Seismic Delineation of the Prairie Evaporite Dissolution Edge in South-central Saskatchewan

H. Hamid 1, I.B. Morozov 1, and L.K. Kreis

Hamid, H., Morozov, I.B., and Kreis, L.K. (2005): Seismic delineation of the Prairie Evaporite dissolution edge in south-central Saskatchewan; in Summary of Investigations 2005, Volume 1, Saskatchewan Geological Survey, Sask. Industry Resources, Misc. Rep. 2005-4.1, CD-ROM, Paper A-8, 11p.

Abstract Approximately 330 km of 2-D seismic data were integrated with well log information to improve the delineation of the southern margin of the Middle Devonian Prairie Evaporite in Saskatchewan. Thirteen seismic lines were re- processed with an emphasis on enhancing high-frequency imaging. The resulting seismic sections show marked improvement in the accuracy and quality of subsurface mapping of the Prairie Evaporite salt edges. Seismic data indicate that salt dissolution structures were created by multistage processes. Thickening of overlying strata related to salt dissolution was observed within both salt-free areas and areas of preserved Prairie Evaporite. Well-log data were combined with seismic results and gridded to create an updated map of the Prairie Evaporite. Different gridding methods provided different interpolations of the data set, especially where the salt layer is thin near its margin. Comparisons with seismic interpretations show that interpolation of well data alone using different interpolation techniques can result in shifts in the delineated position of the salt edges of about 2 to 9 km. Therefore, integration of the seismic and well log data should increase the accuracy of delineating the salt edge.

An attempt was also made to determine whether the effect of the salt edge could be observed in gravity data. An approximately 33 km long gravity profile close to the seismic line NOR-83314 across a known salt collapse was extracted from the national gravity data base. A 2-D model was designed based on the interpretation of seismic line NOR-83314 and well log data. Gravity modeling shows that the salt collapse contributes ~0.4 mgal to the total anomaly (4 mgal). We suggest that performing a high-resolution gravity survey with a station interval of about 100 m might still be useful to constrain the overburden and help detect salt collapses.

Keywords: collapse structures, Devonian, gravity, Prairie Evaporite, salt, salt dissolution, seismic, subsurface mapping, well log, Williston Basin.

1. Introduction Deposits of the Elk Point and Manitoba Groups contain the largest amount of salt in the Western Canada Sedimentary Basin with an approximate volume of salt of about one million cubic kilometres (Zharkov, 1988). Most of the salt is contained within four formations: 1) Lotsberg, 2) Cold Lake, 3) Prairie Evaporite, and 4) Dawson Bay (Meijer Drees, 1986). The Middle Devonian Prairie Evaporite is the most widespread of these deposits and underlies much of southern Saskatchewan (Holter, 1969). Baillie (1953) introduced the term “Prairie Evaporite” for the evaporites found between the Winnipegosis and the Second Red Bed of the Dawson Bay Formation. Subsurface dissolution of the Prairie Evaporite has been reported for over 50 years (Sloss and Laird, 1947; Baillie, 1953; Bishop, 1954). In parts of the study area, located south and south-southeast of Regina (Figure 1), the Prairie Evaporite has been dissolved and removed either partially or completely by groundwater (Holter, 1969). Salt appears to have been dissolved from either the top or bottom of the formation and sometimes from both (Gendzwill and Martin, 1996). The Prairie Evaporite Formation is among the key economic targets in southeastern Saskatchewan. Three main potash members have been recognized within the upper part of the Prairie Evaporite: Esterhazy, Belle Plaine, and Patience Lake (Holter, 1969). Knowledge of the locations and dimensions of salt collapse structures is essential for avoiding mining into hazardous zones and for evaluating hydrocarbon exploration potential. Salt dissolution and the resulting subsidence are important hydrocarbon trapping mechanisms in western Canada (Edmunds, 1980). In the study area, collapse structures may be multistage and located off-salt where all salt has been removed (e.g., the Hummingbird Structure) or on-salt within the Prairie Evaporite itself, such as the Kisbey Structure (Figure 1;

1 University of Saskatchewan, Department of Geological Sciences, 114 Science Place, Saskatoon, SK S7N 5E2; E-mail: haitham.hamid @usask.ca.

