Abstract
Catchment outlet sediments (0–10 cm depth, sieved to <2 mm) collected at a very low density over most of the Australian continent have been analysed using the Mobile Metal Ion (MMI®) partial extraction technique. Of the 54 elements determined, eight are generally regarded as essential nutrients for plant growth: Ca, Cu, Fe, K, Mg, Mn, P and Zn. A further three, Mo, Ni and Se are considered significant micronutrients. Estimation of ‘bioavailability’ from MMI® analysis gives results comparable with standard agricultural measurements for many nutrients. Percentage ‘bioavailability’, operationally defined here as the ratio of MMI® concentration to total element concentration, has been investigated and ranges from 31% for Se to 0.1% for Fe. Smoothed (kriged) colour raster maps for continental Australia have been produced for these 11 nutrients and interpreted in terms of lithology (e.g. presence of carbonates in the MMI® Ca map), mineralization (e.g. known mineral districts in the Cu and Zn maps), environmental processes (e.g. salinity in K map, weathering and acid generation in Fe map) and agricultural practices (e.g. application of fertilizers in the MMI® P map). This first application of a partial extraction technique at the scale of a continent has yielded meaningful, coherent and interpretable results.
- catchment outlet sediment sampling
- MMI® extraction
- ICP-MS analysis
- nutrients
- bioavailability
- geochemical mapping
The Australian landscape is often considered to consist of desert and semi-desert. Whilst approximately half of Australia’s nearly 8 million km2 has a rainfall of less than 250 mm/year and fresh rock outcrop is relatively rare, the truth is that because of its size, climate extremes and widely varying rock ages and types, there is a wide diversity of regolith and landscape types in Australia. Large areas exist with limited outcrop and/or weathered sub-crop as the dominant regolith type. Areas with discernable drainage patterns are prevalent towards the coasts where rainfall is higher. A considerable portion of this regolith supports agriculture and grazing activities, and an understanding of it, including its geochemistry, is important to both agriculture and mineral exploration. Because of the general low relief, due to protracted landscape evolution, stream sediment sampling and, in particular, overbank or floodplain sampling had not – until recently -- been widely implemented in Australia, particularly on a regional or continental scale. This is despite this sampling medium having long been recognized and widely used in Europe for geochemical surveys (e.g. Ottesen et al. 1989). Low density geochemical sampling has been criticized in the past for its perceived lack of repeatability, caused by variation in composition around each sample site. Smith & Reimann (2008) cite a number of cases in which this is shown not to be a limitation on obtaining useful and reliable information on a regional scale. Reedman & Gould (1970) showed marked similarities between low density (1 sample per 200 km2) stream sediment data and higher density (1 sample per 2 km2) data in Uganda. In China, patterns from two overlapping geochemical mapping projects, the EGMON Project, (1 site per 18 000 km2) versus the RGNR project (1 site per 1 km2) were shown to generate compatible information (Xie & Cheng 2001). In the US, radiometric data from the NURE project, essentially at very high density, was compared by Smith & Reimann (2008) with the low density (1 sample per 6000 km2) soil data of Shacklette & Boerngen (1962). In the case of K, the major patterns delineated are very similar. In Europe, comparison of till samples in Finland suggests (Smith & Reimann 2008) that the most cost effective method is to map the whole country at low density (e.g. 1 sample per 500–10 000 km2), then to increase sample density in the most interesting areas at a later stage.
The National Geochemical Survey of Australia (NGSA; see http://www.ga.gov.au/ngsa) was initiated in late 2006 by Geoscience Australia (Caritat et al. 2008; Caritat & Cooper 2011a). The ultra-low density continental geochemical survey targeted transported regolith samples collected on the floodplains of large catchments, as much as possible. Because in many places the influence of aeolian material can not be avoided in Australia, this medium was more generally termed ‘catchment outlet sediment’ and was collected at the lower point of large catchments, be it on the catchment boundary or somewhere more central for internally draining catchments. Coverage of 81% of the Australian continent (>6 million km2) was achieved with some 1315 samples (including field duplicates) collected at an average sample density of 1 sample/5200 km2. In late 2009, 1347 top outlet sediment (TOS) samples (0–10 cm depth, sieved at <2 mm) were provided to SGS, owner of the MMI® Technology, for MMI® analysis (this number included field and laboratory duplicates for QC).
