Anthropogenic subsidence in the Mexicali Valley, Baja California, Mexico, and slip on the Saltillo fault
- First Online:
- Cite this article as:
- Glowacka, E., Sarychikhina, O., Suárez, F. et al. Environ Earth Sci (2010) 59: 1515. doi:10.1007/s12665-009-0137-y
- 159 Views
Deep fluid extraction in the Cerro Prieto geothermal field (CPGF) has caused subsidence and induced slip on tectonic faults in the Mexicali Valley (Baja California, Mexico). The Mexicali Valley is located in the southern part of the Salton Trough, at the boundary between the Pacific and North American plates. The Valley is characterized by being a zone of continuous tectonic deformation, geothermal activity, and seismicity. Within the Cerro Prieto pull-apart basin, seismicity is concentrated mainly in swarms, while strong earthquakes have occurred in the Imperial and Cerro Prieto transform faults, that are the eastern and western bound of the basin. Since 1973, fluid extraction at the CPGF has influenced deformation in the area, accelerating the subsidence and causing rupture (frequently as vertical slip or creep) on the surface traces of tectonic faults. Both subsidence and fault slip are causing damage to infrastructure like roads, railroad tracks, irrigation channels, and agricultural fields. Currently, accelerated extraction in the eastern part of CPGF has shifted eastwards the area of most pronounced subsidence rate; this accelerated subsidence can be observed at the Saltillo fault, a southern branch of the Imperial fault in the Mexicali Valley. Published leveling data, together with field data from geological surveys, geotechnical instruments, and new InSAR images were used to model the observed deformation in the area in terms of fluid extraction. Since the electricity production in the CPGF is an indispensable part of Baja California economy, extraction is sure to continue and may probably increase, so that the problem of damages caused by subsidence will likely increase in the future.
KeywordsCerro PrietoGeothermal fieldFault creepCreep eventsSubsidence modeling
Fluid extraction for electricity production began in the Cerro Prieto geothermal field (CPGF) in 1973 from depths of 1,500–3,000 m. Re-injection of residual water began in 1989, and currently about of 20% of the extracted water is being reinjected at 500–2,600 m depths. The geothermal field is operated by the Mexican Federal Electricity Commission (Comisión Federal de Electricidad, CFE). At present, the CPGF has a 720 MW capacity and is the world’s second largest geothermal field. Cerro Prieto is a high-temperature, fluid-dominated field, contained within sedimentary rocks. The geothermal fluid, with a temperature of 260–350°C, is extracted from gray shales, and is isolated from the unconsolidated rock by a layers of mudstone and brown shales which constitute the cap-rock (CFE 2006). The extracted volume increased from 14 million tons in 1974 to 72 million tons in 2005.
The CPGF reservoir is characterized by having “leaky” boundaries, where the hot geothermal fluids exist in dynamic equilibrium with much cooler waters (Truesdell and Lippmann 1990). It is generally accepted that the field is recharged from two sources, a deep one, to the east, that supplies hot water, and a second source which contributes cold water from shallow aquifers, located to the east, west, and south, fed by the Colorado River (Truesdell et al. 1998).
From the analysis of the leveling surveys done in the area of CPGF and the Mexicali Valley since 1977, Glowacka et al. (1999) noticed that, for the period 1977–1997, the subsidence rate at the center of the field increased after each steam extraction increase. The area of maximum subsidence rate, 12 cm/year, coincided with that where the extraction wells were located, and this subsidence rate was too high to be caused by tectonic activity alone. These facts suggested that subsidence in the CPGF was caused mainly by fluid extraction. Subsidence at CPGF was also measured by Synthetic Aperture Radar (SAR) interferometry by Carnec and Fabriol (1999) and Hanssen (2001), using ERS1/2 images acquired in 1993–1997 and 1995–1997 periods, respectively, and was interpreted as being produced by geothermal fluid extraction.
Surface deformation is an expected consequence of the extraction of geothermal fluids. Surface subsidence rates of dozens of centimeters per year have been measured across several major geothermal fields, e.g., Wairakei, New Zealand (Allis et al. 1998) and Geysers, USA (Mossop and Segall 1997).