Saskatchewan Geological Survey 1 Summary of Investigations 2005, Volume 1 Halabura, 1998). Accurate Regina mapping of the salt margin Sawatzky et al. (1960) contributes to the understanding 15 of the process of salt dissolution, Holter (1969) basement control, and their impact on hydrocarbon Sask. Dept. of Mineral Resources (1961) production. Structures created by SE-84506 salt collapse may influence fluid Sa lt edge interpreted migration, reservoir by this study SE-84524 enhancement, and oil SE-84529 10 NOR-93WN3 entrapment. Knowledge of their Kreis et al. (2003) distribution and nature therefore PCR-11CBY-5W - Weyburn Kisbey structure aids hydrocarbon exploration. CBY-7W NOR-83314 The southern edge of salt beds

S-793 was defined from earlier 5 composite seismic anomaly maps - 9-NS summarized by Holter (1969). R-19 The precise location of the Prairie Evaporite edge is Hummingbird uncertain as it is irregular and Structure CBS-8 SE-850844 1 complex. The Prairie Evaporite 20 15 10 5 ranges in thickness from 0 to 220 m in the central part of the Figure 1 - Study area in south-central Saskatchewan. The seismic lines used in this Williston Basin. In this study, study are shown in blue; those in red were analyzed in the International Energy which extends the results of Agency Weyburn CO2 Monitoring and Storage Project (Hajnal, pers. comm., 2003). Labelled green and black lines indicate the positions of the Prairie Evaporite edge Hamid et al. (2004), additional from previous studies, and the brown line shows the edge interpreted in the present seismic and well-log data are study. The brown polygons in southern and northeastern areas indicate local salt used to improve the accuracy of dissolution. Note the differences in location of the salt-dissolution edge between subsurface mapping of the previous mapping attempts and that of the present study. Numbers along the edges of southern edge of the Prairie the plot indicate townships and ranges. Evaporite Formation. The potential association of salt dissolution edges and collapse structures with basement faulting is also investigated.

2. Objectives The primary objective of this study was to improve delineation of the dissolution edge along the southeastern edge of the Prairie Evaporite Formation (Figure 1). In its utilization and interpretation of seismic and well datasets, the project ties in with regional 2-D seismic studies (Hajnal, pers. comm., 2003) and subsurface geological mapping (Kreis et al., 2003) conducted as part of the IEA Weyburn CO2 Monitoring and Storage Project. Additional objectives were correlation of areas of salt dissolution with underlying structural features such as basement highs, Winnipegosis mounds, and faults, and to document their potential inter-relationships. Specifically, we aimed to:

1) use the available 2-D seismic data acquired by the oil industry to improve delineation of the southeastern margin of the Prairie Evaporite Formation south of Regina; 2) investigate and develop processing and interpretation techniques to help identify thin salt beds and salt collapses near the dissolution edge and seismically evaluate the underlying strata, with particular attention paid to the Precambrian basement, in an attempt to recognize structural features which may have influenced the location of the present-day salt edge; 3) evaluate the effects of different mapping (spatial interpolation) techniques at locations of salt edges; and 4) investigate the potential use of gravity inversion for delineation of salt edges.

3. Seismic Data The approximately 330 km of seismic reflection lines used in this study were surveyed between 1979 and 1984, and were donated to us by Encana Corporation, Petro-Canada, Olympic Seismic Ltd., and Kary Data Consultants Ltd. Thirteen seismic lines were shot using different recording systems and a variety of dynamite and air-gun sources.

