| Fractured reservoirs make up an important part of both hydrocarbon and groundwater reservoirs worldwide,and detailed analysis of fractures and fracture networks is of prime important for utilizing such resources.Fractured crystalline aquifers of mountain watersheds may host a significant portion of the world's freshwater supply.In the western part of United States,groundwater and surface water provided by mountain watersheds can meet the water needs of more than 60 million people?Barnett et al.,2005;Bales et al.,2006?.In many regions of the world,especially in arid and semi-arid regions such as India and Africa,the fractured crystalline rock aquifer is almost the only source of drinking water for local residents?Gustafson and Krásny,1994;Guihéneuf et al.,2014?.In order to properly manage and rationally utilize these resources,it is imperative to obtain updated hydrological knowledge about groundwater in mountain fractured crystalline aquifers.However,groundwater storage and flow in most mountain environments are poorly known.Mountain watersheds,which often consist of granitic or metamorphic rocks,are characterized with rough terrains that are difficult to access.Mountains are often sparsely populated,thus few groundwater monitoring wells exist from which long term water level or characterization data can be obtained.In granite aquifers,fractures can provide both storage volume and conduits for groundwater.Characterization of hydraulic properties in such aquifers is important for predicting flow rate and calibrating models.To effectively utilize water resources in these environments,it is important to understand the hydraulic properties,groundwater storage,and flow processes in crystalline aquifers and field-derived insights are critically needed.The Blair Wallis Fractured Granite Hydrology Research Well Field lies within the Crow Creek Watershed of the Laramie Range which lies within US Forest Service land about 21 km southeast of Laramie,Wyoming.The elevation of the study area is between 2400 and 2600 m.Our study area lies within the so called A-type Sherman Granite which are generated 1.43 Ga ago consisting of microcline,plagioclase,quartz,biotite,hornblende,and ilmenite.Sherman Granite is a coarse-grained,Precambrian aged granite.Local climate data from the Crow Creek SNOTEL station of the last 10 years show that the Blair Wallis well field receives 620 mm of annual precipitation,of which 90% falls as snow.At the Blair Wallis well field,nine bedrock wells?i.e.,BW1,BW2,BW3,BW4,BW5,BW6,BW7,BW8,and BW9?have been completed that are cased to the bottom of the saprolite but remain open boreholes in the fractured granite.In this study,multimethod were used to characterize the permeability of this fractured granite aquifer in multi-scale?i.e.,borehole scale,interwell scale,and large or field site scale?using borehole data from nine wells in the Blair Wallis Fractured Rock Hydrology Research Well field.Borehole hydraulic test data used in this study include: slug tests at all the nine bedrock boreholes;FLUTe liner profiling test at three boreholes?i.e.,BW5,BW6,and BW7?;cross-hole pumping test at five boreholes?i.e.,BW1,BW4,BW6,BW7,and BW8?.Borehole logging data include flowmeter logging under ambient flow condition at three boreholes?i.e.,BW5,BW6,and BW7?,borehole optical and acoustic televiewer,borehole NMR logging at three boreholes?i.e.,BW5,BW6,and BW7?,magnetic susceptibility logging at BW5,BW6,and BW7,and a series of conventional logging,such as caliper,resistivity,gamma ray,etc.Here we emphasize that cores were only used as reference for identifying flowing fractures.This is because?1?only five of nine wells are cored?i.e.,BW1,BW2,BW3,BW4,and BW5?,and?2?granite has the characteristics of high hardness and brittleness and we do find there are a number of drilling-induced fractures in the cores which do not represent the actual borehole condition.These drillinginduced fractures may cause great interference when counting fracture density and characterizing fracture geometry,such as strike and dip angle,etc.In theory,it is possible to distinguish between natural fractures and drilling-induced fractures by identifying weathered surfaces and fresh surfaces in cores.However,due to the large number of drilling fractures observed in the cores,the work of identifying natural fractures in cores is enormous.1.Single hole hydraulic test to characterize borehole scale hydraulic properties and groundwater flowSlug tests are widely used to provide an average Kh estimate in the vicinity of the test well?borehole scale?.However,the test results are susceptible to well skin and nonDarcian flow.Butler et al.?1996?reported a series of guidelines to improve the quality of parameter estimates obtained from slug tests.The key points of these guidelines include:?1?a series of slug tests with different initial displacements?H0?should be carried out at each well during a test Non-Darcian