East African earthquakes below 20 km depth and their implications for crustal structure

We report source parameters for eight earthquakes in East Africa obtained using a number of techniques, including (1) inversion of long-period P and SH waves for moment tensors and source-time functions, (2) forward modelling of first-motion polarities and P and pP amplitudes on short-period seismograms, and (3) determination of pP-P and sP-P differential traveltimes from short-period records. The foci of these earthquakes lie between depths of 24 and 34 km in Archean and Proterozoic lithosphere, and all but one fault-plane solution indicates normal faulting (primarily E-W extension), consistent with the regional stress regime in East Africa. Because many of these earthquakes occurred in areas where the crust may have been thinned by rifting, it is difficult to ascertain whether or not their foci lie within the lower crust or upper mantle. Some of them, however, occurred away from rift structures in Proterozoic crust that is possibly 35–40 km thick or thicker, and thus they probably nucleated within the lower crust. Strength profile calculations suggest that in order to account for seismogenic (i.e. brittle) behaviour at sufficient depths to explain lower crustal earthquakes in East Africa, the lower crust must not only be composed of mafic lithologies, as suggested by previous investigators, but also that significantly more heat (∼100 per cent) must come from the upper crust than predicted by the crustal heat source distribution obtained from a 1-D interpretation of the linear relationship between heat flow and heat production observed in Proterozoic terrains within eastern and southern Africa. Precambrian mafic dike swarms throughout East Africa provide evidence for magmatic events which could have delivered large amounts of mafic material to the lower crust over a very broad area, thus explaining why the lower crust in East Africa might be mafic away from the volcanogenic rift valleys.     

Geochemistry and petrogenesis of the marginal zone of the Mount Lowe Intrusion,central San Gabriel Mountains,California

The marginal zone of the Mount Lowe Intrusion (MLI) is composed of a mineralogically and geochemically distinctive suite of quartz poor, alkali-enriched monzodiorites and quartz monzodiorites. Low initial ⁸⁷Sr/⁸⁶Sr, Rb, and Rb/Sr coupled with high Ab/Or, K/Rb, Ba and Sr suggest that feldspathic continental crust did not contribute significantly to the magma. The inferred original magma composition is most consistent with experimental data for about 20% batch modal melting of eclogite or quartz eclogite of altered oceanic tholeiite composition. The high absolute Ba and Sr abundances require in addition an enriched component, perhaps derived from dehydration processes within a subducted slab. At the level of exposure, the range of observed compositions resulted from fractionation of hornblende, plagioclase, sphene, apatite and magnetite. Fractionation resulted in formation of a peraluminous, Mn — enriched residual liquid which crystallized garnet. The grandite — enriched composition of the garnet is inferred to be due to high pressure (ca. 6.7 kb), high crystallization, relative to typical plutonic garnets. The composition of the marginal zone of MLI has many chemical similarities to early Mesozoic plutonism in the Cordilleran continental margin magmatic arc of western North America.     

Historical variation in guano production from the Peruvian and Benguela upwelling ecosystems

Analysis of the historical record shows that reductions in guano production in the eastern tropical Atlantic Ocean were not related to El Niño events in the Pacific. Guano production in the Atlantic was negatively correlated with guano production in the Pacific. Significant reductions in guano production in the Atlantic occurred 1 to 4 years after increases in guano production in the Pacific. Negative association is consistent with the hypothesis that global shifts in weather associated with the Southern Oscillation affect guano production.     

Ammonia measurements at Niwot Ridge,Colorado and point arena,California using the tungsten oxide denuder tube technique

The applicability of the tungsten oxide denuder tube technique for the measurement of ammonia in the rural troposphere was investigated. The technique is based on selective chemisorption of NH₃ from a gas stream, thermal desorption, conversion to NO, and analysis by NO–O₃ chemiluminescence. Nitric acid, which is also collected and desorbed as NO, was distinguished from NH₃ by differences in desorption temperature. Substituted amines were also collected, but desorbed at a slightly lower temperature than NH₃ in dry air. At high relative humidities, alkylamines may be hydrolyzed to NH₃ on the denuder surface and hence detected as NH₃. Overheating of the denuder tube during the temperature-programmed desorption was found to cause significant irreversible degradation of system performance.The technique was used to measure NH₃ mixing ratios at two rural locations in the United States. At a mountain site in Colorado during the winter of 1984, the average NH₃ mixing ratio was 0.20 ppbv (=0.08 ppbv). At an isolated coastal site in northern California during the spring of 1985, the average NH₃ mixing ratio was 0.36 ppbv (=0.17 ppbv). Correlations of the latter measurements with wind direction and NO _x level suggest that the NH₃ mixing ratio in Pacific marine air at 40°N is <-0.25 ppbv.     