Saskatchewan Geological Survey 2 Summary of Investigations 2005, Volume 1 Spread lengths extended mostly from 1.5 to 3 km. Station intervals ranged from 25 to 67 m, and shot intervals from 75 to 200 m. Record lengths were 3 s at 2 ms sampling intervals. a) Seismic-data Processing Field data were completely re-processed from digital raw field records using PROMAX software (®Landmark Graphics). In the study area, the depth of the Prairie Evaporite ranges from 1600 to 2500 m, and the thickness from 0 to 220 m. At these depths, the higher frequencies are significantly attenuated because of the absorption and scattering effects of the earth, and are also contaminated by noise. Therefore, the primary goal of re-processing seismic data was to recover sufficiently high seismic frequencies to identify thin beds and map the salt dissolution edge of the Prairie Evaporite. In order to increase the resolution and consistency of seismic imaging, all data were re-processed in a uniform manner, with an emphasis on high-frequency enhancement. The processing steps are shown in Table 1. The processing steps can be subdivided into five groups: 1) pre-processing includes SEG-Y input, geometry, trace editing, and refraction statics; 2) pre-stack processing steps such as including f-k filtering, deconvolution, Normal Move Out correction (NMO), residual statics, and Radon filter; 3) common mid-point (CMP) stacking; 4) post-stack processing, including time migration; and 5) signal-enhancement applications included f-x deconvolution and time- variant spectral whitening (TVSW). TVSW boosts the amplitudes of all frequencies within a certain bandpass filter to the same level by applying a series of gain functions to each narrow bandpass-filtered record. The first four of these groups represent the standard processing procedure, while the fifth one contained special steps required for improving the high frequencies. An example from seismic line CBY-5W (Figure 2) shows that reflection amplitudes near the target depth drop by ~50 dB from 10 to 80 Hz because of the strong absorption and attenuation within the thick sedimentary cover. Although the data quality is very good, interpretation is limited by the inherent bandwidth of the data. The dominant frequencies of the seismic data at the target zone are ~30 Hz (Figure 2). The seismic velocity within the Prairie Evaporite is about 2900 m/s. Thus the corresponding vertical resolution is ~25 m based on the 1/4-wavelength criterion (Sheriff and Geldart, 1995).

Due to comparatively low seismic resolution, the internal beds of the Prairie Evaporite are not resolved in the seismic section (Figure 2) and the salt edge is not shown accurately. The fault offset does not appear clearly as a reflector break. Faults would have to exhibit a vertical offset of at least ~25 m or beds a 25 m minimum thickness before they could be reliably imaged as distinct reflectors.

In order to increase the resolution and enhance high frequencies, post-stack TVSW followed by f-x deconvolution were applied. Figure 3 shows the migrated stack after using these procedures. In terms of the dominant frequency, this leads to an improvement from ~30 to ~50 Hz at the formation depth, therefore improving the estimated depth resolution to ~15 m. The resulting stacked section (Figure 3) shows a marked improvement in the detail and continuity of the image, and generally recognizable features include: 1) internal thin beds within the Prairie Evaporite; 2) a more accurately delineated salt edge, and 3) clearer fault displacements that can be measured more accurately from the seismic events. This processing was applied to all seismic lines in this study.

Table 1 - Seismic data processing steps. Process Purpose SEG-Y input Data input from field records. Geometry Providing the geographic reference; correcting logging errors. Trace editing Removing bad traces, reversing polarity as necessary, and muting. Refraction static Time correction for shallow subsurface. f-k filter Attenuating the ground roll in shot gathers. Deconvolution Compressing the input pulse and attenuating reverberations. NMO correction Removal of reflection time moveout. Residual static correction Removing the residual small time shifts of reflected arrivals. Radon filter Suppression of multiple reflections. CMP stacking Increasing signal to noise ratio. Time migration Plotting events and diffraction at their true locations. Time-variant spectral whitening Equalize and flatten the amplitude spectrum. f-x deconvolution Reducing random noise and improving image coherency.

Saskatchewan Geological Survey 3 Summary of Investigations 2005, Volume 1 West 0 5 km East b) Seismic Interpretation Identification of the seismic 1050 horizons was based on the interpretation of the synthetic 1100 Bakken Fm. seismograms produced from sonic logs and verified by the 1150 drill holes in Figures 4, 6, 7, and 8. The Ricker wavelet was used 1200 for generating the synthetic Top of P.E seismograms based on sonic logs

T from the study area. For this i 1250 m Thin bed not detected study, we selected seismic lines e Input frequency data ( P.E. Edge m that crossed the salt edges, so 1300 s