flow can be identified by fitting to the classic solutions,such as Cooper et al.?1967?curve fitting;?2?a series of slug tests should be repeated for a given H0 to evaluate the well skin effect.Therefore,according to the existing equipment conditions,we applied two different initial H0,and for each H0 at least two tests were repeated to ensure the accuracy of the slug test.By analyzing a set of repeat slug tests with different initial displacements and calculating an average Re for either individual or equivalent fracture,we concluded that both skin effect and non-Darcian flow,which may lead to an underestimated horizontal hydraulic conductivity?Kh?,are negligible during the slug tests.In addition,the model used to interpret the data also has a great impact on the estimated hydraulic conductivity.In this study,the Cooper et al.?1967?model,the KGS?1994?model,and the Bouwer and Rice?1976?model are all used to fit the recorded slug test data.Results show that the estimated Kh using Cooper et al.?1967?model for confined aquifer is over a factor of two greater than the estimated Kh using the Bouwer and Rice?1976?model and the KGS?1994?model for unconfined aquifer.However,the estimated Kh using Bouwer and Rice?1976?model and the KGS?1994?model are very close.Moreover,Butler et al.?1996?pointed out that the Cooper et al.model can lead to a significantly overestimated Kh of the formation when a dimensionless storage parameter ? of the formation is moderate to low.Generally,for fractured granite aquifer,? is always low,and in this study,? ranges from 10-7 to 10-10,which suggests that the results obtained using Cooper et al.model are less reliable.FLUTe liner profiling,a recently developed high-resolution Kh profiling method?at the research site,it has ?0.3 m vertical resolution?was used as an alternative way of estimating discrete Kh.For borehole BW5,BW6,and BW7,FLUTe profiling was used to obtain discrete Kh along the open hole.Results show that for all the three boreholes,the FLUTe logs indicate a general trend of decreasing Kh with depth.Borehole Kh varies over?4 orders of magnitude(10-8–10-5 m/s).Kh is consistently higher near the top of the bedrock that is interpreted as the weathering front.Using a cutoff Kh of 10-10 m/s,the hydraulically significant zone extends to ?40–53m depth.Downhole optical and acoustic televiewer logs can clearly identify the visible fractures along the open hole.However,large errors can arise in the interpretation of such logs.On the one hand,micro-cracks are difficult to identify from borehole images,on the other hand,not all fractures identified are necessarily hydraulically active.In this study,we determined the number of flowing fractures by jointly interpreting borehole televiewer logs with spinner flowmeter logs.Total fracture densities were initially estimated from the optical and acoustic logs,then the subset of flowing fractures were filtered by borehole flowmeter?using a cut-off value?.Flowmeter logging can also be used to detect micro-open cracks that provide conduits for groundwater but cannot be identified by the televiewer.However,the number of microcracks cannot be easily determined by flowmeter logging alone,so we assume one equivalent microcrack for the intervals that contain such tight but flowing cracks.For all three wells,flowing fracture density exhibit a decreasing trend with depth,which corresponds to the Kh profiles determined using FLUTe blank profiling.Kh is consistently higher near the top of the bedrock that is interpreted as the weathering front.Moreover,the transmissivity values determined for the open hole using slug tests and FLUTe profiling are proved very close,which further proves the reliability of the two methods used in this study.Based on Kh estimates from slug tests and FLUTe and the determined number of flowing fractures,an average hydraulic aperture is estimated for the entire open hole and for the discrete FLUTe test intervals by employing Cubic Low.Results show that the FLUTe-estimated hydraulic apertures for all the three wells vary over one order of magnitude?14 ?m – 200 ?m?,indicating vertical heterogeneity in fractures.Based on slug test derived Kh,average hydraulic apertures calculated are 92 ?m?BW5?,86 ?m?BW6?,and 105 ?m?BW7?.Moreover,at each well,an average aperture obtained from its FLUTe profile is very close to that obtained from slug test,suggesting that slug test can be used to provide a reliable average aperture estimate.Based on the estimated hydraulic apertures and the determined number of flowing fractures,an effective fracture porosity is estimated to be 4.0×10-4 with a standard deviation of 8.4×10-6,thus fractured crystalline rock in this headwater watershed can host significant quantity of groundwater.Given the estimated Kh,the site-scale horizontal hydraulic gradient and estimated effective fracture porosity,an average linear groundwater velocity can be estimated by employing