffirmed that there were indeed numerous sets of archaeological ruins along the southern coast of modern-day Jordan.

During Whitcomb’s first excavation along the Egyptian wall and gate at Ayla in 1986, he identified several occupational phases as the site, Phases A-E, which included the Umayyad, Abbasid, and Fatimid periods (Table 3.2) (Whitcomb, 1987, 1994a). Ceramic sherds from these trenches, and those from the rest of his archaeological field seasons, testify to an international commerce at Ayla, and include imports from Iran, Iraq, Egypt, and China (e.g. Whitcomb, 1987, 1994a). Exports from Ayla to eastern Africa and the Far East also suggest broad commercial contacts during the Umayyad and Fatimid periods (Whitcomb, 1994a).

Damgaard (2011) published a site plan for Islamic Ayla showing the archaeological phasing of the site based on Whitcomb’s excavations of the city and his own work at Ayla (Figure 3.3). Damgaard’s phasing of the site includes occupational Phases 1-5, and can be roughly divided into a chronological framework based on the ceramic horizon of associated deposits and the morphology of the architecture (Damgaard, 2011). This updated phasing correlates very closely to Whitcomb’s phasing of Ayla (Phases A-E), although there are differences. To summarize these two phasing schemes, Damgaard’s Phase 5 (mid 7th - mid 8th C.) corresponds to Whitcomb’s Phase A (A.D. 650-750) and encompasses the same period of time during the Rashidun and Umayyad periods, the foundation and approximately the first one hundred years at the site (Whitcomb, 1994a; Damgaard, 2011).

The phasing differs slightly between Whitcomb’s Phase B, which spans the years A.D. 750-850, and Damgaard’s Phase 4, which encompasses the mid 8th C. to the early 9th C. (Whitcomb, 1994a; Damgaard, 2011). Whitcomb (1994a) describes these years at Ayla as showing periods of growth and prosperity. Next, according to Whitcomb (1994a), Phase C at Ayla includes the years A.D. 850-950 during the middle Abbasid period and represents continued periods of growth and prosperity. Phase C correlates to Damgaards’s Phase 3 of Ayla, which is described as the Abbasid period from the 9th to the early 10th C. (Damgaard, 2011).

Moving forward in the life of the city, Whitcomb defines Phase D as A.D. 950-1050 during the late Abbasid or Fatimid periods, and Damgaard’s corresponding phase, Phase 2, is described as occurring in the mid-to late 10th C. during the late Abbasid to early Fatimid periods (Whitcomb, 1994a; Damgaard, 2011). It is during this period of time that disruptions and decline start to appear in the archaeological record at Ayla (Whitcomb, 1994a).

Finally, the most recent occupational phase at Islamic Ayla as identified by Whitcomb, Phase E, encompasses the years A.D. 1050 during the Fatimid period to A.D. 1116 when the city is finally taken by Baldwin I during the Crusader period. Damgaard’s earliest phase, Phase 1, spans the 11th C. to the early 12th C. (Damgaard, 2011).

Whitcomb continued to excavate at Islamic Ayla until 1995, and in the process unearthed a good portion of the city and revealed much about the economy and trade happening at Ayla and in the surrounding region. Despite nearly a decade of excavation, however, much of the original city still remains buried, including the majority of the earliest Umayyad occupational layers. The Umayyad materials remain largely unexcavated primarily due to a deep overburden of younger material, a shallow water table which makes accessing the earliest deposits difficult, and limited excavation time each season. Seismic activity at the site has also likely contributed to the overall subsidence of Ayla, making these oldest archaeological remains challenging to access (e.g. Whitcomb, 1987).

