Evaluation of Placer Deposit Within Ghana a Case Study of the Kwabeng Placer Gold Deposits North Western Flank of Atewa Range, Eastern Region, Ghana

The estimation of Mineral Resources/Reserves (MR) together with the quantification of the distribution of metallurgical attributes provide strong evaluation of the large alluvial gold deposits within Kwabeng and its surrounding , which occurs along the north western flank of the famous Atewa Range. This reduces the risk arising from uncertainties within the ore body as far as geological and metallurgical information are concern and also improve decision making process during the production stage of the


INTRODUCTION Background
Alluvial gold deposits occurred throughout the country in favorable areas overlying Birimian metavolcanics and metasediments rocks. The vast majority of historical gold production from Ghana prior to the 20th century came from a myriad of small streams and rivers draining areas with underlying oxide and primary gold deposits (Hilson, 2002). In addition, several of the major rivers have been mined with dredges starting in the early 1900s and in the 1990s, large alluvial operations were started up on river valleys in several alluvial gold districts. With the recent increase in the gold price over the past several years, the amount of small-scale gold mining, primarily of alluvial deposits, has increased dramatically throughout Ghana and especially in places like the Atewa Range, where substantial alluvial gold resources are known. Those of the Offin alluvial gold occurred along the River Offin, north of Dunkwa in the central region of Ghana. As reported by Levich (2010), the lower gold bearing gravel lies on top of weathered Birimian with thickness ranging between 2-6m and is overlain by similar thickness of gravel barren overburden consisting of silt and clay materials. The gold bearing gravels of the Offin River deposit contains large pebbles to cobbles size clast up to15cm in diameter, in matrix of quartz and heavy minerals. The large alluvial occurrences along the flank of the Atewa range also represent some of the largest alluvial gold deposit in the country. The deposits are found within the valleys and are covered by silt and clay-rich overburden. The overburden is massive and is in sharp contact with the underlying gravels which also filled most of the valleys within the area. The overburden within the area is very thick and sometimes can be deep as 4-6m although some few areas may record very thin overburden where the gravel section seems to appear on the surface (Griffis et al 1989). The Kibi area h a s long been known for very significant alluvial gold and although some dredging was attempted within the Upper Birimian, most of the production has come from dozens of the smaller drainages all along the flanks of the Atewa Range. The most famous historical site was at Pusupusu where artisanal mining took place for generations and several sluicing operations were carried out in the early 1900s. Junner (1935) indicated that the PusuPusu area features boulder gravels up to 2.5m thick and these are overlain by clay/silt overburden of a similar thickness; very coarse nuggets (up to 8oz) and recorded production from mining companies during 1920s was more than 267kg from about 298,000 cubic meters of gravel with a recovered grade of 0.8 to 0.9g/m. 3 A very similar pattern seems to prevail in most of the valleys along the Atewa Range; these include the established resources along the Awusu (Kwabeng) Merepong (Pameng), upper Birim (Apapam) and Akusu (Osino/Saaman) drainages. Gold-bearing basal gravels 1-3m thick sit on top of weathered Birimian units, mainly metasediments, and the gravels are invariably overlain by barren sands, silt and clay of comparable thickness. The large alluvial gold resources of Kwabeng Mine are located along Awusu River and its tributaries within favorable geological environments of primary gold deposits. Which is believed to have originated from the weathering of near auriferous quartz veins which are known to occur not only in the headwaters of the major streams from Atewa Range but also in some of the lower valleys (Griffis et al., 1989). This is evident by the large coarse gold recovered from the Goldenrae plant on the Kwabeng concession. Most alluvial gold projects have failed basically due to several factors which are common within the mining industry. But those that have contributed significantly to the failure as far as placer deposit is concern is the poor resource estimation and metallurgical factors reporting (Rae et al, 2006). The issue of placer deposit assessment has resulted in the collapse of many alluvial mines than any other reason. In addressing some of the issues concerning placer deposits assessments, Boostein (2013) outlined some factors one needs to consider when evaluating alluvial deposits. According to him, there must be a relatively large sample size; because most of the deposits are composed of different sizes of gravels which makes it very difficult to obtain a representative sample. Also higher values of gold within certain areas distort the calculation of the resource. In addition the values of the gold recorded must be observed and noted with items associated with gravels such as the boulder sizes, volume of clay, bedrock conditions, and any other physical properties that may have impact on the mining and processing the ore.

