Using Advanced Technologies to More Effectively Utilize Historic Exploration Data

Editor's Notes

• #2: Good afternoon. First off, I’d like to thank the co-authors, Clark Arnold and Steve Van Nort for their kind assistance and work that has gone into the Van Deemen project. The topic of todays talk as you can see is Using Advanced Technologies to more effectively Utilize Historic Exploration Data We’ll touch briefly on techniques that were commonly used in the 80s on a project and then fast forward to what we can do with that same information today.
• #3: But before we begin, I’d like to ask you all a couple of questions….…
• #4: So – raise your hands – how many of you have used a Brunton Planimeter Plane Table/Alidade Hip Chain/Topofil Light Table Gps Portable XRF Leapfrog Mobile Field Device – like your smartphone to collect strik/dip readings, or for field mapping.. Ok that give me a good idea of where we all are
• #5: And what do these items have in common??? They are all tools a geologist has used in the past 50 years or so to gain insight into the question… What is the tonnage/grade of a deposit And is it economic
• #6: Normally you’d define economics in our case by the price of gold, the mining costs, and % recovery. But these days that’s just part of the story, you also have to consider government regulations, permitting, access, environmental concerns and socio economic considerations.
• #7: Technology is changing how we work – this is a screen grab of my phone showing Historic Surface samples anomalous Au colored dots so I can see the Geochem values at a specific location, or I can turn on another layer of a scan of a geologic map and walk up to the outcrop and actually read the notes from the geologist that had been on the ground in the 80s. Or I can send claim location information from my office to a crew while they are still in the field.. But I digress… I think we all have stories about how we are using the current technology
• #8: Thirty years ago, costs for reverse circulation drilling would be in the range of $8.50/ foot, so, for 1,000 feet of it would cost $8,500 for drilling alone. Times have changed. It can now cost over $35,000 to drill that same footage, assuming that there are no permitting issues. Multi-element trace geochemical analysis is a standard tool in detailed and regional geochemical exploration. In the past, geologists would utilize the top 10% of elemental analysis as anomalous. Currently, affordable trace element geochemistry provides the necessary tools to interpret bedrock geology and mineralization systems, a lot of 46 element ICP data resides in company files with limited interpretation. Modern computational tools can rapidly process 3D geological datasets and assist in generating 3D geological models. Implicit modelling generates geological models directly from drill-hole data, using mathematical interpolation functions are used to generate 3D isosurfaces, instead of manual linkage of hand-digitized 2D cross-sections. Many deposits have seen some historic development, whether it is surface sampling, geophysics, drilling, or production. Most current historic data sets are incomplete, with missing drill hole locations, drill logs, assay certificates, level plans or cross sections. How do you incorporate the past work on a project by people who understood the deposit into a coherent package so you can make an informed decision on the potential viability of the project without having to redrill the deposit?
• #9: Which leads us into today’s talk – has anyone been to the Van Deemen?? I continue to be amazed at the number of people who have visited the site at some point.
• #10: The Van Deemen project is located in the northern Black Mountains, Mohave County Arizona, 50 mi northwest of Kingman, Arizona (population ~28,000), which historically is one of the most prolific gold-producing mountain ranges in Arizona, yielding some 2.5 million ounces of gold. That number is most likely higher with the current production from the Moss Mine.
• #11: The Van Deemen property is characterized by Precambrian gneiss and schist which is overlain by Middle Tertiary volcanics and sediments. These two rock formations were brought into contact by a regional, low angle detachment fault with its attendant breccia zone. The top and bottom of this zone is in many cases a sharp contact between broken, but not brecciated, volcanic and sedimentary rocks above, intensely chloritized and moderately to poorly broken Precambrian gneiss and schist below. This picture is from what is referred to as area IV – you can see the detachment fault on the bottom with the yellowish sericite alteration, overlain by Tertiary volcanics.
• #12: Gold mineralization at the Van Deemen occurs primarily in gently dipping zones of quartz-sericite-hematite-pyrite clay alteration of brecciated Precambrian gneiss. The alteration zones are spatially associated with rocks generally exhibiting an open style of brecciation. Stacked sheets of quartz breccia are often present in the gold zones, sometimes forming at the fault contact with the upper plate, and other times forming irregular lenses in the faulted gneiss. These quartz breccias often contain mixed fragment types including brecciated chunks of vein quartz. The matrix supporting the breccia fragments appears to be made up of finely pulverized rock flour subsequently replaced by fine-grained quartz. In these quartz breccia zones, sulphides (pyrite and arsenopyrite) occur in and near late-stage fractures. Mineralization consists of very fine free gold with silica.