• from Allison (2013:82-92)

Besides a very limited number of primary sources available concerning the site, the majority of what is known about Early Islamic Ayla comes from the previous archaeological excavations conducted, as mentioned. As far as pottery is concerned, each of the periods represented at Ayla tends to have its own characteristic style of pottery, and with tight stratigraphic control, pottery recovered from excavations can be used to better understand the history of the region. Depending on the degree of preservation, artifacts can provide clues and act to illuminate the culture of the people who once lived at a site (Khouri and Whitcomb, 1988). Most pottery recovered from an excavation typically consists of undecorated body sherds which are non-diagnostic pieces of a broken ceramic vessel that may give a ceramicist clues to its age. What is preferential to find during an excavation, however, are indicator sherds which reveal a diagnostic aspect of the vessel that is useful for the purposes of reconstruction and dating. A diagnostic sherd can consist of anything from the rim, base, or handle of a vessel, to a diagnostic glaze, paint, or incised pattern specific to a certain time period, for example. It should also be noted that even though an indicator sherd may be found, it may not be specific enough or well enough preserved to be dated accurately by a pottery expert. Still, indicator sherds inevitably reveal more information about the vessel as a whole than do body sherds.

As with any material manufactured throughout history, changes in the production of ceramic pottery occur through time. The ceramics at Ayla most certainly vary through the Islamic period, century to century, and have been found throughout the region signifying extensive trade at the city of Islamic Ayla (Khouri and Whitcomb, 1988). Changes in pottery can occur for several reasons -- as a result of a need for a specific function and/or form (e.g., water vessel vs. container for dry goods), because the available clay source or manufacturing technique of the pottery changes from wheel-made to hand-made, or vice versa, because the firing process changed, or because stylistic changes, which naturally occur in anything man-made, evolve through time as styles still do today. The analysis of pottery on the basis of vessel form, however, is typically not as common as analysis based on stylistic attributes (Sharer and Ashmore, 1979). Paint on the outside of a vessel, for example, served no function for that particular vessel, other than to be aesthetically pleasing or potentially to signify cost or societal status. The decorative paint and patterns used, however, become associated with specific periods and can be traced to determine the extent of trade, which can provide insight into the economics and politics of a historical period.

There are currently two ceramic typologies for the site of Islamic Ayla: the 1986 typology devised by Whitcomb, and a 2009 typology developed by Damgaard (2011). Starting with his excavation in 1986, Whitcomb categorized pottery recovered from stratified loci on the basis of vessel form and, in part, the surface treatment of the ceramics (e.g. Whitcomb, 1987). Rather than creating an entirely different typology, Damgaard (2011) augmented Whitcomb’s 1986 field typology and combined it with the ceramic types collected from the single 2008 Islamic Aqaba Project campaign (Damgaard, 2011). The 2009 ceramic typology, however, categorizes pottery on the basis of ware and fabric instead of form. This change in sorting principles from vessel form to ware was intended to help trace mercantilism at Ayla by distinguishing between locally produced wares and imports (Damgaard, 2011). A complete ceramic typology for the site has not been published, but in his dissertation Damgaard (2011) published the first comprehensive catalog of ceramic wares and types at Ayla, and also included Whitcomb’s 1986 typology for reference.

Since Ayla was founded in or around A.D. 650, the earliest (oldest) pottery found in situ at the site of Islamic Ayla dates to the early Umayyad period (Whitcomb, 1987). The pottery from this period represents the transition from Byzantine to Islamic material culture (Whitcomb, 1989b). Although difficult to reach in a stratigraphic section, as discussed previously, during the 1993 field season Whitcomb was alerted to the fact that there were building foundations that exposed supposed ancient pottery kilns in the modern Radwan neighborhood near Islamic Ayla. After an initial investigation, his planned excavation for the season was expanded to include the excavation of two 7th century kilns, the larger of which was almost 4 m in diameter (Melkawi et al., 1994). This find constituted one of the most important ceramics discoveries made in Aqaba, indicating a vital increase in occupation and commerce, and confirming industrial manufacturing of pottery at Islamic Ayla (Melkawi et al., 1994). Within the pottery found in situ in and around the two kilns at this new site, for instance, were Egyptian imports and the occurrence of ceramics with Coptic attributes, clearly borrowed from Egyptian pottery known as Coptic-glazed ware (Melkawi et al., 1994).

