Geosystem Engineering

ISSN: 1226-9328 (Print) 2166-3394 (Online) Journal homepage: https://www.tandfonline.com/loi/tges20

Quantifying the criteria for classification of

mineral resources and reserves through the
estimation of block model uncertainty using
geostatistical methods: a case study of Khoshoumi
Uranium deposit in Yazd, Iran

Mojtaba Taghvaeenezhad, Mohammadreza Shayestehfar, Parviz

Moarefvand & Ali Rezaei

To cite this article: Mojtaba Taghvaeenezhad, Mohammadreza Shayestehfar, Parviz

Moarefvand & Ali Rezaei (2020): Quantifying the criteria for classification of mineral resources
and reserves through the estimation of block model uncertainty using geostatistical methods:
a case study of Khoshoumi Uranium deposit in Yazd, Iran, Geosystem Engineering, DOI:
10.1080/12269328.2020.1748524

To link to this article: https://doi.org/10.1080/12269328.2020.1748524

Published online: 06 Apr 2020.

Submit your article to this journal

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=tges20
GEOSYSTEM ENGINEERING
https://doi.org/10.1080/12269328.2020.1748524

Quantifying the criteria for classification of mineral resources and reserves

through the estimation of block model uncertainty using geostatistical methods:
a case study of Khoshoumi Uranium deposit in Yazd, Iran
Mojtaba Taghvaeenezhada, Mohammadreza Shayestehfara, Parviz Moarefvandb and Ali Rezaeib
a
Department of Mining Engineering, Shahid Bahonar University of Kerman, Kerman, Iran; bDepartment of Mining and Metallurgical
Engineering, Amirkabir University of Technology, Tehran, Iran

ABSTRACT ARTICLE HISTORY

Investments and progress of mineral projects depend on the quantity (tonnage) and quality (grade) Received 4 August 2019
of mineral resources and reserves. This study examines the impact of various criteria used in the Accepted 25 March 2020
classification of mineral deposits or parameters defining these criteria. The data used in this study KEYWORDS
include the uranium assay analysis from 127 exploratory boreholes, which were then subjected to Variography; block model;
a three-directional variography after statistical studies to identify regional anisotropy. A grade block kriging; uncertainty; JORC
model was built using the optimal parameters of variograms and with the help of kriging estimator. Code
Then, by using different methods of estimating the block model uncertainty including kriging
estimation variance, block error estimation, kriging efficiency and slope of regression, classification
of mineral reserves was carried out in accordance with the JORC standard code. Based on different
cut-off grades, the tonnage and average grade were calculated and plotted. An innovative quanti-
tative method based on the distribution function of the mentioned parameters and the fractal
pattern of separation of populations was used for the classification of mineral reserves. The existence
of the least difference between the use of standard and fractal patterns in the slope of regression
method indicated less error and was a proof of more reliable results.

1. Introduction geological unit (i.e., stationarity decision) as well as the

knowledge of the model parameters (Cressie, 1990). In the
For many decades the mining industry regarded resource
past decades for modeling exploratory indices, geostatisti-
estimation and classification as a mere calculation requir-
cal techniques were introduced based on spatial relation-
ing basic mathematical and geological knowledge. Often
ships between sample locations and sample components in
uncertainty associated with tonnages and grades was either
space (Deutsch et al., 2006). Mineral inventories are based
ignored or mishandled. The economic assessment of
on assessing the grades and tonnages through an estima-
mining projects includes many factors with resource clas- tion procedure such as simple kriging, ordinary kriging
sification being critical at every stage of the mining opera- (OK), lognormal kriging, indicator kriging, co-kriging,
tions. The quality of resource classification is a key universal kriging, residual kriging, moving window regres-
requirement for accurate economic and environmental sion residual kriging, disjunctive kriging, and stochastic
risk evaluation (Silva & Boisvert, 2014). Due to the reduc- simulation such as Sequential Gaussian Simulation (SGS)
tion of mineral resources and increased costs of extraction and Sequential Indicator Simulation (SIS) (Vall´ee, 2000).
and mineral processing, the need for upgrading explora- Since the creation of standards for resource classifica-
tory methods is felt for more accurate and reliable model- tion, a number of different techniques for classification
ing of mineral deposits (Dimitrakopoulos, 2010). In the have been developed (Dohm, 2005; Rossi & Deutsch,
beginning, the conventional methods such as core drilling 2014), with only a few of them used in practice. The
combined with chemical analysis would apply to achieve resource classification index includes the estimated value
an ore grade model. However, too much core drilling of the block and a calibration factor (Arik, 2002).
without considering the spatial dependency is expensive Yamamoto (2000) proposed a classification technique
and time-consuming (Annels & Dominy, 2003). based on interpolation variance, which is defined as the
Geostatistical methods can be used to generate estimates weighted average of the squared difference between the
of the block grades using core drilling, although it requires estimated value of a block and the data values used in its
assumptions of homogeneity of the grades over each estimation. The weights used in this method the ordinary

