GEOLOGY, MINING OPERATION AND SCHEDULING OF THE

PARAGOMINAS BAUXITE MINE

Silva HM, Picanço E, Maurity C, Morais W, Santos HC and Guimarães O*
VALE, Paragominas, Pará, Brazil
Abstract
The Paragominas bauxite mine (VALE) is the first in the world integrating the mine process with beneficiation and transport
of milled bauxite by mineral pipeline. The mine is located in the far north of Brazil, in the State of Pará, part of the eastern
Amazon region. As from 2001, Vale has undertaken exploration and technical characterization of Bauxite over the whole
of the Paragominas Bauxite District, culminating in the opening of the first VALE mine an the Miltônia Plateau, based on
reserves equivalent to 300 MT of beneficiated product.
The geological profile of the deposits consists of seven layers: Topsoil and Clay Overburden (Belterra Clay) Nodular
Bauxite, Ferruginous Laterite, Crystallized Bauxite, Crystallized/Amorphous Bauxite, Amorphous Bauxite and, at the base,
argillaceous sediments of the Lower Cretaceous Ipixuna Formation.
Mine grade control and production monitoring is based on processing data obtained by logging and sampling boreholes
made on a 100 X 100 meter grid and channels made on a 25 X 25 meter grid on the mine fronts.
The mining operation faces two major challenges: a high stripping ratio of 7.6 in volume and a high level of rainfall. The
mining system, therefore, needs to optimise waste stripping and allow for ore selectivity. The system developed to achieve
these results uses hydraulic excavators and bulldozers for waste stripping, and hydraulic excavators and trucks for ore
production.
The mine scheduling developed to design, simulate and sequence the mine on both Paragominas plateaus are described.
Several scenarios were analyzed in order to better understand the relationship between the design parameters and the
geological resources. The use of specific software tools, the appropriate data format and checks, and the experience of
the mine staff were key issues to produce the best results.

Keywords: Stripping mining, Bulldozer, Overburden, Open pit, Stripping ratio

1. Introduction species of palm trees. However, much of the area has been
deforested, and is currently used for agricultural purposes.
Location and Access
The Paragominas bauxite complex is located 220km south of
Belém, Northern Brazil, and 70km southwest of Paragominas,
the closest town (about 3°21’ S, 48°05’ W). Site access is from
Belém, via the BR-010 Belém–Brasília highway, and also from
the town of Paragominas. CVRD stated that local access roads
are due to be asphalt-sealed prior to production start-up. Belém
is the capital of the Pará State, with excellent aerial and river-
borne communication (Figure 1).

Topography, Climate and Vegetation

The mine property is located in the Amazonas sedimentary basin,
between the basins of the Capim and the Gurupi Rivers, and is
characterized by a well-developed drainage system, forming low
valleys that separate isolated plateaus. The Paragominas bauxite
deposits are developed on a 10,000km2 plateau, situated 50 m Figure 1. Paragominas Location Map Geological setting
to 150 m above the drainage-incised valley floors. The average
altitude does not exceed 150 m in the plateaus. Local Geology and Mineralization

The climate is hot and humid. The average temperatures range The Miltônia 3 (M3) and Miltônia 5 (M5) gibbsitic bauxite
from 20°C to 33°C and the natural humidity averages 90%. There deposits have formed by deep tropical weathering of the Ipixuna
are no well-defined seasons, only very wet and less wet periods, Formation. The bauxite layer forms a nearly continuous tabular
locally known as “inverno” (or winters, December to May) and body, less than 5 m thick, but extending 20km north–south, and
“verão” (or summers, June to November). Annual precipitation as much as 8km east–west, beneath the plateau surface.
reaches 2,200 mm, with monthly averages of 180 mm. The geological profile found in drill holes, pits and trenches
The original vegetation is the equatorial latifoliated forest, with indicates that differences in composition, mineralogy, and texture
transitions to a tropical forest, dominated by low and medium have arisen from differences in the original sedimentary protolith,
size plants, and locally with very high trees, among them the as well as from the position within the weathering profile.
cedar (Cedera odera), “ipê roxo” (Tabebuia barbata), “angelim” The main factors in the bauxite formation were as follows:
(Hymenolobium petraeum), and “maçaranduba” (Manikalra
luberi). Other common species include the “cipós” and various • geologic factors: geologic stability over a long period,

