All results estimated by the author unless otherwise noted.
Zone 1 – 169 Mlb Zone 2 & 3 – 214 Mlb (Note 1) Salem – 5 Mlb Ida Dome – 25 Mlb (JORC)
Total – 413 Mlb
Estimates are now rounded to the nearest million pounds. The estimation techniques are not accurate enough to justify more precision.
Note 1: The southern end of Zone 2 is vague and there is no clear indication of where new zones might start. Consequently I have included all the data down to 7500000N as “Zone 2 & 3”, and I have assumed that the uranium mineralisation is contiguous down to this line. It is likely that this uncertainty will continue until it is positively established where the geological splits in the resource are.
Previous Estimates by the Author
Zone 1 - based on drill data to 31 Mar 09 - 107.6 Mlb - based on drill data to 27 Apr 09 - 117.9 Mlb - based on drill data to 28 May 09 - 127.8 Mlb - based on drill data to 31 Aug 09 - 143.6 Mlb - based on drill data to 9 Oct 09 – 155.7 Mlb - based on drill data to 18 Nov 09 – 169 Mlb (latest release by EXT 2 Jul 09 - 145 Mlb)
Zones 2 & 3 - based on drill data to 28 May 09 - 103.6 Mlb - based on drill data to 31 Aug 09 – 178.0 Mlb - based on drill data to 9 Oct 09 – 181.2 Mlb - based on drill data to 18 Nov 09 – 214 Mlb (latest release by EXT 22 Jul 09 - 122 Mlb)
Salem - based on Fig 1 App 1 of EXT report 31 Aug 09 - 6.1 to 10.6 Mlb - based on drill data to 18 Nov 09 – 5 Mlb
Ida Dome – 25 Mlb (JORC)
Total (including Ida Dome and Salem) - 31 Mar 09 – 133.0 Mlb - 27 Apr 09 – 143.3 Mlb - 28 May 09 – 256.8 Mlb - 31 Aug 09 – 353.1 to 357.6 Mlb - 9 Oct 09 - 368.4 to 372.9 Mlb - 18 Nov 09 - 413 Mlb
Resource Estimation
The resource estimate for Zone 1 using data up to 18 Nov 09 came in at 169 Mlb, a typical increase in Zone 1. This shows deeper and higher grade results as more drilling is carried out.
The Zone 2 estimate based on drill data to 18 Nov 09 is 214 Mlb, a significant increase on the 181.2 Mlb on 31 Aug 09. This is partly due to the inclusion of drill reports down to 7500000N, which have greatly expanded the area of Zone 2 by 800 metres in a north-south direction, and between 500 to 1,750 metres in an east-west direction. It is not yet clear that this area is entirely contiguous, there may geological breaks as the zone appears to bifurcate. If it is not contiguous there may be a slight overestimate of the southern 400 metres or so, but as the overall estimation technique is fairly conservative this may have little effect on the total.
The northern part of Zone 2, where the drill results are spaced at approximately 100m, is estimated using a geostatistical program called ILWIS. The southern part, where the drill results are spaced at about 400m, is estimated using a simpler averaging method, as the spacing is too great to use ILWIS.
The Salem estimate was calculated using both ILWIS and the simpler averaging method used in the southern half of Zone 2. The ILWIS program did not give convincing results, and changes to pixel size (which should have had negligible effect) resulted in large changes to the estimate. It is possible that the number of data points (14) is too small for ILWIS to be properly used. The simpler averaging method gave good correlation with earlier estimates using the median of all data in Rossing South (excluding Ida Dome), and appeared to be “about right” to the author based upon examination of the drill results and previous experience gained in using estimation methods for Rossing South.
Detailed Description of Method - Zones 1 & 2
Resource estimation was carried out using a program called ILWIS (v 3.6). Drill data was interpolated using a moving average operation. The moving average operation is a point interpolation which requires a point map as input and returns a raster map as output. The values for the output pixels are the weighted averages of input point values. Weighted averaging is the calculation of the sum of the products of weights and point values, divided by the sum of weights.
The weight factors for the input points are calculated by a user-specified weight function. There are two methods: inverse distance and linear decrease. Both methods ensure that points which are close to an output pixel obtain large weights and that points which are farther away from an output pixel obtain small weights. Values of points which are close to an output pixel are thus of greater importance to output pixel value than the values of points which are farther away.
By specifying a limiting distance, you can determine the distance from any output pixel that points will be taken into account for the calculation a value for that output pixel; for each output pixel, only the values of the points which fall within the limiting distance to this output pixel will be used. Values of points that are farther away from an output pixel than the specified limiting distance, obtain weight zero by the weight calculation, and these values will thus not be used in the output pixel value calculation. This speeds up the calculation and prevents artefacts.