Interestingly, a number of identifiable types of pottery found at Ayla are found nowhere else in the region, and thus, “do not have ready or even vague parallels” (Melkawi et al., 1994). The findings at the 7th century kiln site actually suggest that this medieval city was a separate regional pottery center marked by strong new tradition, and is a feature that may be readily associated with the experience of early Islamic times, according to Melkawi et al. (1994). As discussed, the Red Sea was a route of trade and cultural connections in the early Islamic period, and a special type of 7th century amphora (“Ayla-Axum ware”) was used as a cargo container by the merchants because they were sturdy enough and large enough for carrying quantities of goods some distance (Whitcomb, 1994b). This expansion in Red Sea commerce after the Arab conquest in the 7th century increased the need for such amphorae, which have been found near Aden in present-day Yemen and even as far as Axum in modern Ethiopia (Whitcomb, 1994b). Whitcomb (1994b) also found what he considered to be “startling evidence” of this Ethiopian connection in an Axumite coin, the second recovered from his archaeological excavations.

Starting in A.D. 750, a change of political power from the Umayyads to the Abbasids also resulted in a change of pottery from the amphora mentioned to what is known as “Mahesh ware.” Identified and named by Whitcomb (1989b), Mahesh ware is characterized by cream-colored ware and is easily identified by a wavy pattern of incised combing that circles the vessel near the rim and/or middle of the vessel. While it is not the style of the pottery that is important, Mahesh ware is the pottery most readily associated with the Abbasid period, and it allows archaeologists to study the distinct beginning of new Abbasid traditions (Whitcomb, 1989b).

Whitcomb (1988a) has also identified at Ayla a type of handmade pottery, very utilitarian in nature, dated to the early- to mid-11th century (or perhaps very late 10th C.) for which he coined (or rather borrowed) the name “Tupperware” because it represents an everyday type of pottery that a family might use. The Tupperware also seems to have been made to nest one inside the other. Whitcomb (pers. comm., 2011) suggests that the Ayla Tupperware can act as a gauge of cultural stresses indicative of industrial, economic, and political change. Rautman (1998) explains that the adoption of wheel-based pottery-making in this part of the world is generally considered to reflect rising craft specialization and social complexity, and that the re-emergence of handmade pottery could reflect times of “dramatic cultural change.” Found alongside the handmade pottery in the late 10th - early 11th centuries in the archaeological record at Ayla are also fine luxury ceramics, including imported Chinese wares, as well as an increased trash accumulation pointing to economic disparity, and perhaps population pressures (e.g. Whitcomb, 1994a; Cobb, 1995).

Whitcomb hypothesizes that the Ayla Tupperware was produced at home by women, and perhaps by helpful children (Whitcomb, pers. comm., 2011), unlike most other pottery at Ayla that was wheel-made and produced in bulk, as was evidenced at the 7th century kiln site (Melkawi et al., 1994). Impressions of reed mats are often found on the base of the Tupperware, and some vessels were found to have been repaired (Whitcomb, 1988a). Tupperware pottery was most predominantly recovered from archaeological layers leading up to the A.D. 1068 earthquake, after which ashy destruction layers are commonly found (Whitcomb, 1988a). These handmade rich pottery assemblages indicate, therefore, that even before the city was destroyed by the earthquake of A.D. 1068, times were difficult and making pottery at home was a necessity because the ceramic industry at Ayla had been interrupted or collapsed. The Ayla Tupperware is also sometimes referred to as “abandonment pottery” because it is also found in archaeological deposits after the A.D. 1068 earthquake, although to a lesser extent (Whitcomb, pers. comm., 2011).

One of the most significant archaeological discoveries ever made at Islamic Ayla was a hoard of thirty-two gold dinars (medieval coins), minted primarily in North Africa by rulers loyal to Muslim Spain, found buried only a meter beneath the surface just inside the Syrian Gate (Whitcomb, 1992). This find acts to further illustrate the importance of the port of Ayla as a site of trade and commerce during the early Islamic period. As Whitcomb points out, this financial loss would have constituted a personal tragedy for someone, and he has hypothesized that this coin purse belonged to either a merchant or traveler, perhaps someone on his way to Mecca (Whitcomb, 1992). Further, based on the stratified evidence and the shallow depth of burial near the Syrian Gate, which is one of the four main entrances to the city, Whitcomb (1992) has postulated that the coin purse could have been dropped or temporarily hidden during the sacking of Ayla by Bedouin in A.D. 1024, never to be collected by the owner.

