JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 96, NO. B8, PAGES 13,201-13,224, JULY 30, 1991 EvolvingGeographic Patterns of CenozoicMagmatism in the North American Cordillera: The Temporaland Spatial Association of Magmatism and MetamorphicCore Complexes RICHARD LEE ARMSTRONG Departmentof GeologicalSciences, University of British Columbia,Vancouver, Canada U.S. Geological Survey,Menlo Park, California Four maps are presentedhere that show the location and extent of magmaticfields between eastern Alaska and northernMexico during the successivetime intervalsof 55-40, 40-25, 25-10, and 10-0 Ma, and four others show the distributionof metamorphiccore complexesduring the same Cenozoic time intervals. The maps are based on U.S. Geological Survey and Canadian Cordilleran data bases contining about 6000 isotopic dates and extensive literature review. For nearly 60 Ma the developmentof metamorphiccore complexeshas coincided with the locus of areally extensive and voluminousintermediate-composition magmatic fields. The associationis suggestiveof a close link betweenmagmatism and core complexformation, namely that magma directly and indirectly lowers the strengthof the crust. Magmatism thus controlsthe location and timing of core complex formation. The stressesresponsible may be inherited from Mesozoic crustal thickening,locally createdby uplift and magmatic thickeningof the crust, and imposedby the global pattern of plate motions and driving forces. Since the Miocene, rates of magmatism,extension, and core complex formation have declined. The modern Basin and Range province is not a suitable model for the situation that existed during major magmatic culminations. The singular event of early Miocene time, the merging of two large magmatic fields, extinguishing the Laramide magmatic gap, explains several disconnected observations:the hyperextensionepisode of the ColoradoRiver corridor, rapid reorientationof stress patternsacross much of westernNorth America, and subsequentrapid tectonicmovements in California. Magma-triggeredbreakup of westernNo•h America lithospherecoincided with developmentof the San Andreas transform system. Thermal destructionof the Laramide magmatic gap created a California "microplate"about 22 Ma ago that moved rapidly away from North America. Thus two plate tectonic processes,thermal destructionof the lithosphere"bridge" and northwardgrowth of a transformsystem, interactedto produceMiocene and later tectonicpatterns and events. INTRODU•ON the Cordilleran region and discussedthe significance of those evolving magmatic patterns for the Western Interior There have been many reviews of the Cenozoic magmatic Basin. The mandate of the first paper did not include the evolution of the western United States [Armstrong et al., Cenozoic, but the abundantdata on age and distribution of 1969; McKee, 1971; Lipman et al., 1971; Christiansen and magmatismdid not stop in the early Cenozoic. In fact, the Lipman, 1972;Lipman eta/., 1972;Armstrong and Higgins, time and spacedistribution of Cenozoicmagmatism is richly 1973; Snyder et al., 1976; Cross and Pilger, 1978; and accuratelydocumented. It was almost no extra effort to Armstrong, 1978; Coney and Reynolds,1977; Stewart and continuethe compilation,done up to 55 Ma ago for the Carlson, 1978; Christiansen and McKee, 1978; Luedke and Western Interior Basin paper, to the present. Moreover, Smith [1979a, 1979b, 1981, 1982, 1983, 1984, 1986], there were other interestingstories to tell. The maps that Luedke et al. [1983]; Lipman, 1980; Ewing, 1980; Smith are centralto both paperswere createdtogether from a large and Luedke, 1984; Mutschler et al., 1988], southeastern body of recently assembledinformation. Togetherthey Alaska [Brew, 1988] and some broad overviews of the provide a cinematicimage of evolvingmagmatic patterns, Cenozoic magmatic history of western Canada [Souther, presentedin uniform graphic style, for Late Triassic to 1967, 1970, 1977; Noel, 1979]. Nevertheless, there has presenttime. In discussingmagmatism we include data from never been a single presentation of the distribution of both intrusive and extrusive igneous rocks. For the magmatic activity over the entire Cordilleran region Cenozoic the data base is weighted toward volcanic rocks, extendingfrom southeasternAlaska to the Mexican border whereasin the Mesozoic, intrusiverocks are predominant. (Figure 1). None of the previousreviews has beenbased on Where studied in detail, intrusive bodies and thick lava as large a data baseas we have assembled.Our compilation accumulationsare coincident in time and are essentially for the Cenozoic is an outgrowth of work for anotherpaper coextensive. Distal volcanic ash layers must of course be [Armstrong and Ward, 1991] that synthesizedthe Late disregardedin mappingthe extent of magmaticallyactive Triassic to early Cenozoic(230-55 Ma) magmatichistory of areas. For many years the U.S. Geological Survey has been Copyright1991 by the AmericanGeophysical Union. building a RadiometricAge Data Bank (RADB) [Zartman et al., 1976]. Coveragehas graduallybeen compiled state by Paper number 91JB00412 state. Ward [1986, and unpublishedmanuscripts, 1986, 0148-0227/91/91JB-00412505.00 1988] recently assembled the state files for the entire 13,201 13,202 ARMSTRONGAND WARD: C'EN•IC MAGMATISM IN THE NORTH AMERICANCORD-.I .