Study area
The study area (Fig 1) is located at the northwestern flank of the Atewa Range. The property is measured 44km 2 and is bounded by latitudes 6° 20' ,6° 16' N and Longitude 0°38', 0° 36'W and lies within the Atewa district in the Eastern Region of Ghana.   (Fig 2) is the easternmost belts of the Birimian rocks. It lies next (in east direction) to the wellknown prolific Ashanti belt which host most of the Birimian gold with major active gold mines located in there. According to Simone et al., (2012), the north-east trending Kibi greenstone belt is approximately 60km long with it northern extension overlain by younger, flat lying sediments of the Voltain basin where as the southern extremity is truncated by a large granitoids batholiths. The Kibi Belt geology consists of a tightly folded, slightly overturned, north north-east trending syncline featuring a variety of metavolcanics, mafic intrusions and metasediments. The western margin (where the Kwabeng deposit is situated) of the belt is dominated by steeply dipping, highly deformed and fractured metasediments, often with extensive bands of graphite. These are flanked by metavolcanic flows (mafic to intermediate) and metavolcaniclastics with interbedded metasediments; within these there are numerous shallow level mafic plutons (mostly epidiorites). In the central part of the belt, which more or less coincides with the main Birim River valley, the dominant lithologies are clastic metasediments, some of them quite coarse, whereas on the eastern margin of the belt, along the summit of the prominent Apedwa hills, mafic and intermediate to possibly felsic metavolcanics, are more widespread and appear to be overturned to the northwest. More details on the Kibi gold belt can be found in Griffis (2002). Detailed work carried by Griffis et al., (1989) reported reveals the valleys within the area are underlain by thick sequences of alluvial deposit with several cross-cutting structures observed along the flanks of the Atewa range. With some of these inferred structures corresponding to the valleys hosting significant alluvial gold occurrences which maybe primary lode gold sources. 2.0 MATERIALS AND METHODS. All pits data were derived from the geological database of Kwabeng Mine . A typical hand dug tools consisting of axe, long sharp metal and spade were used for the pitting as opposed to the mechanized one which uses banka drilling machine.

Pitting/processing of samples
A 1m x 1m hand dug pitting on grid density of 200x50m was used in getting through the overburden, gravel section and to the bed rock during the pitting activities. Hand dug pit into placer deposit is one of the oldest technique use to sample alluvial deposits. Pitting is more reliable than drilling especially in areas where the deposit occur close to the bedrock because it gives more reliable information about various sections and also provide more representative samples (Rae et al 2006). Bulk samples from the exploration program were processed using the prospector (a small portable wash plant) where final concentrate were collected and pan. The recorded gold from the pan concentrate are then recorded in grams together with the pit ID. According to Rae et al ., (2006), it may be preferable in some cases, to utilize a cement mixer to wash each of the pit samples in order to ensure that all of the gravels are really well cleaned so that all the gold is liberated and subsequently recovered in portable Knelson bowl. However, for some large pitting programs, this method may be a little impractical to be used on routine basis but it should at least be used to check and confirm results from other pits where the gravels were processed through a Prospector.

Data analysis
Data analysis which forms a small portion of the overall scheme of the work, has the potential of revealing database errors which can be detrimental to the efficient reporting of resource volume , grade and classification. The histogram and box-plot have been used to graphically examine the dataset. In addition to the graphical examination, descriptive statistics were carried out to compute statistical measure such as mean, mode, variance, standard deviation etc.

