Predictive Mapping of Gold Potentials Bay of Exploits

Predictive Mapping of Gold Potentials

Bay of Exploits, Central Newfoundland

GIS Spatial Analysis Applied to Exploration Geology

Michael B. Regular, P.Geo.

Post-Diploma GIS Applications Specialist Program College of the North Atlantic, Corner Brook Campus

June 16, 2014

Predictive Mapping of Gold Potentials, Bay of Exploits, Central Newfoundland: GIS Spatial Analysis Applied to Exploration Geology

Michael B. Regular, GS3210 Capstone Project Report, June 2014 i

TABLE OF CONTENTS ABSTRACT 1. PREFACE 1. INTRODUCTION 2.

Purpose 2. Background 2. Mineral Predictive Mapping 3.

GEOLOGICAL SETTING 4. Area of Interest 4. Regional Geology 4. Gander Zone 5. Notre Dame Subzone 5. Exploits Subzone 5. Overlap Sequences 6. Intrusive Units 6. Gold Occurrences and Models 7.

METHODOLOGY 7. Data Sources 7.

Software 8. Site Suitability Method 8. Modelling Considerations 9.

EVIDENCE LAYERS 10. 16 Geochemical Layers 10. 3 Geological Layers 10. Gravity Layer 11.

Magnetics Layer 11. Potassic Alteration Layer 12. Lineaments Layer 12.

SITE SUITABILITY PREDICTIVE MAP 12. EVALUATION 13. Gold Mineral Occurrence Files 13. DISCUSSION AND CONCLUSIONS 13. Limitations of Performance 13.

Improving Performance 15. ACKNOWLEDGEMENTS 15. BIBLIOGRAPHY 16.


Michael B. Regular, GS3210 Capstone Project Report, June 2014 ii

LIST OF FIGURES

Figure 1. Gold Deposit Models 3. Figure 2. Project Location and AOI 4. Figure 3. Lithology of the Bay of Exploits 5. Figure 4. Tectonic Subdivisions and Gold Mineral Occurrences of the Bay of Exploits 6. Figure 5. Geoprocessing Model 8. Figure 6. Stacked Map Illustration of Site Suitability 9. Figure 7. Predictive Map of Gold Potentials for the Bay of Exploits 14.

LIST OF TABLES Table 1. Classification of the Lithology Evidence Layer 10. Table 2. Classification of the Surficial Geology Evidence Layer 11. Table 3. Classification of the Lineaments and Plutonic Haloes Evidence Layers 12. Table 4. List of Gold Hotspots from the Predictive Map. 13.


Michael B. Regular, GS3210 Capstone Project Report, June 2014 Page 1

Predictive Mapping of Gold Potentials,

Bay of Exploits, Central Newfoundland:

GIS Spatial Analysis Applied to Exploration Geology

Michael B. Regular, B.Sc., P.Geo.

ABSTRACT

The Bay of Exploits, a region of Notre Dame Bay in north central Newfoundland, contains gold mineralization in at least two

known deposit types; Volcanogenic Massive Sulfide (VMS) related deposits and quartz shear/vein hosted hydrothermal

deposits. The Bay of Exploits covers a large area of Newfoundland’s north coast requiring a large amount of traditional

exploration to evaluate. However, the targeting of prospective areas using GIS to define gold potentials is possible. Modern

GIS methods such as Spatial Analysis and Site Suitability has allowed for the gold potential of this area to be delineated and

provide prospective target areas for renewed mineral exploration. Like many GIS spatial analysis methods this technique is

best applied to areas that have good quality data and allow for the most significant outcome. However, it is understood that

at a regional scale this method does have its limitations and pitfalls. The geoprocessing model developed can be applied to

other areas of similar geology or with modification for other commodities and/or depositional models.

Twenty-three evidence layers, including sixteen geochemistry based pathfinder layers from lake sediment and till surveys,

structures from manually generated lineaments maps, total field magnetics, gravity, potassium alteration from radiometric

data, and three geology layers were utilized. The results were validated by comparison with known surface gold occurrences

to determine, in a spatial sense, what areas are positively predicted to host gold mineralization. Combining these twenty

three evidence layers using the site suitability method within a GIS produced a predictive map for gold within the Bay of

Exploits area.

Ten areas were delineated as having a high potential for gold mineralization. These areas are defined as being within the top

two Standard Deviations of eleven resultant classes. These areas coincide with, or are in close proximity to, over forty of the

one hundred sixteen known gold mineral occurrences. This outcome verifies that the site suitability method of generating a

predictive map can define promising areas for gold deposition and new targets for exploration within the Bay of Exploits.

PREFACE

Prospecting for gold in Newfoundland and

Labrador has traditionally been a boots on the

ground exploration activity. In recent years,

with the increased functionality of both

computers and GIS software packages, the

prospecting focus has shifted, in part, to the

desktop. With the release, by the Newfoundland

and Labrador Geological Survey, of numerous

datasets collected and compiled from private

exploration activities new GIS studies seems

only logical. Any tools of the geosciences that

can assist with identifying new mineral

deposits, at a reduced cost, are most welcome,

especially during times when markets are weak.

In the exploration industry, especially among

prospectors who do grass-roots exploration and

find the majority of deposits, a faster and

cheaper means of isolating exploration targets

is paramount. With this knowledge in mind a

project developed around a single commodity,

gold, for a portion of the province known to be

prospective by traditional means is an

opportunistic test and starting point for this

GIS study.



INTRODUCTION

Purpose:

This GIS project fulfilled four needs;

1. It completed a requirement of the GIS

Application Specialist program for a final

capstone project highlighting an understanding

of the GIS skills encountered throughout the

course of study.

2. It demonstrated a predictive mapping

methodology, utilizing spatial analysis and site

suitability analysis, as it can be applied within

the realm of exploration geology.