To quantitatively evaluate the tectonic component of the subsidence recorded in the Mexicali Valley, Sarychikhina (2003) and Glowacka et al. (2005) modeled the elastic deformation caused in the pull-apart basin by the tectonic extension between the Imperial and Cerro Prieto strike slip, right lateral faults, using slip rates determined by GPS measurements of over the plate boundary (Bennett et al. 1996). This modeling estimated a maximum subsidence rate of 0.45 cm/year, over an area coinciding with that of the CPGF, and taking into account compaction and isostasy, this rate could attain ~0.60 cm/year. The estimated tectonic subsidence is of the order of those calculated for nearby areas using other methods: 0.16 cm/year for Laguna Salada, 20 km west of Mexicali (Contreras et al. 2002), and 0.55 cm/year for the Vallecito–Fish Creek basin 50 km north of Mexicali (Johnson et al. 1983).
Thus, the estimated tectonic subsidence rate in the area of CPGF is of the order of only 4–5% of the total 12 cm/year measured rate; leaving some 95% to anthropogenic causes. According to Camacho Ibarra (2006), 82–90% of the observed subsidence in this region is anthropogenic.
Although the extension and the shape of subsidence zone in the Mexicali Valley suggest that these parameters define the pull-apart basin originally controlled by tectonics, the current subsidence rate is mainly related to fluid extraction.
Tectonic faults play an important role in feeding and constraining the geothermal reservoirs. The role of the faults located in the extraction zone was analyzed by Lippmann et al. (1991). According to these authors, faults H and L (Fig. 4) function as conduits for hot and cold water recharging, and as boundaries between reservoirs α, β and γ. As a consequence of fluid extraction and the corresponding pressure drop in the geothermal reservoir, the cold water from a shallow aquifer recharges the geothermal fluid (Lippmann et al. 1991).
Field observations done since the 1980s show that the observed subsidence is not continuous in space and is characterized by abrupt vertical displacements along the known traces of the Imperial and Cerro Prieto faults. Analysis of the extension, amplitude, and time behavior of the slip (relative displacement across the fault) on the Imperial and Cerro Prieto faults, which bound the Cerro Prieto pull-apart basin from the NE and the SW (Glowacka et al. 1999, 2005), suggested that these faults constitute a boundary of the subsided area, and probably are a groundwater barrier.
The objective of the present work is to use existing field (geodetic, geotectonic and geotechnical) data together with two-pass differential Interferometric Synthetic Aperture Radar (InSAR) technique to model the current spatio-temporal evolution of the surface deformation in the CPGF area.
This rapid subsidence and the associated fault slip (more than 2 m accumulated since the 1970s) in an otherwise flat region, cause damage to roads, railroad tracks and irrigation systems (both major and local irrigation channels). The step in ground level causes problems to farming fields irrigated from channels: in the subsiding side it causes a health hazard because it is not possible to drain waste agriculture water from the subsidence basin; on the upper side, it causes economic problems in farms which do not receive enough water for agricultural production. It is also a source of hazard because, in the event of even a moderate earthquake, the shaking could cause damage with resulting floods if the one of the, currently barely functional, main irrigation channels gets broken.
Some of the features observed on the surface related to the subsidence, and damage caused by subsidence to local infrastructure, have been described by González et al. (1998), Glowacka et al. (2006) and Suarez-Vidal et al. (2008). The name Imperial Fault was used for Saltillo Fault until distinction between the two branches was made by Suarez-Vidal et al. (2008).
To evaluate the elastic deformation caused by fluid extraction, when the depth of the subsidence source is large compared with the dimension of the affected area on the surface, mathematical models used for volcanology and hydrofracturing can be used. The one most commonly used is Mogi’s (1958) model of a spherical hydrostatic pressure source imbedded in an elastic half-space. This model was used by Mossop and Segall (1997) to model subsidence in the Geysers geothermal field, and by Carnec and Fabriol (1999) and Hanssen (2001) to model subsidence in the CPGF. The model of deflation of a triaxial ellipsoidal cavity in an elastic half-space (Davis 1986) was used for subsidence modeling in the Coso geothermal field (Fialko and Simons 2000). The subsidence induced by fluid extraction was evaluated by Segall (1989) using a poroelastic model of an axisymmetric reservoir. All these models have some kind of symmetry: three- or two-dimensional, which does not seem appropriate for the CPGF. Because the CPGF reservoirs are located in sedimentary layers and bounded by faults, the decision was made to use the mathematical model of rectangular tension fractures (Yang and Davis 1986) as the one which best represents the geometry of reservoirs.