Prior to this, and in order to confirm the applicability of MMI® to transported (overbank) sediments at such low densities, two pilot studies had been undertaken. Samples from the Thomson Pilot study in New South Wales conducted by Geoscience Australia and the Geological Survey of New South Wales (Caritat & Lech 2007) were analysed after MMI® extraction. Significant variations and meaningful, coherent anomalies were noted. Separately, SGS conducted an overbank MMI® survey of the Swan-Avon drainage system in Western Australia at a density of 1 sample/2200 km2. The results of this survey also suggested that the combination of low-density sampling and partial extraction analysis could be used to infer lithology and locate mineralized sources. These two pilot studies, and the continental distribution of non-nutrient elements will be the subject of a separate, but similar paper to this one; preliminary results for these other elements have already been presented (Caritat et al. 2011b).
The analytical suite for standard MMI® analysis includes eight of the ten nutrients considered essential to plant growth plus three important micronutrients. As such, this method of analysis can provide information relevant to agriculture. The nutrients routinely analysed after MMI® extraction are: Ca, Cu, Fe, K, Mg, Mn, Mo, Ni, P, Se and Zn. MMI® is a partial extraction based on a multi-ligand extraction solution (Mann 2010). The soil matrix is not dissolved, the analytes in solution being obtained from the portion of elements adsorbed (as ions) on the surface of soil particles. Samples for MMI® analysis are normally taken from a depth of 10–25 cm in true soil (Mann et al. 2005), believed to be the zone of maximum evapo-transpiration in most soils. The study by Bajc (1998) obtained erratic results after sampling from deeper in the regolith. Given that soil pore waters also contain a number of inorganic and organic species capable of complexing metal (and non-metal) ions, while leaving the soil substrate relatively unaffected, MMI® analysis has the potential to provide a measure of bioavailability for a number of analytes in a similar manner to, for example, ammonium acetate-EDTA (AA-EDTA) discussed by Albanese (2008). Although bioavailability is relatively simple to parameterize under controlled simulated conditions, this is not true for field conditions (Iskander & Kirkham 2001). Iskander & Kirkham (2001) state ‘more recently the scientific community has come to the consensus that, although the ‘total’ content of nutrient and contaminant elements has been infrequently well correlated with element uptake by organisms, the ‘more bioavailable’ or ‘labile’ form has more merit since that is the form that can be physically, biologically and chemically described. The urge to measure the ‘bioavailable’ forms of chemical compounds in an environmental setting has resulted in numerous and diverse techniques’.
A number of partial extraction chemicals have been used or have the potential to be used for assessment of bioavailability. Hall (1998) examines a number of these, in relation to their attack on specific soil phases and relevance to mineral exploration rather than agriculture or environmental monitoring. For example NaOAc/HOAc is a preferred reagent for dissolution of the carbonate phases, H2O2 or NaOCl is used for selective dissolution of ‘soluble’ organics and hydroxylamine hydrochloride or Enzyme Leach are used for manganese oxide phases. Tamm’s reagent (ammonium oxalate in oxalic acid) and hydroxylamine hydrochloride are commonly used for release of elements attached to amorphous Fe oxide. The use of sodium pyrophosphate for treatment of humic substances and the issue of re-adsorption prior to analysis are also examined by Hall (1998). Magnesium and calcium chloride, sodium nitrate and ammonium nitrate provide unbuffered neutral salts for determination of many ‘exchangeable’ mobile species. An example of the application of calcium chloride (and sequential extraction) to Polish soils is given by Siebielic et al. (2006). Partial digestions have long been utilized in agriculture; individual techniques such as the Colwell method for P (Colwell 1967a, b) and the DTPA method for Cu, Fe, Mn, and Zn (Lindsay & Norvell 1978) have been designed to closely approximate the amount estimated to be obtained from a soil by plants by comparison with plant uptake studies (e.g. Podlesakova et al. 2001). Hall et al. (1998) found that 1M NH4Cl and 0.1M Na4P2O7 extractions applied to chernozem and podzol soils on the Canadian prairies could be used to predict uptake of Cd in durum wheat, in contrast to a 1M NH4NO3 extraction which suffered from severe problems of readsorption of elements during leaching. Further work by Garrett et al. (1998) modeled the uptake of Cd in wheat from the ploughed soil horizon (<2 mm) using results obtained by the pyrophosphate leach on the soil and the organic carbon content. The current paper examines the results for 11 nutrient elements extracted by MMI® from catchment outlet sediment samples collected as part of the NGSA project. ’Bioavailability’ for each of these nutrients is examined by comparison with either XRF or total digestion ICP-MS for the same samples, and where possible compared with standard agricultural analysis procedures.