3. It evaluated the GIS site suitability method as

applied to predictive mapping for gold within

the Bay of Exploits, Notre Dame Bay,

Newfoundland and identified the method’s

limitations. Recommendations for more refined

results were discerned.

4. It identified areas within the Bay of Exploits

region that are predicted to have an increased

potential for hosting gold deposits.

Background

Mineral exploration is conducted to discover

deposits with economically viable

concentrations of minerals and metals, for the

end purpose of mining. The four main stages of

mineral exploration are: area selection, target

generation, resource evaluation and the

definition of reserves (Carranza, 2009). GIS can

be a very powerful and useful tool for Geologists

during the target generation phase of

exploration. The focus of this initial phase is to

define prospective areas for further

investigation into the potential of a workable

mineral deposit (Bonham-Carter, 1994). The

required geological mapping process typically

takes place over several scales, ranging from the

regional to the property/deposit scale. The

geological method, often incorrectly applied,

involves a complex modeling process, with

numerous factors controlling if and where a

deposit might form. This inevitably requires the

input and analysis of additional geological,

geophysical and geochemical data (Bonham-

Carter, et al 2000).

This project generated a regional-scale gold

potential map, using the site suitability method

within ArcGIS. Using freely available geological

data from the Newfoundland and Labrador

Department of Natural Resources

(Newfoundland and Labrador Geological

Survey) in conjunction with a current

knowledge of natural controls influencing the

presence of gold deposits in the region a

predictive map for the presence of new gold

mineralization in the Bay of Exploits region was

generated.

Gold deposits and showings in the Central

Newfoundland area are associated with Cu-Pb-

Zn deposits (Volcanogenic Massive Sulfides

(VMS)) such as those of Tally Pond, Buchans,

Rambler and Tilt Cove or hosted in later

occurring shear sets with hydrothermal quartz

veining like those at Nugget Pond, Hope Brook,

Pine Cove and the Hammerdown deposits.

GIS based predictive mapping for the Bay of

Exploits area utilizing GIS and Site Suitability

suggest new areas of exploration interest and

new exploration methodologies. Collecting,

organizing and analyzing currently available

geological, geochemical, geophysical and

structural datasets provided invaluable insight

into the best practices for conducting such a site

suitability analysis. Outcomes will assist in

developing future GIS project plans applied to

other geological regions and for other valued

commodities.

Geological features including lithology,

structure and stratigraphy are known to control

the presence of gold deposits in Newfoundland

(Sandeman, 2010). Each feature layer is given a

weighted “score” based on their ascertained



influence with respect to known gold deposit

types encountered for the Bay of Exploits

region. Within the predictive map the highest

scores indicate a greater likelihood of gold

mineralization. The final predictive map was

compared to known gold mineralization for the

Bay of Exploits region to gauge whether the GIS

model is successful. A discussion of the GIS

spatial analysis limitations is included following

a review of validation results. The final map and

methodology are useful to prospectors and

exploration companies as a general predictor of

deposit location on the regional scale and allow

for targeted field planning. It must however be

understood that these maps are based on a

simplified conceptual and regional GIS model,

and will need further refining in order to

increase overall accuracy and gold target

definition. Like many tools used in mineral

exploration it cannot be used alone.

Mineral Predictive Mapping

Predictive mapping within a GIS can utilize any

number and combination of spatial analysis

techniques from Weights of Evidence statistical

methods, Fuzzy Logic methods, Site Suitability

methods and a host of others. The basic premise

is to use available spatial data to produce a map

that indicates areas that are most likely to

contain economic concentrations of the metal

or commodity being explored. These predictive

maps, which are a result of spatial data analysis

and modeling provide the explorer sound

statistical information for financial and

tenement management decision making.

Geological, geochemical, and geophysical

exploration data in conjunction with spatial

data modelling techniques were used to create

predictive maps that represent aspects of a

particular mineral system as defined by the

mineral depositional model. Predictive maps

are made up of two or more classes that will

have either a positive or a negative association

with the mineralization.

Spatial data modelling is one of the best

techniques available to assess mineral

prediction, it allows for the combination of all

important predictive variables related to a

mineral deposit model (Figure 1.) into one map.

The model is also based on statistics, this means

that it is not bias to previous ideas or current

exploration trends. Instead the model is based

on what's been measured on the ground and

which of these measurements are most related

to the mineralization model.

Although the genesis for the formation of an ore

body can be simplified to geology,

geochemistry, and geophysics the combination

of predictive maps that can be created from

base datasets are many and varied. The

predictive maps must have some relationship to

the processes that formed the mineral deposits

in question.

Figure 1. Gold Deposit Models (After Sillitoe and

Bonham, 1990 and Hannington et al, 1999).

Regional scale geological mapping and

geophysical data sets are excellent data sources

for modelling as they provide continuous data

coverage, minimizing problems associated with

missing data. Point geochemical data are also

valuable and need to be analyzed for anomalous



geochemical associations before they can be

used in spatial data modelling. Most of the data

used here is from historical exploration and is

freely available through government Geological

Surveys. This data is often available in a digital

format ready for use in GIS and spatial

modelling.

The geoprocessing model allows you to plan for

the new and or more detailed data that can be

collected for the prospective areas. The model

can be rerun to assess the effectiveness of the

new data in enhancing the prospectivity of the

area being tested.

GEOLOGICAL SETTING

Area of Interest

The area chosen for this study was based upon

nine coincident 1:50,000 scale NTS sheets

(Figure 2.).

These NTS Sheets are;

002E01 (Weir’s Pond)

002E02 (Gander River)

002E03 (Botwood)

002E06 (Point Leamington)

002E07 (Comfort Cove-Newstead)

002E08 (Carmanville)

002E09 (Fogo)

002E10 (Twillingate)

002E11 (Exploits)

This area was chosen for a number of reasons

related to testing the site suitability GIS method

and include; a range of land coverages (forested

areas, islands and ocean), a varied geological

base, and largely because of currently available

spatial data.

Figure 2. Project Location and AOI.