Proceedings of the 8th International Alumina Quality Workshop • 2008 11

AQW2008_Proceedings_FINAL-1-120.11 11 4/08/2008 2:54:53 AM

 • petrographic factors: the presence of the porous and Sampling, Sample Preparation and Assaying
aluminum-rich Barreiras formation rocks, After logging, the core was split in two, with hammer and chisel
• climatic factors: a tropical climate, with abundant when needed. Sample intervals were marked off in nominal
precipitation, 0.5 m intervals. However, main geological contacts were honored;
• morphological factors: the depth of the phreatic layer, the maximum and minimum sample lengths were usually 0.75 m
presence of old elevated surfaces (plateaus). and 0.30 m, respectively, although sometimes those figures
were exceeded. All material was placed in large, 300 μm thick
Prolonged weathering and oxidation of aluminum-rich sediments
plastic bags, with the drill hole number and sample number on
allowed the formation of gibbsite, an aluminum hydroxide, which
can occur in a microcrystalline, porcelain-like form (amorphous the outside and the inside. Out of every 20 samples, one sample
bauxite), or as fine-grained euhedral crystals that occasionally was weighed to compare it with the weight obtained later by the
reach 1 mm in size. laboratory.

Minerals within the deposit display either ‘detrital’ or ‘secondary’ Only samples from the bauxite section were submitted to the
character. Minerals that are ‘detrital’ were originally deposited laboratory, although one or two samples from the overlying and
as part of the host water-borne sediment (e.g., quartz and underlying material were also included. The crude sample and
anastase); those that are ‘secondary’ are weathering products of the +20 mesh fraction were assayed in the lab. The assay suite
kaolinite and feldspar. includes ICP determinations for Total Al2O3, Total SiO2, Fe2O3
and TiO2; and Loss on Ignition (LOI). In addition, a Parr sodium
2. Exploration hydroxide bomb was used to measure Available Al2O3 (by titration
with ZnSO4) and Reactive SiO2 (by ICP).
Drilling
Drilling at the Paragominas bauxite district started in July 2002. Quality Assurance and Quality Control
The drilling was made using triple-barrel air-flush drills. Core People from CVRD had detailed written procedures for every
drilling at the Miltônia 3 was initiated on a 1,600 m x 1,600 m operation, from drilling and sampling to sample preparation
grid, and progressed to a 200 m x 200 m centered grid (a 200 and assaying, including recommendations to minimize errors. A
m square with a central hole). In some areas, the grid was even quality control procedure was implemented during the Miltônia 3
more detailed, 100 m x 100 m and 25 m x 25 m. The Miltônia and Miltônia 5 exploration programs. The program included the
5 grid went from 800 m x 800 m to 200 m x 200 m with center following control operations and control samples:
holes. Table I lists the drilling meterages per plateau.
• mass control after crushing (5%)
Core Logging • sieve tests after pulverization (5%)
Drill core in the PVC pipe was transported to a logging facility, • coarse duplicates (5%): conducted after splitting the original
where the entire hole was laid out and logged. Measures of core sample, before the granulometric separation (sent blind to
recoveries were made by inserting a measuring tape in each end the laboratory)
of the PVC pipe and then were compared to the recovered length • pulp duplicates (5%): conducted after pulverization (blind to
of the drilled length. the laboratory)
• check samples (5%): submitted for external analysis to a
The PVC pipe was then cut longitudinally with a saw; the core secondary laboratory.
was placed into the underlying tray with a minimum amount of
disruption, and was then photographed with a digital camera 3. Resource Estimation
(one picture per drill hole). Geological Model
Logging intervals were measured to the nearest centimeter. The The geological modeling was prepared using two different
logs were general, mainly describing the rock type and color. The general mining software packages. Two methodologies were
distinction between ore and waste units was quite clear in the used for building Miltônia 3 and Miltônia 5 models where eight
drill core. lithological units were interpreted and modeled.

