Mineral reserve and mineral resource overview

Due to the subsequent decline in the gold price, the Argonaut Project was put on hold pending an improvement in the gold market. With this improvement now realising, the Argonaut Project again presents itself as an opportunity. The Argonaut Project has an Inferred Mineral Resource of 26.33 million ounces. Management believes that the Argonaut Project provides a huge opportunity for all shareholders to participate in future “blue sky” gold mining in an optimistic gold market.

9 GROWTH POTENTIAL

DRD’s strategy remains that of growth and diversification through discovery and/or acquisition of new Mineral Resources and Mineral Reserves. DRD has established specific objectives that will ensure sustainable, profitable growth for the Company within acceptable risk parameters. Acquisitions will be considered at any stage on the development curve ranging from greenfields projects to mature operating mines. Of paramount importance in the growth strategy is the search for quality assets. The minimum requirements for acquisition is the enhancement of the DRD Mineral Resource and Mineral Reserve base through DRD management being able to effect an improvement of the assets performance through implementation of expansion and renewal programmes supported by capital expenditure.

A hybrid strategy towards growth will be adopted which considers corporate acquisitions, producing asset acquisitions, advanced exploration asset acquisitions, strategic exploration partnerships and in-house expansion, “organic growth”. This strategy towards growth is preferred as it spreads the risks involved and lowers costs.

Apart from the Argonaut Project, growth in South Africa is largely limited to operational organic growth or mature mines that are rationalised by other companies. The growth opportunities within South Africa have become very limited due to the completion of the restructuring process that has taken place within the South African gold mining industry. The growth potential for DRD lies largely offshore with the greatest opportunities currently presenting themselves in the Australasian region. This is therefore the region where DRD believes efforts should be focused.

DRD has become increasingly active in pursuing appropriate projects for gold exploration and acquisitions. Several acquisition projects have been considered and pursued, however most have proved to offer limited return. This refusal to acquire low return operations has contributed to our strong stock performance over the last year. DRD will continue to build on its position as a leading South African gold producer and strive to establish and entrench its position as one of the world’s premier international gold mining companies.

ASSESSMENT CRITERIA USED IN THE COMPILATION OF THE MINERAL RESOURCE AND MINERAL RESERVE – SOUTH AFRICAN OPERATIONS

Data density – underground
On-reef development is sampled on a two-metre grid. The mined ore-body is sampled on a grid varying from 3 m x 5 m to 6 m x 10 m, depending on the ore-body. This information is used to project reef characteristics for Mineral Resource into and beyond development. Argonaut data was digitised by Rand Mines in the 1980’s, regularised to a 100 m x 100 m grid and this regularised data used to project reef characteristics for Mineral Resource beyond stoping and development.

Data density – surface
All sand and slime sources are drilled and sampled to a grid pattern sufficient to clearly define the physical, metallurgical and grade structure of the deposit. Additional drilling and bulk sampling then tests the valuation model. Rock dumps are valued from bulk samples and historical records.

Data density – open-pit
Mineral Resources are based on a combination of exploration drill holes and blast hole sampling. Selective mining of the ore-body takes place based on the blast hole results and geological controls. The Mineral Reserve figures quoted are based on historical selective mining and grade-control efficiencies.

Geological interpretation
The ore-body has been classified into geo-zones with similar grade characteristics by its macro features. These geo-zones can be recognised in exploration drilling and development, and grade characteristics.

Geological interpretation – surface dumps
The specific grade zones, metallurgical variances, sedimentological changes and contamination structures identified from drilling and test work, together with historical information, are used to create a model for reclamation planning.

Sampling technique
Underground sampling is by means of hammer and chisel sampling averaging 1,5 kg samples of mineralised material. This is followed up for ore accounting by broken ore sampling (“BOS”) and “go-belt” sampling. Surface deposits are mainly sampled by means of auger holes. Samples are taken at 1,5 m increments, or at such specific intervals as deemed necessary to clearly define the deposit. Archive and silt deposits are sampled by trenching.

Quality of assay data
Independent and Company assay laboratories are used. Underground chip samples are assayed by fire assay using 25 g charges, applying discounts for silver-by-silver discount chart. At the operations, 10% of all chip samples are re-assayed and parted to confirm the validity of the silver discount chart. All other samples are on 50 g charges completed by parting with nitric acid to account for the silver content. Monthly re-assays and checks on standards, mill products and mine samples are conducted with external laboratories. Throughput at the laboratories is generally in excess of 300 samples per day. Turnaround time is generally within 12 to 24 hours.

Estimation techniques – underground
Normal and lognormal kriging is used for current mining areas. Measured, Indicated and Inferred blocks are estimated using regularised data in different sized blocks. All kriging is done within clearly defined geo-zones. Pillars or ground left within old mining areas have been evaluated using simple weighted average regression techniques. Surface deposits are evaluated also using these techniques.

Estimation techniques – open-pit
Lognormal third parameter normal kriging is done using the blast hole sampling and limited exploration drilling. Measured and Indicated blocks are estimated using 5 m x 5 m and 25 m x 25 m regularised data respectively. All kriging is done within clearly defined geo-zones. Inferred blocks re-allocated grades based on historical mined values within the same geo-zones.

Estimation techniques – surface
Mineral Resource categories are based on drilling density and metallurgical test work.