Regional Geology

The underlying regional geology for the Bay of

Exploits area within the project AOI consists of

five broad tectonic subdivisions (Figure 5.);

Dunnage Zone - Notre Dame Subzone

Dunnage Zone - Exploits Subzone

Gander Zone

Overlap Sequences

Intrusive units.

These subdivisions span the Early Cambrian

(542 Ma) to Late Devonian (360 Ma) age with

some later intrusions during the Middle

Jurassic to Early Cretaceous (162-131 Ma).



Figure 3. Lithology of the Bay of Exploits.

Gander Zone

The Gander Zone represents the Gondwanan

Margin, an ancient continental margin similar

to that seen as our Grand Banks today. The

underlying lithologies consist of siliciclastic

marine sediments of the Jonathan’s Pond

Formation and some very minor intermediate

intrusions. The Gander Zone forms the

southeast portion of the project area and some

962 km2.

Notre Dame Subzone

The Dunnage Zone - Notre Dame Subzone

represents some of the ancient Iapetus Ocean

island arc volcanic sequences like those of

Japan today. The underlying lithologies consist

of mafic and felsic marine volcanics with

siliciclastic marine sediments and minor chert,

felsic plutonics (Twillingate Pluton), mafic

hypabyssal dikes, and a mélange. These

volcanics belong to the Chanceport, Cottrell’s

Cove, Morton’s Harbour and Sleepy Cove

Groups. The Notre Dame Subzone forms the

north edge of the project area where the geology

disappears under Notre Dame Bay with a total

area of 229 km2, 37 km2 of Twillingate Pluton

and 143 km2 of Volcanic units.

Exploits Subzone

The Dunnage Zone - Exploits Subzone

represents another portion of the ancient

Iapetus Ocean island arc volcanic sequences.

The underlying lithologies consist of mafic and

felsic marine volcanics with siliciclastic marine

sediments, minor chert, limestone, ultramafic

to felsic plutonics, and mélange. These marine

volcanics and related marine sediments belong

to the Badger, Davidsville, Duder, Exploits,

Gander, Hamilton Sound, Summerford and

Wild Bight Groups as well as the Boone’s point,

Gander River, Phillip’s Head, and South Lake

Igneous Complexes. The Dunnage Mélange is a

centralized geological feature under the Bay of

Exploits and a number of smaller black shale

and plutonic sequences are noted. The Exploits

Subzone forms the largest portion of underlying

lithologies for the project area with a total area

of 2233 km2. Marine siliciclastics have an area

of 1475 km2, the Dunnage Mélange 210 km2,

marine volcanics 205 km2, plutonics at 188 km2

and black shales at 154 km2.

In geology, a mélange is a large to regional-scale

breccia, a mappable body of rock characterized

by a lack of continuous bedding and the

inclusion of fragments of rock of all sizes,

lithologies and from varied sources. Large-scale

melanges typically form in active continental

margin settings. The mixing mechanisms in

such settings may include tectonic shearing

forces, ductile flow of a water-charged or

deformable matrix (such as serpentinite),

sedimentary action (such as slumping, gravity-



flow, and olistostromal action), or some

combination of these.

An olistostrome is a sedimentary deposit

composed of a chaotic mass of heterogeneous

material, such as blocks and mud, known as

olistoliths, that accumulates as a semifluid body

by submarine gravity sliding or slumping of the

unconsolidated sediments. It is a mappable

stratigraphic unit which lacks true bedding, but

is intercalated amongst normal bedding

sequences. Olistostromes are mélanges formed

by gravitational sliding under water and

accumulation of flow as a semi fluid body with

no bedding.

Overlap Sequences

The Overlap Sequences represent non-marine

or terrestrial sediments that occur in shallow

marine to non-marine environments. In the Bay

of Exploits project area these sediments

represent the rise of oceanic environments

above sea level during orogenesis (Mountain

Chain Building) and continental collision. The

underlying lithologies consist of non-marine

mafic to felsic volcanics, non-marine

siliciclastics and minor shallow marine

siliciclastics, shales and limestone. Much of

these volcanics and sediments belong to the

Botwood and Indian Islands Groups. The

Overlap Sequences form a central corridor

portion of underlying lithologies for the project

area with a total area of 890 km2 and non-

marine volcanics at 186 km2.

Intrusive Units

Intrusive units are represented by hypabyssal to

small plutonic plugs through to batholith and

layered intrusive scale bodies, typically the

youngest units within the project area. These

lithologies comprise ultramafic to felsic

plutonics with minor hypabyssal equivalents.

Some older units may represent magmas

chambers below volcanic centers while others

may be termed suture plutonics. These plutonic

rocks are seen as the Fogo batholith, Hodges

Hill Intrusive Suite, and Mount Peyton

Intrusive Suite (Layered Mafic Intrusive in

part). Smaller bodies occur throughout the

project area. The youngest units for the project

area are the Budgell’s Harbour and Dildo Pond

plutons at 162-131 Ma. Plutonic bodies are

scattered around the project area with a total

area of 750 km2 with Mount Peyton at 440 km2,

Fogo at 214 km2, Hodges Hills at 14 km2, and all

others combined at 282 km2.

Figure 4. Geology and Gold Mineral Occurrences of the

Bay of Exploits.



Gold Occurrences and Models

Mineral Occurrence data for the Bay of Exploits

region consists of 240 records with 6 indicated

as past-producers, 2 developed prospects, 85

indications, 21 prospects, and 119 showings. Of

these occurrences 30 occur within the VMS

range of deposit types and 116 occur as gold

occurrences. 35 of these gold hits are described

as having insufficient data to classify; the

remainder, are described as being structurally

controlled vein systems. Within the 116 mineral

occurrences there are 87 noted as being

showings, 17 as prospects, 11 as indications and

one as a past producer (Little Harbour Mine,

Twillingate Island).