Darcy's law.At the Blair Wallis well field,the hydraulic gradient ranges from 0.03 to 0.05,and the magnitude of hydraulic gradient of 0.04 was averaged from September 2015–September 2016 and is used to calculate the groundwater velocity.Natural groundwater velocity is estimated to range from 0.4 to 81.0 m/day,implying rapid pathways of fracture flow.It is noteworthy that in snow melting season in spring,the hydraulic gradient at Blair Wallis can reach 0.05,which suggest groundwater velocity will be higher in spring.Flinchum?2017?determined three zones at Blair Wallis field site using seismic P-wave velocity: an upper weathered zone?saprolite?,a middle fractured zone,and a lower and often less fractured zone.The P-wave velocity corresponding to the interface between saprolite and fractured granite is about 1.2 km/s,and this velocity interface is basically consistent with the casing depth of the borehole.The boundary between fractured granite and protolith corresponds to the P-wave velocity of 4.0 km/s.Based on all available data,including cores,borehole logs,LIDAR topography,and a seismic P-wave velocity model,a three dimensional geological model of the site was built.The model shows that the fractured bedrock is ?4× the thickness of the saprolite.The average ambient water table position follows the interface between saprolite and fractured bedrock and dips to the east.This indicates the groundwater system at the site appears to be topography driven.2.Hydraulic Conductivity Calibration of Logging NMRFor three wells?i.e.,BW5,BW6,and BW7?,at which both NMR logging data and high resolution FLUTe Kh are acquired,we demonstrate the potential of using NMR logging measurements to provide estimates of hydraulic conductivity for this fractured granite aquifer by calibrating the Schlumberger Doll Research?SDR?model.Bootstrap,a non-parametric approach is used to get the optimized parameters in the SDR equation at the three test wells.The uncertainty of the estimated Kh can also be evaluated.Results show that at Blair Wallis field site the difference of the optimized parameters for the SDR equation at the test three wells exceeded a factor of three,which means that if calibration of the SDR equation parameters was performed at a one of the three boreholes and used to estimate Kh at another well,the resulting Kh would be more than a factor of 3 from the estimated Kh using the location optimized value.This is due to strong heterogeneity with widely varying hydraulic apertures at different locations at this site,which leads to different diffusion regimes.We found that it is unrealistic to expect that one set of empirical parameter can be identified for a fractured aquifer.However,we interpret the factor of 3 difference observed here to be acceptable for doing a “local” calibration with one high resolution Kh log that would be transportable to all wells at a wellfield.We also demonstrate that when high-vertical resolution Kh logs are unavailable,it would also be possible to achieve acceptable calibration using only a slug test in each well to obtain a bulk aquifer Kh value.However,this is a case for the Blair Wallis wells but its general applicability to other fractured rock sites remains to be tested.We conclude that porosity does not improve the fit to the data,which is consistent with the findings of Maurer and Knight?2016?.For generalized pore geometry,Dlugosch et al.?2013?presented the Kozeny-Godefroy Model?KGM?.Compared to the SDR model,the KGM model is physically based and all KGM parameters are known or can be obtained except tortuosity and surface relaxivity.More importantly,the KGM model is free from the diffusion regime assumption.In this study,the KGM is also used to estimate Kh at BW5,BW6,and BW7.We demonstrate that KGM is also feasible for fractured rock,although Dlugosch et al.?2013?speculate that the KGM will fail under the planar geometrical fracture case due to the lower influence of diffusion.Using the magnetic susceptibility logging data from the three boreholes,we found that “intermediate diffusion'' dominates the relaxation time and ‘‘slow diffusion'' during the NMR relaxation time may occur in the flowing fractures when hydraulic aperture are sufficiently large,therefore assuming ‘‘fast diffusion'' in the interpretation of NMR data from fractured rock may lead to inaccurate KNMR estimates.A similar effect was also observed by changing the exponent on the T2ML term—associated with sensitivity to the diffusion regime—from a value of 2 to 1 during the Kh calibration.3.Pumping test and 2D hydraulic tomographyThe foregoing permeability estimating methods,no matter direct and indirect,are limited to borehole scale.In order to predict permeability in larger scale?i.e.,inter-well scale and field site scale?,a series of cross-hole pumping tests are needed.A total of five constant rate pumping tests at five individual boreholes have been carried out at the study site,and