CONTACT Mohammadreza Shayestehfar [email protected]

© 2020 The Korean Society of Mineral and Energy Resources Engineers (KSMER)
2 M. TAGHVAEENEZHAD ET AL.

kriging weights. Mwasinga (2001) gives a brief description methods of estimating the block model uncertainty
of some other geostatistical classification approaches such including the kriging estimation variance, block error
as variogram range, kriging variance, confidence limits estimation, kriging efficiency, and the slope of regres-
based on normal and lognormal models, block efficiency, sion. The classification of mineral reserves is carried out
Isobel Clark’s classification index, and linear regression in accordance with the JORC standard code. This study
slope. Also, several international classification systems aims at describing the application of these approaches in
have been developed in the past decades (Rendu & a real case study for a uranium deposit in Iran and tries
Miskelly, 2001), the main ones being the American to compare them in terms of tonnage evaluation based
United States Geological Survey (USGS) Circular 831 on geometrical or geostatistical considerations.
(USGS, 1980) and the Society for Mining, Metallurgy,
and Exploration, Inc. (SME Guide) (SME, 1999), the
South African Code for the Reporting of Exploration 2. Materials and methods
Results, Mineral Resources and Mineral Reserves 2.1. Statistical analysis
(SAMREC Code) (SAMREC, 2000), the Canadian of
exploration information, mineral resources and mineral Determining the statistical properties of data to be used
reserve (CIM) guidelines (CIM, 2000) and National for geostatistical evaluation is very important. The
Instrument 43–101 (CSA, 2001), the European Code Khoshumi deposit is located in northeast of Yazd. The
(EURO, 2002), and the Joint Ore Reserves Committee Khoshumi deposit was explored by drilling 127 bore-
(JORC) Code in Australia (JORC, 2004). holes and samples were collected from these locations at
The significance of the ore grade uncertainty in open- different depth intervals. The boreholes were drilled in
pit mining has received significant attention in recent a grid pattern and the spacing between boreholes varies
years. Kriging, as a group of geostatistical methods, is an from 20 to 40 m with an average spacing of 25 m. The
interpolation technique that considers both the degree of first step in exploratory data analysis includes the detec-
variation and the distance between known data points in tion of all possible outliers and duplicated samples. The
estimating the values in unknown areas. The question now representation of the frequency histogram and the
is how much you can trust the proposed reserve based on cumulative distribution function of the data are one of
the geostatistical methods. One of the methods proposed the most common methods in evaluating the frequency
in relation to the geostatistics is the error distribution distribution and also in data normalization. In addition
function, although this function cannot be applied well in to the histogram, the probability plots are also used to
ore classification and mineral blocks (Blackwell, 1998). For identify outlier data (Rezaei et al., 2019b, 2019a).
this reason, using the parameters that can be obtained Statistical analysis was conducted on the composite
through geostatistical methods, the quantification of samples. The probability–probability (P-P) Plot dia-
exploratory blocks is made which is one of the needs and gram after correcting the outlier values shows that all
strengths of this research. The error estimation and uncer- variables follow a normal distribution (Figure 1).
tainty of the estimated space on the one hand can be used
to evaluate the performance of the geostatistical method 2.1.1. Data composition
(Chil`es & Delfiner, 1999) and, on the other hand, allow One of the most important factors in preparation of the
the spatial classification of the estimates to be made by assay data is the uniformity of the nature and statistical
quantitative and mathematical methods, can help the mine weight of the samples used. The first step in geostatis-
designer or expert geologist in the detailed explorations tical techniques is to use a dataset that has equal sup-
and offer the proposing sites for future sampling. port, and the sampling community is to create
Uncertainties including ore grade of deposit would signifi- a homogeneous environment and the same probability
cantly affect the technical and financial aspects of the plans. in terms of the size of the samples taken. For the samples
The traditional methods for investigating the uncertainty collected in this deposit, it is assumed that all conditions
and classification of deposits in Australia and the United among the samples are the same except the length of
States are as follows: a) error quantification; b) Kriging each sample.
efficiency; c) regression coefficient (correlation) that is
used in the present work for borehole data (Wilde, 2009).
Also, the integration of these data has been used to propose 2.1.2. Compare composites with different lengths
the prospecting sites and to classify blocks and estimated Composites with different lengths of 0.1, 0.5, 1 and,
spaces. 5 m were constructed and compared in terms of an
In this research, we focus on the evaluation and average grade of the sample and variance of the data.
classification of mineral resources using different The results are shown in Table 1. Based on the obtained
 GEOSYSTEM ENGINEERING 3