) that the seismic interpretation would help to delineate the 1350 positions of the edges more accurately. The data indicated a 1400 Base of P.E. major salt dissolution of the Prairie Evaporite located off-salt, 1450 and local salt dissolution within the Prairie Evaporite. In the 1500 study area, seismic sections indicate that salt dissolution Precambrian basement surface occurred in Mississippian, 1550 Triassic, Jurassic, and more recently as it disturbs all of the 1600 overlying seismic horizons. Figure 2 - Segment of the final stacked section from seismic line CBY-5W (Olympic Seismic Ltd.) processed without spectral whitening. Note the slumping of the strata caused by salt dissolution in the western part of the section. Also note that the depth Seismic line CBY-5W (Figure 4) resolution of the image is relatively low (modified from Hamid et al., 2004); P.E., shows salt dissolution that Prairie Evaporite. occurred during the Mississippian. Salt dissolution West 0 5 km East structures, basement uplift, and vertical fault throws are well 1050 imaged in this line. Some of the West 5 km East faults are deep, rooted in the basement, and appear to extend 1100 Bakken Fm. to near the surface; others are smaller and appear to be 1150 Frequency enhanced essentially confined to within data Devonian strata. The seismic line 1200 displays a relatively wide salt-

Internal beds of P. E. edge front of about 3 km. The )

s 1250 Top of of P. E. image from line CBY-5W m ( P.E. Edge

e indicates that the Prairie m

i 1300 Evaporite Formation decreases in T thickness from 110 m in the east 1350 to near-zero near the western part of the section. All layers above Thin bed Base of of P. E. 1400 where the Prairie Evaporite is absent appear to have collapsed up to and including the Late 1450 Devonian Bakken Formation. The layers above the Bakken 1500 Precambrian basement surface Formation between ~980 and 1100 ms are thickened. These 1550 thickened layers mainly represent the Mississippian Frobisher 1600 Formation, suggesting that the Figure 3 - The same section as in Figure 2 with f-x deconvolution and post-stack salt was dissolved and removed spectral whitening applied. The internal thin beds of the Prairie Evaporite are more during the time of Frobisher visible and are clearly thinning toward the west. Note the interpreted salt collapse deposition. that perturbed the strata beneath the Bakken Formation (modified from Hamid et al., 2004); P.E., Prairie Evaporite.

Saskatchewan Geological Survey 4 Summary of Investigations 2005, Volume 1 West 0 10 km East 21/04-01-009-14W/0 Seismic lines CBY-7W, NOR- 83314, and S-793 (Figures 5, 6, 500 and 7) represent salt collapse that may have occurred more recently. Lines CBY-7W and NOR-83314 (Figures 5 and 6) Fault display a well defined depression 750 that has affected all the seismic horizons above the Prairie Evaporite and is related to the Thickened layer Frobisher Regina Trough, indicating that 1000 the trough is associated with T i m faults rooted in the basement. Subsidence Kisbey e

( Thickened layers are not m Bakken Fm. s observed in the seismic sections ) Internal beds of P. E. Top of P. E. P.E. Edge but could be present in Upper 1250 Cretaceous or Cenozoic strata which are not imaged. Seismic line S-793 (Figure 7) shows Thin bed Base of P. E. regional salt dissolution located 1500 off-salt in the eastern part of the seismic section, and a localized Precambrian basement surface salt collapse about 800 m in diameter within the body of 1750 Prairie Evaporite near the western end of the seismic section. Seismic lines show that the salt edge is relatively sharp, Figure 4 - Salt-dissolution-induced subsidence that did not affect strata shallower with thickness changing from than about 980 ms (seismic line CBY-5W); P.E., Prairie Evaporite). ~70 to ~0 m over a distance of about 800 m. Such steep 0 15 km West East dissolution relief could be an indication of its association with activated faults.

Lines CBS-8 and SE-850844 750 (Figures 8 and 9) represents salt dissolution that occurred during Triassic and Jurassic times. Features which generally can be Frobisher seen on the sections and related 1000 to the salt dissolution are faulting T

i and disruption of the normal m

e Phanerozoic column in the ( m Fault s Williston Basin. Anomalous ) Bakken Fm. thickening of sedimentary strata 1250 such as those of the Watrous P.E. Edge Top of of P. E. Formation that overlie Prairie Evaporite thins indicates areas of contemporaneous salt dissolution. No evidence of Base of P. E. further salt dissolution was 1500 Precambrian basement surface identified from these seismic sections.

