S U M M A R YWe have obtained three-layer crustal models of shear wave Q (Q µ ) for several surface wave paths in China and peripheral regions using a single-station, multimode method in which amplitude spectra of fundamental-mode and higher-mode Rayleigh waves, computed for known source depths and mechanisms, are compared to measured spectra. The three layers, totaling 60 km in thickness, roughly comprise the entire crust in the Tibetan plateau and include both the crust and part of the uppermost mantle in other regions where the crust is thinner. Q µ in the shallowest layer, 10 km thick, is lowest (about 40) in the western portion of the Tibetan plateau and highest (about 250) in southeastern China. In the middle layer, at 10-30 km depth, Q µ is lowest (60-80) beneath Tibet and the Pamir thrust system and highest (120-140) in central China and parts of the Sino-Korean platform. Uncertainties of Q µ in the deepest layer, at 30-60 km depth, are much greater than in the upper two layers but available results suggest that Q µ is lowest (about 80) under the Pamir thrust system and highest (about 180) under southern Mongolia. The densities of event and station coverage, although wanting in some regions, allow us to develop the first regionalized maps of crustal Q µ variation for continents. The maps, for depth ranges 0-10 km and 10-30 km, generally show good agreement with Q results obtained earlier with other phases and good correlation with the tectonics of this active region.

No abstract

No abstract

Crustal attenuation across the East African plateau in Tanzania, an area of uplifted and rifted Precambrian crust, has been investigated using seismic data from regional earthquakes recorded by the 1994-1995 Tanzania broadband seismic experiment. We use 1 Hz Lg coda waves from the 17 events, together with the energy flux model of Frankel and Wennerberg (1987), to obtain estimates of intrinsic (Q I) and scattering (Q S) attenuation for East Africa. Q I values across the plateau are fairly uniform, ranging from a low of ∼300 to a high of ∼600. Q I values for the Tanzania craton, in the middle of the plateau, are similar to those for the mobile belts, which form the sides of the plateau. Q I of 300 to 600 is somewhat lower than the average crustal Q for Precambrian terrains elsewhere. Heat flow from the Tanzania craton and surrounding mobile belts is not elevated; therefore, we attribute the lower-thanaverage Q values not to elevated crustal temperatures, but instead to rift faults in the crust that are interconnected and filled with fluids. Q S ranges from ∼1000 in the mobile belts and along the eastern margin of the Tanzania craton to ∼2200 in the north central part of the plateau just south of Lake Victoria. We attribute the variability in Q S to scattering of Lg by surface topography, in particular, rift basins along the eastern side of the plateau.

Starting with fundamental-mode Rayleigh-wave attenuation coefficient values (c R ) predicted by previously determined frequency-independent models of shear-wave Q (Q l ), we have obtained frequency-dependent Q l models that explain measured values of c R as well as of Lg coda Q and its frequency dependence at 1 Hz (Q o and g, respectively) for China and some adjacent regions. The process combines trial-and-error selection of a model for the depth distribution of the frequency dependence parameter (f) for Q l with a formal inversion for the depth distribution of Q l at 1 Hz. Fifteen of the derived models have depth distributions of f that are constant, or nearly constant, between the surface and a depth of 30 km. f distributions that vary with depth in the upper 30 km are necessary to explain the remaining seven models. f values for the depth-independent models vary between 0.4 and 0.7 everywhere except in the western portion of the Tibetan Plateau where they range between about 0.1 and 0.3 for three paths. These low f values lie in a region where Q Lg and crustal Q l are very low and suggest that they should also be low for high-frequency propagation. The models in which f varies with depth all show a decrease in that value ranging between 0.55 and 0.8 in the upper 15 km of the crust and (with two exceptions where f ¼ 0.0) between 0.3 and 0.55 in the depth range 15-30 km. The distribution of f values between 0.6 and 0.8 (the higher part of the range) in the upper crust indicates that high-frequency waves will propagate most efficiently, relative to low-frequency waves, in a band that includes, and strikes north-northeastward from the path between event 212/97 and KMI to the path between event 180/95 and station HIA in the north.