from OK weights that account for both data config-

uration and data values. These estimation variances
are then used to classify resources based on confi-
dence levels and the results are compared with those
obtained by OK variance. The methods are illustrated
using an exploration borehole data set from the
Khoshumi uranium deposit. After exploration of
data analysis, spatial continuity was explored by con-
structing semivariogram models. In the spatial studies,
both directional and Omni-directional variograms are
constructed. The spatial structures revealed
a significant contribution from the nugget, indicating
difficult variogram modeling conditions (Rezaei,
2019c). The directional variogram model provided
a better understanding of the deposit and helped in
identifying the anisotropies. Presence of anisotropy
was taken into account while conducting OK. Three
main directions of the search ellipsoid were chosen
from the experimental directional semi-variograms of
uranium to create the best variography model for the
OK estimation. The variogram of the data is com-
Figure 1. P-P plot of raw data after replacing the outlier values. posed of three isotropic spherical models with ranges
of 50, 76 and 63 m. The directional semi-variogram
Table 1. Comparison of composites with different lengths based model parameters are shown in Table 2. The main
on raw data. directions resulting from variography for three princi-
NO. Mean Variance Minimum Maximum pal directions of the search ellipsoid are presented in
RQW 29,610 119.42 16,393.48 0.01 2645 Figure 2. After variography based on the parameters
0.01 204,901 118.41 29,345 0.01 2645
0.05 39,747 117.12 29,030 0.01 2645 obtained for the variogram, the Three-dimensional
0.1 19,543 117.5 28,654 0.01 2645 (3D) modeling of data was carried out. The
0.5 3377 117.2 26,409 0.01 2645
1 1847 107.53 21,256 1.2 1886 Datamine Studio software was used to prepare the
3D models (Datamine Studio 3, 2006). 3D modeling
of grade in an ore deposit has a lot of advantages
results, 0.5-m composite length has been selected for the (Rezaei, 2019c). If this process is performed carefully,
next steps. evaluations and judgments about different parts of ore
deposit would be more accurate. The block size to be
used for interpolation and reporting were assigned
2.2. Geostatistical analysis
according to the estimation space and sampling dis-
In order to model the deposit and to make an estima- tance. The size of the blocks selected to be 5 (m) × 5
tion of the grade, geostatistical methods – experimen- (m) × 5 (m). The OK method was used to estimate
tal semivariograms, fitting the semi-variogram models the grade of the exploration blocks using calculated
and kriging were used. Among geostatistical methods, variogram and interpolation parameters, and then, the
ordinary kriging (OK) is the one most used for ore block model was plotted. Miners can interpret
mineral resource estimation (Krige, 1996). As which part is ore and which part is waste. Grade
a general rule, the increase in the amount of data maps show that the deposit can be broadly distributed
often leads to a decrease in the uncertainty and into three classes depending on the estimated ura-
improves the confidence on the estimates; however, nium grade. It is noticeable that the threshold grade
geologic data are often spatially correlated and the used for modeling purposes is referred to as
reduction in uncertainty caused by the increased
amount of data is not only a function of the number Table 2. Results directional variography and variogram model
of data points, but it is also dependent on the spatial fitted to data.
distribution of the dataset and the spatial continuity Range
of the attributes under study. This research investi- Model Nugget effect Sill R1 R2 R3
gates alternatives for a computing estimation variance Spherical 916.03 25,501.62 50 76 63
4 M. TAGHVAEENEZHAD ET AL.