Political turmoil most certainly interfered with the yearly pilgrimage to Mecca along both the Nile and Red Sea routes. Peters (1994) refers to the Crusaders entering the area at this time as “Western intruders” who at the beginning of the 12th century held Jerusalem, large areas of coastal Syria, and even areas east of the Jordan River. As discussed, Ayla is known to have finally fallen when King Baldwin I took the city in A.D. 1116 (Khouri and Whitcomb, 1988). By this year, the Seigneurie of Montreal extended down through the Gulf of Aqaba and included the “fortress of Ayla” (Peters, 1994).

Given the long history of ancient Aqaba, it is telling that modern Aqaba functions in much the same way today as it has throughout the centuries -- it is still primarily a town that people visit on vacation or to pass through on their way to Israel to the west or to Saudi Arabia in the south. Known for very hot summers and temperate winters, it is located in an oasis-like setting with date palm trees and the stunning Gulf of Aqaba as a backdrop, Jordan’s only access to the Red Sea and Indian Ocean. The Gulf of Aqaba also functions as a vital port for the import and export of goods. So while the number of permanent residents in Aqaba continues to grow as the city itself expands, the city swells significantly on the weekends and during the holidays with tourists, and also with visitors from Amman and other near-by cities looking for a place to relax.

The archaeological evidence of the Islamic periods at Ayla in Aqaba provides important information on the sequence of the cultural changes that happened between late antiquity and the formation of early Islam, as well as the development between the Abbasid and the Fatimid periods (Whitcomb, 1988b). The pottery assemblages at Ayla, combined with Muqadassi’s first-hand 10th century accounts of the port, also provide evidence of the prosperity of Ayla and its extensive trade connections with North Africa, Palestine, the Hejaz, and the Far East, including China (e.g. de Goeje, 1906; Khouri and Whitcomb, 1988). More than anything, it was Ayla’s geographic location at the head of the Gulf of Aqaba at the intersection of the land and sea routes (namely the North African/Egyptian and Palestinian/Syrian roads), as well as the abundant freshwater sources (Whitcomb, 1994b), that turned Ayla into a wonderful stop-over town for Muslim pilgrims on their way to Mecca (Khouri and Whitcomb, 1988).

• from Allison (2013:89-92)

Work at Islamic Ayla and in the city of Aqaba has also included an evaluation of the liquefaction susceptibility of the site and region. The phenomenon of liquefaction is an earthquake hazard usually associated with granular sediments like sand and some gravels, although finer-grained sediments such as silts and clays can liquefy as well (McCarthy, 2002; Boulanger and Idriss, 2006). Especially susceptible where the water table is shallow, seismic ground shaking can cause sediment grains to push apart and shift, settling past one another in an attempt to move into a denser and more compact arrangement as the groundwater rises to the surface acting to reduce the load capacity of the affected sediment (Marshak, 2009). The presence of pore water interferes with the cohesion and subsequent rearrangement of grains, however, and much of the soil’s shear strength is also temporarily lost during seismic ground shaking (McCarthy, 2002). As the once-sturdy ground weakens because of the rising groundwater and sinking sediment, structures built on the surface of the affected region can sustain damage ranging from toppled chimneys, cracked foundations, and leaning or sinking walls to the complete failure and loss of a structure (Ambraseys, 2009).