Several methods were trialled using the 27 Mar 09 data set for Zone 1 (see next section for the reason why this data set was chosen). This included both weight factor methods (inverse and linear decrease), different weight factors (0.5, 1 and 2), different limiting distances (up to 300m), and different pixel sizes (1x1 m and 5 x 5 m). Other interpolation methods were trialled, including Moving Surface (Plane), Moving Surface (2nd degree linear) and an attempt at Ordinary Kriging.
As a result of all these trials, the best method was determined to be Moving Average with a linear decrease (factor 1) and a limiting distance of 150 m, with a pixel size of 5 x 5 m (smaller pixel sizes made little difference and greatly increased processing time).
Ordinary Kriging (OK) is a more sophisticated form of averaging often used for resource estimation by professional geologists. Instead of just using a linear or a inverse decrease for the weight function, more complex weight functions can be used. These functions are determined by statistical analysis of the data set. This analysis requires rather sophisticated interpretation of the data set, and relies on the (geological) experience of the user. The simpler nature of my techniques did not seem to justify the use of OK, and I am not nearly qualified enough to make appropriate interpretations of the data set when determining the weight functions.
I considered it safer to use a simpler interpolation method that did not rely on interpretation. This also was better suited to the areas of the zones where less data was available.
After interpolation was carried out, the resulting raster map was converted to a point map. Pixels containing very small amounts (less than 1 lb of uranium per square metre) were discarded. These were mostly pixels on the outer edge of the resource and were probably an effect of the interpolation software than an actual picture of the resource. The pixels were then totalled to arrive at a resource estimate.
Data Set
All drill data released by EXT was used in the resource estimate. Corrections were made to radiometric data in accordance with the formula given in the NI 43-101 report paragraph 17.3.1. The specific gravity (SG) is the same as that assumed by the NI 43-101 report. Corrections were made to account for drill angle (usually 60 degrees for most data).
An attempt was made to recreate the data set used when the initial resource was released in the NI 43-101 report. If I had such a data set, I could compare my own method of estimation with the known value of the JORC estimate of 108.3 Mlb. The NI 43-101 report made an estimate based upon 11 diamond and 174 RC drill holes (paragraph 17). Using drill data to 27 Mar 09, 154 drill holes are found of which 12 are diamond. It's not quite right but is the closest match.
Using the map from the NI report (page 33) showing all the drill holes used to determine the resource, the data set to 27 Mar 09 was overlaid on top of this. There is a close correlation between the 27 Mar 09 data set and the NI report. Using this data set, a result of 107.6 Mlb was calculated for Zone 1, remarkably close to the NI 43-101 calculation of 108.3 Mlb. Later calculations showed reasonable correspondence with the official reports issued by Extract.
Extract uses radiometric data when calculating resource estimates. This is available to the company quickly and is not usually released separately. Later, chemical analysis of drilled holes is released once it has been completed - usually two months after the radiometric data. As a result, it is impossible to exactly correlate the data used by Extract for resource estimation with the data released.
To ensure a firm edge to the data, a series of false data points with zero uranium is spaced 50 m around the edge of the resource. This forces the program to sharply delineate the resource near the edges of the current drilling.
Detailed Description of Method – Zones 2 (southern end) & 3, and Salem
The southern part of Zones 2 & 3, where the drill results are spaced at about 400m, is estimated using a simpler averaging method, as the spacing is too great to use ILWIS. Salem also uses this estimation method as the number of data points (14) seem to be too few to get consistent results from ILWIS.
The averaging method looks at each line of data points running in an east-west direction. the amount of uranium for each data point is averaged and an amount per square metre is calculated. Between each pair of east-west lines, the uranium per square metre is averaged again, and the width east-west is averaged between the two lines. The area is usually a trapezium (UK English) or a trapezoid (US English).
A buffer of 50 m is allowed for in an east-west direction to match the 50 m edge used by the ILWIS program. No buffer is used in a north-south direction.
The result is usually conservative because the east-west lines are only to a medium depth as a guide to further exploration. Infill drilling is usually to greater depths and discovered areas of greater mineralisation are usually targeted by the drilling program. This results in infill drilling data points being higher in uranium content and, once the ILWIS program can be used, result in a greater total for the resource estimate.
Disclaimer
It is still possible that there are errors in my calculations, which will lead to errors in estimation of the resource size. However, the results obtained so far have been reasonably close and it is assumed that the differences are caused by the use of different data that is released later by Extract. The estimates given here are intended to be a general, quick look at what might be possible for Rossing South, with no claims as to their accuracy.
EXT Price at posting:
$8.34 Sentiment: Buy Disclosure: Held
(20min delay)