Figure 2. Three axis of anisotropy of the experimental variogram in perpendicular directions fitted to each other.

a geological cut-off grade and is not calculated like an Lagrange coefficient, and the average function Fisher
economic cut-off. can be quantitatively categorized errors and calculated
relationships (Cressie, 1990). These classifications will
be in accordance with standard JORC codes (JORC,
3. Discussion
2012). The categorization of mineral resources in the
In the mineral industry, the information available for JORC standard generally depends on the geological
modeling is limited and represents a very small fraction model of the deposit, the sampling quality and the
of the domain of interest. Fortunately, geological data data spacing (Rivoirard & Renard, 2016). Other para-
are often spatially correlated, which enables the infer- meters available in the estimation, such as the number
ence of attributes at non-sampled locations with of data contributed to the estimation of a block or the
a quantifiable degree of uncertainty. The estimates number of bumps involved in estimating a block, can
were kriged (OK) into a sub-celled block model. The also be used to perform innovative quantifications.
information provided were fed to geostatistical techni-
ques such as block error estimation, kriging estimation
3.2. Kriging estimation variance
variance, kriging efficiency, and the slope of regression
against each other. Also, tonnage and grade estimates Kriging variance (KV) provides an indication of relative
according to differing degrees of geological confidence estimation quality mainly in terms of data density and
and economic evaluation are performed. Fractal models geometry and so is the kriging efficiency (KE) (Vann
belonging to nonlinear mathematics are effective tools et al., 2003). The KV is a measure of uncertainty in
for describing the natural variability and skewed distri- predictions which is a function of the variogram, sample
bution of geological objects as used for separation of structure, and sample support (defined as the area in
assay communities. which an observation is made, which may be estimated
as a point or maybe an area) (Journel & Huijbregts,
1978). The KV accounts for data redundancy and uses
3.1. Investigating the uncertainty grade estimation
information about the spatial correlation of the data and
of blocks and classification of them based on JORC
which leads to more accurately capturing the complex
codes
relationships between data availability and confidence
Mineral resource classification is vital in uncertainty on estimates. There are a number of methods for
assessment and risk analysis of mineral resource devel- resource classification, mostly based on ordinary kriging
opment. Using parameters obtained from geostatistical variance. The relative kriging standard deviation,
relations such as estimation variance, block variance, defined as the ratio between kriging standard deviation
 GEOSYSTEM ENGINEERING 5

Figure 3. Cumulative frequency plot of the Kriging logarithmic estimation variance for separation of Communities by fractal method.

and the estimated value of a block is one option (David, blocks with kriging variance below 35% of the block
1988). As shown in Figure 3, using the statistical dis- variance are classified as measured while the blocks
tribution of the variance estimation causes different with kriging variance below 65% of block variance, but
groups of the population to be separated which makes higher than 35% are classified as indicated. The results
comparing similar populations possible. The classifica- show the differences in tonnages within each class of
tion based on kriging variance is performed by defining resources when different measures of uncertainty are
thresholds for each category. It is obvious that co- used.
dimension data are in a population (separation of sta-
tistical population by fractal technique). Using different
3.3. Block error estimation
statistical methods helps in separating communities
with higher precision. It is noteworthy that the blocks Block error estimation is another parameter for quanti-
were selected for classification that has less error estima- fying the uncertainty of the geostatistical process. Since
tion. Based on the obtained results of the amount and this parameter depends on the spatial distribution and
distribution of kriging estimation error variance, 65% of the dimension of data, the use of a particular standard
the estimated reserve is considered to be the proven for it cannot be reliable (Sayyadi et al., 2012). Therefore,
reserve category (Figure 4). It can also be observed in addition to using the available standard, it has been
that the tonnage–grade curve based on the variance tried to use the analytical and numerical parameters.
estimation uncertainty parameter is also found in the Based on the statistical distribution of these parameters,
proven reserve class (Figure 5). In other words, the as described in the previous section, and by comparing