4. Subsurface Map 1750 Interpolation Seismic and well-log data Figure 5 - Effect of a salt collapse on seismic events (seismic line CBY-7W). Note provide linear readings and that compared to the collapse feature shown in Figure 4, this collapse structure must have formed more recently as it disturbs all of the overlying strata; P.E., Prairie points that need to be Evaporite. interpolated to produce subsurface maps. However, with

Saskatchewan Geological Survey 5 Summary of Investigations 2005, Volume 1 West a relatively sparsely spaced well- 10 km East 0 41/04-06-006-13W/0 log data set as in this study area 500 (Figure 1), such interpolation Regina Trough depends on the method employed. This is particularly important where the data are interpolated or extrapolated in 750 poorly constrained areas, such as the salt edges of this study. In order to evaluate the Fault effectiveness of the interpolation Watrous techniques on isopach contours, 1000 T Frobisher

i well picks of Middle Devonian m

e strata in the IEA Weyburn CO2 ( m Monitoring and Storage Project

s Bakken Fm.

) area (Kreis et al., 2003) were 1250 used. Isopach maps of Prairie Top of P. E. Evaporite created in spatial P.E. Edge interpolation programs from Surfer, Matlab, and Generic Mapping Tool (GMT) software 1500 Base of of P. E. packages were compared to evaluate the dependence of the resulting isopach map on the interpolation techniques. After Precambrian basement surface several experiments, GMT was 1750 chosen as the preferred option. GMT programs offer a choice of “spline tension” parameter Figure 6 -Seismic line NOR-83314 showing the conventional section of Regina 0 ≤T≤1, with tighter splines Trough; P.E., Prairie Evaporite. resulting in smoother maps (Smith and Wessel, 1990). West 01/12-33-005-23W/0 0 10km East Additional advantages of GMT are the free availability of the programs, possibilities of modifications, scalability, use of 750 simple ASCII data formats, and high-quality of PostScript images produced on a variety of cartographic projections. Watrous Different values of spline tension 1000 parameter T were used to Frobisher generate isopach maps of the Fault Prairie Evaporite. The value of T=0 (i.e., the tightest spline) was T Bakken Fm. i

m finally chosen as the preferred e

( 1250 mapping option, and a Prairie m

s Local salt dissolution Top of P. E. salt isopach map based on well ) P.E. Edge data alone was created (see Hamid et al., 2004).

1500 Further, combining the well logs Base of P. E. with seismic data allows the construction of a map which is more detailed and accurate in the areas crossed by the seismic lines 1750 (Figure 10). The results of seismic interpretations fill in the gaps where well-log data are not available. Figure 7 - Seismic line S-793 displays both regional and local salt dissolution of the Prairie Evaporite (P.E.). The interpretations of the position of the salt dissolution edge of seismic lines CBY-5W

Saskatchewan Geological Survey 6 Summary of Investigations 2005, Volume 1 West 0 15 km 01/08-05-002-14W/0 East and CBY-7W indicate important differences, in places as large as ~9 km, from the edge shown by 750 Kreis et al. (2003). Connecting the positions of the salt edge of the two seismic lines is difficult because of the large distance between the seismic lines. 1000 Fault However, the salt edge of the Thickened layer northern part of the study area Watrous can be estimated based on the interpretation of the 29.7 km long, north-south seismic line 1250

Time(ms) Frobisher SE-84524, which lies close to, and is nearly parallel to, the salt- Subsidence P5_salt Bakken Fm. dissolution edge (this is suggested by the Prairie 1500 Evaporite thickness, which, P.E. Edge along the entire line, ranges from Top of P. E. ~15 to 40 m). Assuming that the salt-dissolution edge has a constant dip, the salt edge along 1750 the seismic line SE-84524 could Base of P. E. be extrapolated west of this line, Precambrian basement surface as shown in Figure 10.

2000 In the southern part of the study area, the positions of the salt Figure 8 - Seismic line CBS-8 indicates strata collapse due to salt removal, basement edge interpreted from seismic uplift and several deep faults. Loss of Prairie Evaporite (P.E.) continuity in the west lines CBS-8 and SE-850844 is due to salt dissolution and faults. generally agree with 0 2km interpretation by Kreis et al. West East (2003) whereas to the west, seismic lines S-793 and 9-NS show that the positions of the salt 750 edge are shifted by ~2 km to the east (Figure 10). Thus, inclusion of additional seismic information significantly improves mapping of the Prairie Evaporite salt edge. 1000 Prairie Evaporite was not found Thickened layer along seismic lines R-19, PCR- Watrous 11 and SE-84506. Time(ms)

