ISSN:
1420-9136
Keywords:
Q
;
surface waves
;
crust
;
upper mantle
Source:
Springer Online Journal Archives 1860-2000
Topics:
Geosciences
,
Physics
Notes:
Abstract Fundamental-mode Rayleigh wave attenuation data for stable and tectonically active regions of North America, South America, and India are inverted to obtain several frequency-independent and frequency-dependentQ β models. Because of trade-offs between the effect of depth distribution and frequency-dependence ofQ β on surface wave attenuation there are many diverse models which will satisfy the fundamental-mode data. Higher-mode data, such as 1-Hz Lg can, however, constrain the range of possible models, at least in the upper crust. By using synthetic Lg seismograms to compute expected Lg attenuation coefficients for various models we obtained frequency-dependentQ β models for three stable and three tectonically active regions, after making assumptions concerning the nature of the variation ofQ β with frequency. In stable regions, ifQ β varies as ωξ, where ζ is a constant, models in which ζ=0.5, 0.5, and 0.75 satisfy fundamental-mode Rayleigh and 1-Hz Lg data for eastern North America, eastern South America, and the Indian Shield, respectively. IfQ β is assumed to be independent of frequency (ζ=0.0) for periods of 3 s and greater, and ζ is allowed to increase from 0.0 at 3 s to a maximum value at 1 s, then that maximum value for ζ is about 0.7, 0.6, and 0.9, respectively, for eastern North America, eastern South America, and the Indian Shield. TheQ models obtained under each of the above-mentioned two assumptions differ substantially from one another for each region, a result which indicates the importance of obtaining high-quality higher-mode attenuation data over a broad range of periods. Tectonically active regions require a much lower degree of frequency dependence to explain both observed fundamental-mode and observed Lg data. Optimum values of ζ for western North America and western South America are 0.0 if ζ is constant (Q β is independent of frequency), but uncertainty in the Lg attenuation data allows ζ to be as high as about 0.3 for western North America and 0.2 for western South America. In the Himalaya, the optimum value of ζ is about 0.2, but it could range between 0.0 and 0.5. Frequency-independent models (ζ=0.0) for these regions yield minimumQ β values in the upper mantle of about 40, 70, and 40 for western North America, western South America, and the Himalaya, respectively. In order to be compatible with the frequency dependence ofQ observed in body-wave studies,Q β in stable regions must be frequency-dependent to much greater depths than those which can be studied using the surface wave data available for this study, andQ β in tectonically active regions must become frequency-dependent at upper mantle or lower crustal depths.
Type of Medium:
Electronic Resource
URL:
http://dx.doi.org/10.1007/BF00881746
Permalink