• #13: There are three distinct gold zones at the Van Deemen prospect; Area II, III, and IV, Each gold zone occurs within the brecciated crystalline assemblage along the Van Deemen Fault. However, it is also evident that all the gold zones have a pronounced northeasterly trend. In Areas I, III, and IV the northeasterly trend of the gold zones appears to be related to a northeasterly-trending high-angle deformation zone, especially in Areas III and IV. Area I is an exploration target and is off the map.
• #14: Gold production on the Van Deemen Mine is believed to date from the 1930s. A section of mineralized rock was mined by open cut. Approximately 350 feet of exploratory adits, along with two shallow shafts, were driven in what is now the central part of the property
• #15: More recently, the Van Deemen area has been actively explored for both copper and gold. Copper exploration was conducted mostly in the 1970's and was directed toward deciphering a highly faulted and sliced Laramide (?) quartz monzonite porphyry copper system. The Van Deemen area again received attention in 1979/1980, but this time as a gold play rather than copper and has essentially been explored for gold since then. As you can see there has been extensive exploration by experienced mining people / companies With all the previous work that’s been done on the property, you’d think there would be quite a large data base to draw from…
• #16: Technology in use in the 80s included Using Cooper Aerial to fly the area Creating a local mine grid based on a known 0,0 point – these were x,y coordinates no projection information High quality detailed geologic mapping at 1”=100’ by Bud Hillemeyer, Jim Faulds Detachment Model represented current thinking at the time. F/W produced computer generated maps with topography, geochemical samples Geochem analysis was usually limited to Au, Ag, Cu, Zn, Pb, W, Sn – in this case we had Au assays and only limited Ag assays from Amselco Sectional Resource hand generated using a 50 area of influence Some of the techniques we will be discussing later were actually developed in the 80s, for example Clark Arnold developed techniques as a consultant for Freeport, and I developed statistical techniques in for Duval during that time.
• #17: Historic data available for this case study at the Van Deemen gold deposit near Kingman, Arizona was incomplete, with missing drill hole locations, assays, drill logs, cross-sections 2 detailed 1’=100’ geologic maps from Fisher-Watt 5 cross sections from Fisher-Watt 140 Drill hole locations Drill logs from 1987 drilling by Fisher Watt Assay Certificates for Au only for all drilling Amselco data Miscellaneous reports and maps Historic resource of 32,000 oz Au
• #18: Cross sections were generated by hand This slide shows 4 of the 5 sections we received hung in space so we could check continuity of lithology
• #19: 2011 Technology relies on computer assistance, everything from data entry, to resource estimation. GIS has replaced a draftsman and light table. GPS units have replaced plane tables and alidades. Multi element Geochem packages are the industry standard. Drilling costs have increased. Price of gold has gone up from $288 in the 1989 to $1500 in 2011. Deposits that were not worth looking at in the 80s are economic at the 2011 gold price. International Star, a publicly traded company on the OTC markets took out an option on the property in 2011. Their focus was to demonstrate that the historical exploration data confirmed that the project merits additional work. Towards that end, they hired GeoGRAFX to create a geological data base, calculate a resource using the 140 drill holes that were available at the time using the same parameters that had been used in the 1980s resource. The 2013 the check resource served to confirm the historic resource, ILST elected to move forward with the project.
• #20: In 2013 they obtained over 1000 pulps from the 1986 drilling and surface sampling. The pulps were used to confirm the historic drill hole assays. ILST also collected a bulk sample to check historic metallurgical results. In 2015, additional information was recovered from the Arizona Geological Survey Archives. This included reports, geochemical, drill hole location maps for the 1986 and 1989 holes which added 44 holes to the data base. With the additional information, ILST was able to regenerate the resource, engage a mine engineer Joe Bardswich who has worked with several other projects in the Black Mountains, to create a preliminary mine cost analysis. Is the project economic? At $1200 Au yes. Based on that information ILST commissioned a SEC Guide 7 Technical Report as an aid in raising funds to move the Van Deemen towards production.
• #21: This is the block model with the additional drill holes showing a 0.014 cutoff. You can see that quite a bit of the deposit is either at or near surface. In 2017 ILST defaulted on the claims. 3 of us that had worked on the project for ILST elected to move the project forward.
