Magmatic Evolution of the Coso Geothermal Area, California
Geothermal energy in the Coso field owes its origin to basaltic magmatism. Volcanism commenced ~3.5 Ma ago, coincident with a widespread Pliocene outburst in eastern California. Although most basalts associated with this event are highly potassic, those at Coso are not. Pliocene volcanic rocks at Coso (erupted between 3.5-2 Ma) range from basalt to rhyodacite, show abundant petrographic evidence for open-system behavior (e.g., quartz xenocrysts in basalts), and have compositions consistent with mixing. In contrast, Pleistocene rocks, erupted <1 Ma ago, comprise a strongly bimodal suite of mildly alkalic basalt and high-silica rhyolite. Pleistocene basalts differ from their Pliocene counterparts in generally having more depleted 87Sr/86Sr and εNd values (0.703, +7 vs. 0.704, +4); higher TiO2 and Nb; lower MgO; greater stalling depths in the crust. Pliocene rocks are distinctly arc-like even though they were erupted ~10 Ma after subduction ceased. In contrast, Pleistocene basalts have a distinctly OIB-like geochemical signature, with undepleted high field strength elements and plume-like radiogenic isotope ratios; these characteristics are shared with late Cenozoic basalts across the western U.S. Rare Pleistocene basalts that were erupted from within the footprint of the rhyolite field have notably high TiO2 contents (>3 wt%), similar to basalts from the Columbia River and Snake River Plain fields. Unlike Pliocene rocks, which scatter toward isotopic values of local basement with increasing SiO2, Pleistocene rhyolites generally have high and consistent εNd (+1 - +2.5). Producing this signature by AFC processes involving basalt and basement rocks requires remarkably consistent mixing and fractionation at small-volume volcanic centers separated by several km. Alternatively, high εNd values in the rhyolites could have been produced by partial melting of Pliocene basalts and andesites, which have very similar Nd isotopic compositions. Increasing εNd in silicic rocks as the geothermal production area is approached suggests that the magmatic flux is highest there even though erupted volumes are significantly larger outside the geothermal area. One scenario consistent with the above data is as follows. Post-subduction tectonic events triggered magmatism at 3.5 Ma, tapping fertile, subduction-metasomatized lithospheric mantle. Basalts stalled in and partially melted the mid-crust, generating a mixed-magma series and copious volcanism. Depletion of the mantle source by 2 Ma led to a hiatus in magmatism. A change in basalt chemistry to OIB- affinity in the last 1 Ma suggests a profound change in magma source – likely involving decompression melting of ascending asthenospheric mantle, perhaps related to lithosphere delamination. Injection of such magmas into the lower crust, would have generated rhyolites by remelting of earlier emplaced mafic bodies – imparting a juvenile isotopic signature in the late rhyolites. Precursory Pliocene magmatism is a common feature of other western U.S. geothermal areas, including Twin Peaks, The Geysers, and Long Valley.
Radioisotopic Age Constraints on Crystallization, Crystal Inheritance, and Eruption of Coso's Pleistocene Rhyolites: Tracking the Evolution of a Silicic Magma System
Radioisotopic dating at Coso provides a direct assessment of the rates at which large silicic magma reservoirs develop and whether upper crustal reservoirs remain thermally viable for protracted amounts of time. We dated a representative suite of Coso extrusions (8 units) ranging in age from ~230 ka to ~56 ka using Ar/Ar techniques, with additional analyses forthcoming. Accurate eruption ages are required to track secular geochemical and petrological changes within the magma system. We also dated zircon from a granodiorite core sample from injection well 46A-19RD and in favorable cases we were able to separate and date accessory minerals (zircon and allanite) from the crystal-poor extrusions. Application of accessory mineral dating is a robust approach for quantifying the time scales associated with physicochemical changes in magmas reservoirs. For example, the age distribution and character of zircons from Coso's ~600 ka Devils Kitchen rhyolite led Miller and Wooden (2004) to suggest that it was emplaced in the crust as a near-solidus crystal mush over an ~200 ka period prior to eruption. Our U-Th ages for zircon and allanite obtained by ion microprobe analysis, when compared to their respective ~115 ka (Dome 24) and ~85 ka (Cactus Peak and Sugarloaf Mountain) Ar/Ar ages, yield near-eruption ages and little evidence of recycled zircon (or allanite) from their older Pleistocene predecessors. Likewise zircon ages in the granodiorite