Abstract:
Disposal of hydraulic fracturing flowback and produced water into Ordovician and Cambrian formations of the Fort Worth Basin (FWB), coupled with an increase in observed seismicity in the Dallas-Fort Worth area, necessitates an understanding of the geologic character of these disposal targets. More than 2 billion barrels (Bbbls) of wastewater have been disposed into the Ordovician Ellenburger Group of the FWB over the past 35 years. Since the implementation of the TexNet Earthquake Catalog (1 January 2017), more than 20 earthquakes of local magnitude ML2.0 or greater have been detected in the area, with depths ranging from 2 to 10 km (approximately 6500–33,000 ft). The cited mechanism for inducement of these earthquakes is reactivation of basement faults due to pore pressure changes, either directly related to proximal disposal or due to disposal volume buildup over time. Here, we present a stratigraphic and petrophysical analysis of FWB disposal targets and their relation to basement rocks. The Ellenburger consists of alternating layers of limestone and dolomite, with minor siliciclastics above the basement toward the Llano Uplift. Matrix porosity averages <5 porosity units (p.u.), with higher porosity in dolomitic layers than in limestone. Dolomite dominates at the top of the Ellenburger, which was exposed at the end of both the Lower and Upper Ordovician. Where crystalline basement rocks are penetrated, the composition ranges from granitic to chlorite-bearing metamorphosed lithology. The basement-sediment interface is frequently marked by increased porosity. An updated map of structure on top of basement indicates elevations ranging from outcrop at the Llano Uplift to more than −12;200 ft (−3.7km) subsea toward the northeast. The disposal zone pore volume is estimated from thickness and porosity maps and ranges from <0.1 to >0.60 billion barrels per square mile (Bbbl∕mi2). Introduction In the Fort Worth Basin (FWB), Texas, flowback and produced water associated with Barnett Shale gas production is disposed into the underlying Ellenburger Group and has been linked to increased seismic activity since 2008 (Frohlich, 2012; Hornbach et al., 2015, 2016; Frohlich et al., 2016; Scales et al., 2017). The mechanism linking disposal and induced seismicity is based on the hydraulic connectivity of an overpressured disposal formation and the seismogenic basement (e.g., Zhang et al., 2013; Frohlich et al., 2014; Hornbach et al., 2015; Walsh and Zoback, 2015; Scales et al., 2017; Hincks et al., 2018). From 2000 to 2017, more than 2 Bbbls of saltwater have been disposed into the locally unproductive Cambrian-Ordovician-aged formations of the FWB, including the Ellenburger Group, via 166 disposal wells (Figure 1a). These formations are the primary disposal targets, with increased disposal volumes after 2008 (Figure 1b), spatially and temporally coincident with Barnett Shale production. Disposal well completion methods are plug and perf (50%), openhole (48%), and existing open-hole zones with new perforations (2%); Cambrian-Ordovician disposal depths range from Viola-Simpson to near the top of the basement. The stratigraphic architecture and rock properties of the disposal intervals, and their relation to basement rocks, are key in understanding the disposal reservoir, the flow of injected fluid, and the potential for induced seismicity (e.g., National Research Council, 2013; Zhang et al., 2013; Shah and Keller, 2017; Hincks et al., 2018). This type of geologic analysis is integral to any attempt to model not only historical disposal and induced seismicity but also to predict areas of concern based on potential pore pressure increases and reactivation of basement faults. Several outcrop studies have been undertaken on the Ordovician Ellenburger Group and Cambrian Moore Hollow Group near the Llano Uplift (e.g., Cloud et al., 1945; Crowley and Hendricks, 1945; Cloud and Barnes, University of Texas at Austin, Bureau of Economic Geology, Austin, Texas, USA. E-mail: katie.smye@beg.utexas.edu (corresponding author); casee.lemons@beg.utexas.edu; ray.eastwood@beg.utexas.edu; guin.mcdaid@beg.utexas.edu; peter.hennings@beg.utexas.edu. Manuscript received by the Editor 26 October 2018; revised manuscript received 11 February 2019; published ahead of production 06 August 2019; published online 15 October 2019. This paper appears in Interpretation, Vol. 7, No. 4 (November 2019); p. SL1–SL17, 13 FIGS., 1 TABLE. http://dx.doi.org/10.1190/INT-2018-0195.1. © 2019 Society of Exploration Geophysicists and American Association of Petroleum Geologists. All rights reserved. t Special section: Wastewater and CO2 injection Interpretation / November 2019 SL1 D ow nl oa de d 10 /2 2/ 19 to 1 28 .6 2. 