--> An Empirical Determination of the Minimum Number of Measurements to Estimate the Mean Random Vitrinite Reflectance of Disseminated Organic Matter, by M. Pawlewicz and C. Barker; #90986 (1994).

Datapages, Inc.Print this page

Abstract: An Empirical Determination of the Minimum Number of Measurements to Estimate the Mean Random Vitrinite Reflectance of Disseminated Organic Matter

Mark Pawlewicz, Charles Barker

Statistical methods for coal samples recommend 100 measurements to estimate mean vitrinite reflectance (Rv-r) to within +/-2 percent and 50 measurements to estimate mean vitrinite reflectance (Rv-r) for dispersed organic matter (DOM). The smaller sample size for DOM poses a problem because the standard deviations for DOM reflectance are typically larger. However, in studies of DOM thermal maturation, attaining 50 measurements can be unrealistic given the small amount of vitrinite in some samples. Our empirical approach to how many measurements are needed to adequately define the mean is to compute reflectance distribution statistics (mean, standard deviation, skewness, and kurtosis) in increments of ten measurements. This study compares these cumulative incremental computations of Rv- statistics with one computed using all measurements for that sample.

Vitrinite reflectance was measured on mudstone and sandstone cuttings with a wide range of thermal maturation and a comparable humic DOM with depth, from borehole M-25 in the geothermal system of Cerro Prieto, Mexico. The results suggest that after 30 measurements the mean Rv-r is generally known within 5 percent and always within 12 percent of that calculated using all particles measured. Thus, the precision after measuring only 30 particles is in good agreement to one decimal place as is recommended for DOM Rv-r. The co-efficient of variation (V=standard deviation/mean) is proposed as a statistic to indicate the reliability of mean Rv-r estimates made at n^Lt20. This preliminary study suggests a V<0.1 indicates a reliable mean and V>0.2 an unreliable mean for small samples. Fo a value of V between 0.1 and 0.2 the predictability is variable and dependent on the sample OM homogeneity.

AAPG Search and Discovery Article #90986©1994 AAPG Annual Convention, Denver, Colorado, June 12-15, 1994