• #22: What do we have and how do we move the project forward?? We have a new company made up of geologist, mine engineer, and business manager Incomplete technical report In-situ resource has increased contained ounces to 97,000 oz. No geology. The reason we could increase the resource was based on the additional drilling information and the variography showed the continuity of 65 feed in a horizonal direction and 35 feet in the vertical rather than the 50 feet that had been previously used Deposit inside ACEC area which requires a mine plan of operations for any additional drilling Missing data for 100 Kunkes holes and unknown number drilled by Red Dog and Frisco No drill logs for 1986 drilling 5 cross sections Over 1000 pulps from 1986 drilling and surface sampling
• #23: N&B is Looking for funding to move towards production Need to Finish Technical Report Find a way to include geology in the resource Use drill hole pulps to assign lithologic signatures to samples No map with surface sample IDs found so ignore surface sample pulps Use additional information to create geologic model Assign Munsell color signature to pulps
• #24: We focused on the pulps Pulps were rebagged prior further work 932 Samples + Standards and Blanks analyzed using Olympus Delta Handheld XRF in Geochem Mode 38 elements analyzed
• #25: The Munsell Code of was designed as a visual color system for classification of rocks, sediments and soils. Steve van Nort had found this to be useful at Picacho. Took picture of each pulp sample with smartphone Used Soil Analysis Pro app to assign color to the sample Created db with ddh info + color
• #26: The main goal of the statistical analysis was to find a way to use the Geochem analysis from the pXRF to classify the data by known lithology. I used ioGAS™ software developed for geochemical data analysis. It’s advantage over standard statistical packages such as SPSS, Statistica, or excel spreadsheet add-ons is that it has additional useful tools such as Classification diagrams, and the tool I was interested in specifically Discriminant Projection Analysis. It also reads pXRF data and integrates with the discover software platform I was using to manage the data. As with any statistical applications, there are certain rules that must be followed to optimize your results. For example, in order for a data set to be use for statistical analysis it is assumed to be taken from a random population, normally distributed, and it is assumed that the sample set we are working with represents the population we are describing.
• #27: 24 elements with most values above detection limit (including Au from 1980s assays) The first thing to look at is the data distribution. It is common for elements such as Au, Cu, Zn, etc to show a lognormal distribution, while elements such as Al, Ca, K, Na show a normal distribution. Looking at the distribution – that is apparent here – the histograms are colored to reflect known lithology in our sample set with the black values representing unclassified lithologic values.
• #28: This slide shows the elements transformed to log values and you can see an improvement in the distribution of some of the elements, for example Cu. Based on a visual inspection and skewness/kurtosis values, 16 of the elements were transformed to log values for further work. Log of Au, S, Ti, V, Cr, Mn, Fe, Ni, Cu, Zn, As, Sr, Y, W, Pb, Th
• #29: Next, I created probability plots of the elements. What I’m looking for here is an indication of the number of populations I may be dealing with and what’s happening with the left and right tails of the plots. If the data is from a single population the graph should be a relatively straight line, if there are bends or breaks in the line then it gives me an idea that there may be several population present. Flat line tails on the left side show me the number of samples at or below detection limit. Scatter at the top of the right tail may indicate an additional population, possibly related to mineralization.
• #30: Discriminant Analysis technique creates a set of rules to assign a sample to one of a set of groups. In order to come up with these 'rules', a training dataset is used for which the group memberships are known ahead of time 14 drill holes from the 1986 drilling with lithology derived from cross sections – training set 289 lithology values 4 lith classifications, Qt(3-elim), Tba(8), Pca(211), Pcs(67) Qt – Quaternary gravels Tba – Tertiary volcanics Pca – PreCambrian acidic gneiss, metarhyolite Pcs – PreCambrian mixed gneiss schist
• #31: Discriminant Projection Analysis is similar to Principal component analysis in that it tries to maximize the variance between groupings, except that here we already know what our grouping variables are going to be – in this case lithology. The projection step uses the training data to calculate linear discriminant functions, which are linear combinations of the original variables that maximize the differences between the predefined lithology groups. Discriminant Projection 1 (DP1) calculation is a combination of elements that explain the majority of the variance in the population, DP2 explains the next amount of variance and so on. It creates a value for each Discriminant Projection - in this case DP1 and DP2 for each data point. These functions allow the samples to be plotted in the discriminant space so that group separation can be visualized and investigated. UR the graph in the upper right hand of the screen displays a scatter diagram of Discriminant Projection Axes 1 vs DP 2 – Tba samples are in blue, Pca samples in red, and Pcs in orange. You can see that the Tertiary volcanics separate well (except for one lone sample) from the Precambrian units, while in the Precambrian units there is some overprint of the acidic and gneissic units. UL – displays the same graph with the classification colored in the background, so for any new sample that falls in the blue area would be classified as Tba LL – Probability plot of all data colored by lithology – in the case DP1 you can see Tba samples shown differently than the Precambrian units LR - A split probability plot subdivides the data according to the color attribute groups (lithology) and plots a normal probability plot for each group for the selected variables, on the same diagram (with the respective color from the attribute dialog).  You can see in the case of DP2 it’s doing a better job of separating the lithologies