Geospatial interpolation
Geospatial technique was used for the interpolation of the gold grade and gravel thickness of the deposit to determine the weighted average. The method used in this study is the inverse distance weighting (IDW) method which estimates spatial variability of parameters based on the weighteing of neighborhood sampled points. Several authors (e.g. Sarala The idea behind IDW interpolation is based on Tobler's first law of geography (1970) which states that "All places are related, but nearly places are more related than distant places". The estimation of values for unsampled areas are primary based on the nearest point than further away sampled points. To satisfy tobler's law, IDW uses another exponential number which gives more weighting to the close points during its estimation. For the purpose of this study, exponential power of two( 2) is used for the interpolation of the various parameters. This value is known to be the best exponential values that best fit the IDW estimation ( e.g Al-Hassan et al., 2015; Isaaks et, 1989). To evaluate the accuracy of the estimation cross validation using the root mean square error (RMSE) was initiated. This validation has been recommended by authors

Resource estimation
The resource potential was estimated to know the amount of fine gold within the study area after the estimation of the mean gold grade and gravel thickness. The method used for the resource potential here is given as P=VG after Chirico et al (2010), where P is the estimated resource potential of the study area, V is the volume of the gravel section (deposit) and G is the mean (average) grade of the deposit. The mean grade used for the final calculation of the gold resource potential was derived from the generated grade estimation from the IDW model.

Pit lithology
The various sections of the pits were logged to determine the overburden thickness, gravel thickness, depth to bedrock, the overburden materials as well the gravel section lithology. The depth of most of the pits ranges from 2-6m with most section being the overburden. Overburden ranges from 0-4.8m whiles the mineralized section ranges from 0.4 to 2.8m. The uppermost layer is mostly clay with laterite whiles the middle section comprises of quartz pebbles with silty clay or with laterite in some areas. The layer which usually ends the pit is the saprolite zone which continue as the bedrock .The various layers or sections encountered conform well with other alluvial deposits studied elsewhere (e.g Eberle et al, 2016; Jonas et al; 2017). They normally consist of three lithologically inhomogeneous units in vertical succession.

Statistical results
About 324 pits have been analyzed for the optimal resource estimation. Since the technique of determining normal distribution of datasets is either through graphical representation of the data or descriptive analysis, the histogram and box-whisker is used as the best graphical representation of the normal distribution of this geological datasets. It is important to have a normally distributed datasets before performing any inferential statistical analysis.
The graphical representation of the gold grade using the histogram and box plot show some skewness in the gold grade data. The Au grade is positively skewed as shown in Fig 3 with major outliers indicated in the box-whisker plot (Fig 5). Somewhat normal distribution was achieved after it was log10 transformed (Fig 4). The standard deviations, variance for the Au grade and mean grade include 0.3g/m 3 , 0.155g/m 3 and 0.49g/m 3 respectively. The minimum and maximum grades obtained values were 0g/m 3 and 2.73g/m 3 with a range of 2.73g/m 3 as indicated in Table 1.
Gravel section (ore zone ) shows some skeweness and various outliers as shown in the graphical histogram and box-whisker plot (Fig 7& 8). Natural log of the gravel thickness reveals normal distribution in the datasets (Fig 9). The arithmetic mean of the gravel thickness within the area is 2.5m.

Resource estimation
All log transformed data were back transformed for the purpose of resource estimation.The results of the gold resource are shown in the equation below. The gravel thickness and Au grade used for the calculation were derived from the IDW using the surfer software with the power of 2 as indicated in Table 2. The area was calculated from the polygon around the pit holes in Fig  10 using the MapInfo software. The area covered is 7.55Km 2 . P=GV, P is the resource potential, G is the mean grade of the gold grade model which is 0.46g/m 3 and V is the volume of the deposit. Volume of deposit=Area X mean gravel thickness V= 7550000m2 X 2.56m V=19328000m 3 P=19328000m 3 X 0.46g/m 3 P=8890880g This estimated resource with value 8890880g of gold may contain some impurities, so it was converted based on the purity of gold reported by Kwaben Mine as well as those purity of gold reported by Precious Minerals and Marketing Company (PMMC) from the study area. The purity from the area is noted to be 92%. Using a purity value of 92%, the gold resource was estimated as indicated below; At 92.0%, P=8890880g X0.92 P=8179609.6g P= 261,728 oz 4.0 DISCUSSION The hand dug pitting has provided the best technique for sampling the alluvial deposit. For major alluvial gold deposit evaluation in the country and elsewhere , Banka drilling has been used which sometimes give poor sampling representation (Rae et al., 2006). The hand dug pit used in this study is seen as the best method for sampling alluvial deposit and this can be use elsewhere especially where the depth of the deposit ranges to a maximum of 8m. Beyond this depth, it is not preferable to use this method. The grid density 200mx50m for the pitting has recorded the same standard error as indicated in the Fig  (11) with other high dense grids (100mx50m, 50mx50) and is term as optimum grid density for evaluating placer deposit. It is only preferable, may be under some consideration to reduce to much closer grid in upstream areas where the grade seem to be high. The various section which span from the overburden, gravel section to the bedrock comprises of different geologic materials. The reddish brown clay/silty component of the alluvial deposits may have been formed by the weathering of the rocks of the Kibi belts rocks which comprises of metabasalts, phyllites, greywackes and quartz veins. The clay may be characterized by the minerals smectite, halloysite, chlorite and vermiculite as they are the main products of weathering of such rocks as discussed by Richard et al., (1987).