Determined to be highly susceptible to liquefaction in studies conducted by Mansoor (2002) and Mansoor et al. (2004), the modern city of Aqaba, Jordan is currently at risk of liquefaction should a sizeable earthquake occur within the region. Islamic Ayla lies in an area of especially high liquefaction susceptibility due to the presence of saturated sands at shallow depth and its close proximity to the Gulf of Aqaba (Mansoor, 2002; Mansoor et al., 2004). The water table at the site of Islamic Ayla is approximately 3 m below the ground surface except in the wadi that now runs through the site where the water table is at a depth of 2 m (Al-Homoud and Tal, 1998). The city was also built on top of primarily unconsolidated alluvial sediments washed down from the large Wadi ‘Arabah and Wadi Yutim drainages located to the north of the city (e.g. Al-Homoud and Tal, 1998). Allison and Niemi (2010) conducted a drilling campaign in the so-called Circular Area, a portion of the Aqaba coastal plain located approximately 200-800 m inland from the Gulf, in order to reconstruct the paleoenvironment of ancient Aqaba. Thirteen sediment cores drilled to a depth of 10 m in the Circular Area define a basal marine transgression overlain by a regressive sequence of coastal lagoon deposits and backshore lake, sand dune, and distal alluvial fan deposits (Allison and Niemi, 2010). Brückner et al. (2002) also report the stratigraphy of a core recovered from the Mövenpick Hotel construction site, located 50 m from the shoreline and just north of our study site at Islamic Ayla. This core revealed that a 6-m-thick alluvial sequence overlies littoral sands encountered at a depth of 1.35 meters below sea level.

Al-Homoud and Tal (1998) conducted geotechnical investigations at Islamic Ayla where three borings were made to a depth of 12 m in order to obtain a soil profile with depth. They found that the archaeological site is underlain by relatively uniform subsurface conditions, and that the archaeological sediments are supported by up to 5.5 m of medium dense silty and clayey sand. The basal sediments located even further underneath Ayla are fluvial granitic gravels and sands (Al-Homoud and Tal, 1998). Two types of fill layers were found at Islamic Ayla: (1) a non-homogeneous fill lacking texture and structure and containing remnants of ash and other unidentified objects, and (2) poorly-stratified silt, sand, and gravel flood deposits from the wadi that cuts through the site (Al-Homoud and Tal, 1998). Al-Homoud and Tal (1998) also reported tilting and sinking of exterior walls that they interpreted as slumping due to horizontal ground acceleration in an earthquake.

Al-Tarazi and Korjenkov (2007) also studied the liquefaction of the Aqaba area during an archaeoseismic investigation of the city of Islamic Ayla and determined that Ayla has been damaged by liquefaction due to seismic ground shaking in the historic past and that repairs were made as a result, as has been documented by Whitcomb (1994, 1997). They confirmed that an earthquake damaged the city in the 8th century at the end of the Umayyad period, which was likely the January 18, 746/749 earthquake, although Al-Tarazi and Korjenkov (2007) cite this event as A.D. 748. They also report observing substantial damage including the tilting of walls from the A.D. 1068 earthquake during the Fatimid period. Finally, they found that the modern November 22, 1995 earthquake in Nuweiba, Egypt, which had a Mw = 7.1-7.3 (e.g. Klinger et al., 1999; Hofstetter, 2003) and was the largest to affect the area since the A.D. 1068 event, also left clear traces of liquefaction damage in the excavated buildings at Ayla. The 1995 Nuweiba epicenter was located at least 100 kilometers from the Islamic city, and a Modified Mercalli Intensity (MMI) of VIII was observed by Al-Tarazi (2000) along the beach in the Aqaba coastal zone for that event.

Finally, evidence of liquefaction typically includes sand blows and leaning or toppled structures on the surface, and dikes of fluidized sediment (usually sand) in cross-section. Noller (2001) discusses how the compactness of archaeological deposits and floors can actually lead to greater overpressure in areas with seismically accelerated sand, causing it to liquefy and form dikes and sills in cross-section. At Islamic Ayla in 1992, Whitcomb actually identified a small sand dike in cross-section as it was exposed in a trench wall during the course of his excavation of Area G near the wadi (Figure 3.4). In addition to the studies by Al-Homoud and Tal (1998), Mansoor (2002), Mansoor et al. (2004), and Al-Tarazi and Korjenkov (2007) which have shown that the Aqaba area and the site of Ayla are liquefiable, this photograph (Figure 3.4) and associated excavation log are evidence that liquefaction has occurred in the historic past in the city of Ayla, and that the strata upon which the modern city of Aqaba is built are capable of being liquefied in the future.