Metal mining technical guidance: revised guidance for sample sorting, chapter 4

Calculation of Subsampler Correction Factors, Precision and Accuracy

• Calculation of subsampler correction factors
• Determination of subsampling accuracy and precision

Calculation of subsampler correction factors

Subsample correction factors and calculation of estimates of whole sample densities:

The correction factor and equation to convert subsample number to estimates for the entire sample are as follows:

1. Correction Factor (CF) = total sample area, volume or weight / subsample(s) area, volume or weight sorted to achieve the minimum 300 organism count
2. Whole density estimate = number of organisms in subsample(s) x CF

Examples:

• Area-based methods
• Weight based methods
• Volume-based methods

Area-based methods

Correction Factor = total area of sieve (415 cm2)/ # of subsamples sorted x area of each subsampled (41.5 cm2).

• 2 subsamples sorted to achieve 300 count
• final count upon sorting the whole 2nd subsample in it's entirety is 347
• Correction factor = 415 / 2 x 41.5 ml = 5 Estimate of total number of organisms in sample = 347 x 5 = 1735

Weight based methods

Correction Factor = total weight of sample 30.5 g (i.e., sample + sieve wgt - sieve wgt) / total weight of all subsamples sorted (11.7 g)

• 3 subsamples sorted to achieve > 300 count
• final count upon sorting the whole 3th weight fraction in it's entirety is 419
• Correction factor = 30.5 g / 11.7 g = 2.6

Estimate of total number of organisms in sample = 419 x 2.6 = 1089

Volume-based methods

Correction Factor = total volume of sample in Imhoff cone (1000 ml)/ # tubes sorted x tube volume (55ml)

• 4 tubes sorted to achieve > 300 count
• final count upon sorting the whole 4th tube in it's entirety is 365
• Correction factor = 1000ml / 4 x 55ml = 4.55 . Estimate of total number of organisms in sample = 365 x 4.55 = 1661

Determination of subsampling accuracy and precision

The effects of subsampling on abundance estimates should be examined on a minimum of 10% of the samples. If the error exceeds 20% for any group of samples, all samples within that group of samples should be completely sorted to assure the subsampling process is not compromising data integrity. This requires that 10% of samples which have been subsampled are randomly selected and the remaining unsorted material is sorted in it's entirety. The estimates (calculated as above) are then compared to the actual counts from the sample and the accuracy of the estimates and the precision between subsamples can calculated as below:

• Accuracy of the subsampling estimate
• Precision between subsamples

Accuracy of the subsampling estimate

% Error in the estimate = [1-(estimated # in sample/actual # in sample)]x100

Example (repeated from Section 3.2)

1. a count in subsample A = 289, representing 15% of the sample by volume, for an estimate of the total in the sample of 1927
2. a count in subsample B = 316, representing 15% of the sample by volume, for an estimate of the total in the sample of 2106
3. the count in the remainder of the sample = 1359, for a actual total of 1964
4. the reported precision would be the same as in the first example, 8.5 % the reported accuracy would be -1.9% and +7.2% for sample A and B respectively.

Precision between subsamples

% Difference between two subsamples (A& B) = [1- (count in subsample A / count in subsample B)] x 100

1. Example (repeated from from Section 3.2)
2. a count in subsample A = 289
3. a count in subsample B = 316
4. the reported precision between theses two subsamples would be 8.5% (1-(289/316))x100.