exhibit evidence for a single crystallization event at ~200 ka. However, U-Pb dating of a significant subset of zircon in Coso's ~85 ka rhyolites and in the granodiorite core yield Mesozoic ages between ca. 100 and 200 Ma. The latter includes grains that previously yielded U-Th isotope ages within error of secular equilibrium. Two end-member cases may account for the bimodal distribution of zircon ages and evidence for assimilation and crustal contamination of low temperature (≤800 °C) rhyolite, as well as secular changes in their trace element compositions. Post ~230 ka extrusions may have tapped a persistent and integrated magma chamber. If a single long-lived magma reservoir applies to Pleistocene Coso then trace element variations (e.g., La/Nd) track its chemical differentiation. Alternatively trace element changes may reflect separate magma generation events with distinct source components. The observed decrease in the La/Nd ratio of post ~115 ka extrusions, which is often caused by fractionation of allanite or monazite, can be used as a monitor of fractional crystallization differentiation. However this trend is also consistent with auto-assimilation of highly evolved intrusions. Increases in the La/Nd ratio in Coso's ~230 ka to ~115 ka rhyolites likely reflect an episode of magma replenishment by less evolved melt, but could be due to incorporation of allanite-rich cumulate materials. Fluctuating La/Nd ratios recorded by the succession of Coso rhyolite extrusions, therefore, either represents: (1) a single long-lived reservoir that incorporated and/ or was rejuvenated by a less evolved component starting at ~230 ka, then was contaminated by a more evolved component by ~115 ka, and finally incorporated Mesozoic crust, as indicated by the xenocryst ages showing up in the ~85 ka rhyolites, or (2) the existence of a number of distinct magma bodies that formed and evolved more or less independently.
Regional Tectonics of the Coso Geothermal Area: Three-dimensional Vp and Vp/Vs Models, Spatial-Temporal Seismicity Patterns, and Seismogenic Deformation Along the Intracontinental Plate Boundary in Central Eastern California
We use regional earthquake data to synthesize the tectonics of the southern Walker Lane belt and Coso Range in central eastern California. We calculated three-dimensional models of the Vp and Vp/Vs structure of the upper and middle crust. Using these models, we also determine three-dimensional Vs and Poisson's ratio models. The changes in seismic velocities across the region are small, except for low velocities in sedimentary basins and a ~2-km positive elevation of the basement velocities (Vp > 6 km/s) beneath the southern Sierra Nevada. Localized low-Vp and low-Vs zones beneath the central Coso Range image a geothermal reservoir at 0- to 3-km depth, as well as distinct low-velocity anomalies in the depth range of ~8 to ~12 km. Because the Vp/Vs has average crustal values within this broader zone, we interpret the anomaly to indicate a zone of few percent geothermal brines extending from 8 to 12 km depth. In addition, an embedded highly localized poorly resolved zone (possibly as small as 1 km3) of slightly above average Vp/Vs and higher Poisson's ratio is a tentative suggestion of a small volume percent of magma present at depth of ~10 km. We also relocated the seismicity in the Coso region using absolute traveltimes and differential traveltimes determined from waveform cross correlation. The relocated seismicity forms several spatially clustered lineaments along the southeast side of the Sierra Nevada and in the Indian Wells Valley and vicinity of the Coso geothermal field, which coincide with mapped late Quaternary faults in the region. The base of seismicity shallows from a regional depth of about ~11 to ~5 km beneath the central Coso Range, which we interpret as evidence for shallowing of the brittle-ductile transition zone beneath the geothermal field. In addition to abundant background seismicity, two large earthquake swarms, located 5 to 8 km to the west of Coso, occurred in April to May 1992 and May to June 2001. Two dual main shock-aftershock sequences also occurred as follows: the 1994 sequence near Ridgecrest and the later Coso earthquake sequence from late 1996 to early 1998, with the pairs of main shocks spaced 47 days and 16 months apart, respectively. Kinematic analysis of the focal mechanisms indicates that the crustal stress loading process varies across the region. The low-Vp anomaly, abundant seismicity, and crustal thinning provide quantitative evidence for the Coso region being an extensional releasing step over between two northwest-striking dextral faults: The Airport Lake fault zone to the south, and the Owens Valley fault to the north.