Model parameters: z is the depth to the center of fracture or fault
Color (Fig. 4)
Interferometric pair used in this study
The effects of subsidence are clearly seen along the faults that bound the subsidence zone. The Colorado River was the principal source of the very high sedimentation in the Mexicali valley that kept the active faults covered, but since it was dammed in 1934, sedimentation from this source stopped (Suarez-Vidal et al. 2008) and, since then, surface traces of active faults have appeared in places when there is no agricultural activity. These surface breaks have been studied and published by González et al. (1998), Glowacka et al. (1999), and Lira (2005), and summarized recently by Suarez-Vidal et al. (2008).
Interferometric synthetic aperture radar (InSAR), is a radar technique used in geodesy and remote sensing. This geodetic method uses two or more synthetic aperture radar (SAR) images to generate maps of surface deformation or digital elevation, using differences in the phase of the waves returning to the satellite (Hanssen 2001). The technique can potentially measure centimeter-scale changes in deformation over time spans of days to years. It has applications for geophysical monitoring of natural and induced hazards, for example earthquakes, volcanoes, landslides, subsidence and structural stability.
Synthetic Aperture Radar images (single-look complex) from the European Space Agency (ESA), were used to generate interferograms, using the Doris InSAR processing software (Kampes et al. 2003) and precise satellite orbits from Delft University (Scharoo and Visser 1998). The topographic effects were removed from the interferograms, using the Shuttle Radar Topography Mission (SRTM) 3-arcsec digital elevation model. Unwrapping was performed using the SNAPHU software (Chen and Zebker 2001). Preliminary results of InSAR data analysis for the study area over the period 2004–2005 have been presented in Sarychikhina et al. (2007a, b). In this paper we present an analysis of one ENVISAT interferogram, which was selected from 20 interferograms from the period 2004–2006 based on good correlation and lack of obvious atmospheric artifacts. The selected interferogram spans a 35 day period (2004/12/19–2005/01/23) and is from descending track images (track 84, frame 2961). The spatial coverage of this interferogram is presented in Fig. 1.
The interferogram has good coherence at the CPGF zone and its surrounding area, and the deformation pattern is clearly visible. High deformation rates define a roughly elliptical area oriented NE–SW, featuring two subsidence basins. The elliptical area, and the two basins coincide with those from the 1994–1997 leveling; but whereas for the older study the highest subsidence rate (~12 cm/year), was east of the evaporation basin at the center of the CPGF, in the present study we find not only that the subsidence rate at this point has increased to ~1.2 cm/month (14 cm/year), but that the highest subsidence rate ~1.7 cm/month (20 cm/year) is found now at the other basin, located to the NE, close to the Saltillo fault in what was previously proposed as a region of recharge.
A recapitulation of facts resulting from new field, instrumental and InSAR observations is: The subsidence area is limited by faults, and this phenomenon is mainly seen along the Saltillo fault. During the last few years the subsidence rate increased over all the area, but this increase was significantly larger towards the NE, at the eastern subsidence basin, located outside of CPGF. Slip on the Saltillo fault, which corresponds to about half of the total subsidence, occurs mainly during slip events. The vertical slip rate on the fault increased from 5.3 to 7.3 cm/year around second half of 2003. We will now proceed to interpret and discuss these facts.
Three processes are probably at work to produce the observed fault behavior. According to Sheng and Helm (1998), existing faults often are poorly permeable in the direction perpendicular to the fault, and play the role of a ground water barrier, increasing differential displacement of both sides of the fault. Additionally, the process of fault breaking is accelerated by fluid extraction, when different thicknesses of sediments in both sides of the fault (as seen in Fig. 3) induce differential compaction and, hence, tensional and shear stresses on the ground mass, thus producing ground failures leading to vertical displacement on the faults which limit the affected aquifer (op cit.). Thirdly, a diminution of the normal stresses is to be expected at the fault, due to extension at the edge of a subsidence basin, which reduces the friction that locks the fault (Segall 1989).