Table I. Drill Hole Meterages by Deposit For Miltônia 3, a Gridded Seam Model was used. The methodology
Deposit Year No.Holes Meterage (m) Grid
consists of defining the lithological contact points in the drill hole
Miltônia 3 9 131 1,600 m x 1,600 m database, then interpolation of the seam thickness, and finally
43 567 800 m x 800 m corrections of the surfaces that auto intersect. Surfaces were
2002- 741 10,223 400 m x 400 m
2005 1,087 14,724 200 m x 200 m
created and then a block model was defined.
580 8,466 200 m x 200 m Centered
28 457 100 m x 100 m For Miltônia 5, a traditional method of vertical section
128 2,159 25 m x 25 m interpretation was used. Sections north–south and east–west
Subtotal 2,616 36,727 -
were interpreted where drill holes exist. A manual interpretation
Miltônia 5 91 1,174 800 m x 800 m
2004- 561 7,638 400 m x 400 m of the lithological units was done, by “snapping” the line point to
2005 735 10,695 200 m x 200 m an existing drill hole intersection, thereby building a straight line
112 1,252 200 m x 200 m Centered
between two drill hole points. This contact line is then smoothed
Subtotal 1,499 20,759 -
using a GMP command tool that will create intermediate points
The following lithology types were logged: CAP, capeamento along the string segments and interpolate a curve, or smoothed
(overburden); BN, bauxita nodular (nodular bauxite); BC, bauxita line in between the existing points, honoring the contact
cristalizada (crystallized bauxite); LF, laterita ferruginosa information. The strings, north–south and east–west, are then
(ferruginous laterite); BM, bauxita maciça (massive bauxite), BA, used for building three-dimensional surfaces. Where lenses of
bauxita amorfa (amorphous bauxite), ARV, argila variegada material occur in between the seams, a three-dimensional solid
(mottled clays), with mixed or transition zones, like BNC, BCM modeling technique was put in place.
and BCBA. These attributes were hand-entered by a technician
into the in-house database management system. Models were extrapolated laterally up to 400 m from the last
existing drill hole, just to impose a limit for the block model.

12 Proceedings of the 8th International Alumina Quality Workshop • 2008

AQW2008_Proceedings_FINAL-1-120.12 12 4/08/2008 2:54:53 AM

 Resource Model Validations Bauxite (BCBA), Amorphous Bauxite (BA), and Variegated Clay
A statistical validation was done comparing averages from the (ARV). According to Kotshoubey et al. (2006) the oldest rocks
composites and estimated blocks. The comparison considered of the District pertain to the Itapecuru Group (south and central
separate blocks that were estimated in each run, compared to the portions) and the Ipixuna Formation (north portions) belonging
average of the total number of composites available. The graphs to the Upper Cretaceous, according to Santos Jr. and Rosseti
showed a very good correlation of the average from estimated (2002).
blocks in the first run and the composites, being within less than
3% for all variables. This difference is higher for blocks estimated
during second and third runs.

Swath plots were also used as another grade estimation /

validation tool. Slices of 500 m along north–south and east–
west directions and 10 m in the vertical axis were generated. The
averages of the blocks and composites within these slices were
then compared in a graph where the number of composites used
for estimation was also considered.

Based upon examination of drill core, test pits, and trenches,

CVRD planners calculate that losses of the BCM (ore) horizon Figure 2. Typical Lateritic profile at the Miltônia Plateau
during mining will be minimal. The contact with underlying BCBA
from crystalline bauxite (BCM) is gradational and of only slightly The Nodular Bauxite is characterized by nodules which increase
inferior quality to the main ore horizon. Therefore, it is calculated in size towards the base. This material forms quite discontinuous
that 20 cm of BCBA below the BCM will be incorporated into bodies. At the top, there are millimetres to centimetre sized
run-of mine ore. Given that the BMC horizon averages about a bauxitic nodules, coloured from yellow to lilac, associated with
meter thick, the dilution is equivalent to about a 13% increase ferruginous pseudo-pisolites. At the base the nodule become
in tonnes, with only a slight effect on the ore quality parameters. concretions, with a diminishing quantity of pseudo-pisolites and
This twenty centimeters of dilution was considered for the block increased degree of crystallization of the gibbsite.
model used at mine planning, and the built ore blocks include
the dilution obtained from the BCBA blocks. With increasing depth, the nodular bauxite grades into Crystallized
Nodular Bauxite, constituted by nodules and/or concretions of
4. Resource Classification reddish colour, with a relative absence of ferruginous pseudo-
The resource classification of each block was based on the pisolites. This material sits discordantly on the Ferruginous
anisotropic distance between the block centroid and the nearest Laterite level. Together, the thicknesses of Nodular and
sample used to estimate the block, see Table II. The distance Crystallized Nodular can be up to 2 meters.
cutoff was selected by reference to the BCM thickness variogram, The Ferruginous Laterite is predominantly in the form of pseudo-
since this variable was considered to have the greatest impact on pisolites or nodules, at the top and in concretions at the base.
the reliability of the in situ resource estimate and presents the Sometimes, this horizon contains medium to coarse grained
shortest range of variograms when compared to other variables. crystalline bauxite, associated with ferruginous concretions.
The distance cut-off corresponds to one half of the distance Thickness of this layer varies from a few centimetres to 2
at which the first spherical component of the BCM thickness meters.
variogram reaches a level of 98% of the first structure sill, The Crystallized Bauxite horizon is that which has the best lateral
termed D98/2. Measured resources comprise blocks with a continuity. It consists of reddish coloured bauxite in the form of
BCM thickness value estimated in the first pass and in which the blocks or concretions. Frequently, the top is enriched in iron,
anisotropic distance to the nearest sample was less than D98/2. due to contact with the Ferruginous Laterite, but iron grades
Indicated resources comprise blocks interpolated in the second decrease towards the base. The contact with the underlying
pass. Inferred resources comprise all remaining blocks and are horizon, Crystallized/Amorphous Bauxite is represented by a very
estimated in the third run. irregular transition. The average thickness of the horizon is about
Table II. Resources Statement 1.5 meters.