Figure 4. 3D block model based on the distribution of statistical communities of the estimation variance parameter.
6 M. TAGHVAEENEZHAD ET AL.

calculated, accordingly. As shown in Figure 6, using the

statistical parameters and separation analysis of fractal
communities made it possible to calculate the definite,
probable, and possible reserve amount. Figure 7 illus-
trates a 3D block model derived from the separation of
statistical communities by the distribution of the block
error estimation. Based on the obtained results from the
comparison of the two methods, the proven reserve
amount in relation to the total reserve amount in the
standard method is 33% and based on the statistical
analysis of the error estimation distribution is 72%,
which is closer, more tangible and reliable to the reality
than the available standard.
Figure 5. Grade–Tonnage curve; proven reserve class based on
distribution of block estimation variance.
3.4. Kriging efficiency
the obtained results of the standardized and quantifiable
The kriging efficiency (KE) is the best parameter to
method the deposit is classified into three proven, prob-
check the error of modeling. The KE is directly related
able and possible categories. The formula for calculating
to kriging variance, and these two parameters have
the block error estimate parameter is shown in
a correlation coefficient of 1. The KE is between 0 and
Equation 1:
1, with the value of 1 showing a full estimate. Kriging
σ E  t95% efficiency is calculated by the following formula
BE ¼ pffiffiffi (1) (Equation 2):
Z n
ðBV  KV Þ
where BE: block error estimation; σ E : kriging estima- KE ¼ (2)
BV
tion variance; Z: estimated grade; n: number of samples;
t: the 95% coefficient of confidence equals to 1. 96%. where BV = theoretical variance of blocks within the
For a reliable modeling and characterization of domain; KV = kriging variance; and KE = kriging
a given ore deposit, measurement uncertainties should efficiency;
be included as an intrinsic part of the geo-modeling The kriging efficiency of the assay data in the
procedure. The 3D block model of the deposit based Khoshumi deposit was calculated. Figure 8 shows the
on the estimation error uncertainty parameter was built, cumulative frequency curve of the KE parameter, which
and the current standard was used. The estimation can be a reasonable quantitative criterion to estimate the
uncertainty and the amount of the proven reserve are actual separation threshold of the statistical community

Figure 6. Cumulative frequency curve of logarithmic block error estimation for separation of communities by fractal method.
 GEOSYSTEM ENGINEERING 7

Figure 7. Uncertainty representation of a block model in a quantitative method based on block error estimator parameter.

Figure 8. Cumulative frequency curve of KE parameter for separation of communities by fractal method.

and the reserve classification in accordance with the resources by means of conditional simulations of the
actual distribution of data. grades that can also be used to classify each block as
measured, indicated, or inferred resource. The SLP
method was applied to examine the uncertainty of the
3.5. Slope of regression data in this study. Its formula is as follows (Hassani Pak,
2003):
The slope of regression (SLP) parameter is useful in
assessing the quality of the estimates. This method is BV  KV þ jμj
one of the most up-to-date tools for quantifying uncer- R¼ (3)
BV  KV þ j2μj
tainty and reserve classification parameters due to the
use of various parameters for uncertainty. The SLP is where BV: sill value; KV: kriging variance; and μ:
a method that helps evaluating the accuracy of the lagrange multiplier
estimation through computation of a set of geostatistical The value of the slope of regression, if closer to 1,
computational parameters (Jalali et al., 2010). The SLP is indicates the reliability of the estimate presented as the
referring to the unknown true value of a random vari- grade of that block. The mean slope of regression for our
able Z with estimation volume V against the known analysis was calculated as 0.65 and the blocks were
estimate of the random variable Z* with the same plotted and colored according to the slope of regression
volume V. The SLP approximates the conditional bias value. After plotting the cumulative frequency curve and
of the kriging estimation results. based on the separation of the communities by fractal
In this section, we propose an alternative approach method (Figure 9), it is observed that the values of the
that quantifies the overall uncertainty on the mineral SLP parameter more than 0.8 are classified as the proven
8 M. TAGHVAEENEZHAD ET AL.