The underlying lithologies associated with these

116 gold mineral occurrences include 19 within

the Dunnage Melange, 17 within Badger,

Exploits and Davidsville Group sandstones, 16

within Mortons Harbour and Summerford

Groups, 16 within the Mount Peyton Intrusive

Suite, Thwart Island Gabbro and Gander River

Complex, and 11 within Badger Group

Conglomerates. The remaining occurrences lay

within siliciclastic marine sediments or rarely

within non-marine sediments (Figure 3.).

METHODOLOGY

Geoprocessing spatial data using the Site

Suitability method involved the design of

multiple evidence layers focused on

determining the prospectivity of gold. Each

evidence layer was converted to a raster image

of equal cell size (20m X 20m) and was

classified based on the relevance of each class to

gold deposition. For the Bay of Exploits AOI

this produced 23 evidence layers form 5 vector

datasets and 3 raster/gridded datasets.

Combining evidence layers using weighted sum

geoprocessing tool was based upon weighting

related to gold depositional models, this

produced the final predictive map.

Favorable criteria utilized for the prediction of

gold around the Bay of Exploits was limited to 8

specific pathfinder elements from both lake

sediment and till samples, total field magnetic

data, bouguer gravity data, potassium

radiometric data, and data pertaining to

lineaments (faulting/fracturing), geology-

lithology, surficial geology, and metamorphic

haloes surrounding intrusive bodies.

1. Input Datasets

3 raster images and 6 vector files for the Bay of Exploits.

2. Derived Datasets

Resample Bouguer Gravity, Total Field Magnetics and Potassium Radiometrics to 20m cell size. Lineaments polyline and Plutons polygon vector files are buffered. Geochemical point vector files create 8 elemental files and are Inverse Distance Weighted. Geology and Surface Geology polygon vector files are categorized. All vector files are then converted into raster files with 20m cell size.

3. Reclassified Datasets

Each of the 23 derived datasets are reclassified to 9 classes based on the strength of each class and its importance to gold deposition.

4. Site Suitability Predictive Map

All 23 reclassified datasets are assigned a weight based on the strength of the dataset for predicting gold deposits. These datasets are run through the Weighted Sum tool to result in a final site suitability gold predictive map.

Data Sources

GIS vector data for this predictive study

includes a number of recently updated (2014)

datasets available for download from the

Newfoundland and Labrador Geoscience Atlas



(geoatlas.gov.nl.ca) and as a minimum

included;

Detailed Bedrock Geology (NFLD2616

v.7)

Regional Surficial Geology

Mineral Occurrences

Regional Lake Sediment Sites

Till Sediment Sites

Topographic vector data was downloaded from

the Natural Resources Canada (CANVEC) site

(geogratis.gc.ca/geogratis) including all 9 NTS

sheets as listed above in the Area of Interest.

The NTS index vector data was acquired from

the College of The North Atlantic datasets.

Raster data was sourced largely from the

GeoBase website (www.geobase.ca) and

included;

CanDigElevData (DEM) for each of the 9

NTS sheets at 20m resolution

LandSat L7 imagery (003025 &

003026) at 30m resolution

SPOT Imagery (05347, 05410, 05418,

05430, 05548, 05455, 05507, and

05515) at 20m resolution.

Gridded spatial data was sourced at Natural

Resources Canada via the Earth Science Sector

webpage

(gdr.agg.nrcan.gc.ca/gdrdap/dap/search-

eng.php). This included;

Gravity data at 2000m resolution for

Island of Newfoundland

Magnetic-Radiometric-EM data for the

North Central Newfoundland area at

200m resolution.

Software

For this project the Windows 7 operating

system was utilized with the Microsoft office

suite. GIS software included ArcGIS 10.2 for

Desktop including ArcCatalog, ArcMap,

ArcGlobe and ArcScene. This was a fully

functional version including all licensed

modules. PCI Geomatica 2013 was used to best

manipulate raster imagery where necessary.

Within ArcGIS 10.2 ModelBuilder is an

interface used to create, edit, and manage

geoprocessing models (Figure 5.) which are

workflows that string together sequences of

geoprocessing tools, feeding the output of one

tool into another tool as input. ModelBuilder

can also be thought of as a visual programming

language for building workflows that can be

shared and modified for use in future projects.

Figure 5. Geoprocessing Model.

Site Suitability Method

Within ArcMap there are numerous tools that

allows for spatial analysis and site suitability to

be represented in the form of a model. This

model visually portrays what spatial files are

being utilized and created and what geospatial

tools are being used in a flow diagram layout.

For this Bay of Exploits gold predictive mapping

project the model started with nine spatial files

(6 vector and 3 raster) and utilized over 15

tools.

In the end the model consisted of over 90

spatial files, both vector and raster and over 85

tool uses to compile the final predictive map.

http://www.geobase.ca/



Site Suitability Analysis in a GIS context is a

geographic or GIS-based process used to

determine the appropriateness of a given area

for a particular use. The basic premise of GIS

site suitability analysis is that each aspect of the

landscape has intrinsic characteristics that are

in some degree either suitable or unsuitable for

the activities being planned. Site suitability is

determined through systematic, multi-factor

analysis of the different aspect of the terrain.

Model inputs can include a variety of physical,

cultural, and economic factors. The results are

often displayed on a map that is used to

highlight areas from high to low suitability.

A site suitability model typically answers the

question, Where is the best location?—whether

it involves finding the best location for a new

road or pipeline, a new housing development,

or in this case the most prospective ground for

gold deposition. ArcGIS Spatial Analyst derives

new information from the overlay of multiple

layers, which can then be used to determine the

best location.

Modelling Considerations (Favorable

Criteria)

The Site Suitability method utilizes multiple

evidence layers based on a single theme to

determine a final output. These layers are

converted to raster images of equal cell size

then through weighting and addition reveal

areas that have incurred more favorable

evidence responses. Layers used depend on the

project at hand and can consist of spatial data

from just about any source. However, only the

data that is both unique and directly related to

the project characteristics should be used. Too

many redundant or non-related layers will only

create a less than ideal result and poor

outcomes.