Methods
Sampling
The National Geochemical Survey of Australia (NGSA, see http://www.ga.gov.au/ngsa) collected transported regolith samples in 1186 large catchments covering c. 81% of Australia (Caritat & Cooper 2011a). Parts of northeastern Western Australia and northwestern South Australia could not be sampled within the time frame of the project due to access difficulties. Transported sediment was considered to be the best natural proxy for the average composition of all major rock and soil types within a catchment, and is deposited by receding floodwaters when material erosion and transport is most intense, and immediately thereafter. These ‘catchment outlet sediments’ are thus similar to floodplain sediments where alluvial processes dominate, but can also be strongly influenced by aeolian processes in many parts of arid and semi-arid Australia (e.g. Gawler Region; see Caritat et al. 2008). In all instances, the lowest point of each catchment, as determined by hydrological modelling, was the target site for sample collection in the NGSA project, whether it be near the catchment boundary (here floodplains typically were the targeted landform to sample) or, in the case of internally draining catchments, toward the centre of the catchment (here dune swales were the targeted landform).
At the target site a surface (0-10 cm deep) ‘Top Outlet Sediment’ or TOS, and a deeper (on average 60–80 cm deep) ‘Bottom Outlet Sediment’ or BOS samples were collected. Both samples were taken as composite samples either from a shallow c. 1 m2 soil pit (TOS) or from generally at least 3 auger holes within an area of c. 100 m2 (BOS). Where augering was not possible, soil pits were dug. Because pedogenesis typically has occurred on these landforms, the sampled materials can generally be considered equivalent to topsoil and subsoil samples formed on alluvial (± aeolian) parent material. Site descriptions, GPS coordinates and digital photographs were recorded in the field as were texture, dry (where possible) and moist soil Munsell colours and field pH (Cooper et al. 2010; Caritat et al. 2011a). The interested reader is referred to Lech et al. (2007) for more details on sampling. In total 1315 TOS and 1315 BOS samples (including c. 10% field duplicates) were collected with randomised sample numbers pre-allocated to each site. The average sampling density of the survey is 1 site per 5200 km2. Sampling locations for the NGSA project are shown in Figure 1.
Sample locations, National Geochemical Survey of Australia. All maps in Lambert Conformal Conic projection with Central Meridian 134 °E and Standard Parallels 18°S and 36 °S (GDA 1994); graticule in ° longitude E and ° latitude S.
Sample preparation
All samples were prepared in a central laboratory (Geoscience Australia, Canberra). The samples were oven-dried at 40 °C, homogenised and riffle split into an archive sample for future investigations and an analytical sample for immediate analysis. The latter was further riffle split into a bulk subsample, a dry-sieved <2 mm grain size fraction subsample, and a dry-sieved <75 μm grain size fraction subsample using nylon screens (Caritat et al. 2009). Each subsample was further split into aliquots of specific weight as per analytical requirements. Only the TOS <2 mm subsample was used for MMI® analysis, an aliquot of which was sent in a single batch to the SGS laboratory in Perth.
Sample analysis
At the SGS laboratory in Perth, a 50 g riffle-split aliquot of each sample was treated with 50 ml of MMI®-M (multi-element) extract solution, shaken for 1 hour and allowed to stand for 24 hours. The characteristics and advantages of this procedure have been described previously (Mann 2010). An aliquot of this solution was then analysed by Inductively Coupled Plasma-Mass Spectrometry (ICP-MS) using a Dynamic Reaction Cell™ (DRC II™), thus allowing very low Lower Limits of Detection (LLDs). The solution was analysed for the following 54 elements: Ag, Al, As, Au, Ba, Bi, Ca, Cd, Ce, Co, Cr, Cs, Cu, Dy, Er, Eu, Fe, Ga, Gd, Hg, K, La, Li, Mg, Mn, Mo, Nb, Nd, Ni, P, Pb, Pd, Pr, Pt, Rb, Sb, Sc, Se, Sm, Sn, Sr, Ta, Tb, Te, Th, Ti, Tl, U, V, W, Y, Yb, Zn and Zr; results pertaining to non-nutrient elements and other applications will be reported separately.