•RA EXPLANATION BOUNDARIES OF THICK-SKINNED LARAMIDE DEFORMED BELT EAST BOUNDARY OF LATE MESOZOIC FOLD AND THRUST BELT WEST BOUNDARY OF OLDER PRECAMBRIAN BA8EMENT OF NORTH AMERICA EAST LIMIT OF TERRANES ACCRETED IN PALEOZOIC TO EARLY TRIASSIC TIME EAST LIMIT OF TERRANES ACCRETED IN LATE TRIASSIC TO JURASSIC TIME EAST LIMIT OF TERRANES ACCRETED IN LATE JURASSIC TO CRETACEOUS TIME - WRANGELL TERRANE IN NORTH - FRANCISCAN COMPLEX IN SOUTH EAST LIMIT OF TERRANES ACCRETED IN CENOZOIC TIME BARBS ARE ON UPPER PLATE OF MAJOR TECTONIC CONTACTS AND THUS INDICATE DIP DIRECTION OF THOSE CONTACTS. BARBS ON BOTH SIDES INDICATE THAT DIP OF TECTONIC CONTACT IS UNKNOWN, Fig.1. Mapshowing tectonic belts and geographic base for Plates 1-4 [after Armstrong and Ward, 1991]. Cenozoic tectonicand magmatic features are labeled on subsequentFigures. western United States and produced a series of computer- There are 11,500 age determinationsplotted on the maps plotted maps showingdistribution of dated rocks during a that we preparedspanning the last 230 Ma in the Cordillera. sequenceof time intervals. He discussedthe observed About 50% of those are within the 55-0 Ma time interval magmatichistory and possibleplate tectonicexplanations, that we focus on here. The data presented in this paper with particular emphasison the Cenozoicpatterns in the representa fourfold increasein the numberof isotopicdates western United States. compiled since the last similar syntheses[Snyder et al., The Canadian Isotopic Date Data File was convertedto 1976; Cross and Pilger, 1978], but there are certain computer-readableform by B entzen [1987], with support necessarydisclaimers. The data are not completefor many from the GeologicalSurvey of Canadaand British Columbia reasons. The intent of the compilers of RADB was to Ministry of Energy, Mines and PetroleumResources, during include only dates published in major journals with the winter of 1986-1987. We merged the two data banks to supportingdetails on analytical techniques,decay constants provide reasonably comprehensivecoverage for the entire used, and observedconsistency with other dates and geologic North American Cordillera between Mexico and Alaska relationships. The Canadianfile, in contrast,contains many (Figure 1). unpublished,or grey literature, items. Some large regions, ARMSTRONGAND WARD: CENOZOICMAGMATISM IN TIlE NORTH AMERICAN CORDII I F.RA 13,203 such as northern Washington, northern British Columbia, of magmatic belt boundaries from the previous interval to and the Yukon, are less well studiedand may not be properly the one being illustrated. Solid symbols show age representedin this type of summary. Young depositsmay determinationsbelonging to the early part of time intervals, obscure the older record in areas such as the Intermontane and open symbolsshow the age determinationsin the last 5- Belt of British Columbia, Columbia Plateau, Snake River 10 Ma of the intervals. This allows changes within any Plain, and Puget-WillametteTrough. Neither file is entirely interval to be directly visualized. These within-interval up to date or complete. For example, the Canadianfile on changes are highlighted by larger outlined arrows. Some magnetictape has been updatedonly to late 1984, did not larger arrows are based on literature discussions of the have unpublished data from universities other than the magmatic evolution of specific areas. Where open and University of British Columbia or from the Geological closed age determination symbols mingle on the maps, Survey of Canada, and was known to be incompletewith magmatism persisted throughout the time interval for the respectto datesfor ash layers in the WesternInterior Basin. map. The outlines of magmatic areas on Plates 1-4 have The RADB was last updated in 1986. There are unknown been extensively edited to conform to literature descriptions amountsof grey literature, unpublisheddata, and overlooked of magmatic activity in many areas where dates are not data that result in incomplete representation of the shown. information that exists at any time. Probably another 20-30 It has not been practical to compile volumetric and % could be added today to the data collection. Both large petrochemicaldata for this review. The