Each evidence layer generated is reclassified to

a specific project parameter to include an equal

number of classes. This project utilized a total

of 9 classes. Each class is based on Standard

Deviation statistics and reviewed to be assigned

a best value of 9 downward to a low value of 1

sometimes hosting the No Data as a value of 1.

Favorable criteria for this project was, in part,

limited to specific pathfinder elements from

both Lake Sediment and Till samples. Magnetic

data, Gravity data, Radiometric data and data

pertaining to Lineaments (faulting), Geology-

Lithology, Surficial Geology, and Metamorphic

Haloes surrounding intrusive bodies completed

the evidence layer list.

Figure 6. Stacked Map Illustration of Site Suitability.

With all 23 evidence layers properly classified

into 9 classes each, these images were combined

using the Weighted Sum tool to generate the

final predictive map. Within this Weighted Sum



tool a weight was assigned to each evidence

layer based on its importance to delineating

gold deposit favorability.

EVIDENCE LAYERS

Predictive mapping of gold potentials in the Bay

of Exploits area using site suitability generated

a total of 23 evidence layers. These evidence

layers are essentially classified raster images

over the same area of interest with the same cell

size. A generalized review of these evidence

layers and their generation are provided below.

Using the NL_NTS50K NTS map index polygon

file an area of interest (AOI) was created to use

for clipping purposes. From this file the NTS

sheets 002E/01, 02, 03, 06, 07, 08, 09, 10, and

11 were selected to create the Exploits_NTS file,

this file was then merged to form the NTS_AOI

file and Dissolved to form the final

Exploits_AOI file used for clipping.

16 Geochemical Layers

Both Till Sediment and Lake Sediment point

data vector files were sourced from the

Geoscience Atlas. These datasets both consisted

of multi-element analysis for each point source.

Following a review of pathfinder elements usage

in gold exploration, and the statistics of these

elements within each dataset, it was decided

that 8 elements would be used from each

dataset. These 8 elements consisted of Gold

(Au), Arsenic (As), Barium (Ba), Copper (Cu),

Lead (Pb), Zinc (Zn), Antimony (Sb), and Iron

(Fe).

The NL_LakeSediments and NL_TillSediments

vector files were first clipped with the

Exploits_AOI to create the EX_LSeds and

EX_Tills files. Using the select tool to isolate

each of the 8 elements LSed and Till elemental

files were generated. All 16 elemental files were

run using the IDW (Inverse Distance Weighted)

tool to create contoured images based on

elemental assay values with a 20m cell size.

Each of the 16 elemental files was reclassified to

9 classes with the highest values based on

standard deviation within class 9. This created

16 reclassified elemental images, with 8

elements for each of the tills and lake sediment,

to be used in final weighting.

3 Geological Layers

From the NFLD2616v7 Newfoundland Detailed

Geology vector file and the NL_Surficial

Geology vector file, both sourced from

Newfoundland and Labrador Department of

Natural Resources via the Geoscience Atlas, 3

geological layers were created.

Table 1. Classification of the Lithology Evidence Layer.

Lit h ology Cla ssifica t ion

Ca r bon a te lim eston e 1

Ch er t 1

Hor n fels 1

Sedim en ta r y 1

Silicicla st ic n on -m a r in e con g lom er a te 1

Hy pa by ssa l m a fic 2

Silicicla st ic a r g illite 2

Silicicla st ic n on -m a r in e 2

V olca n ic in ter m edia te n on -m a r in e 2

Hy pa by ssa l in ter m edia te 3

Hy pa by ssa l u ltr a m a fic 3

Plu ton ic 3

Silicicla st ic con g lom er a te 3

Plu ton ic in ter m edia te 4

Silicicla st ic 4

V olca n ic felsic m a r in e 4

V olca n icla st ic m a fic 4

Silicicla st ic n on -m a r in e sa n dston e 5

V olca n ic felsic n on -m a r in e 5

V olca n ic m a fic n on -m a r in e 5

V olca n ic n on -m a r in e 5

Hy pa by ssa l felsic 6

Plu ton ic felsic 6

Plu ton ic m a fic 6

Plu ton ic u ltr a m a fic 6

Silicicla st ic bla ck sh a le 7

Silicicla st ic m a r in e sa n dston e 7

Silicicla st ic m a r in e sh a le 7

Mela n g e 8

Silicicla st ic m a r in e con g lom er a te 8

Silicicla st ic m a r in e 9

V olca n ic m a fic m a r in e 9



The Clip tool was used to reduce the size of the

NL_SurficialGeology file to the Exploits AOI

creating the Exploits_SurfaceGeology file. This

Exploits_SurfaceGeology file was converted

into a raster with 20m cell size generating the

Surface_Geo file. The Surface_Geo file was then

reclassified to 9 classes based on the

importance of each of the 11 classifications

provided. (Table 2.)

Table 2. Classification of the Surficial Geology Evidence

Layer.

The NL_Geology vector file (NFLD2616v7) was

used to create 3 additional files; the Lithology

file, a non-plutonic mask, and the buffered

halos around plutons where metamorphic

aureoles are known to generate gold veining

fluids and structures. The NL_Geology is first

clipped with the Exploits_AOI to generate the

Exploits_Geology file. This feature class was

transformed into a 20m cell sized raster of

Lithology. This file is reclassified to 9 classes to

create the Lithology file for use in weighting.

(Figure 3.)

This Exploits_Geology attributes were selected

to isolate non-plutons and create a

PlutonicMask for use in the Alteration file

generation.

Another selection was completed to create a

Exploits_Plutons file that was multi-ring

buffered to create the Plutons_Buffer. This

Plutons_Buffer file was converted to a 20m

raster to create the Pluton_Buff file and finally

reclassified to 9 classes to create the

Pluton_Halo file for use in weighting. (Table 3.)

