Introducing Improved Performance of Boxplot (New Method) in Estimating the Threshold (Separating Anomalies from the Background)

Hanza district is located in the southern part of Urumieh-Dokhtar Metallogenic belt in southeastern Iran. This region includes fourteen significant Cu anomalies associated with porphyry/vein-type copper mineralizations. This research is aimed to examine effective various processing techniques (i.e., univariate statistical methods) in the analysis of stream sediment geochemical datasets for separating anomalies from the background. The threshold values were estimated by various ways, including 1. [Mean +/- 2 x Standard deviation (X + 2S)], 2.Median Absolute Deviation (MAD), 3. Exploratory Data Analysis (EDA or Tukey's boxplot) and 4. Improved Performance of Boxplot (IPB). The latter method is introduced as an original method for the first time. The results derived from these methods; have been mapped in four classes associated with the first quartile, third quartile, and threshold values, respectively. The primary approach of this research is to investigate the effectiveness of each method in the identification of fourteen significant copper anomalies. The efficiency of each method is determined based on the "success rate". To compare threshold estimation methods, the Dot Frequency Distribution Histogram was used, and the position of threshold values of various methods was displayed on the histogram. The results of threshold estimation have shown that MAD and IPB methods have the least threshold values for most elements compared to other methods. The efficiency of MAD and IPB methods for the threshold estimation is high. The threshold value estimated by EDA and [Mean +/- 2Sdev] methods, in comparison with other methods, is high. Indeed, their efficiency is abysmal because using these methods will eliminate many of the anomalous areas. Thus, using the [Mean +/- 2Sdev] and EDA methods is not suggested.