Product The Crystallized/Amorphous Bauxite occurs as nodules and

Tonnage Al_Av Si_React concretions, having yellow to lilac coloured micro to macro-
Plateau Category (Dry Mt) (%) (%) crystalline gibbsite, which is associated (or not) with light grey
Measured 161.90 49.39 3.49 to light brown porcelaneous (micro or crypto-crystalline) gibbsite.
Miltônia 3 Indicated 59.50 48.80 3.77 The contact with the Amorphous Bauxite is characterized by an
Measured & Indicated 221.40 49.23 3.57 increase in the clay matrix material found between nodules and
Measured 130.60 45.69 4.48 strong indications of kaolinization, appearing as variegated clay.
Miltônia 5 Indicated 48.00 42.14 5.39
The average thickness of this horizon is about 1 metre.
Measured & Indicated 178.60 44.73 4.73
Measured 292.60 47.74 3.93 The Amorphous Bauxite is characteristically formed of elongated
Indicated 107.50 45.87 4.48 nodules and blocks of micro-crystalline gibbsite, of yellowy colour
Total Measured & Indicated 400.10 47.23 4.08
set in a clay matrix, which is typically variegated. This horizon is
Inferred 18.10 44.60 5.48
the part of the bauxitic zone with the highest silica content.

5. Mine Geology The Variegated Clay, or kaolinized sapprolite, characterized by

multicoloured kaolinitic clay, is the basal limit of the Laterite
The geological profile of the Miltônia deposit has eight horizons profile.
(Figure 2), with lateral variation in thickness to characterize the
following lenticular horizons: Clay Overburden (CAP), Nodular The graph on Figure 3 shows that there are two zones of alumina
Bauxite (BN), Crystallized Nodular Bauxite (BNC), Ferruginous enrichment in the bauxite/Laterite profile found at Miltônia.
Laterite (LF), Crystallized Bauxite (BC), Crystallized/Amorphous These are denominated Upper and Lower Enrichment Zones by

Proceedings of the 8th International Alumina Quality Workshop • 2008 13

AQW2008_Proceedings_FINAL-1-120.13 13 4/08/2008 2:54:53 AM

 the authors of this paper. They correspond to the BNC and BC 6. Mining Methodology
horizons, respectively.
The mining sequence consists of deforestation, waste stripping,
Despite the gradational contact between BN and BNC, the BNC and ore scarification, bauxite mining (loading and hauling), waste
exhibits strong alumina enrichment, with consequently strong fill (from subsequent strip) and reforestation. The geometry of the
suppression of iron content. Silica is slighthy depressed coming operation is characterized by the continuous flat tabular bauxite
to grades of less than 5.5%. The LF acts like as inversion zone for deposit, on average 1.59 m thick for Miltônia 3 (M3) and 1.22 for
iron grades, while silica is slightly enriched here. Miltônia 5 (M5), under overburden that averages 11.45 m on M3
and 10.18 m on M5.
The top of the BC is marked by the presence of ferruginous
blocks and nodules, usually forming concretions less than 20 The mining operation has two major conditions to consider: a
centimetres thick. This explains why iron oxide grades are about high stripping ratio (7.6 in volume) and a high rainfall, thus it
15% in this portion, dropping off below. Visually, the gradational is necessary to have a mining system that minimizes the cost
contact between BC and BCBA is marked by a change in texture of waste stripping and aids selectivity. The appropriate system
of the reddish concretions with clearly crystalline bauxite. The comprises hydraulic excavators and bulldozers for waste stripping,
nodules become yellow and lilac; the crystals are less well defined and hydraulic excavators with trucks for ore production. These
and frequently show evidence of degradation. Chemically, this are supported by auxiliary equipment. Figures 4 and 5 show the
contact is marked opposite trends in the silica and iron curves. steps of the mining methodology.

