Dispersal Trains Research Articles

Late Wisconsinan buildup and wastage of the Innuitian Ice Sheet across southern Ellesmere Island, Nunavut

During the Late Wisconsinan, a precursor of the Prince of Wales Icefield, southern Ellesmere Island, formed a prodigious ice divide of the Innuitian Ice Sheet. Initial buildup occurred after 19 ka BP, when the icefield advanced west (inland) across Makinson Inlet from margins similar to present. Subsequent reversal of flow to the east required ice divide migration to the west onto a plateau that is largely ice-free today. From this divide, a trunk glacier flowed eastward through Makinson Inlet to join the Smith Sound Ice Stream en route to nothern Baffin Bay. Westward flow from this divide filled Baumann Fiord, depositing a granite dispersal train that extends a further 600 km across the archipelago to the polar continental shelf. Deglaciation of most of Makinson Inlet occurred catastrophically at ~9.3 ka BP, forming a calving bay that thinned the Innuitian divide, thereby triggering deglaciation of most of Baumann Fiord by 8.5 ka BP. Ninety14C dates on Holocene shells and driftwood constrain deglacial isochrones and postglacial emergence curves on opposite sides of the former Innuitian divide. Isobases drawn on the 8 ka BP shoreline rise northwest towards Eureka Sound, the axis of maximum former ice thickness. Ice margins on Ellesmere Island were similar to present from ~50–19 ka BP (spanning marine isotope stages 3 and 2). However, significant regional variation in ice extent during this interval is recorded by ice rafting from the Laurentide Ice Sheet into Baffin Bay. Later buildup of the Innuitian Ice Sheet occurred during the low global sea level that defines the last glacial maximum (18 ka BP). We also suggest that the Innuitian Ice Sheet was influenced by the buttressing and subsequent removal of the Greenland Ice Sheet along eastern Ellesmere Island.

A method for detecting glacial dispersal trains in till geochemicaldata

A new algorithm for identifying potential glacial dispersal trains in till geochemical data is introduced. The algorithm, referred to as the dispersal train identification algorithm (DTIA), requires a set of user input parameters used for simulating dispersal train characteristics. The algorithm employs a succession of wedge-shaped search regions (each one shaped like an isosceles triangle with the search point at the principal vertex) to ‘look’ in user-specified directions from a set of search points that may be mineral prospects, geochemical anomalies in rock or a homogeneous grid of points. The values of the till sample points occurring within the wedge are analysed as candidates for a dispersal train using a set of criteria. The geologist specifies the length and angle of the wedge, where the length reflects the distance of down-ice dispersal and the angle is used to model a ribbon or fan-shaped dispersal train. For each search position, DTIA calculates the average of the values within the wedge (pmed), the difference between wedge average and the local average (diff), and fits an exponential model to the within-wedge points (geochemical value as a function of distance from the search point). Candidate search points for the heads of dispersal trains are selected based on high values of pmed and diff, as well as best fit lines (fitted to the natural logarithms of geochemical values) that show high negative slopes (b), significant values of a test statistic for the slope magnitude (ts), and large negative correlation coefficients (R). The slope statistic, ts, is particularly useful in identifying candidate trains, but care must be taken to eliminate situations where the closest sample point is relatively far from the search point. DTIA was tested on two datasets: a dataset in Cape Breton where documented till dispersal trains exist and in the Swayze greenstone belt where no documented trains are known. The effects of varying the input parameters (e.g. length, angle and direction of wedge) were studied as well as different search strategies with the objective of identifying the most appropriate parameters for successfully identifying points that represent the heads (or source zones) of potential till dispersal trains. The algorithm performed well with the Cape Breton data. The results are particularly sensitive to both direction of search and the wedge angle. By systematically searching from a set of grid points, then screening out search points (and associated directions) that have (1) pmed values less than a threshold, (2) diff values less than a threshold, (3) distance of point closest to search point less than a threshold, (4) slope values less than zero and (5) ts values usually less than –2.0,a suite of candidate search points can be plotted on the till geochemical map, colour coded by search direction. These candidate points (and associated wedges) can then be visually appraised as potential dispersal trains. Application to the Swayze data revealed candidate trains whose directions are consistent with glacial transport directions known from data on glacial striations. Although this method shows promise as a data exploration tool, more work is needed to improve the algorithm and to test it under a greater range of conditions.