Figure 9. Cumulative Frequency curve of SLP for Separation of communities by fractal method.

Figure 10. 3D model of uncertainty by the SLP based on quantitative analysis.

reserves, between 0.6 and 0.8 as probable reserve, and 0 tonnage and the average grade curve in terms of grade
to 0.6 as a possible reserve. Figure 10 presents the block (Figure 11).
model of the deposit based on the SLP uncertainty In this research work, we have presented results from
estimation parameter. Using the obtained results from Kriging estimation variance, block error estimation,
the quantification of the SLP measure we can obtain the Kriging efficiency and the slope of regression. The

Tonnage-Grade Histogram(Measured) Tonnage-Grade Histogram(Measured)

1200 350 450 350
400
Average grade (ppm)

1000 300 300

Tonnage- Tone

Tonnage- Tone

250 350
800 250
Average grade (ppm)

300
200 200
600 250
150 200 150
400 150
100
100
200 50 100
50 50
0 0
0 0
0 50 100 150 200 250
0 50 100 150 200 250
Cut- off grade (ppm)
Cut-off grade (ppm)
Tonnage Ave u
Tonnage Ave u

Figure 11. Grade–Tonnage curve; proven reserve class based on the SLP method (left side: quantitative; right side: standard).
 GEOSYSTEM ENGINEERING 9

Table 3. The comparison of the obtained results of the current proposed standards.
Kriging estimation variance Block error estimation Kriging efficiency Slope of regression
Uncertainty index Quantitative analysis Standard Quantitative Standard Quantitative Standard Quantitative
Proven reserve ratio to total reserve (%) 60 35 48 64 69 76 79