Figure 4. Mining methodology up to 8m of overburden

The methodology uses a hydraulic excavator (Liebherr 994) when

the productivity of the bulldozer (D11R) decreases. This happens
when great distances and steeper up-slopes leave an optimal
working height for the excavator.
Figure 3. Progression of Available Alumina, Reactive Silica and Iron Oxide in
the Typical Miltônia Profile Depending on the overburden thickness, there are two different
ways to remove the waste material. Areas with up to 8m are
The lower horizons are marked by successive enrichment of totally removed by bulldozer Cat D11R (Figure 4) and the other
reactive silica concomitant with suppression of iron and available areas are excavated by bulldozers and hydraulic excavators
alumina. (Figure 5). In both cases the bulldozer scarifies the ore, which is
mined using backhoe excavators of 4.5m3 to load trucks of 35
The variation in mineralogy is relatively small over Miltônia. The metric tons.
principal mineral constituents are gibbsite, kaolinite, quartz,
hematite, anatase, goethite and, in lesser quantities, micas, In areas with more than 8m of overburden, the bulldozer (Cat
zircon, tourmaline, ilmenite, pyrophyllite, and locally a little D11R) does the pre-stripping Figure 5: (I). Afterwards, the
boehmite. excavator (15 m3) digs the approximately 5m of overburdenleft
on the strip, Figure 5: (III). The remaining material is dug and re-
Characteristics of the Ore handled as showed in Figure 5 (II).
The Crystallized Bauxite and the upper portion of the Crystallized/
Amorphous Bauxite are the principal components of the ore, The mining schedule was constructed from block models,
because of the high grades of available alumina and lower delimiting reserves by the maximum proximity to the plateau
quantities of reactive silica found here. This combination also border (40m) allowed for environmental preservation.
has an almost constant presence over the entire plateau and The uniformity of bauxite quality throughout the deposit, and
affords the best thickness for mining. The Upper Enrichment sourcing of run-of-mine (ROM) bauxite from a minimum of three
Zone offers a subordinate option in places where grades and faces at any time assures a homogeneous supply of material
thickness are adequate and the intervening Ferruginous Laterite from the mine.
is not too thick to be included without upsetting overall grades, or
it is thick enough to be separated with efficient use of the mining The mining schedule considered both plateaus, Miltônia 3 and
equipment. Miltônia 5; the sequence for them shows that during the first
four years only Miltônia 3 will be in production. During year 5,
production will start in Miltônia 5. At which point the plant will

14 Proceedings of the 8th International Alumina Quality Workshop • 2008

AQW2008_Proceedings_FINAL-1-120.14 14 4/08/2008 2:54:53 AM

 reach the maximum capacity. Also the detailed investment
profile was determinate for the Life of Mine Plan, which includes
equipment acquisition and replacement; and installations; etc.

One of the main issues for fleet calculations is productivity,

especially of the CAT D11 bulldozer, because this equipment has
a large influence on the mine cash cost.

Figure 6. Bulldozer productivity

The main factors for estimating the productivity are: the operator
skills (0.5-1.0), type of material (0.6-1.2), Slot Dozing (1.2),
side by side dozing (1.15-1.25), job efficiency (.67-.83). The
grade factor has a big influence on the productivity and can be
calculated from Figure 7.

Figure 5. Mining methodology above 8m of overburden

7. Production Schedule
A simple way to calculate this productivity is to use the supplier
manual. After selecting the appropriate blade according to the
material (mainly clay), the next step is to estimate the dozing
production by the following formula:

Production (Lm3/hr) = Maximum x Correction Figure 7. Grade factor

(LCY/hr) production factors
Another method of estimating the dozer productivity is to use
The bulldozer production curves give maximum uncorrected the supplier dozer simulator (DOZSIM). The first step is to define
production for universal, semi-universal, and straight blades a geometry that is compatible with the geology of the deposit.
and are based on several conditions. Figure 6 shows a typical Figure 8 shows the parameters used to simulate the correct
productivity chart. geometry.

The operation parameters are entered to the system in order to

calculate the productivity. Since the estimations are quick, it is
possible to evaluate the impact on the results of changing the
different parameters.