Dispersion of gold in Slesse Creek, British Columbia

Because of the high costs and difficult logistics of geochemical surveys in remote regions, sediments collected from large streams must often suffice to determine the presence or absence of mineralization. Here we study this problem in Slesse Creek, a gold-rich river that drains north from the Mt. Baker District, Washington, USA, to enter the Chilliwack River, British Columbia, c . 90 km east of Vancouver. Anomalous Au values are derived from the abandoned Red Mountain Mine just south of the USA–Canada border. Sediments were collected from Slesse Creek and its tributaries, and from other large streams in the Chilliwack River basin. Concentrations of Au in tributaries from the Red Mountain Mine range from 250 to 2330 ppb and are much greater than in other tributaries. Sediments from Slesse Creek also have anomalous, but extremely erratic, Au values. This erratic distribution of Au is generally similar to that of magnetic grains and associated elements (e.g. V) but is negatively correlated with elements, such as Na, associated with light minerals. Results are interpreted as an evolution in sediment geochemistry from the tributaries to Slesse Creek: composition of the former appears to be controlled by source (i.e. geology) whereas geochemistry of sediments in Slesse Creek has been strongly modified by fluvial processes that concentrate Au and other heavy minerals. This shift from source- to process-related geochemistry invalidates the traditional geochemical dilution model and gives long anomalous dispersion trains for Au. These provide suitable targets for low-density regional surveys, but are difficult to interpret because of their erratic character. The contribution of fluvial processes to the concentration of Au can, however, be evaluated by comparison with elements such as V, that are associated with ubiquitous heavy minerals, or by plotting data on x – y – z source–process plots.

Indicator mineral and till geochemical dispersal patterns associated with the Ranch Lake kimberlite, Lac de Gras region, NWT, Canada

The Ranch Lake kimberlite, in the Contwoyto Lake area of the northwest Territories, consists of volcaniclastic crater facies kimberlite which contains thousands of kimberlite indicator mineral grains in a 10 kg sample. Cr-diopside and pyrope are the most abundant indicator minerals in the kimberlite, accompanied by much less abundant chromite and rare Mg-ilmenite. Indicator minerals in till samples collected around the Ranch Lake kimberlite define a spectacular ribbon-shaped glacial dispersal train trending west for 70 km. At its head, the train is 500 m wide and gradually widens 2 km at 20 km down-ice. The lateral edges of the train are sharply defined by the presence or absence of indicator minerals in till. Background (up-ice) concentrations of all indicator minerals are zero. Patterns for Cr-diopside and pyrope abundance are similar and concentrations increase down-ice, reaching their highest levels between 15 and 19 km down-ice. Of the two size fractions of heavy minerals examined, most indicator minerals occur in the finer, 0.25 to 0.5 mm, size fraction. The <0.063 mm fraction of the till matrix displays weak geochemical signatures of the kimberlite down-ice, most notably for Ba, Cr, Ce and Th. The results of this study of indicator mineral and geochemical methods around a known kimberlite will aid in the design of mineral exploration programs for kimberlite-hosted diamond deposits in glaciated terrain around the world.

Influence of geology, glacial processes and land use on soil composition and Quaternary landscape evolution in The Burren National Park, Ireland