55 .1 5. R ed is tr ib ut io n su bj ec t t o SE G li ce ns e or c op yr ig ht ; s ee T er m s of U se a t h ttp :// lib ra ry .s eg .o rg / 1948; Hendricks, 1952; Barnes et al., 1959; Barnes and Bell, 1977; synthesized in Wright, 1962, 1979; Bradfield, 1964; Hendricks, 1964; Watson, 1980; Collier, 1983), with limited attempts to correlate formations in the subsurface based on physical properties. Similarly, subdivision and correlation of units within the Ellenburger of the Delaware and Val Verde Basins has proven difficult (Ijirigho, 1981). This paper contains a geologic characterization of Ordovician and Cambrian formations used for fluid disposal in the FWB, as well as an understanding of the basement-sediment interface and depth-to-basement. Interpretations are based on stratigraphic and petrophysical analyses of wireline well logs. We show that the Ellenburger of the FWB consists of alternating layers of limestone and dolomite, with minor porous siliciclastics at the base of the section toward the Llano Uplift. Due to uplift and erosion, the uppermost Ellenburger is only observed in the subsurface of the FWB and has a high dolomite fraction with increased porosity. The sediment-basement interface contains granite wash in some wells; elsewhere, carbonates directly overlie basement rocks. The lithology of the basement ranges from granitic to metamorphic composition. These findings provide an understanding of the geology of the disposal formations in the FWB, including their stratigraphic architecture and petrophysical properties. The characterization of properties that influence flow, such as porosity, and their facies associations, lateral continuity, and geometry, provides needed geologic context for the flow of injected fluid and the potential for induced seismicity. Geologic background The FWB is an asymmetric, north–south elongated basin bounded by structural features of the Ouachita thrust front to the east, Muenster and Red River arches to the north, the Llano Uplift to the south, and the Bend Arch to the west. It is one of several foreland basins, including the Appalachian, Val Verde, and Anadarko, which formed during the Paleozoic in front of the Ouachita-Allegheny-Marathon Foldbelt. The basin contains up to 12,000 ft (3.6 km) of preserved sediment fill (Walper, 1982; Montgomery et al., 2005), including the Mississippian-age Barnett Shale, which has been widely targeted for natural gas production. Most of the basement of the FWB is part of the Texas Craton, consisting of plutons — predominantly granite and diorite — emplaced into metasedimentary hornblende and biotite-schist, gneiss, and marble. Plutonic rocks make up most of the Texas Craton, with metasedimentary rocks of secondary importance (Flawn, 1956). In the subsurface, basement lithology has been interpreted through gravimetric anomalies (e.g., Olorunsola et al., 2015). The Abilene Gravity Minimum in the western FWB has been interpreted to reflect a granitic batholith 4–16 km (approximately 13,000– 20,000 ft) thick that probably represents a Middle Proterozoic continental margin arc batholith, such as the Sierra Nevada, with an age of 1.4–1.34 Ga (Adams and Keller, 1996). The Precambrian basement surface was exposed and eroded for more than 500 million years (Clabaugh and McGehee, 1962), and exhibited local relief of up to 800 ft (240 m) (Barnes and Bell, 1977). Initiation of a Wilson Cycle — opening and subsequent closing of Figure 1. (a) Distribution of Cambrian-Ordovician saltwater disposal (SWD) wells (white dots) within the core Barnett producing area (dashed red line) and greater FWB study area (black line), along with cumulative Cambrian-Ordovician SWD volumes for each 100 mi2 area and earthquake locations (pink dots). (b) Distribution of SWD volume and monthly count of active wells with and earthquake activity (pink dots). Earthquakes were identified by combining the SMU Earthquake Catalog (Quinones et al., 2019) and the USGS Earthquake Catalog (USGS, 2018). SL2 Interpretation / November 2019 D ow nl oa de d 10 /2 2/ 19 to 1 28 .6 2. 55 .1 5. R ed is tr ib ut io n su bj ec t t o SE G li ce ns e or c op yr ig ht ; s ee T er m s of U se a t h ttp :// lib ra ry .s eg .o rg / an oceanic basin — led to formation of the protoAtlantic (Iapetus) Ocean in this region. The ancestral divergent plate margin is evidenced by the Delaware, Southern Oklahoma, and Reelfoot Aulacogens (Walper, 1977; Adams and Keller, 1996). Deposition of siliciclastics, shelf facies (carbonates), and deeper Ouachita basinal facies occurred throughout the Late Cambrian and Early Ordovician (Figure 2). The extent of siliciclastic deposition during Cambrian sea transgression is unclear, with some studies (e.g., Barnes et al., 1959) suggesting that the Hickory — the basal member of the Riley Formation (Figure 3) — laps out northeast of the Llano uplift, and other studies (e.g., Bradfield, 1964) hypothesizing that it extends into the FWB. Barnes and Bell (1977) suggest pinchout of sandstones away from the Llano Uplift, with the zero thickness line extending from Shackleford to Eastland and Erath Counties. Local thickness variations are related to Precambrian basement topography. Although Cambrian siliciclastics are
