JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 96, NO. Bll, PAGES 18,323-18,334, OCTOBER 10, 1991 Thermoelastic Stress in Oceanic Lithosphere Due to HotspotReheating ANNING ZHU AND DOUGLAS A. WIENS Department of Earth andPlanetarySciences, Washington University, St.Louis, Missouri We investigate the effect of hotspot reheating on the intraplate stress field by modeling the three- dimensional thermal stress field produced by nonuniform temperaturechangesin an elastic plate. Temperature perturbations are calculated assuming that the lithosphere is heated by a source in the lower part of the thermal lithosphere. A thermal stress model for the elastic lithosphere is calculated by superposing the stress fields resulting from temperature changes in smallindividual elements. The stress in an elastic plate resulting from a temperature change in each small element is expressed as an infinite series, whereineachterm is a source or an image modifiedfrom a closed-form half-space solution. We apply the thermalstress solution to midplateswells in oceanic lithosphere with various thermal structures andplate velocities.Our results predict a stress field with a maximum deviatoric stress on the order of 100 MPa (1 kbar)covering a broad area around thehotspot plume. The predicted principalstress orientations showa complicated geographical pattern, with horizontal extension perpendicular to the hotspot track at shallow depths and compression along the track near the bottom of the elastic lithosphere. Although stress data near oceanic swellsare limited, the source parameters of intraplate earthquakes near several hotspots are consistent with the thermal stress model. These results indicatethat thermal stress due to reheating may be an important contributor to stress fields nearhotspots in old oceanic lithosphere. INYRODUCTION Temperature changes have long been recognized as potentially important sources of stress which may affect intraplate tectonics [e.g., Turcotte, 1974a]. Seismicity rates and stress orientations derived from oceanic intraplate earthquakes in young lithosphereindicate that the intraplate stress field is controlledby thermal stresses from lithospheric cooling [Wiens and Stein, 1983, 1984; Bergman and Solomon, 1984]. Bratt et al. [1985] were able to fit these observations using an elastic half-space model in which thermal stresses were relieved on time scales shorter than the age of the lithosphere. The effects of thermoelastic stress in oceanic lithosphere have alsobeen identified at fracture zones. Parmentier and Haxby [1986] and Haxby and Parmentier [1988] found that geoid profiles across fracture zonesshowed clear evidence of bending induced by thermal stresses associated with lithospheric cooling. It has also been suggested that transform faults may result from thermal stresses [Turcotte, 1 9 7 4 a; Sandwell, 1986]. Hotspot swells represent large-scale thermal anomalies within the lithosphere and upper mantle [e.g., Detrick and Crough, 1978; Crough, 1983]; thus it is natural to question whether thermalstress due to reheating is a dominant factor in the intraplate stressfield near hotspots. In this paper we discuss the possible influence of lithospheric reheating on the intraplate stress pattern in mechanically strong oceanic lithosphere near hotspots. Our primary goals are to estimate the order of magnitude of thermal stress induced by temperature perturbation from hotspot reheating and to predict the thermal stress orientations. Our models predict that thermal expansion due to reheating will producea complexpattern of stress orientations, with maximum deviatoricstresses as large as 100 MPa (1 kbar)near the base of the elastic lithosphere. Copyright1991 by the AmericanGeophysical Union Paper number 91IB01907. 0148-0227/91/911B -01907 $05.00 MODEL FORMULATION Temperature Model Several lines of evidence suggest a thermal origin for midplate swells. Most swells are associated with chains of hotspot volcanos formed as the plate moves over a mantle thermal anomaly [Wilson, 1963]. Geoid anomalies over swells are correlated with topography and indicate compensation within the middle or lower regions of the thermal lithosphere [Crough, 1978; McNutt and Shure, 1986]. Midplate swell topography subsides according to a predictable thermalcooling curve after passage over the locus of volcanism [Detrick and Crough, 1978; Crough, 1983]. Swells are also associated with heat flow anomalies [Von Herzen et al., 1982; Detrick et al., 1986; Courtney and White, 1986], although recent studies suggest that some anomalies are smaller than previously thought[Von Herzen et al., 1989; Stein and Abbott, 1991]. Seismological evidence for temperature anomalies beneath hotspots is equivocal; Zhang and Tanimoto [1990] find distinct low-velocity zones beneath hotspots, whereas Woods et al. [1991] find no Rayleigh wave velocity anomalies along the track of the Hawaiian hotspot. In order to calculatethe stresses within the strongpart of the lithosphere inducedby hotspotreheating,we must assume a temperature model for the swell. Severalgeophysical models have been developed to explain the observations associated with midplate swells. The simplest of these are the lithospheric reheating models, in which the lithosphere is thinned due to the reheating of the lower part of the lithosphere as it passesover a mantle thermal anomaly [Detrick and Crough, 1978; Von Herzen et al., 1982]. Replacement of the lower lithosphere by material at asthenospheric temperatures and the conductive heating of the upper lithosphere resultsin uplift. Lithospheric reheating models can fit the observed geoid and topography data and predict the subsidence after the plate passes over the hotspot. However, thesemodels do not attempt to explain the process by which the lithosphere is 18,323