Total content determinations were obtained by X-Ray Fluorescence (XRF) analysis of a lithium-borate fused bead at Geoscience Australia for six of the 11 nutrients (Ca, Fe, K, Mg, Mn and P; original oxide data converted to elemental concentrations). For Cu, Mo, Ni and Zn, fragments of the glass beads were fully digested in HF and HNO3, then analysed by ICP-MS at Geoscience Australia. For Se, the ‘total’ concentrations were determined by an aqua regia digestion followed by ICP-MS analysis. More details on the analytical methods of the NGSA are given in Caritat et al. (2010).
Quality control
Quality control consisted of (1) randomisation of sample numbers; (2) analysis in a single batch; (3) insertion of c. 10% blind field duplicates; (4) insertion of blind laboratory replicates; and (5) insertion of laboratory standards SRM-18 and SRM-19 in the analytical stream. The results of detailed quality control for the NGSA project can be found Caritat & Cooper (2011b), and those pertaining to the MMI® analysis for the elements of interest here are briefly summarised in Table 1. Overall, precision quantified as the Relative Standard Deviations (RSDs) for the laboratory standards are acceptable, except for Se, Fe, P and Mo, for which SRM-18 has too low concentrations (<8 × LLD). Bias ranges from 73–104% and averages of replicate analyses are well within tolerance of the recommended values. Total precision, including an estimate of sampling, sample preparation and analysis errors, is shown by the RSDs of field duplicates to range from 30 to 115%; of these, the worse values are obtained for Se (115%) and Mo (57%), the latter element being affected by many censored (<LLD) results (<LLD values were replaced by 0.5 × LLD for statistical analysis). Thus, quality assessment of the MMI® data for the 11 nutrients under investigation here suggests reliable data for all, except for Se and Mo results which should be regarded as unreliable and qualitative only. Further, results for Fe and Zn should be treated with a cautionary notice due to their borderline bias (73 and 87%) and total precision (both 50%) results.
Quality control data for 11 nutrients by MMI® extraction of NGSA catchment outlet sediment samples (0–10 cm, <2 mm). LLD, Lower Limit of Detection; RSD (Std), Relative Standard Deviation for 22 analyses of laboratory standard SRM-18; RV, Recommended Value; Tolerance, Acceptable range about the Recommended Value; Bias, 100% × Average/Recommended Value; RSD (FD), Relative Standard Deviation for analyses of field duplicates (all FD data had <25% of values <LLD, except for Mo, which had <50% of values <LLD)
The reproducibility of MMI® values through time is documented, for instance, by the results from the pilot study carried out in the Swan-Avon catchment. In this study, a site adjacent to a salt lake in the wheat-belt of Western Australia was sampled on three separate occasions (twice in duplicate), by two separate sampling crews, over a two-year period. The five Au results obtained were: 1.3 parts per billion (ppb), 1.1 ppb, 1.1 ppb, 1.2 ppb and 1.5 ppb, using a method with a LLD for Au of 0.1 ppb.
Data analysis
Location coordinates were appended to the geochemical dataset and Australia-wide raster maps were plotted using ordinary kriging using a spherical semivariogram model, an output cell size of 0.01 ° and a variable search radius (12 nearest points) in ESRI’s ArcGIS 10 package. Kriging is the recommended approach for generating raster maps from point source data (Reimann 2005).
Results
Comparison of MMI® and total analysis results for 11 nutrients
In a manner similar to that described by Albanese (2008), the percentage ’bioavailability’ of nutrients was calculated by comparing the MMI® with the total concentrations from XRF or ICP-MS analyses for the dataset. Percentage bioavailability, B, for each sample was thus operationally defined as:
The average percentage bioavailability for each nutrient obtained on the NGSA TOS <2 mm samples after removal of field and laboratory duplicates is shown in Table 2.
Average ‘bioavailability’ (with 1 standard deviation) determined for 11 nutrients by MMI® extraction of NGSA catchment outlet sediment samples (0–10 cm, <2 mm)
For a small number of samples, the MMI® and/or the total analyses produced results below the LLD and these samples were ignored for the bioavailability computation (because the replacement values, 0.5 × LLD, hardly make sense when used in the numerator and/or denominator of a division).
Selenium, calcium, copper and magnesium show the highest percent bioavailability using MMI® as an extractant, Fe the lowest. The results for Cu and Zn here are lower than those obtained by Albanese (2008) using AA-EDTA as an extractant, where percentage bioavailability was estimated in most cases to be over 50% for Cu and over 40% for Zn. The present results are probably closer to extraction capabilities of plants, given the high concentrations of complexing agents used in the AA-EDTA extraction.