Proceedings of the 8th International Alumina Quality Workshop • 2008 15

AQW2008_Proceedings_FINAL-1-120.15 15 4/08/2008 2:54:54 AM

 The productivity obtained by both methods were simulated and The influence of the resource classification, ore mining, washing
compared with the estimates calculated using the Performance and transportation costs and the price of the bauxite were also
Manual. Figure 9 shows the simulation of the situation where the simulated and estimated. This type of study could help to indicate
overburden are less then 8m. For areas, where the strip ratios where VALE must invest in drilling, and evaluate the impact that
were high, the simulated results did not get as close to those efforts to reduce costs have on the final results.
estimated using the Manual. Further tests and studies must be
made with the dealer. 8. Conclusions
The mining methodology is highly dependent on the thickness
Another aspect that can be examined using the simulator is
of the waste and ore of the deposit, the effective productivity
estimating productivity for different types of bulldozer (CD),
of the dozer and the hydraulic excavator and their respective
different sizes of strips, different strip ratios and different
costs. Studies to compare the mining costs with other stripping
contributions by the hydraulic excavators.
methods (dragline, bucket wheel, etc.) are very important and
under constant review.

The use of a specific software tool for scheduling, the appropriate

data format and the experience of the staff developing the mine
sequence, together with field checks are critical to producing the
best results.

The key indexes and results obtained in current mining operations

consolidate the methodology applied, in terms of estimating the
equipment fleet required and the grade variation throughout the
life of the mine.

Figure 8. Design of the geometry

Figure 9. Use of the simulator DOZSIM

References
1. Canadian Institute of Mining, Metallurgy, and Petroleum, “CIM Standards on Mineral Resources and Reserves, Definitions and Guidelines: CIM
Standing Committee on Reserve Definitions”, (December 2005).
2. Caterpillar, “Caterpillar Performance Handbook”, (34 Ed. Caterpillar Inc., Peoria, Illinois, U.S.A. 1999), 1282.
3. Chiles, J.P., and Delfiner, P., Geostatistics; (New York, Wiley Press, 1999), 695.
4. Companhia Vale do Rio Doce. “Avaliação Geoestatística de Recursos – Projeto Bauxita Paragominas Depósitos Miltônia 3 e Miltônia 5”, (February
2006).
5. JORC, “Australasian Code for Reporting of Exploration Results, Mineral Resources, and Ore Reserves (the JORC Code): The Joint Ore Reserves
Committee of The Australasian Institute of Mining and Metallurgy,” (Australian Institute of Geoscientists and Minerals Council of Australia”, December
2004).
6. Kotschoubey B., Truckenbrodt W. & Hieronymus B. 1997. Bauxite deposits of Paragominas. In: A. Carvalho, B. Boulangé, A.J. Melfi & Y. Lucas (eds.)
Brazilian Bauxites. USP/FAPESP/ORSTOM, São Paulo, p. 75-106.
7. Kotschoubey, B., Calaf, J.M.C., Lobato, A.C.C., Leite, A.S., Azevedo, C.H.D. Caracterização e Gênese dos Depósitos de Bauxita da Província Bauxitífera
de Paragominas, Nordeste da bacia do Grajaú, Nordeste do Pará/Oeste do Maranhão. In: Queiroz, E.T., Mrini, O.J., Ramos, B.W.(Org.).Caracterização
de depósitos minerais em distritos mineiros da Amazônia. Ouro Preto: DNPM/CT-MINERAL/FINEP/ADIMB, 2006, v. p. 688-782.
8. MRDI, “Update to the Reserves Audited in 2000 for the F-3 and 20-F Filing of 31 December 2000 Reserves with the SEC”, (Unpublished Internal Report
to Companhia Vale do Rio Doce, June 2001).
9. Pincock, Allen, and Holt, “Audit of Reserves”; (Unpublished Internal Report to Companhia Vale do Rio Doce, May 2005).
10. SANTOS, Jr. A. E. & ROSSETTI. D.F. 2002. Influência estuarina na Formação Ipixuna, porção média do rio Capim, Sub-bacia de Cametá. In: Simp.
Cretáceo do Brasil, 6, Boletim, p. 1-5.
11. Sinclair, A.J., “Evaluation of errors in paired analytical data by a linear model,” Exploration and Mining Geology, (7) (1) (1999), 167–173.

16 Proceedings of the 8th International Alumina Quality Workshop • 2008

AQW2008_Proceedings_FINAL-1-120.16 16 4/08/2008 2:54:54 AM