The Burren in County Clare, western Ireland is a glaciated karstic landscape with a fragmented soil and vegetation cover on a predominantly limestone substrate. A representative area approximately 4 km 2 in extent located within The Burren National Park was studied in order to determine the distributions of soil types and relate these to geology, glacial erosion and deposition, post-glacial mass movements and land use. Mineralogical and particle size analyses indicate the presence of two distinct soil types, one shale-bearing, the other granite-bearing. Geochemical analyses confirm this division and also reveal soils of mixed composition and a distinctive subtype of granite-bearing soil that occurs exclusively on drumlin-like landforms. The soils on the drumlins formed by the decalcification of limestone-rich glacial diamicton (till) containing granite and metamorphic rocks derived from County Galway to the north. The occurrence of mineralogically similar though geochemically heterogeneous soils throughout the area suggests that there was once an extensive thin cover of diamicton, which has almost entirely vanished because of pedogenesis and erosion. Soils containing shale fragments occur in a glacial dispersion train extending south from outcrops of a small body of shale-rich limestone breccia. Isolated patches of shale-bearing soil occurring up to 1.2 km from these outcrops may have formed by the weathering of ice-transported boulders. The presence of relatively deep soil cover on outcrops of the shale/limestone unit is attributed to Holocene weathering to form an erosion-resistant clay-rich soil. Tree charcoal, 14C-dated to 3300 BP and buried to a depth of 2 m by mixed granite- and shale-bearing soil, indicates that Bronze Age deforestation triggered mass movement and soil erosion. Denudation processes are still active in the Burren landscape.

The application of heavy indicator mineralogy in mineral exploration with emphasis on base metal indicators in glaciated metamorphic and plutonic terrains

Abstract Indicator mineralogy is used to explore for a wide variety of mineral commodities. The method utilizes minerals which are sufficiently heavy to be readily concentrated in the laboratory, often colourful and possess other useful physical and chemical properties. The minerals also must be source specific. Some indicator minerals are true resistate minerals. The others, although less resistant, are stable in oxidized glacial drift and many non-glacial sediments. A few, such as gold grains, are silt sized but most are coarse grained. Grain size has a major impact on indicator mineral dispersal patterns in glacial drift. The coarse-grained indicator minerals are of two main types: (1) kimberlite indicator minerals (KIMs); and (2) metamorphosed or magmatic massive sulphide indicator minerals (MMSIMs). KIMs are enriched in Mg and Cr and most MMSIMs are enriched in Mg, Mn, Al or Cr. These indicator elements cannot be diagnosed geochemically in anomalous heavy mineral concentrates because the concentrates contain other, more plentiful non-indicator minerals containing the same elements in the same chemical form. Chalcopyrite is also a very useful MMSIM but the number of surviving grains in a dispersal train is too low for detection by selective geochemical analysis. MMSIMs are derived from three main types of base metal deposits and their associated alteration or reaction zones: (1) volcanosedimentary massive sulphides (encompassing volcanogenic, Sedex and Mississippi Valley subtypes) in medium to high grade regional metamorphic terrains; (2) skarn and greisen deposits; and (3) magmatic Ni-Cu sulphides. The variety of MMSIMs associated with Ni-Cu deposits is astonishing, apparently reflecting mineral hybridization related to assimilation of sulphurous sedimentary rocks by ultramafic magmas. Cr-diopside is one of the best indicators of fertile Ni-Cu environments although not necessarily of the actual Ni-Cu deposits. Heavy indicator mineralogy is much more sensitive than heavy mineral geochemical analysis and offers many exploration benefits in regional exploration programs including: (1) sampling efficiencies; (2) enlargement of both the bedrock target and dispersal train; (3) coverage of a wider range of mineral commodities; (4) undiminished sensitivity in areas of overabundant non-indicator heavy minerals; (5) visual evidence of points of origin of dispersal trains; and (6) indications of the economic potential of the source mineralization. It is most effective as a reconnaissance exploration tool and is particularly well suited for testing gneissic volcanosedimentary and plutonic terranes where base and precious metal deposits are highly modified and difficult to recognize by other prospecting methods.