obtained results were analyzed and the impact of ore total reserve is obtained with the SLP method. Thus, the
grade uncertainty was identified. Available standard SLP method can be considered for resource classifica-
criteria for classification of mineral reserves regardless tion for improved accuracy of final classification mod-
of the state of the reserve and the statistical distribution els. The comparison of the obtained results between the
of estimation, undoubtedly have problems that make it current proposed standards and the quantitative inno-
difficult to rely on the results of the estimation. Using vative methods based on the distribution function
quantitative analyses based on the statistical distribution reflects the fact that in all the deposits with different
of data and the separation of communities by fractal conditions, it is impossible to use the results of existing
technique will effectively help to better understand the standards and, certainly, the need for quantitative ana-
reserve conditions and estimate the uncertainty para- lysis depends on the conditions of the deposit will be
meters and, finally, the classification of mineral reserves felt.
(Table 3). It can be understood from the achieved results
that multiple grade models of a deposit enable mine
planners to assess the best estimation method in the Acknowledgements
classification of mineral resources and reserves. The The authors would like to thank the Shahid Bahonar
results obtained from the analyses show that which University of Kerman for supporting this research. Thanks
SLP method can be used to develop the mining strate- are also due to the editor and editor in chief of the journal for
their review and handling of this paper. The authors would
gies that have less risk in relation to grade uncertainty.
like to thank the reviewers for their very helpful and con-
However, to show the difference between the methodol- structive reviews of this manuscript. The contributions of
ogies, the tonnages are calculated within the realizations Nima Jabbari are highly appreciated.
obtained from four methods depending on the cut-off
grades of uranium.
Disclosure statement
No potential conflict of interest was reported by the authors.
4. Conclusions
The main goal of this research was to compare the most Funding
common techniques and the proposed methods for
estimating the block model uncertainty using sub- This work was supported by the Shahid Bahonar University of
Kerman.
surface data from Khoshumi uranium deposit in order
to highlight the advantages and disadvantages of each
method in mineral resource classification. Kriging esti- References
mation variance, block error estimation, kriging effi-
Annels, A. E., & Dominy, S. C. (2003). Core recovery and
ciency, and the slope of regression parameter methods
quality: Important factors in mineral resource estimation.
were compared and employed in the mineral resource Applied Earth Science, 112(3), 305–312. https://doi.org/10.
classification of a uranium deposit. The obtained results 1179/037174503225011306
of the actual and estimated grades were all good, and Arik, A. (2002). Comparison of resource classification meth-
these are usually acceptable measurements of the errors odologies with a new approach. APCOM 2002: 30th
used in ordinary kriging. Using a fractal analysis model International Symposium on the Application of Computers
and Operations Research in the Mineral Industry (pp.
in that it recognizes the parameter distribution pattern 57–64).
and separates the groups from each other can be an Blackwell, G. H. (1998). Relative kriging errors- A basis for
innovative approach and closer to reality. However, mineral resource classification. Exploration and Mining
the results of the SLP method are more reasonable and Geology.
satisfactory in practice, and the use of this method is Chil`es, J. P., & Delfiner, P. (1999). Geostatistics: Modeling
spatial uncertainty. Wiley.
encouraged for the classification of the mineral
CIM. (2000). CIM standards on mineral resources and
resources in the underlying uranium deposit into mea- reserves– definitions and guidelines, standards of disclo-
sured, indicated and inferred resources based on the sure for mineral projects (NI 43-101). Canadian CIM
JORC code. The maximum proven reserve ratio to the Guidelines.
10 M. TAGHVAEENEZHAD ET AL.