P5_salt 1250 Frobisher 5. Gravity Signature of Prairie Evaporite Fault Bakken Fm. In addition to the seismic study, Subsidence we also attempted to investigate 1500 the relationship between P.E. Edge Top of P. E. basement structures and salt collapses of the Prairie Evaporite using gravity data available from the Geological Survey of 1750 Base of P. E. Canada. The gravity data have high density in the east and northwest of the study area while the centre is poorly covered by only a few scattered gravity Figure 9 - Seismic line SE-850844 shows a salt collapse that appears to have readings. From the interpolated occurred in Jurassic and Triassic times. 2-D gravity grid, a ~33 km long east-west profile just south of the centre of a 3-D gravity anomaly

Saskatchewan Geological Survey 7 Summary of Investigations 2005, Volume 1 was extracted. The profile is parallel to, and ~2 km to the north of, the seismic line NOR- 5550 83314. The location of the profile has been chosen because SE-84506 it is well covered with gravity readings and is near the seismic line NOR-83314 (Figure 11). SE-84524 5525 Two-dimensional modeling of SE-84529 Thickness (m) gravity data was performed using GM-SYS software. The gravity model was constructed using the ) PCR-11 CBY-5W known and confident seismic m interpretation of seismic line k

( 5500

NOR-83314 and assumptions of g the average densities of the n CBY-7W i

h Williston Basin sedimentary t

r column and basement rocks. The o NOR-84431 densities of various formations N S-793 5475 were assigned based on density logs. Assigning densities to the 9-NS formations using well-log information was considerably R-19 complicated because of the lithological variability between 5450 and within formations. In order to facilitate the interpretation, average densities of the various CBS-8 formations were estimated from SE-850844 the density logs. As a result of this procedure, five main blocks were established within the 500 525 550 575 Phanerozoic rocks (Figure 12). Easting (km) Blocks 1, 2, 3, 4, and 5 represent 3 Figure 10 - Interpolated isopach map of the Prairie Evaporite (PE) Formation Mesozoic-Cenozoic (2.4 g/cm ), using both well and seismic data. Dashed yellow and solid white lines show the salt Middle Devonian-Mississippian edge of the Prairie Evaporite interpreted by this study and Kreis et al. (2003), (2.6 g/cm3), Middle Devonian respectively. Salt edge shown as solid yellow lines west of seismic line SE-84524 was Prairie Evaporite (2.2 g/cm3), determined by extrapolation of PE thickness in this line. Coordinates are UTM. Middle Ordovician-Middle Devonian (2.71 g/cm3), and Cambrian-Middle Ordovician rocks (2.45 g/cm3) respectively. Modeling of the observed anomaly suggested adding a small sub-block of low density (2.30 g/cm3) within the first block. Mesozoic rocks show highly variable densities when compared to more consistent densities of the other formations. From logs, the Mesozoic densities appear to range from about 1.9 to 2.6 g/cm3. Using 2.3 and 2.4 g/cm3 for the average density therefore seems appropriate. These values are considered reasonable for gravity modeling purposes. However, the scatter is quite large suggesting that there could be significant density variation within the shallow subsurface. Basement densities were assigned based on the studies by Leclair et al. (1993 and 1994) and increased from 2.73 g/cm3 in the west (the granitic complex of the Wyoming Craton) to 2.82 to 3.03 g/cm3 in the east (within the Trans-Hudson Orogen). The Precambrian basement represented by blocks 6, 7, and 8 has densities ranging from 2.73 to 2.9 g/cm3. Blocks 6 and 7 are in the Trans-Hudson Orogen while block 7 represents Wyoming Province. The observed gravity data were fitted by adjusting the densities of the Phanerozoic strata and Precambrian basement and the position of the bottom of the Precambrian basement block. The observed gravity profile shows a major positive anomaly (~4 mgal) occurring in the vicinity of a known salt collapse west of the gravity profile. At a more regional scale, the observed gravity decreases sharply near the margin of the Trans-Hudson Orogen and Wyoming Province (Figure 11; see also Hamid et al., 2004). Modeling of the observed anomaly suggests a depression in the basement rock which affects all of the sedimentary layers (Figure 12). After careful application of the constraints from detailed seismic imaging of the sedimentary cover, gravity modeling reveals strong density contrasts within the Precambrian basement (Figure 12). It may be speculated that the interpreted high-density basement between kilometres 0 and 22 of our gravity profile (Figure 12) could be associated with the localized gravity high observed north-northwest of the profile (Figure 11). Because of