Glacial transport and secondary hydromorphic metal mobilization: examples from the southern interior of British Columbia, Canada

Abstract Glacial transport of trace elements was studied at seven mineral occurrences in the southern interior of British Columbia; a region where mineral exploration is hampered by the scarcity of bedrock outcrop and by a variable sediment thickness associated with the Cordilleran Ice Sheet. The till deposited in the region was, for the most part, a product of the last glacial period. A review of previous geochemical studies conducted by the mineral exploration industry provides an indication to the variable configuration of the local dispersal patterns in the area. Dispersal trains in till are short, generally < 1–2 km, rarely exceeding 10 km in length and are usually proximal to bedrock source. They are commonly ribbon-shaped and rarely exceed 1 km in width. Observed dispersal patterns suggest that drastic changes in topography might have affected basal ice velocity which increased the distance of glacial transport. In addition, the distance that separates bedrock mineralization from its surficial geochemical expression in till varies with drift thickness and topography. Several glacial dispersal trains have been modified in shape by secondary hydromorphic dispersion.

Indicator mineral and geochemical methods for diamond exploration in glaciated terrain in Canada

Abstract This paper provides a summary of selected diamond exploration techniques used in the glaciated terrain of Canada, focusing on indicator mineral methods and till geochemistry but also including geochemistry of lake sediments and vegetation. Diamond exploration in Canada focuses on kimberlite, the primary host rock for diamonds in this country. Kimberlite is a mineralogically and chemically distinct point source which may yield discrete dispersal trains in glacial sediments. Understanding the ice flow history and depositional history of glacial sediments and identifying multiple till sheets in areas covered by thick glacial sediments are essential for successful sampling, interpretation and follow-up of indicator mineral and geochemical anomalies related to these rocks. Orientation studies over known kimberlites provide important information on the mineralogical and geochemical signatures of kimberlite, and the size fractions of glacial sediments that are best suited to indicator mineral and geochemical analysis. Kimberlite indicator minerals survive glacial transport over long distances and the relative abundance of each mineral in till is a function of the primary mineralogy of individual kimberlites. Indicator mineral distributions observed at a regional scale define the net effect of glacial dispersal, often along different ice flow directions. Local scale distributions define individual dispersal trains. The finer (0.25 to 0.5 mm) fraction of heavy mineral concentrates prepared from till samples is best suited for indicator mineral surveys. Till geochemistry is gaining popularity in diamond exploration because it is significantly cheaper than indicator mineral analysis and it can be performed quickly. Important kimberlite pathfinder elements that provide good contrast in till geochemical surveys include Ni, Cr, Ba, Co, Sr, Rb, Nb, Mg, Ta, Ca, Fe, K, Ti and REE, the relative importance of which will depend on kimberlite composition as well as that of the surrounding bedrock. Biogeochemical studies over kimberlites in Canada reveal geochemical signatures in vegetation despite the glacially transported substrate.

The glacial transport and physical partitioning of mercury and gold in till: implications for mineral exploration with examples from central British Columbia, Canada

Abstract Mercury glacial dispersal was measured in the clay-sized fraction (< 0.002 mm) and heavy mineral concentrate (0.063–0.250 mm, specific gravity > 3.3 g/cm 3 ) of till in a region of bedrock cinnabar occurrences, in central British Columbia, Canada. Most of the Hg in till occurs as sand-sized cinnabar (HgS) grains. A longer dispersal train was measured with the heavy mineral concentrates because Hg concentrations in heavy minerals yielded a higher ratio between anomalous and background concentrations when compared to the clay-sized material. It is proposed that geochemical or mineralogical analyses on a specific grain size fraction or density fraction of till, where the desired metal resides, result in a higher contrast between anomalous and background concentrations. Such a great contrast translates into a longer detectable dispersal train and hence, a larger target for mineral exploration. Therefore, in drift exploration programs, it is crucial to identify the mode of occurrence of a sought commodity in till; this can be achieved in part with a simple partitioning study whereby metal concentrations are measured in specific grain size fractions of till. Physical partitioning results for Au in the study area indicate that close to the bedrock source, large metal concentrations in some cases are present in the sand- (0.063–2 mm) and granule-sized (2–4 mm) fractions. Therefore, the significance of a regional Au anomaly, commonly defined in the silt plus clay-sized fraction of till could be evaluated by further determining the Au content of coarser size fractions (sand and granule).