Cressie, N. (1990). The origins of kriging. Mathematical Rezaei, A., (2019c). Modeling the effect of structural pattern on
Geology, 22(3), 239–252. https://doi.org/10.1007/ mineralization in Sangan Central Iron Ore Mine [PhD
BF00889887 Dissertation]. Amirkabir University of Technology,
CSA. (2001). Standards of disclosure for mineral projects: Department of Mining and Metallurgy Engineering, p. 364.
National Instrument 43–101. Canadian Securities Rezaei, A., Hassani, H., Fard Mousavi, S. B., Hassani, S., &
Administration, p. 22. Jabbari, N. (2019b). Assessment of heavy metals contam-
Datamine Studio 3. (2006). Version 3.24.25.0.,Users Tutorial. ination in surface soils in Meiduk copper mine area, SE
David, M. (1988). Handbook of applied advanced geostatistical Iran. Journal of Earth Sciences Malaysia.
ore reserve estimation. Elsevier. Rezaei, A., Hassani, H., Fard Mousavi, S. B., & Jabbari, N.
Deutsch, C. V., Leuangthong, O., & Ortiz, M. J. (2006). A Case (2019a). Evaluation of heavy metals concentration in
for geometric criteria in resources and reserves Jajarm Bauxite deposit in Northeast of Iran using environ-
classification. CCG Paper 2006-301. Centre for mental pollution indices. Malaysian Journal of Geosciences,
Computational Geostatistics, University of Alberta. 3(1), 12–20. https://doi.org/10.26480/mjg.01.2019.12.20
Dimitrakopoulos, R. (2010). Advances in orebody modelling Rivoirard, J., & Renard, D. (2016). A specific volume to
and strategic mine planning I. Spectroscopy Series, 17, 345. measure the spatial sampling of deposits. Mathematical
https://www.springer.com/gp/book/9783319693194. Geosciences, 48(7), 791–809. https://doi.org/10.1007/
Dohm, C. (2005). Quantifiable mineral resource classification- s11004-016-9647-9
A logical approach. In O. Leuangthong & C. V. Deutsch Rossi, M. E., & Deutsch, C. V. (2014). Mineral resource esti-
(Eds.), Geostatistics Banff 2004 (Vol. 1, pp. 333–342). mation. Springer.
Springer. SAMREC, (2000). South African code for reporting of mineral
EURO, (2002). Code for reporting of mineral exploration resources and mineral reserves (The SAMREC Code). Report
results,mineral resources and mineral reserves (the prepared by the South African Mineral Resource
European code). Report prepared by the Institution of Committee SAMREC under the auspices of the South
Mining and MetallurgyWorking Group on Resources and African Institute of Mining and Metallurgy, p. 38. www.
Reserves in conjunction with the European Federation of samcodes.co.za.
Geologists and the Institute of Geologists of Ireland, p. 34. Sayyadi, A. H., Fathianpour, N., & Mousavi, A. (2012).
Hassani Pak, A. (2003). Error and risk management in explora- Application of conditional sequential Gaussian simulation
tion (1st ed.). Tehran University Press. (in Persian). in uncertainty assessment of the estimated blocks grade in
Jalali, M., Rahimipour, G. R., Dianti, M. R., & Niyazad, M. Sordid Phosphate Mine. Journal of Analytical and
(2010). Investigating the validity of linear and nonlinear kri- Numerical Methods in Mining Engineering.
ging estimates in the zonation of ore and tile blocks in Silva, D., & Boisvert, J. (2014). Mineral resource classification:
Sarcheshmeh mine. A comparison of new and existing techniques. Journal of
JORC, (2004). Australasian code for reporting of exploration the Southern African Institute of Mining and Metallurgy,
results, mineral resources and ore reserves (the JORC Code, 114(3), 265–273. ISSN: 2225-6253. https://www.saimm.co.
2004 Edition). Report prepared by the Joint Ore Reserve za.
Committee of the Australasian Institute of Mining SME. (1999). A guide for reporting exploration information,
and Metallurgy, Australian Institute of Geoscientists and resources, and reserves. Report prepared by SME Resources
Minerals Council of Australia, p. 21. and Reserves Committee, Society for Mining, Metallurgy
JORC, (2012). Australasian code for reporting of identified and Exploration, p. 17. http://www.crirsco.com/usa_sme_
mineral resources and ore reserves (The JORC Code). The guide_2007.pdf.
Joint Ore Reserves Committee of the Australasian Institute USGS. (1980). Principles of a resource/reserve classification
of Mining and Metallurgy, Australian Institute of for minerals. U.S. Geological Survey Circular 831, p. 5.
Geoscientists, and Minerals Council of Australia. Vall´ee, M. (2000). Mineral resource + engineering, economic
Journel, A. G., & Huijbregts, C. H. J. (1978). Mining geosta- and legal feasibility = ore reserve. CIM Bulletin, 93(1039),
tistics. Centre de Geostatistique Fontainebleau. 53–61. pascal-francis.inist.fr.
Krige, D. (1996). A practical analysis of the effects of spatial Vann, J., Jackson, S., & Bertoli, O. (2003). Quantitative kriging
structure and of data available and accessed, on conditional neighbourhood analysis for the mining geologist –
biases in ordinary kriging. Geostatistics Wollongong, 96(2), A description of the method with worked case examples.
799–810. In: Proceedings Fifth International Mining Geology
Mwasinga, P. P. (2001). Approaching resource classification: Conference (pp. 215–223). The Australasian Institute of
General practices and the integration of geostatistics. Mining and Metallurgy.
APCOM 2001: 29th International Symposium on Computer Wilde, B. (2009). Minimizing error variance in estimates by
Applications in the Mineral Industries (pp. 97–104). optimum placement of samples. CCG Anuual Report 11,
Rendu, J. M., & Miskelly, N. (2001). Mineral resources and Paper 406.
mineral reserves: Progress on international definitions and Yamamoto, J. K. (2000). An alternative measure of the relia-
reporting standards. Transactions of the Institution of bility of ordinary kriging estimates. Mathematical Geology,
Mining and Metallurgy, Section A: Mining Technology, 32(4), 489–509. https://doi.org/10.1023/A:100757791
110, 133–138. https://doi.org/10.1179/mnt.2001.110.3.133. 6868