Given the above, the second question is why the slip is episodic. The presence of displacement in form of slip events is probably related to the stick-slip mechanism during a subsiding process. A similar process was suggested, by Pennington et al. (1986), as the cause of seismicity induced by petroleum production. The presence of slip events suggests that the subsidence is a relatively deep phenomenon, and slip events are located between the depth of the seismogenic zone (high normal stresses) and the upper part of the fault, where normal stresses, caused by closing of both aquifers (L.R and s.r), are small enough to allow the slip to be continuous.
The occurrence of aseismic slip events is well documented on active, strike-slip faults in California, and has been observed and documented since the 1960s (e.g. Wesson 1988). Bilham and Behr (1992) analyzed aseismic slip phenomena on the Superstition Hills strike-slip fault, and suggested that stable sliding should extend from the surface to the zone of episodic slip, which ranges from ~300 m to ~3 km depth where the seismogenic zone begins, and the different slip behaviors are due primarily to the lithostatic increase in fault normal stress.
However, seismicity in the Mexicali valley ranges in depth between ~1 km and 15 km, with most earthquakes occurring between 4 and 6 km (Glowacka et al. 1999), near the southern part of Imperial fault (Glowacka et al. 2002) seismicity occurs below the 2 km depth. Near the southern part of the Imperial fault unconsolidated sediments reach depths of ~2.5 km west of the fault (Fig. 3). A depth of ~2.5 km agrees well with the upper limit of seismicity, while the estimated depth of the deep s.r. fracture and width of the SF fault (modeled above) suggest a slightly shallower limit of ~1.5 and ~1.1 km, respectively. Within the uncertainties in these data we speculatively propose the depth range of slip events to be from ~1 to ~2.5 km.
As mentioned above, the main difference between the subsidences measured using InSAR for 2004–2005, and those from the 1994 to 1997 leveling, is that the maximum subsidence rate shifted eastwards from the older part of the CPGF towards the edge of the basin, close to the Saltillo fault. Similar results were obtained for different pairs of SAR images from 2004 to 2005, and were attributed to changes in the spatial distribution of steam extraction (Sarychikhina et al. 2007a, b). In 2000 a new power plant began operating in the newest, eastern part of the CPGF, named Cerro Prieto IV (CPIV), shown in Fig. 2, extracting steam from a depth of around 3.5 km (Ocampo-Diaz 2005).
Talwani and Acree (1984/1985) proposed that seismicity induced by dam filling and fluid injection occurs with a time lag related to the distance from the source of the pressure change, by a process of pore pressure diffusion, with a “seismic” hydraulic diffusivity in the range 5 × 103–5 × 105 cm2/s. To explore whether the subsidence rate change recorded by the crackmeter on the Saltillo fault, that began around mid 2003, could be attributable to the change in production location, we calculated the value of hydraulic diffusivity that could account for the time delay of about 3 years, given that the Saltillo fault is about 5 km to the east of CPIV. We obtained a hydraulic diffusivity of the order of 104 cm2/s, close to the 3 × 104–3 × 105 cm2/s, determined by Glowacka (1996) for seismicity related to the fluid extraction (pressure decrease) in the CPGF in the same area.
Data from leveling and geological surveys, geotechnical instruments and satellite observations, together with model estimates on the active tectonics of the area, indicate that fluid extraction in the CPGF is the major cause of subsidence in the area and this subsidence can influence vertical displacement on the Saltillo fault through several mechanisms.
The distance and time relation between the extraction changes in the CPGF and displacement rate changes on the Saltillo fault suggests that the fault is affected through diffusive transmission of pore pressure changes with a characteristic hydraulic diffusivity.
Since the electricity production at the CPGF is an indispensable part of Baja California economy, economic pressures will surely result in future increased extraction and consequently, increased subsidence with a higher rate and more extended spatial area.
The results and conclusions presented in this work are solely the author’s and do not necessarily express the points of view of CFE or CICESE. This work greatly benefited from comments done by two anonymous reviewers. This research was sponsored in part by CONACYT, project number 45997-F and CICESE internal founds. O. Sarychikhina benefited from the scholarship from Secretaria de Relaciones Exteriores de México, and aditional scholarships from CONACYT and CICESE. The European Space Agency’s (ESA) ENVISAT satellite has been used to collect the interferometric data. The data were obtained as part of ESA Cat-1 Project (ID - C1P3508).