Geoscience Reference
In-Depth Information
In all cases, a detailed audit trail should be documented,
corresponding to the level of detail of the resource model
and the development stage of the mining project or mine.
Question 1: Plot the histogram over 10 realizations and
compare the mean and variance to from the
realizations to the original declustered data
used in the simulation. Comment on the
results.
Question 2: Plot the variogram from 10 realizations (in
one principal direction) and compare to input
model. Comment on the results.
Uncertainty has a very precise meaning—80 % probability
means that a proportion of 0.8 of the values should fall with-
in the 80 % probability interval. For the data in red.dat :
Question 3: Construct distributions of uncertainty of the
normal scores transform of grade in cross
validation mode.
11.8
Exercises
The objective of this exercise is to review cross valida-
tion and to experiment with ways of checking uncertainty.
Some specific (geo)statistical software may be required. The
functionality may be available in different public domain or
commercial software. Please acquire the required software
before beginning the exercise. The data files are available for
download from the author's website—a search engine will
reveal the location.
Question 4:
Calculate the proportions of true values within
ixed intervals of the Gaussian distributions 
of uncertainty and plot an accuracy plot.
11.8.1
Part One: Cross Validation
References
Provided with enough data, cross validation is a useful ex-
ercise. The drillholes are left out one at a time and re-esti-
mated from surrounding drillholes. Blunders in the data or
estimation may be detected, as well as obtaining an initial
appreciation for the expected degree of uncertainty.
Question 1:
Clark I (1986) The art of cross-validation in geostatistical applications.
Proceedings 19th APCOM, pp 211-220
Davis BM (1987) Uses and abuses of cross-validation in geostatistics.
Math Geol 17:563-586
François-Bongarçon DM (1998c) Due-diligence studies and mod-
ern trends in mining. Unpublished internal paper, Mineral
resources development, Inc
Journel AG, Rossi ME (1989) When do we need a trend model? Math
Geol 22(8):715-738
Leuangthong O, McLennan JA, Deutsch CV (2004) Minimum
acceptance criteria for geostatistical realizations. Nat Resour Res
13(3):131-141
Parker HM (2012) Reconciliation principles for the mining industry.
The Australasian Institute of mining and metallurgy. Mining Tech
121(3):160-176
Rossi ME, Camacho VJ (1999) Using meaningful reconciliation infor-
mation to evaluate predictive models, Preprint, SME Annual meet-
ing, March 1-3, Denver
Schofield NA (2001) The myth of mine reconciliation. In: Edwards
AC (ed) Mineral resource and ore reserve estimation—the AusIMM
guide to good practice. Vic., AusIMM, Melbourne, pp 601-610
Vaughan WS (1997) (July) Due diligence issues for mining investors
post Bre-X. Randol conference on sampling and assaying of gold
and silver, Vancouver
Setup to perform cross validation with the
largedata.dat used in previous exer-
cises. Plot a scatterplot of the estimate versus
the true value and a histogram of the errors.
Look for data/drillholes where unusually
large over- or under-estimates have occurred.
Comment on the results.
Question 2:
Perform between 2 and 4 reasonable sensi-
tivity runs and document. You could vary the
search, the number of data or the variogram
model. Comment on the results.
11.8.2
Part Two: Checking Simulation
The goal of simulation is to reproduce the input histogram
and variogram. Use the same set of realizations constructed
in previous exercises. You may have to recreate some SGS
realizations.
 
Search WWH ::




Custom Search