A Quaternary geological perpective on geochemical exploration in glaciated terrain

Abstract The application of Quaternary geology and glacial sedimentology is given as a broad guide for geochemical exploration in glaciated terrain. Predictive models of glacial dispersal provide an important basis for tailoring drift prospecting methods to suit regional variations in ice flow history and dynamics. The models relate compositional variations in glacial dispersal trains to ice flow direction, glacial history and subglacial processes. They are continually refined with reference to the geological and physical properties of the ice bed; new empirical field evidence constraining particle trajectories; and knowledge of subglacial processes affecting glacial erosion, transport and deposition. Transport at the ice bed leads to an exponential decrease in indicator concentrations with increasing distance of glacial transport, whereas linear decrease is associated with englacial transport, and may be characteristic of ice streams. The partitioning of rock and mineral fragments through subglacial comminution leads to compositional differences among size fractions that can reflect intensity of subglacial process, distance of transport, and provenance; hence, the choice of size fraction is important to drift prospecting by geochemical methods.

Regional and local-scale gold grain and till geochemical signatures of lode Au deposits in the western Abitibi Greenstone Belt, central Canada

Abstract This paper is an overview of drift exploration methods for lode Au deposits in areas of thin and thick cover of glacial sediments within the Abitibi Greenstone Belt of central Canada. It summarizes a large volume of data produced by government regional surveys and case studies as well as that from industry-led gold exploration programs. Regional till surveys can be used as targeting mechanisms for further Au exploration. Anomalies are defined by a series of samples with elevated Au concentrations that lie along significant bedrock structures, occurring in clusters or as isolated samples in areas of low sample density. Thresholds between background and anomalous Au grain abundances or Au concentrations are variable and depend on location within the Abitibi Greenstone Belt. Case studies around known deposits provide examples of geochemical and mineralogical signatures of Au deposits that can be expected in till down-ice. These serve as sources of information on appropriate sampling methods and size fractions to analyse, and on ice flow patterns, local glacial stratigraphy and suitable till units for sampling. Two methods for measuring the Au content of till are commonly used: (1) a count of visible Au grains and (2) geochemical elemental analysis. Close to source, till contains thousands to hundreds of thousands ppb Au and several hundred Au grains. The Au grains vary from coarse sand to silt sizes and have pristine shapes. The presence of high Au concentrations in till indicates that the ore zones subcrop and that glacial processes have produced Au dispersal trains down-ice.

Contrasting styles of glacial dispersal in Newfoundland and Labrador: methods and case studies

Abstract A review of practical approaches to drift exploration intended for use by exploration geologists working in drift covered areas is presented. The contrasting styles of glacial dispersal between Labrador, dominated by the effects of the Laurentide ice sheet, and the Island of Newfoundland, affected by small, coalescing ice caps at the glacial maximum and smaller topographically-controlled ice centres during deglaciation, are described. The effect has been to produce longer, ribbon-shaped dispersal trains in Labrador, except in the Labrador Trough near the centre of the Labrador sector of the Laurentide Ice Sheet, and shorter more diffuse dispersal patterns in Newfoundland.

Late Wisconsinan glaciation of southern Eureka Sound: evidence for extensive Innuitian ice in the Canadian High Arctic during the Last Glacial Maximum

Southern Eureka Sound was originally proposed as the centre of an Innuitian Ice Sheet in the Canadian High Arctic at the Last Glacial Maximum (LGM) based largely on the pattern of Holocene emergence. This paper focuses on the glacial geological evidence for such an ice sheet in the region. Granite dispersal trains and ice-moulded bedrock record regional, westward flow of warm-based ice into Eureka Sound from SE Ellesmere Island. Regional ice was coalescent with local ice domes on inter-fiord peninsulas. Marine limit in the form of raised deltas, beaches and washing limits formed during deglaciation of the regional ice. Throughout southern Eureka Sound, marine limit dates ⩽9.2 ka BP, indicating that ice commenced retreat during the early Holocene. Ice-divides were located along the highlands of central Ellesmere and Axel Heiberg islands, from which ice inundated Eureka Sound, flowing north and south along the channel. Regional radiocarbon dates on marine limit show that deglaciation occurred in two steps. Initial break-up and radial retreat of ice from Eureka Sound to the inner fiords was rapid and preceded stabilisation along adjacent coastlines and at fiord heads. Two-step deglaciation is also reflected in differences in glacial geomorphology between the inner and outer parts of many fiords. A prominent belt of fiord-head glaciogenic landforms, long proposed to mark the last glacial limit, is re-interpreted to record initial, stabilisation of ice margins due predominantly to bathymetric control.