Continental-scale mapping of bioavailable nutrients
Coherent regional patterns were obtained for most of the 54 elements determined (Pt was an exception), reflecting favourably on the choice of sample density, sampling medium and analytical technique. A number of contributing factors need to be considered when interpreting regional patterns. Underlying lithology is a very important factor governing the distribution of many elements in the regolith. In many cases (e.g. Au, Cu), local mineralization can also potentially be important as a source. However, since the combination of techniques is measuring mobility, in some cases (e.g. Fe, Al, Mn) weathering, and in particular rainfall, plays an important part. Finally, contributions from environmental factors (e.g. mine waste dumps) and/or agriculture (e.g. fertilizer application) may be a determinant in element distribution; in the NGSA sampling strategy, however, sites showing signs of man-made influence, such as proximity to mining operations, were deliberately avoided. It must be remembered that each sample yields an integrated result of several processes operating within a catchment, and interpretations must be tailored accordingly. The Australia-wide maps for the 11 nutrients after MMI® extraction and ICP-MS analysis are discussed in the following paragraphs. Nutrients are grouped according to common processes and/or spatial distribution patterns.
Phophorous and potassium
The three major nutrients for plants are N, P and K; nitrogen is not available from the MMI® technique. The Australia-wide maps for P and K, generated after MMI® extraction are shown in Figures 2 and 3, respectively. Many Australian soils with their high residual Fe content in the form of Fe oxides and oxyhydroxides, are notably P deficient (e.g. Singh & Gilkes 1991). This is evident in Figure 2, with vast areas of inland Australia below 2 parts per million (ppm) MMI® P in catchment outlet sediments. Exceptions are the agricultural areas along the eastern seaboard, southern Tasmania and the extreme southwest coast of Western Australia, where agriculture is carried out, Fe is mobile and the use of high P fertilizers is prevalent. The P recorded in catchment outlet samples in these areas is in all probability a result of P released from fertilizer which is not taken up by plants, adsorbed to fine fraction soil particles and transported down catchments. There are a few ‘anomalies’ above 4 ppm in central Australia (Northern Territory and northern Queensland) in areas prospective for phosphate deposits, likely to be related to mineralization.
Kriged distribution of P by MMI® in NGSA catchment outlet sediments.
Kriged distribution of K by MMI® in NGSA catchment outlet sediments.
Potassium is a very soluble and mobile ion, with high concentrations in some granites. The map for K after MMI® analysis (Fig. 3) shows that high K is mostly associated with saline lake and river systems. The Murray-Darling system in eastern Australia, with its salinity and drought issues and which exits to the ocean on the south coast of South Australia, is well defined in the map. High K is also evident along the southern coast along the edge of the arid Nullarbor Plain. Potassium is notably poor in the soils of northern Australia where tropical rainfall promotes extensive leaching.
Calcium and magnesium
The Australia-wide maps for bioavailable Ca and Mg, generated after MMI® extraction and analysis, are shown in Figures 4 and 5. Figure 4 shows high Ca values (>800 ppm) on the southern coast in the vicinity of the Nullarbor Plain where Mesozoic limestone is the underlying lithology, in central Western Australia and western Northern Territory where the climate is very arid and secondary surface carbonate is ubiquitous, and in the Lachlan Fold Belt (southeastern Australia) along the eastern seaboard wherever limestone or dolomite is the prevalent lithology within a catchment.
Kriged distribution of Ca by MMI® in NGSA catchment outlet sediments.
Kriged distribution of Mg by MMI® in NGSA catchment outlet sediments.
Most high MMI® Mg values (Fig. 5) are associated with high MMI® Ca. Thus in the Lachlan Fold Belt of the eastern states, high Mg values are likely to be associated with carbonate lithology; and in the interior of Western Australia and the Northern Territory with secondary carbonates of the arid and semi-arid zones. High MMI® Mg values in the eastern goldfields of Western Australia are more than likely a result of the high Mg content of mafic and ultramafic sequences in the Yilgarn Craton (also shown by the total Cr distribution in the NGSA samples; see Reimann et al. 2012).
