Stalled eruption or dike intrusion at Harrat Lunayyir, Saudi Arabia?

Since 2007 there have been periodic small earthquake swarms recorded at seismic stations located around Harrat Lunayyir in northwest Saudi Arabia. In response to a longer-lived and larger seismic swarm, in early May 2009 the Saudi Geological Survey (SGS) installed a network of broadband seismometers surrounding the Harrat, with real-time data transmitted by VSAT back to its Earthquake and Volcano Center in Jeddah. The new network captured data from the swarm, which culminated with a M5.4 earthquake on 19 May that caused very minor structural damage in the town of Al Ays, 40 km to the southeast. Because of the strength of ground-shaking and resulting structural damage, the Saudi government evacuated more than 30,000 people from a 40 km radius surrounding Harrat Lunayyir, including the town of Al Ays. At the request of the SGS and the U.S. Consulate in Jeddah, the Volcano Disaster Assistance Program (a partnership program of the USGS and USAID’s Office of Foreign Disaster Assistance) sent two advisers to work with the SGS to determine if the unrest was volcanic in origin and to assess hazards associated with the unrest. The SGS-USGS team collected geologic, geodetic and seismic data that confirmed the volcanic nature of the activity. 1) A 3-km-long NW-trending surface rupture appeared in the central part of the Harrat before 19 May, during a time when maximum earthquake magnitudes exceeded M4. 2) The length of this rupture extended to 8 km with the M5.4 earthquake on 19 May. 3) The rupture opened 0.5 m and had an absolute motion of 0.9 m to N35W with a 63 degree plunge to the NE- a tensional offset consistent with that expected from an M5.4 earthquake in a rift-shoulder setting. 4) InSAR data showed about a meter of extension and uplift (as of July 2009) with the majority of deformation associated with the 19 May M5.4 earthquake. The InSAR-derived deformation field is best modeled by intrusion of a narrow (~2 m wide), 8-km-long dike to shallow crustal levels. 5) Finally seismic data showed features similar to other volcanic swarms, including high rates of occurrence of small events, shallow event locations that clustered beneath the lava field and were coincident with the area of maximum deformation, RSAM values that peaked during periods of increased InSAR inflation, and a mixture of event types- high frequency, low frequency, possible very low frequency, and tremor. Based on our analysis of similar active volcanic systems worldwide and associated earthquakes in rift settings, we determined that for the three month period following 19 May 2009 there was a moderate probability that the activity could culminate in an eruption, and a very low probability of additional large earthquakes of M6 or greater. Subsequent to this hazard assessment and once the stability of structures in Al Ays was adequately addressed, the evacuees were allowed to return to their homes and daily lives.