• #32: Once we had the calculations from the training set we applied those calculations to the other samples ioGAS classified 640 samples with unknown lithologies into 3 groups based on the training set parameters Tba – 239 items Pca – 145 items Pcs – 256 items
• #33: Scatter diagram showing DP1 and DP2 on the x,y axes similar to what was shown in the training set. As you can see it did a fairly decent job of classifying the unknown values, however there is still a bit of scatter with the samples in each grouping
• #34: Split probability plot of DP2 and again you can see that there is a visual difference between the lithologies
• #35: So now I’m fairly comfortable with the results, however I do need to check the results. 640 samples were checked against known database of lithologies, Munsel color designation Discriminant Function only classifies against groupings it knows, if there is a lithology that is not in the training set it will continue to classify it as best it can. Several holes showed discrepancies – Hole 31 in area II near QMP intrusive, the pulp were a relatively light gray in color so I’m thinking that those values could be from the intrusive Area I Pca/Pcs order reversed. Those are exploration holes outside the resource area, possible new lithology? In all, 32 drill holes were added to the data base That data was merged with the existing lithologic data base. So now I have a database with additional lithologies that I can use to create a geologic model
• #36: It seems that everyone wants to jump onto the implicit modelling bandwagon. Implicit modelling uses mathematical tools to derive the model from the data. A definition from the micromine web site says that An Implicit Model is a continuous mathematical representation of an attribute across a volume.. "This method is not only efficient, it eliminates the personal perceptions and interpretations of geologists because it is a numerical process that is free of bias, which means less subjectivity and greater reliability." Traditional modelling methods (those relying heavily on manual digitizing) as ‘explicit’. Implicit modeling techniques were used to create a 3D geologic model of the deposit. We used Leapfrog to modeled Area III and IV separately from Area II. Those results were then tied back to historic observations and existing cross sections to confirm the validity of the model.
• #37: We modeled just 3 lithologies, the Tertiary upper unit shown here in green, the detachment fault shown in red, and the PreCambrian basement shown in brown. You’ll see that there are also yellow quaternary units displayed on some of the drill holes. We started out using only the surface geology, and drill holes to create the model. We did include strike and dips from the historical mapping. We then cut sections along the same section lines as the 4 historical sections we had previously recovered. The results showed marked dissimilarities in contacts between units To resolve the discrepancies, we digitized the boundaries on the sections back into the model. The result was more in line with what was encountered in the drill logs, sections and diagrams in reports.
• #38: Area II was more complex to model. We had the same lithologic units as in Areas III and IV, but we also had faulting that ran thru the area. We modeled the fault blocks separately. Once we’d resolved the issues involved with merging the surface geology with drill holes, leapfrog did a good job of handling the fault blocks.
• #39: As of September 1 that’s what we’ve done up to this point to move the Van Deemen forward. There is still more to do. our goal is to move the project forward towards production. As far as science goes, I need to see if we can tie the Munsell color assignments from the pulps to lithology to gold values to predict gold mineralization. This is something that Steve Van Nort was able to do at Picacho I’d like to do additional multivariate work with the Geochem data to see if I can get alteration signatures and test if it’s important to mineralization. I need to update the resource to include the geologic domains. Once that is complete, we can create a preliminary Pit design – from there it involves the mine engineer to move the project forward
• #40: Try to find as much information as possible so you don’t have to regenerate it – as we said earlier, – Thirty years ago, costs for reverse circulation drilling would be in the range of $8.50/foot – now it can cost $35/foot or more and there may now be access and permitting issues. The AZGS provides a wonderful service with it’s document repository. Statistical techniques need to follow standard practice. Treat each data set separately, they are different Check your data to confirm results - this applies to everything from data entry, to modeling to resource estimation Implicit modeling needed addition of historical cross sections to refine the model While these techniques provide tools to aid in interpretation, i.e. boot leather on the ground, they are not a replacement for field observation and understanding the geology and mineralization of the deposit.
• #41: Thank you for your time. For additional information or questions please feel free to contact me at 520 275-6173 or bcarroll@geografxworld.com