Fig11. Optimization plot of number of grid densities against standard error
The lateritic sections occur through the overburden and also within the gravel section. The small quartz pebbles found within the laterites maybe due to the quartz veins resistant to weathering found within the phylites or the metabasalts of the Kibi belt rocks. Laterites formation have been favored by the various rocks types, climatic conditions and the geomorphic features within the area. The composition of the laterites could likely be hematites, gibbsite and limonite which reflects the direct weathering/alteration of the minerals olivine, pyroxene, ca-pagioclase, K-feldspar and biotite. The hematite may have been formed from the alteration of the minerals; Olivine, pyroxene and biotite whiles that of gibbsite may have been formed from the weathering of Ca-plagioclase, and K-feldspar. The gravel section is characterized by quartz pebbles together with silty clayey or laterites in some areas. This is the mineralization section and contains gold of different sizes as observed by earlier prospectors (e.g Griffis et al., 1989& Rae et al ., 2006. The mineralization of gold within such layer brings to mind two different types of enrichment. Gold particles eroded and transported together with the gravels and later deposited within the are. Secondly, gold mineralization which formed as a result of continues weathering of mineralized rocks in in-situ. These two mineralization maybe differentiate based on the nature of the encountered gravels within the section. Those gravels carried from elsewhere may have developed roundness whiles those formed in-situ or near to the source may have some angular shapes. The estimated resource potential is based on the 92% purity of gold within the area that has been reported by Kwaben Mine themselves and also from the PMMC reports of gold purity from the area. At this purity, the gold resource within the study area has been estimated to be 261,728 oz. This estimation has been solely based on the 200mx50m pitting interval which is term as the optimal pitting interval for the Kwabeng placer gold T deposit. The determination of the processing parameters from logging of the deposit is very important as they render processing difficulties. The clay materials associated with alluvial deposits is one of the minerals considered before setting up the wash plant. Too much clay slows the throughput rate and thus lower the amount of gold recovered in an hour per washing leading to reduction in general quarterly production rate.

Conclusion
The Kwabeng placer deposit evaluation has primarily been based on the logging parameters from the hand dug pits. The hand dug pit provided representative samples for the purpose of the resource evaluation and deriving key processing parameters like gold grain sizes, clayey content, gravel sizes and silica cemented medium. Also the 200mx50m was used for the estimation of the resource which is seen as the optimal sampling density. In the overall evaluation process, it is established that the Kwabeng deposit contains much clay materials within the gravel sections (ore zone) with gold grain sizes decreasing downstream. The total resource estimate for the study area include 19,328,000m 3 volumes of gravels with average grade of 0.46g/m3 and estimated fine gold of 261,728 fine ounces.

Recommendations
It is recommended that additional studies be conducted within the area to provide a spatial model of the variation of the clay content within the mineralized zones as well as

International Journal of Engineering Research & Technology (IJERT)
ISSN: 2278-0181 http://www.ijert.org the spatial distribution of the gold grain sizes, as it provide the basis of building the alluvial wash plant. It was observed that some areas recorded much lower values of gold grade which affected the final estimated mean gold grade value. In future studies , such low grade areas must be block and resource estimate provided separately for such areas . This will provide a more detailed resource estimate of the area under study amidst high variation of gold grade within the area.