Effects of A Weak Crustal Layer in a Transtensional Pull-Apart Basin: Results from a Scaled Physical Modeling Study
Results of scaled physical models of a releasing bend in the transtensional, dextral strike-slip Coso geothermal system located in the southwest Basin and Range, U.S.A., are instructive for understanding crustal thinning and heat flow in such settings. The basic geometry of the Coso system has been approximated to a 30? dextral releasing stepover. Twenty-four model runs were made representing successive structural iterations that attempted to replicate geologic structures found in the field. The presence of a shallow brittle-ductile transition in the field known from a well-documented seismic-aseismic boundary, was accommodated by inclusion of layers of silicone polymer in the models. A single polymer layer models a conservative brittle-ductile transition in the Coso area at a depth of 6 km. Dual polymer layers impose a local elevation of the brittle-ductile transition to a depth of 4 km. The best match to known geologic structures was achieved with a double layer of silicone polymers with an overlying layer of 100 µm silica sand, a 5° oblique divergent motion across the master strike-slip faults, and a thin-sheet basal rubber décollement. Variation in the relative displacement of the two base plates resulted in some switching in basin symmetry, but the primary structural features remained essentially the same. Although classic, basin-bounding sidewall fault structures found in all pull-apart basin analog models formed in our models, there were also atypical complex intra-basin horst structures that formed where the cross-basin fault zone is situated. These horsts are flanked by deep sedimentary basins that were the locus of maximum crustal thinning accomplished via high-angle extensional and oblique-extensional faults that become progressively more listric with depth as the brittle-ductile transition was approached. Crustal thinning was as much as 50% of the original model depth in dual polymer models. The weak layer at the base of the upper crust appears to focus brittle deformation and facilitate formation of listric normal faults. The implications of these modeling efforts are that: 1) Releasing stepovers that have associated weak upper crust will undergo a more rapid rate of crustal thinning due to the strain focusing effect of this ductile layer; 2) The origin of listric normal faults in these analog models is related to the presence of the weak, ductile layer; and, 3) Due to high dilatency related to major intra-basin extension these stepover structures can be the loci for high heat flow.
Time-dependent Seismic Tomography of the Coso Geothermal Area, 1996-2006
Measurements of temporal changes in Earth structure are commonly determined using local earthquake tomography computer programs that invert multiple seismic-wave arrival time data sets separately and assume that any differences in the structural results arise from real temporal variations. Such an assumption is dangerous because the results of repeated tomography experiments would differ even if the structure did not change, simply because of variation in the seismic ray distribution caused by the natural variation in earthquake locations. Even if the source locations did not change (if only explosion data were used, for example), derived structures would inevitably differ because of observational errors. A much better approach is to invert multiple data sets simultaneously, which makes it possible to determine what changes are truly required by the data. This problem is similar to that of seeking models consistent with initial assumptions, and methods similar to "damped least squares" can solve it. We are developing such a program, dtomo. This program inverts multiple epochs of arrival-time measurements for hypocentral parameters and structural change in the inter-epoch period. We are applying this work to data from the seismically active Coso geothermal area, California. The permanent network operated there by the US Navy, supplemented by temporary stations, provides excellent earthquake arrival-time data. Furthermore, structural change is expected in the area as a result of geothermal exploitation of the resource. We have studied the period 1996 through 2006. Our results show, for a 2-km horizontal grid spacing, an irregular strengthening with time of a negative Vp/Vs anomaly in approximately the upper 2 km of the reservoir. This progressive reduction in Vp/Vs results predominately from a increase of Vs with respect to Vp. Such a change is expected to result from effects of geothermal operations such as decreasing fluid pressure and the drying of argillaceous minerals such as illite. http://cosomeq.wr.usgs.gov
Development of Genetic Occurrence Models for Geothermal Prospecting
