A brief Oligocene period of flood volcanism in Yemen: Implications for the duration and rate of continental flood volcanism at the Afro-Arabian triple junction

Abstract

40 Ar 29 Ar dating of mineral separates and whole-rock (WR) samples has established that basaltic continental flood volcanism (CFV) began between 30.9 and 29.2 Ma in northwestern and southwestern Yemen, respectively. Rhyolitic volcanism commenced at 29.0–29.3 Ma throughout Yemen. Lower basaltic lavas were erupted every 10–100 kyr, whereas upper bimodal volcanic units were erupted every 100–500 kyr, which reflects generation of rhyolitic magmas from basalts that resided for longer periods in lithospheric magma chambers than during the early phase of exclusively mafic magmatism. The youngest dated flood volcanic units were erupted between 26.9 and 26.5 Ma throughout Yemen. The duration of preserved CFV defined by 40 Ar 29 Ar dating (4.4 myr) contrasts with the wide range of WR K Ar dates previously obtained in Yemen (> 50 myr). 40 Ar 39 Ar "> 40Ar39Ar step-heating studies of WR samples has shown that this discrepancy is due to the disturbed Ar systematics of volcanic samples. Most samples have experienced post-crystallization loss of radiogenic Ar and/or contain excess Ar, with only ca. 25% of the WR K Ar dates within 1–2 myr of true crystallization ages. WR K Ar data can be screened for reliability using the radiogenic Ar yield and 40 K 36 Ar "> 40K36Ar ratio, which reflect the Ar retentivity of the sample, the likelihood that alteration has disturbed a sample's Ar systematics, and the susceptibility of the sample to a finite amount of Ar loss or the presence of a finite amount of excess Ar. Examination of existing WR KAr data in the Ethiopian part of this flood volcanic province, using these parameters, suggests that much of these data are also misleading. Two phases of flood volcanism are inferred in Ethiopia and Eritrea at 38-30 Ma and ca. 20 Ma. The older phase is equivalent to that in Yemen, and is consistent with the progression in basal volcanic ages obtained in Yemen moving from north to south. The younger phase is related to the onset of upper crustal extension and incipient Red Sea-Gulf of Aden rifting. The sequence of events — surface uplift (?), flood magmatism and subsequent upper crustal extension — in Yemen is consistent with the involvement of a mantle plume at the Afro-Arabian tripe junction. However, the overall eruption rate for this flood volcanic province is only 0.03 km 3 /yr, much slower than that postulated for other plume-related provinces such as the Deccan or Siberian Traps, but perhaps comparable to the Paraná-Etendeka province, which also contains significant amounts of rhyolitic volcanic products like those of Yemen-Ethiopia. The highly variable eruption rates in individual provinces must reflect the very different character of individual plumes, or the control of lithospheric structure and plate tectonic stresses on the surface manifestations of plumes. The long duration of CFV and large amounts of rhyolitic volcanism at the Afro-Arabian triple junction may be attributed to the relatively slow separation of the African and Arabian plates compared with, for example, the rifting of India and the Deccan Traps.