Gravity Layer

The gravity data was sourced through Natural

Resources Canada via the Geoscience Data

collection. The dataset consisted of 7 gridded

data files covering Newfoundland with a 2000

m (2 km) crude point distribution. For this

study the Bouguer gravity file was utilized.


Bouguer gravity file to the Exploits AOI creating

the EX_GRAV file. This EX_GRAV file was

resampled to the project 20m cell size to create

the Boug_20m file which in turn had its

statistics defined with the statistics tool to form

the Boug_20m (2) file. The Boug_20m (2) file

was reclassified to 9 classes with the highest

values making the highest value of 9. This

created the Bouguer file available for final

weighting.

Magnetics Layer

The magnetic data was sourced through Natural

Resources Canada via the Geoscience Data

collection. The Gander-Botwood Radiometric-

Magnetic survey from 1987 was selected to

cover the Bay of Exploits area. The dataset

consisted of 7 gridded data files with a 200 m

crude point distribution. For this study both the

Potassium and Total Field Magnetics were used.


Total Field Magnetics file to the Exploits AOI

creating the EX_MAG file. This EX_MAG file

was resampled to the project 20m cell size to

create the MAG_20m file which in turn had its

Su rficia l Geology Cla ssifica t ion

Bog 1

A llu v iu m 2

Ma r in e cla y , sa n d, g r a v el a n d dia m icton 3

Gla cioflu v ia l g r a v el a n d sa n d 4

Hu m m ocky ter r a in 5

Ridg ed t ill 5

Collu v iu m 6

Till bla n ket 7

Till v en eer 7

Con cea led bedr ock 8

Ex posed bedr ock 9



statistics defined with the statistics tool to form

the MAG_20m (2) file. The MAG_20m (2) file

was reclassified to 9 classes with the highest

values making the highest value of 9. This

created the Total_Field_Mag file available for

final weighting.

Potassic Alteration Layer

From the same source as Total Field Magnetics

noted above the Potassium from the

Radiometric survey of Gander-Botwood area

was used to define potassic alteration.


Gander_Botwood_Pot file to the Exploits AOI

creating the Clip_GammaK file. This

Clip_GammaK file was resampled to the project

20m cell size to create the GammaK file. To

reduce the unwanted potassium signature

associated with felsic plutonics a mask file

created from the NL_Geology was used to

remove those areas to create the Alteration file.

This alteration file was reclassified to 9 classes

with the highest values making the highest

value of 9. This created the K_Alteration file

available for final weighting.

Lineaments Layer (Structural)

Several attempts were made to create a

lineaments map from existing DEM data using

both PCI Geomatica and ESRI ArcMap. Both

resulted in data that did not identify obvious

fault sets and generated a large number of short

lineaments including ridgelines unsuitable for

this project. In the end using the hillshade

image generated from the DEM data lineaments

were entered manually where they were known

to occur and postulated to exist.

Table 3. Classification of the Lineaments and Plutonic

Haloes Evidence Layers.

This Exploits_Lineaments vector file was then

buffered with multi-rings to generate 8 rings at

a spacing of 125m each from linear extents and

created the Lineaments_Buff file. This

Lineaments_buff file was then converted into a

raster image with 20m cell size creating the

Buff_Lines file that was reclassified to 9 classes

with closest buffered ring having the highest

value for weighting.

SITE SUITABILITY PREDICTIVE MAP

Weighting of evidence layers was done logically

through characterization of layers as being fixed

or transported, and a further characterization of

their relevance to gold deposition models. The

following schema was used to define the weight

of each layer.

F1 – Fixed High Relevance

F2 – Fixed Moderate Relevance

T1 – Transported High Relevance

T2 – Transported Moderate Relevance

T3 – Transported Low Relevance

The final site suitability predictive map for gold

within the Bay of Exploits area is presented as

Figure 7. There are a total of 10 regions

containing predicted areas for gold each listed

in Table 4.

Ev idence La y er Linea m ent s Plu t onic Ha loes

Unit s m et ers m et ers

Cla ssifica t ion T y pe Dist a nce Dist a nce

Cla ss 1 NoDa ta NoDa ta

Cla ss 2 8 7 5 - 1 000 8 7 5 - 1 000

Cla ss 3 7 5 0 - 8 7 5 7 5 0 - 8 7 5

Cla ss 4 6 2 5 - 7 5 0 6 2 5 - 7 5 0

Cla ss 5 5 00 - 6 2 5 5 00 - 6 2 5

Cla ss 6 3 7 5 - 5 00 3 7 5 - 5 00

Cla ss 7 2 5 0 - 3 7 5 2 5 0 - 3 7 5

Cla ss 8 1 2 5 - 2 5 0 1 2 5 - 2 5 0

Cla ss 9 0 - 1 2 5 0 - 1 2 5



EVALUATION

Evaluation of the site suitability method and the

generated predictive map for gold in the Bay of

Exploits required a means of validation.

Isolating the mineral occurrence files associated

with gold to see where they plotted in relation

to the predictive map’s prospective areas

(Figure 7.) allowed for this validation. This use

of validation points did not require any ground-

truthing which was not possible for this project

within the given time frame. It is hoped that

maybe an adventurous prospector/geologist

might pick up a prospective lead and discover

some new gold mineralization as a result of the

map and provide additional validation.

Table 4. List of Gold Hotspots from the Predictive Map.

Gold Mineral Occurrence Files

Within the Bay of Exploits AOI there are 240

total mineral occurrence points. Of these 116 are

related to gold directly. One of the occurrences

for gold is a past producer at Little Harbour

Mine, Morton’s Harbour (1897-1900), 11 are

classed as indications(a mineral deposit upon

which no known development work has been

done, and for which there exists only an

"indication" of its existence), 17 as prospects (a

mineral deposit upon which enough

development work has been done to provide

data for the making of a reasonable estimate of

the spatial extent of the deposit, but not enough

to estimate the amount of any commodity

present), and 87 as showings (a mineral deposit

upon which some development work may have

been done, but the extent of such work was not

adequate to provide enough data to estimate its

spatial dimensions).