the time order of each test is as follows: BW4,BW7,BW8,BW1,and BW6.We first used multiple pumping test solutions to estimate large scale hydraulic parameters.The used solutions include Theis?1935?,Cooper and Jacob?1946?,Neuman?1974?,Moench?1984?,and Barker?1988?.Results show that the estimated Kh from pumping test ranges from 9.5×10-6 to 1.5×10-5m/s,and the estimated Kh using different models are proved very close.Moreover,it is found that the estimated Kh from pumping test is over one order of magnitude greater than the estimated Kh from slug test,which suggest the existence of scale effects at Blair Wallis well field.It should be pointed out that pumping test at BW4 was carried out in November,2015,during which BW6,BW7,BW8,and BW9 haven't been drilled.So BW1 is the only monitoring well while pumping BW4.Moreover,due to the well configuration and the pump used for BW1,the pumping period at BW1 cannot last long,and no drawdown can be observed at all the monitoring wells.Thus,pumping test at BW1 is not used for the subsequent 2D hydraulic tomography?HT?experiment.In this HT experiment,our goal is to characterize the heterogeneity of Kh and Ss within the enclosed area by BW1,BW6,BW7,BW8,and BW9.Cross-hole responses were observed and all the water level data at both the pumping well and the observation wells were recorded using a pressure transducer.Then,a two-dimensional HT experiment was performed by employing the Sim SLE?Simultaneous SLE?algorithm?Xiang et al.,2009?to interpret the four cross-hole pumping tests together,and the intermediate scale?i.e.,interwell?hydraulic parameters can be estimated.All the invertion process was carried out in the finite element numerical simulation software of VSAFT2?Yeh et al.,1993?.Here we employed uniform grids and a two-dimensional domain?160×80m?.Every grid is square with a size of 2×2m.The total model contains 3321 nodes and 3200 elements.Constant head boundary conditions were assigned to the west and east boundaries,while the north and south were assumed as no-flux boundaries.Moreover,the mean K and Ss as well as their correlation scales are assumed to be known a priori.In this study,the inversion starts with homogenous mean fields of K= 1.0×10-5 m/s and Ss= 1.0×10-5 m-1,which are almost the same as the values derived from traditional pumping test.Exponential correlation functions were employed to describe the unconditional spatial correlations of K and Ss,and the correlation scales were set to 15 m in both x and y directions.In order to get both Kh and Ss,the head?or drawdown?data over both transient flow and steady state flow is necessary.Specifically,for pumping BW7,four heads are selected from BW1 hydrograph,twelve heads are selected from BW6 hydrograph,eleven heads are selected from BW8 hydrograph,and nine heads are selected from BW9 hydrograph.For pumping BW8 and BW6,four heads are selected from each hydrograph.Results show that HT is a useful tool to image continuous high K and low Ss zones,which represent fast flow pathways.These high K zones may also reflect the connectivity of fractures between boreholes.Specifically,the HT analysis yielded highest K zone between BW7 and BW9,and another high K zone between BW7 and BW8.There is a low K zone?or barrier?between BW7 and BW1.These inversion results are consistent with the observations,for example,pressures at borehole BW9 respond first while pumping BW7 and BW7 shows the fastest response comparing to the other boreholes while pumping BW8.It is observed that the Ss maps are negatively correlated to corresponding K tomogram,and we suppose that area with high K and low Ss is the area that flowing fractures developed.Further study reveals that the patterns of the K and Ss tomograms do not change significantly due to the change of correlation scales.Those large-scale high K zones and low K barriers remain the same.Here we emphasize that only four pumping tests are available in this 2D HT analysis,detailed results of the K and Ss heterogeneity cannot be drawn clearly and only preliminary results can be obtained.Even so,the results of this study still have some significance for characterizing heterogeneity of interwell permeability of Blair Wallis well field.In this study,the permeability of a mountain fractured granite reservoir at Blair Wallis Fractured Rock Field Research Site is characterized comprehensively using various testing methods in multi-scale,including borehole scale,interwell scale and large or field site scale.The results of this study have a certain significance for enriching the understanding of groundwater reserves and groundwater flow in fractured aquifers in mountainous areas,and also have a certain reference value for rational development and utilization of such groundwater resources.Furthurmore,the results of this research have important reference significance for reservoir characterization of fractured petroleum reservoirs. |
PDF Full Text Request