Carbonate dispersal trains, secondary till plumes, and ice streams in the west Foxe Sector, Laurentide Ice Sheet

Regional carbonate dispersal trains cross Melville Peninsula in the northeastern Canadian Arctic. These trains are 50–125 km across and 100–300 km long. The northern dispersal train crosses upland and lowland areas, while the Rae Isthmus train follows the lowlands. The distribution of carbonate in boulder, pebble and matrix till fractions indicates long‐distance glacial transport of limestone, a low rate of dilution, and comminution from pebble to silt sizes. Dispersal patterns and geotechnical properties of the till suggest deforming bed and basal gliding mechanisms of ice flow associated with ice‐streaming. Regional ice streams may have been ‘normative’ in carbonate terrain during the main Foxe (Wisconsin) Glaciation. Till plumes within regional dispersal trains, showing little dilution of carbonate down‐ice from bedrock sources, are oriented towards lowlands. These features, sharp‐edged and 300 mwide, mark zones of augmented flow. They are late‐glacial features that formed after calving bays developed in Committee Bay. The dimensions of the dispersal trains, and ice‐flow styles and mechanisms, are similar to those of modern Antarctic ice streams. The sharp edges of till plumes may delineate flow boundaries marked by lateral crevasse zones in the late‐glacial ice sheet. Secondary streams‐within‐ice streams have not been reported previously in the literature.

Carbonate dispersal trains, secondary till plumes, and ice streams in the west Foxe Sector, Laurentide Ice Sheet

Carbonate dispersal trains, secondary till plumes, and ice streams in the west Foxe Sector, Laurentide Ice Sheet

The application of glacial dispersal models to the interpretation of till geochemistry in Labrador, Canada

In western Labrador, Canada, bedrock composition and glacial history are both reflected in till geochemistry. There, compositional contrasts between igneous and metamorphic bedrock of the Canadian Shield and comparatively unmetamorphosed bedrock of the Labrador Trough allow regional-scale (100s of km) definition of glacial dispersal trends through lithological and geochemical analyses. Compared to till derived from shield terrane, Labrador Trough till is depleted in chromium, but enriched in zinc, iron, and manganese. Further, the study area is situated near one or more dispersal centers of the Laurentide Ice Sheet and has been subjected to a complex sequence of ice flow, defined by striations and indicator erratic distributions. Three glacial events serve as the principal controls on drift composition, although their relative significance varies. Compositional profiles along the axis of two glacial dispersal trends show systematic changes in indicator erratic concentrations with distance along the path of ice flow. In the profile having one prominent glacial transport direction, indicator concentrations decrease exponentially with increasing distance of glacial transport. From dispersal models, Renewal Distances representing the compositional `half-life' tend to increase with decrease in grain size, i.e. 15 to 25 km for pebbles, and 30 to 40 km, or more, in silt- and finer-sized fractions. The differences between geochemical and lithological profiles reflect differences in: (1) the intensity or duration of subglacial process affecting the milling of indicator detritus to finer fractions; and (2) the physical and chemical properties of the rocks and minerals hosting the trace metals. Where regional glacial dispersal trains extend across contrasting geological terranes, indicator properties can be used to resolve till composition into `local' and `transported' components, providing a better approximation of bedrock composition for mineral exploration. The compositional expression of bedrock can be masked by transported overburden, with the extent of masking dependent on the trace metal examined. Linear relations between metals (Cr, Zn) and lithophile elements (e.g. K, Mg) indicate that Mg-bearing minerals, likely phyllosilicates, have a strong mineralogical control on till geochemistry, and metal : lithophile element ratios can be used to distinguish bedrock provenance in glacially transported materials of western Labrador.