DISCUSSION AND CONCLUSIONS

There appears to be a moderate to strong

correlation between prospective gold areas on

the predictive map and known gold mineral

occurrences. Of the 116 current gold mineral

occurrences 7 (6%) appear to be related to the

top Classification, 23 (20%) appear to be related

to the second highest Classification, 22 (19%)

appear to be related to the third highest

Classification, and 15 (13%) appear to be related

to the fourth highest Classification.

Limitations of Performance

The generation of the predictive map using the

selected classifications and final weighting does

suggest the methodology is sound, albeit

requiring some modifications for a more refined

result. These modifications would be sourced

from both refined data and expert knowledge of

the gold depositional models for the Bay of

Exploits.

Much of the vector and raster data has

resolution issues that have introduced errors

into the site suitability analysis. Accuracy of the

lithological, surficial geological contacts, point

sources for mineral occurrences, real or

erroneous lineaments all produce

misinformation when geoprocessed. Gridded

data had resolutions greater than that used for

the analysis and needed resampling to be useful

within the site suitability analysis. Resampling

will also produce misinformation especially

when resampling from 2000m to 20m cell

sizes.

Gold Prediction Hot Spots (West to East)

1 Lock’s Harbour – Budgell’s Harbour – Little Northwest Arm

2 Strong Island – Tea Arm – Saunder’s Cove

3 Fortune Harbour New Bay Head

4 Embree – Salt Pond Cove

5 Morton’s Harbour – Hillgrade

6 Duder Lake – Burnt Lake

7 Gander River – Salmon River – Rocky Pond

8 Weir’s Pond

9 Carmanville – Eastern Arm – Shoal Pond

10 Indian Bay – Little Bear Cave Pond – Four Mile Pond



Figure 7. Predictive map of gold potentials for the Bay of Exploits.



The bulk information from Radiometric, gravity

and magnetic surveys was useful but would be

more meaningful at a better resolution. In the

case of VMS style deposits the gravity and

magnetic surveys were more important because

of their magnetic and density properties when

discovered, however for the bulk of vein style

hydrothermal gold deposits gravity and

magnetics are not of value since these deposits

have no magnetic or density properties that can

be diagnostic.

From a more generalized point of view these

gold deposits have been dumped into 2 genetic

models where it is likely many occur. Finding

more expert knowledge about these deposits

might shed new light on how the classifications

could be revamped, producing better results.

Similarly, with only limited data for structures

within the Bay of Exploits, more expert

knowledge about where structures occur, how

they are related to gold mineralization and

especially if certain trends are more important

than others would better define the

classification for lineaments overall.

Improving Performance

The regional nature of this study has inherent

spatial and data issues that result in a varied

model performance. For raster images one issue

is the resolution of 2000m for the gravity

survey and 200m for magnetics and

radiometric surveys. This is partially resolved

by resampling the data using a cubic resampling

technique however images with a finer

resolution would no doubt have yielded a better

result.

For the vector point files like till and lake

sediment sites the survey data is again spaced to

approximately 1000m. Like the raster images a

finer spacing between the data would have

yielded a more precise result. Although

structural geology data like faults and contacts

was available it was not complete resulting in

the need to generate lineament maps to cover

the area of interest.

In future studies using this site suitability

method additional evidence layers could be

used or removed depending on the parameters

of that study. Spending more time to delineate

better lithology classifications, using more or

less classifications overall and utilizing data

with a finer resolution would be paramount.

ACKNOWLEDGEMENTS

This report is a result of a capstone project

required to complete the GIS Applications

Specialist program at the College of the North

Atlantic in Corner brook, Newfoundland and

Labrador during 2013-2014. Help with

computer applications and spatial data analysis

was provided by instructors Darin Brooks,

Richard Wheeler, and Neala Griffin. My

classmates, who have taken similar adventures

over the past months, have to be acknowledged

for being available when times were chaotic and

advice was needed. Spatial and research data

was primarily sourced through the

Newfoundland and Labrador Department of

Natural Resources via the Geoscience Atlas and

literature searches. Additional data was

acquired from Natural Resources Canada and

GeoBase.ca. A Thank You to these data sources

for information that made this project possible.

Finally, a huge thank you to my wife and three

wonderful kids, who had to endure my absence

over the past 9 months while I attended this

post-diploma program, it is for them that I have

worked so tirelessly.



BIBLIOGRAPHY

2003: Airborne gamma ray spectrometry compilation, Island of Newfoundland, Newfoundland

and Labrador. Geological Survey of Canada, Open File 4463, 8 sheets.

Brushett, D.,

2012: Quaternary Geology and Till Geochemistry of the Carmanville (NTS 2E/08), Wesleyville

(NTS 2F/04) and Musgrave Harbour (NTS 2F/05) Map Areas. Current Research (2012)

Newfoundland and Labrador Department of Natural Resources Geological Survey, Report 12-1,

pages 179-190.

Brushett, D.,

2011: Quaternary Geology of Weir’s Pond and Surrounding Areas (NTS Map Areas 2D/15,

2E/01, 2E/02 and 2F/04). Current Research (2011) Newfoundland and Labrador Department of

Natural Resources Geological Survey, Report 11-1, pages 33-42.

Bonham-Carter, G.F., Agterberg, F.P., and Wright, D.F.,

1988: Integration of Geological Datasets for Gold Exploration in Nova Scotia. Photogrammetry

and Remote Sensing 54, pages 1585-1592.

Bonham-Carter, G.F.

1994: Geographic information systems for geoscientists: modelling with GIS. In Computer

Methods in the Geosciences. Edited by D.F. Merriam. Elsevier Science Inc., Tarrytown, New

York, 398 pages.