Saskatchewan Geological Survey 8 Summary of Investigations 2005, Volume 1 their proximity to both the Prairie Evaporite and salt collapse, such strong density contrasts could make detection 3-D gravity of the gravity signatures of the anomaly mgal salts problematic. 5500 Although the gravity anomaly is caused mainly by the high T densities within the crystalline H A A’ O crust (with the wavelength of /W Seismic line NOR-83314 gravity anomaly exceeding ) y o m m ~40 km and likely associated k ( in g 4750 with the varying crustal g n i

h P

t Weyburn thickness), the effect of the r r

o o v Phanerozoic cover was also N i n c examined. Seismic interpretation e of the line NOR-83314 and well- b o u log data were used to determine n d the effect of the Phanerozoic a 4500 r cover and especially of the salt y Trans-Hudson dissolution edge of the Prairie Orogen Evaporite. As the maximum Wyoming thickness of Prairie Evaporite in Province the proposed model is 60 m, modeling shows that the Prairie 520 540 560 580 600 620 salts have an effect of 0.4 mgal, Easting (km) or ~10% of the total anomaly. Thus, given the uncertainties in Figure 11 - Bouguer anomaly map of Saskatchewan obtained using GMT characterization of its (minimum-curvature spline with T=0) interpolation of gravity readings from the overburden, and particularly in Geological Survey of Canada database. Black line shows the edge of the Prairie Evaporite interpreted by Kreis et al. (2003) (Figure 1). A-A′ is the line of the gravity the Precambrian basement, cross section extracted from the data and modeled in Figure 12. Black dots indicate unambiguous recognition of the the locations of gravity readings. Seismic line NOR-83314 is shown in blue. The gravity signature of the Prairie major tectonic domains are indicated. Coordinates are UTM in kilometres. salts does not appear to be Wyoming Trans-Hudson Orogen A A’ West East

Low-density rocks Mesozoic-Cenozoic 1

2 M.Devonian-Mississipian 3 Prairie Evaporite 4 5 Cambrian-M.Ordovician 8 Precambrian Precambrian 6 Precambrian 7

M.Ordovician-M.Devonian

Figure 12 - Gravity model along the ~33 km long gravity profile across the salt collapse (Figure 11). Note the strong density contrast within the Precambrian basement below ~2.1 km depth.

Saskatchewan Geological Survey 9 Summary of Investigations 2005, Volume 1 possible. However, high-resolution gravity surveys with station spacing of 100 to 200 m might still constrain this overburden and help detect salt collapses and thin beds in the study area.

6. Conclusions 1) Inclusion of seismic results allows adding additional detail and improves confidence of interpretations. 2) High-frequency enhancement seismic data processing techniques help improve the resolution and result in improved images of the salt edge and thin beds. 3) In the seismic sections analyzed to date, indications of salt-collapse events of Mississippian, Mesozoic, and more recent age were observed. 4) The interpretations of the positions of the Prairie Evaporite edge in the seismic lines processed in this study generally agree with the interpretations by Kreis et al. (2003) and Holter (1969) in the southern part of the study area, but also indicate important differences in the northern part. 5) Gravity anomalies within the region are mostly related to lateral variations within the deep-seated basement or mantle structures. However the effect of Phanerozoic cover has been noticed. High-resolution gravity surveying with station spacing of ~100 m, in combination with seismic imaging, could still be useful in detecting salt collapses in the study area.

7. Acknowledgments This research was made possible through a research grant from Saskatchewan Industry and Resources to the University of Saskatchewan. Additional financial support for H.H. was also provided by University of Saskatchewan. We thank Drs. Z. Hajnal, B. Pandit, and D. Gendzwill for many valuable discussions and advice. Dr. Hajnal’s support was also critical in obtaining the seismic data and formulating the initial goals of the project. This work was facilitated by software grants from Landmark Graphics Corporation, Schlumberger Limited, and Hampson-Russell Limited. GMT programs (Wessel and Smith, 1995) were used in preparation of some of the illustrations.