Some advances in understanding and application of stream sediment geochemistry to mineral exploration in SE Asia

Stream sediment geochemical surveys are a widely used method of mineral exploration in tropical rainforests worldwide. The basic premise of such surveys is that a stream sediment is representative of the products of weathering and erosion upstream ofthe sample site and can thus detect anomalous concentrations of metals released from a mineral deposit within the drainage basin. Ideally, representative samples must be collected from streams throughout the survey area at a sample density sufficient to detect anomalies related to mineral deposits. Representativity of the sample must then be maintained through all stages of sample processing and analysis. There are many situations in which these ideals cannot be achieved. For example, a flood plain may isolate the stream from the valley sides or supply of new sediment to a stream may be from a few point sources (e.g., landslides), the location and activity of which varies with time. Also, fluvial processes modify the composition of the sediment as it moves along the channel with important consequences for elements, such as gold and cassiterite, associated with heavy minerals. Detailed sedimentological studies in the S. Petai, Malaysia have shown that sediments (SG-2.7) finer than about 100 pm tend to be rapidly swept downstream in suspension whenever sediment transport occurs. This causes both the greatest and most consistent enrichment of cassiterite to occur in the finer fractions of the sediments. Heavy mineral content of coarser size fractions becomes increasingly erratic because their accumulation is strongly dependent on local hydraulic conditions on the streambed. Based on the S. Petai studies, the relatively uniform enrichment of heavy minerals in the finer fractions « 100 pm) of the sediments best represents the geochemistry of the drainage basin as a whole. Use of fine fractions also reduces the nugget effect during sampling, so that these fractions typically give the most consistent anomalous dispersion trains. Anthropogenic disturbance, by logging or agricultural activities , can also influence sources of sediment and rates of supply. For example, in Thailand increased soil erosion, caused by deforestation and planting of maize, has resulted in a gold anomaly being diluted to the extent that it is not detected reliably by conventional stream sediments. However, effects of disturbance are not always so severe and results from a mature rubber estate in Peninsular Malaysia suggest that conditions return to near­ normal once a groundcover of vegetation has been reestablished.

Contamination from forestry activities: implications for stream sediment exploration programmes

Although it is well established that background geochemical patterns in stream sediments generally depend on drainage basin geology, little is known of the effects of watershed disturbance on such patterns. Here we report on the effects of forestry activities on sediment geochemistry on the Interior Plateau of British Columbia. Stream sediment samples were collected from six small streams in 1996, before logging, and in 1997 after clear-cut logging. Samples were sieved and the −212 μm fraction analysed by multi-element ICP after both strong acid and total decomposition procedures. Field and analytical duplicates, and samples from unlogged control sites were used to estimate variations not associated with logging activities. Results show that prior to logging each stream had a distinct multi-element geochemical signature related to drainage basin geology. There was no overall effect of logging on these signatures within or downstream of the cut-blocks — probably because unlogged buffer zones along the stream channels prevented the development of new sediment sources. Logging roads did, however, cause local changes of sediment geochemistry in two ways. (1) Abrasion of zinc from galvanized culverts results in zinc anomalies, with concentrations > 200 ppm. These anomalies can extend several hundred metres downstream and persist for at least five years after installation of culverts. (2) Construction of stream crossings and roadside ditches create new sources of sediment supply to the streams. Although of local origin, the newly exposed and eroding surficial deposits contain more fine-grained material and have multi-element geochemical signatures that differ significantly from those of the mature fluvial sediments found in the stream channel before logging. Input of the new material results in geochemical anomalies that can extend several hundred metres downstream from the source. In this study Co, Mg and Ni were found to have the largest concentration differences between sediment sources and sediments, and hence to give the longest anomalous dispersion trains. The effects of logging roads on sediment geochemistry thus relate to creation of new sources of local sediment as well as to introduction of exotic materials into the stream. No geochemical response could be attributed to the clear-cut forest block. Such changes might develop eventually if soil erosion increased or if changes to the groundwater regime modified input of dissolved elements to the stream channel.