Carranza, E. J. M.,

2009. Geochemical Anomaly and Mineral Prospectivity Mapping in GIS. Handbook of

Exploration and Environmental Geochemistry; Vol 11. Elsevier, UK, 351 pages.

Colman-Sadd, S.P. and Crisby-Whittle, L.V.J.

2002: Partial bedrock geology dataset for the Island of Newfoundland. Newfoundland and

Labrador Department of Mines and Energy, Geological Survey Branch, Open File NFLD/2616, v.

4.

Davenport, P.H., Nolan, L.W., Butler, A.J., Wagenbauer, H.A. and Honarvar, P.

2002: The geoscience atlas of Newfoundland. Newfoundland and Labrador Department of

Mines and Energy, Geological Survey Branch, Open File NFLD/2687, v. 1.1.

Harris, J.R,

2006, GIS Applications For The Earth Sciences: Special Publication, Geological Association of

Canada, Special Publication 44, 616 pages.



He, B., Chen, J., Chen, C., and Liu, Y.,

2012: Mineral Prospectivity Mapping Method Integrating Multi-Sources Geology Spatial Data

Sets and Case-Based Reasoning. Journal of Geographic Information System, 2012, volume 4,

pages 77-85.

Honarvar, P. and Squires, G.C.,

2006: Preliminary Analysis of Mineral Potential Modeling of the Victoria Lake Supergroup

Volcanic Rocks: A Weights of Evidence approach. Current Research (2006) Newfoundland and

Labrador Department of Natural Resources Geological Survey, Report 06-1, pages 25-44.

Honarvar, P., Nolan, L.W., Crisby-Whittle, L., and Morgan, K.,

2013: The Geoscience Atlas. Current Research (2013) Newfoundland and Labrador Department

of Natural Resources Geological Survey, Report 13-1, pages 1-3.

Harris, J.R., Wilkinson, L., Heather, K., Fumerton, S., Bernier, M.A., Ayer, J. and

Dahn, R.

2001: Application of GIS processing techniques for producing mineral prospectivity maps - a

case study: mesothermal Au in the Swayze Greenstone Belt, Ontario, Canada. Natural Resources

Research, Volume 10, No. 2, pages 91-124.

Hinchey, J.G.,

2007: Volcanogenic Massive Sulphides of the Southern Tulks Volcanic Belt, Central

Newfoundland: Preliminary Findings and an Overview of Styles and Environments of

Mineralization. Current Research (2007) Newfoundland and Labrador Department of Natural

Resources Geological Survey, Report 07-1, pages 117-143.

Hinchey, J.G.,

2008: Volcanogenic Massive Sulphides of the Northern Tulks Volcanic Belt, Central

Newfoundland: Preliminary Findings, Overview of Deposit Reclassifications and Mineralizing

Environments. Current Research (2008) Newfoundland and Labrador Department of Natural

Resources Geological Survey, Report 08-1, pages 151-172.

Kemp, L.D., Bonham-Carter, G.F., Raines, G.L. and Looney, C.G.

2001: Arc-SDM: Arcview extension for spatial data modelling using weights of evidence, logistic

regression, fuzzy logic and neural network analysis. http://ntserv.gis.nrcan.gc.ca/sdm/.

Piercey, S.J.,

2007: Volcanogenic Massive Sulphide (VMS) Deposits of the Newfoundland Appalachians: An

Overview of their Setting, Classification, Grade-Tonnage Data and Unresolved Questions.

Current Research (2007) Newfoundland and Labrador Department of Natural Resources

Geological Survey, Report 07-1, pages 169-178.

http://ntserv.gis.nrcan.gc.ca/sdm/



Sandeman, H., Hull, J.R., and Wilton, D.H.C.,

2014: Geology, Lithogeochemistry and Mineralization at the South Wood Lake Gold Prospect

(Staghorn Property), Exploits-Meelpaeg Subzone Boundary, Western-Central Newfoundland.

Current Research (2014) Newfoundland and Labrador Department of Natural Resources


Sandeman, H., Rafuse, H., and D. Copeland D.,

2010: The Setting of Orogenic, Auriferous Quartz Veins at the Golden Promise Prospect, Central

Newfoundland and Observations on Veining and Wall-rock Alteration. Current Research (2010)

Newfoundland and Labrador Department of Natural Resources Geological Survey, Report 10-1,

pages 77-92.

Squires, G.C.,

2005: Gold and Antimony Occurrences of the Exploits Subzone and Gander Zone: A Review of

Recent Discoveries and their Interpretation. Current Research (2005) Newfoundland and

Labrador Department of Natural Resources Geological Survey, Report 05-1, pages 223-237.

Stapleton, G., Smith, J.L. and Adams, T.

2013: Mineral occurrence data system. In Current Research. Newfoundland Department of

Mines and Energy, Geological Survey, Report 2013-1, pages 199-204.

Stapleton, G., Smith, J.L., Pollock, J.C. and Way, B.C.

2000: Mineral occurrence data system. In Current Research. Newfoundland Department of

Mines and Energy, Geological Survey, Report 2000-1, pages 341-348.

Swinden, H.S., Jenner, G.A., Kean, B.F. and Evans, D.T.W.

1989: Volcanic rock geochemistry as a guide for massive sulphide exploration in central

Newfoundland. In Current Research. Newfoundland Department of Mines and Energy,


Williams, H., Colman-Sadd, S.P. and Swinden, H.S.

1988: Tectonic-stratigraphic subdivisions of central Newfoundland. In Current Research, Part B.

Geological Survey of Canada, Paper 88-1B, pages 91-98.

Wright, D.F., Bonham-Carter, G.F. and Rogers, P.J.

1988: Spatial data integration of lake sediment geochemistry, geology and gold occurrences,

Meguma Terrane, eastern Nova Scotia. In Prospecting in Areas of Glaciated Terrain - 1988.

Edited by D.R. Macdonald and K.A. Mills. CIMM Geology Division, pages 501-515.

Predictive Mapping of Gold Potentials Bay of Exploits

Documents