8. References Baillie, A.D. (1953): Devonian System of the Williston Basin area; Man. Dept. Mines Nat. Res., Mines Branch Publ. 52-5, 105p.

Bishop, R.A. (1954): Saskatchewan exploration progress and problem; in Clark, L.M. (ed.), Western Canada Sedimentary Basin Symposium, Amer. Assoc. Petrol. Geol., Ralph Leslie Rutherford Memorial Volume, p474- 485.

Edmunds, R.H. (1980): Salt removal and oil entrapment, Can. Soc. Petrol. Geol., Mem. 6, 988p.

Gendzwill, D.J. and Martin, M. (1996): Flooding and loss of the Patience Lake Potash Mine; CIM Bull., v89, no1000, p62-73.

Halabura, S.P. (1998): Salt Collapse: The key to new Saskatchewan Devonian Oil? Some Initial Musings; North Rim Exploration Ltd, Saskatoon; http://www.northrim.sk.ca/NorthRim/exad/free_stuff/SaltCollapse_.PDF, accessed 20 Feb 2005. Hamid, H., Morozov, I.M., and Kreis, K. (2004): Seismic delineation of the southern margin of the Middle Devonian Prairie Evaporite in the Elk Point Basin, south-eastern Saskatchewan; in Summary of Investigations, Volume 1, Saskatchewan Geological Survey, Sask. Industry Resources, Misc. Rep. 2004-4.1, Paper A-5, 9p. Holter, M.E. (1969): The Middle Devonian Prairie Evaporite of Saskatchewan; Sask. Dept. Miner. Resour., Rep. 123, 134p. Kreis, L.K., Thomas, P.L., Burke, R.B., and Whittaker, S.G. (2003): Prairie Salt Isopach (Prairie Formation); in Devonian Isopach and Structure Maps, IEA Weyburn CO2 Monitoring and Storage Project Area (test version), Sask. Industry Resources/N. Dak. Geol. Surv., CD-ROM. Leclair, A.D., Lucas, S.B., Scott, R.G., Viljoen, D., and Broome, H.J. (1994): Regional geology and geophysics of the sub-Phanerozoic Precambrian basement south of the Flin Flon–Snow Lake–Hanson Lake Belt, Manitoba- Saskatchewan; in Current Research, Part-C, Geol. Surv. Can., Pap. 94-1C, p215-224.

Saskatchewan Geological Survey 10 Summary of Investigations 2005, Volume 1 Leclair, A.D., Scott, R.G., and Lucas, S.B. (1993): Sub-Paleozoic geology of the Flin Flon Belt from integrated drill core and potential field data, Cormorant Lake area, Manitoba and Saskatchewan; in Current Research, Part C, Geol. Surv. Can., Pap. 93-1C, p249-258. Meijer Drees, N.C. (1986): Evaporitic Deposits of Western Canada; Geol. Surv. Can., Pap. 85-20, 118p. Saskatchewan Department of Mineral Resources (1961): Structure Contour Maps E-166 to 173. Sawatzky, H.B., Agarwal, R.G., and Wilson, W. (1960): Helium Prospects in Southwest Saskatchewan; Sask. Dep. Min. Resour., Rep. 49, 26p. Sheriff, R.E. and Geldart, L.P. (1995): Exploration Seismology (second edition); Cambridge University Press, Cambridge, U.K., 592p. Sloss, L.L. and Laird, W.M. (1947): Devonian system in central and northwestern Montana, AAPG Bull., v31, p1404-1430. Smith, W.H.F. and Wessel, P. (1990): Gridding with continuous curvature splines in tension; Geophys., v55, p293- 305. Wessel P. and Smith, W.H.F. (1995): New version of the Generic Mapping Tools released; EOS Trans. Amer. Geophys. Union, v76, p329. Zharkov, M.A. (1988): Devonian Evaporite Basins; in McMillan, N.J., Embry, A.F., and Glass, D.J. (eds.). Devonian of the World: Proceedings of the Second International Symposium on the Devonian System, Volume 2 - Sedimentation, Can. Soc. Petrol. Geol., Mem. 14, p415-427.

Saskatchewan Geological Survey 11 Summary of Investigations 2005, Volume 1