Estimating monthly streamflow values by cokriging

Cokriging is applied to estimation of missing monthly streamflow values in three records from gaging stations in west central Virginia. Missing values are estimated from optimal consideration of the pattern of auto- and cross-correlation among standardized residual log-flow records. Investigation of the sensitivity of estimation to data configuration showed that when observations are available within two months of a missing value, estimation is improved by accounting for correlation. Concurrent and lag-one observations tend to screen the influence of other available observations. Three models of covariance structure in residual log-flow records are compared using cross-validation. Models differ in how much monthly variation they allow in covariance. Precision of estimation, reflected in mean squared error (MSE), proved to be insensitive to this choice. Cross-validation is suggested as a tool for choosing an inverse transformation when an initial nonlinear transformation is applied to flow values.     

High-pressure crystal chemistry of phenakite (Be2SiO4) and bertrandite (Be4Si2O7(OH)2)

Compressibilities and high-pressure crystal structures have been determined by X-ray methods at several pressures for phenakite and bertrandite. Phenakite (hexagonal, space group R \(\bar 3\) ) has nearly isotropic compressibility with β=1.60±0.03×10^?4 kbar^?1 and β=1.45±0.07×10^?4 kbar^?1. The bulk modulus and its pressure derivative, based on a second-order Birch-Murnaghan equation of state, are 2.01±0.08 Mbar and 2±4, respectively. Bertrandite (orthorhombic, space group Cmc2₁) has anisotropic compression, with β _a =3.61±0.08, β _b =5.78±0.13 and β _c =3.19±0.01 (all ×10^?4 kbar^?1). The bulk modulus and its pressure derivative are calculated to be 0.70±0.03 Mbar and 5.3±1.5, respectively. Both minerals are composed of frameworks of beryllium and silicon tetrahedra, all of which have tetrahedral bulk moduli of approximately 2 Mbar. The significant differences in linear compressibilities of the two structures are a consequence of different degrees of T-O-T bending.     

Boninite petrogenesis and alteration history: constraints from stable isotope compositions of boninites from Cape Vogel,New Caledonia and Cyprus

¹⁸O values of unaltered olivine and pyroxene phenocrysts in boninites from several areas range from 5.8 to 7.4 and indicate that the source for most boninites is more ¹⁸O-rich than MORBs and other oceanic basalts. The source for oxygen and other major elements is most likely a refractory portion of the mantle having a ¹⁸O value of up to 7.0 to which must be added a small amount of H₂O-rich fluid to induce partial melting. This fluid, which is derived from subducted crust, is the vehicle for LREEs including Nd. The variable, normally low _Nd values typical of boninites do not correlate with the ¹⁸O values.Post eruptive exchange of oxygen in the glass of boninites with that of sea water at low temperatures (<150° C) produces ¹⁸O values of >10 in optically fresh glass. Hydration of the glass has increased the water contents of most boninites from estimated magmatic values of 1–2 wt% to 2–4 wt% and produced D values of < –80, which may be lower than the original magmatic D values. In contrast to most submarine pillow basalts, the magmatic volatile composition of boninite lavas has been extensively modified as a result of post eruptive interaction with seawater.     

Thick- and thin-skinned tectonics in the Pyrenees

A comparison is made between the Gavarnie thrust and the Mérens Fault in the Axial zone of the Pyrenees. The former has a gentle dip and quite a large displacement (at least 12 km) but does not cut through either Hercynian or Alpine isograds. The latter has a smaller displacement (~ 5 km) but dips steeply and cuts through both Hercynian and Alpine isograds at a high angle. On this basis and on the basis of shear zone geometries immediately north of it, it is proposed that the Mérens Fault nucleated as a steeply (65°–80°) dipping structure, while the Gavarnie thrust nucleated with a shallow attitude. The Mérens Fault is not a backward-rotated thrust fault, nor is it the root zone for any major nappe structure. Similar steep ductile structures occur within the Gavarnie nappe and may reflect considerable internal strain in basement lithologies.The relationship between steep and shallow structures is not yet clear; the shear zones may pre-date the thrusting in which case they may be thick-skinned structures affecting the whole lithosphere, or they may be contemporary with thrusting reflecting only local thickening above a décollement.Rheological models can be used to test proposed geometrical and kinematic models for the lithosphere-scale evolution of the Pyrenees. Suggested models are dominated by a cool, rigid, high-level mantle wedge beneath the North Pyrenean zone which probably controlled the location of north-dipping thrust faults. Thick-skinned shortening is possible in thick crust in the Axial zone but is very unlikely in the North Pyrenean zone where steeply rooted structures would have to cut through the strongest part of the lithosphere.