Exploration for utility-grade geothermal resources has mostly relied on identifying obvious surface manifestations of possible geothermal activity, e.g., locating and working near steaming ground or hot springs. This approach has lead to the development of over 130 resources worldwide, but geothermal exploration done in this manner is akin to locating hydrocarbon plays by searching for oil seeps. Confining exploration to areas with such features will clearly not discover a blind resource, that is, one that does not have surface expression. Blind resources, however, constitute the vast majority of hydrocarbon plays; this may be the case for geothermal resources as well. We propose a geothermal exploration strategy for finding blind systems that is based on an understanding of the geologic processes that transfer heat from the mantle to the upper crust and foster the conditions for hydrothermal circulation or enhanced geothermal exploration. The strategy employs a genetically based screening protocol to assess potential geothermal sites. The approach starts at the plate boundary scale and progressively focuses in on the scale of a producing electrical-grade field. Any active margin or hot spot is a potential location for geothermal resources. Although Quaternary igneous activity provides a clear indication of active advection of hot material into the upper crust, it is not sufficient to guarantee a potential utility-grade resource. Active faulting and/or evidence of high strain rates appear to be the critical features associated with areas of utility-grade geothermal potential. This is because deformation on its own can advect sufficient heat into the upper crust to create conditions favorable for geothermal exploitation. In addition, active deformation is required to demonstrate that open pathways for circulation of geothermal fluids are present and/or can be maintained. The last step in the screening protocol is to identify any evidence of geothermal activity, including high heat flow, anomalous temperature water wells, high-temperature indications from aqueous geothermometry and geochemistry, Pliocene or younger ages from low-temperature thermochronometers, as well as more obvious factors such as geysers and fumaroles (which by definition will be missing for blind resources). Our occurrence-model strategy inverts the current approach that relies first on obvious evidence of geothermal activity. We evaluated our approach by retrospectively applying the protocol to the characteristics of producing geothermal fields, and in all cases, known resource areas fit the parameters identified from a genetic perspective.
The geo-scientific basis for the geothermal evolution in Iceland
More than half of the primary energy use in Iceland is economically produced from geothermal resources. The main reasons for this unique success in Iceland are favourable geological conditions and highly developed technology in geosciences and engineering. Iceland is a sub-aerial part of the ocean floor, located where the central axis of the Mid-Atlantic Ridge intersect the Iceland hot-spot, resulting in abnormal crustal thickness and complicated tectonic patterns. The ridge axis crosses the island from South-West to North East forming a volcanic rift zones that is characterized by many active central volcanoes and associated high temperature geothermal fields (T more than 200°C at 1 km depth). The rift zone is highly faulted and the uppermost 1 km is composed of permeable young basaltic material. Outside the volcanic zone the crust is normally made of altered basaltic lavas of low primary permeability due to secondary mineralization. However, recent tectonic activity, probably due to glacial rebound and relative movement of the ridge axis and the hot spot, has formed permeable fractures that are pathways for geothermal fluid and result in numerous low temperature geothermal fields ( T less than 150°C at 1km depth). The background heat flow in Iceland varies with age from 70 to 250 mW/m2 and the crustal thickness varies from 20 to nearly 40 km. Geothermal exploration is done with a multidisciplinary approach where geological mapping, geochemistry and geophysics interact. The geological mapping with emphasis on tectonic structure, stratigraphy, hydrothermal alteration and eruption history is usually the first step. If hot springs or fumaroles exist, chemical methods are used to predict the reservoir temperature and the fluid properties prior to drilling. Geophysical surveys are the most widely used methods to detect subsurface high temperature fields and to estimate their size and properties. Resistivity soundings, mainly based on TEM and MT measurements, play the key role, but analysis of natural seismic events, aeromagnetic and gravity surveys are also helpful. In case of exploring for low temperature fields heat flow measurements are also important as well as geophysical methods to detect water-bearing fractures. The exploratory work leads to a conceptual model of the geothermal field. During the exploratory drilling phase borehole geology supported by geophysical well logging are the main tasks, which together with results from well testing make the basis for a revised geological and hydrological model of the reservoir. During exploitation geo-scientific research is still important to understand the response of the geothermal field to exploitation. This includes monitoring of temperature, pressure, induced seismicity as well as changes in fluid chemistry, geodesy and gravity.
Seismicity Induced by Water Injection for Geothermal Reservoir Stimulation 5 km Below the City of Basel, Switzerland
To stimulate the reservoir for a proposed "hot dry rock" geothermal project in the city of Basel, approximately