Treatment of pillars
Pillars that have been assessed as mining opportunities are included in the Mineral Resource and Mineral Reserve. An entire mining-evaluation team who will take into account costs, access, rock mechanics and site investigation does assessment.

Surface material – allocation of costs
The surface Mineral Resource is quoted at in situ tonnes and grades. The surface Mineral Reserve is quoted at delivered-to-the-plant tonnes and grades. Dump material screening is regarded as a mining cost and the Mineral Reserve is quoted at post-screening tonnes and grades. Pre-concentration of sand dump material is regarded as a metallurgical cost. That material is quoted at delivered-to-the-plant tonnes and grades.

Development waste
Dilution includes waste from on-reef development.

Cut-off calculation
Cut-off is based on the stoping or mining, transport and milling cost over the previous 12 months, and the production plan for the next 12 months. A different operational cut-off is applied to each surface deposit, depending on its composition, location and reclamation method.

Resource cut-off grades
The cut-off grade used for exclusion of blocks from the Mineral Resource was based on a gold price of R102 500 per kilogram ($290 per ounce) at an exchange rate of R11: US1.

ASSESSMENT CRITERIA USED IN THE COMPILATION OF THE RESOURCE AND RESERVE – TOLUKUMA OPERATION (JORC Terminology)

Data density – underground
All on-vein horizontal development faces are sampled, which provides data at 1.4 m intervals along strike. The vertical interval between levels varies between 8 m and 20 m. Exploration diamond drill holes provide data beyond development on a grid ranging from 30 m x 30 m to 100 m x 100 m.

Data density – open pit
Pit floors are channel sampled across the entire pit floor at 5 m intervals along strike. Every second flitch is sampled, which provides data on a 5 m x 5 m grid.

Geological interpretation
The ore-body has been classified into geological zones based on mineralogy, vein strike and dip and the Ag:Au ratio. These are used to define estimation domains.

Sampling techniques
Underground faces are duplicate-sampled by two rows of random chip samples over geologically defined intervals across the face. The results are length-weighted averaged for calculation purposes. Pit floor samples are taken across geologically defined intervals from channels excavated by shallow ripping by bulldozer. Face and pit samples have an average mass of 1.5 kg. Drill core samples are from half-core from geologically defined intervals. Core recovery varies from 25% to 100%. Core samples have an average mass of 1.0 kg. Muck-pile and stockpile sampling is used for metal accounting and mill feed grade and metallurgical control. These samples have an average mass of 5 kg.

Quality of the assay data
All samples are dried and crushed, with 200 g split out and pulverised. A second split is taken from every tenth sample for quality control of sample preparation procedures. Assaying is by “aqua regia digest” with an “AAS finish”, using standard procedures. Every tenth pulp is duplicate-assayed for quality control of analytical procedures. All samples are routinely assayed for gold, silver, antimony and mercury. Site laboratory results are checked against fire assays of random mine sample pulps by outside laboratories. These show that the “aqua regia digest” consistently under-estimates gold by 12% and silver by 7%.

Tonnage volume factor
The ore density is highly variable, ranging from 1.8 tpm3 to 3.1 tpm3. It is not practical to determine the density of every sample so an average factor of 2.24 tpm3 based on drill core samples is used.

Estimation techniques
All model blocks are 5 m vertical x 5 m north south. The block thickness (east-west dimension) is determined from DTM models of the hangingwall and footwall of the vein defined from sampling. Vein thickness variability is the largest single source of error in resource estimation. Grade estimations for gold, silver, antimony and mercury are by inverse distance squared within hard-boundary geological domains. The domains are based on geological fundamentals (such as vein orientation and mineralogy) but also data density (development versus drill hole data). All assay data are tagged by domain, as are all model cells, and only data specific to a domain is used to model the grade of that domain. Use of hard domain boundaries causes sharp model grade changes, but in reality most domain boundaries are only fuzzy over a distance of 2 m to 10 m, i.e. one or two model cells, so this does not introduce significant errors.

Resource classification
Measured Resources are developed resources, plus a 10 m down-dip projection. Indicated Resources are based on a 10 m down-dip projection beyond Measured Resources, or areas that have been drilled on a 30 m x 30 m grid. Inferred Resources are based on widely spaced drilling and geological projection within the drilled area, but also include remnants within and adjacent to mined areas where mining problems may exist. Some areas of mineralisation that have sufficient data for classification as Measured or Indicated Resource have been downgraded to Inferred because of known mining problems.

Mineral Reserve estimation
Mineral Reserves are derived from the Measured and Indicated Resource by applying call and dilution factors derived from mining history. Dilution is applied as a skin around the resource (with the thickness depending on intended mining method), with an additional factor applied for waste extracted with broken ore in backfill stopes. The call factor applied to contained gold varies by mining method and stope height for back-filled stopes. Mineral Reserve contents are estimated by mining area based on the 5 m x 5 m model intersected with the planned mining perimeters. No publishable Mineral Reserves are based on the Inferred Resource.

Reserve classification
In general terms, Proven Reserves are derived from the Measured Resource and Probable Reserves are derived from the Indicated Resource. However, after application of dilution and recovery factors some of the resource is excluded from the reserve. Revision of the mine plan could cause some of this material to be reclassified as reserve.