REVIEWS OF GEOPHYSICS AND SPACE PHYSICS, VOL. 22, NO. 1, PAGES 1-36, FEBRUARY 1984 Azimuth, Energy, Q, and Temperature' Variations on P Wave Amplitudes in the United States RHETT BUTLER 1 Cooperative Institutefor Research in Environmental Sciences Universityof Colorado/NOAA Mean relative P wave (1 s) amplitude characteristics are determinedfor standardized seismic stations in the United States. Data are compiled from three azimuths:to the north, explosions in Russia;to the northwest, earthquakes along the Asian-Pacificarc; and to the south-southeast, earthquakes in South America. Characteristic receiver amplitudes are found to be azimuthally dependent.High-amplitude areasoccur in the central United States, and lowest amplitudes occur along the Rocky Mountain front range and the Rio Grande rift zone. Amplificationeffects of low-velocity sediments are significant in the central United States. Azimuthal amplitude variationsin the eastern United States are consistent with an anomalous attenuating or defocusing body in the lower mantle beneath Colombia, South America. Nonazimuthal amplitude studies are tied to the azimuthal data set to increase station amplitude cover- age. The regional and local stability of amplitudes is analyzed.Amplitudesare stable with an average variation of a factor of about 1.3 around the regional mean. For seismicstations situated on hard bedrock, the squaredamplitude variations are related to energy variations in the P wave fronts. Ac- knowledging the likely importance of elastic focusing/defocusing and scattering effectsat individual stations, variations of anelastic attenuation are derived that are consistent with the observedenergy- amplitude variations. Assuming that the variationsof anelastic attenuation occur in the upper 400 km of the earth, high-amplitude stations may be unattenuated, low-amplitude stations are consistent with Q •, 110, and a mean amplitude implies Q •, 220-240. The range of variation of Q is mapped into temperature variations,assuming a thermally activated Q mechanism and an olivine activation energy for the upper mantle. The apparent Q variations are consistent with temperaturevariations of 135 ø- 287øC between high- and low-amplitude seismic station sites. Geophysicaland physiographic corre- lations with the P wave amplitude variations are discussed. Accounting for the motion of the North Americanplate, low P wave amplitudes along the Rocky Mountain front may be related to the uplift of the Colorado Plateau. Low amplitudesextending from northern Texas to New England lie along the approximate boundary of the continental craton. CONTENTS Introduction and review ............................ 1 Present course ................................... 3 Data set 5 Response of instruments .......................... 7 Amplitude data reduction ........................... 8 Amplitude results for three azimuths .................. 10 Stationmean relativeamplitudesummary.............. 14 Mean amplitudepattern for the United States and southern Canada .............................. 15 Azimuthal comparisons ............................ 16 Amplitude stability................................ 18 Introduction ................................... 18 Array studies .................................. 18 Regionalvariations .............................. 19 Azimuthal variations ............................ 20 Confidence intervals of amplitudeestimates ........... 21 Seismometer site effects ............................ 22 Amplification effects of shallow sediments ............ 22 Focusing/defocusing and scatter effects at seismometer sites ............................ 23 Seismometer calibration .......................... 24 Effectof frequency-dependent Q ...................... 25 Energy, amplitudes, and Q .......................... 25 Temperature variations inferred from Q variations ........ 27 •Now at HawaiiInstitute of Geophysics, University of Hawaii. Copyright 1984 by the AmericanGeophysical Union. Paper number 3R1817. 0034-6853/84/003R- 1817515.00 Tying LRSM amplitudes to the WWSSN array .......... 28 WWSSN and LRSM amplitude variations .............. 29 Geophysical and.physiographic correlations ............ 30 Summary ....................................... 32 Appendix....................................... 33 1. INTRODUCTION AND REVIEW Seismic stations exhibit characteristic amplitudes thai are a property of station sites in much the sameway that travel time residuals are. Determined by averaging over many earth- quakes and explosions and corrected ['orgeometric spreading, these characteristic relative amplitudes have been found to vary both among different seismometer sitesand as a function of azimuth at a given seismometer. Theseamplitude variations are fundamental in their relation to energy variations in the seismic wave fronts. Beyond the measurement of P wave am- plitude variations lie the uncertainty and nonuniqueness of their interpretation. This paper• pre•ents an analysis of P wave (1 s) amplitude variationsin the United Statesfor three azi- muths. The observed amplitude variations may be interpreted hypothetically as anelastic attenuation variations, acknowl- edging the likely contribution of other mechanisms such as elastic focusing and scattering, and then temperature vari- ations in the earth may be inferred with the assumptionof a thermally activated Q mechanism.Amplitude variations at World-Wide Standard Seismograph Network (WWSSN) and the Long Range SeismicMeasurement (LRSM) stations are tied together and discussed in relation to geophysical and physiographic correlations. Following the studies of Gutenberg and Richter, which used a variety of seismometers, amplitudesof earthquakes and ex-