Dispersion of gold and associated elements in stream sediments under semi-arid conditions, northeast Brazil

A stream sediment orientation survey was carried out for gold mineralization in the semi-arid northeast of Brazil. Eleven stations were sampled in duplicate along a 3000-m stream. The stream is ∼1.5 m wide at the upper station and ∼8 m wide at the lower station. Estimated overall stream gradient is 2%. Mineralization at the headwaters occurs as centimeter- to meter-thick gold-sulfide-bearing quartz veins located in a NNE–SSW shear zone and hosted by sillimanite–garnet–biotite schists. Sampling was carried out in summer when the stream was dry. Sediments were dry-sieved on site to collect two 12-l samples of −2 mm material, composited from three to five high-energy sites at each station. In the laboratory samples were dry-sieved to obtain the −2+0.500 mm, −0.500+0.180 mm, −0.180+0.106 mm, −0.106+0.063 mm and −0.063 mm fractions. All fractions less than 500 mm were analyzed for Au, As, Hg, Sb, Cd, Co, Cu, Fe, Pb and Zn. Cursory evaluation indicates that sampling and analytical variabilities are negligible compared to between-station variability. The Au anomaly, which was only consistently detected in the two finest fractions, shows a downstream decay pattern with a decrease of Au values from 40–55 ppb to <5 ppb over a distance of ∼1.3 km. Erratic high and low values are superimposed on this trend. Anomalous downstream dispersion patterns are observed for As, Cu, Pb, Zn, Co and Cd in all size fractions analyzed. These are less erratic than for gold and extend ∼3 km downstream to the limit of the study. Sediment size fraction data show that distribution of gold is closely related to sediment texture, with gold concentrations decreasing as the abundance of −0.063 mm material increases. This relation is observed between adjoining stations but, insofar as abundance of the −0.063 mm increases down the stream, has probably also magnified the anomaly in the headwaters while diluting it further downstream. We attribute the relation between gold and sediment texture, which is not shown by any of the base metals, to gold behaving as a heavy mineral that is concentrated on the streambed by washing away of fine sediments, or diluted by their deposition. Based on the results reported here, it seems possible that in arid regions the limited development of deeply weathered regolith and retention of fines within the drainage basins can hinder development of long dispersion trains for gold. In such situations the importance of pan concentrates and pathfinder elements may be increased.

Comparison of vegetation and stream sediment geochemical patterns in northeastern New South Wales

This study compares the geochemical response of stream sediments and adjacent vegetation samples, with variations in drainage catchment lithology and the occurrence of mineralisation, within a 14,000-km 2 block of the northeastern region of New South Wales, Australia. The area contains a range of lithologies within a Devonian–Permian accretionary complex and Mesozoic sedimentary basin, as well as a wide range of mineral deposits. Sampling was designed to confine each sub-catchment to a single lithological group (mafic and ultramafic rocks, acid intrusives, volcanics, metasedimentary and sedimentary rocks or alluvium). Leaves of over 20 genera, dominated by (Allo-)Casuarina, Eucalyptus, Acacia, Callistemon and Melaleuca, and the <250-μm fraction of the stream sediments were analysed by INAA. The uptake of most trace elements varied between genera, with Callistemon displaying the highest median As and La contents and Eucalyptus the highest Co contents. The stream sediment and vegetation geochemistry reflect both hydromorphic and mechanical dispersion within sub-catchments, with regional patterns dominant over local influences. The vegetation appears to be influenced to a greater extent by hydromorphic dispersion, as indicated by differences in the ratio of leaf to sediment Cr concentrations in sub-catchments draining serpentinites and basalts. Although most known mineral deposits in the region produced anomalies in at least one medium, there was little correlation between the trace element concentrations of the vegetation and stream sediments on a site-by-site basis. A number of Au targets were only detected on the basis of the biogeochemistry and others were only reflected in the stream sediment geochemistry. In general, vegetation displayed more extensive dispersion trains away from mineralisation than did the stream sediments. Differences in the response of the two sampling media suggest their joint use in exploration or environmental surveys to maximise the probability of detecting mineralisation.