Iron and manganese
The distribution map for Fe (Fig. 6) after MMI® extraction and analysis is dominated by weathering, more specifically high rainfall. Areas rich in residual Fe (e.g. the Pilbara) are not highlighted in this map as it is strongly bound in oxides and oxyhydroxides. Areas with the highest concentrations of mobile Fe are those in which processes akin to lateritisation and bauxitisation are believed to be currently operating (Bettenay et al. 1979). Bauxite deposits, which form by removal of Fe (as Fe2+) from the laterite profile, as well as conversion of kaolinite to gibbsite, are present in the Darling Ranges of southwestern Western Australia, at Gove in the Northern Territory and at Weipa on the Cape York Peninsula of northern Queensland – all areas showing high MMI® Fe concentrations. Many of the river systems along the east coast, also with high MMI® Fe, have environmental problems due to acidification (see Discussion).
Kriged distribution of Fe by MMI® in NGSA catchment outlet sediments.
Manganese exhibits a distribution pattern (Fig. 7) similar to that of Fe, i.e. indicating the dominance of weathering processes. Rocks and soil surfaces of the arid and semi-arid interior of Australia frequently exhibit coatings rich in Mn illustrating its immobility in this environment under present-day conditions. By contrast, Figure 7 shows that Mn is mobile where rainfall is higher and is represented to a far greater degree as an adsorbed species on the surface of soil particles in wetter, coastal catchments.
Kriged distribution of Mn by MMI® in NGSA catchment outlet sediments.
Copper and zinc
The distribution pattern for Cu after MMI® extraction is shown in Figure 8. The highest MMI® Cu value is on the northern coast, south of Darwin, in the Finniss River catchment, where the Cu-Pb-Zn-Ni-Co-U mines of the Rum Jungle area are located. In northern Queensland, in the well-known Cu mining district of Mt Isa and Cloncurry, high MMI® Cu is also evident in streams and rivers draining north into the Gulf of Carpentaria. Here secondary Cu mineralization (malachite and azurite) is common at or near surface. In New South Wales, Victoria, South Australia and Western Australia many of the MMI® Cu anomalies are again associated with known Cu mineralization. It is likely this level (>2 ppm) of bioavailable Cu is in excess of normal nutritional plant requirements.
Kriged distribution of Cu by MMI® in NGSA catchment outlet sediments, overlain by distribution of Cu deposits/mines.
Zinc after MMI® extraction (Fig. 9) shows a similar distribution pattern to MMI® P. It is unlikely that this is due to its use as a trace element supplement in fertilizers, but, particularly in eastern Australia (eastern New South Wales, eastern Victoria and Tasmania), due to the presence of Zn-rich bedrock and Zn mineralization in the Lachlan Fold Belt. Sulphides high in Zn are known, for example, at Cobar (New South Wales), Captains Flat (Australian Capital Territory), Benambra (Victoria) and Rosebery, Queenstown and Hobart (Tasmania) where sulphides are treated. Sphalerite is a known constituent of ores in the Northampton mining district north of Perth on the west coast in Western Australia. The high MMI® Zn values in far northwest Queensland and in the northern part of the Northern Territory appear to be related to mineralization at the Century Mine (Australia’s largest open pit Zn mine) and to deposits in the Pine Creek Geosyncline respectively. High MMI® Zn values in many of these cases (i.e. >2 ppm) may well be in excess of the (trace) values required for plant nutrition. A vast area of inland Australia has low bioavailable Zn on this scale (<0.5 ppm MMI® Zn).
Kriged distribution of Zn by MMI® in NGSA catchment outlet sediments, overlain by distribution of Pb-Zn deposits/mines.
Molybdenum, nickel and selenium
Molybdenum, nickel, selenium, sulphur and boron are also often considered nutrients (Handreck & Black 1994); the latter two are not in the present analysis suite. The Australia-wide maps for Mo, Ni and Se, generated after MMI® extraction are shown in Figures 10, 11 and 12, respectively. Some of the anomalous values of Mo shown in Figure 10 are coincident with high MMI® W and are therefore more likely to be associated with mineralization than fertilizer application. Similarly there is a correlation of some of the high Se values in Figure 12 with those for MMI® Te. The distribution pattern for Ni (Fig. 11) is not related directly to mineralization but to lithology; high values for Ni occur uniformly in the Lachlan Fold Belt along the eastern coast of Australia on mafic and less commonly ultramafic rock types.
Kriged distribution of Mo by MMI® in NGSA catchment outlet sediments.
Kriged distribution of Ni by MMI® in NGSA catchment outlet sediments.
Kriged distribution of Se by MMI® in NGSA catchment outlet sediments.
Discussion and Conclusions
‘Bioavailability’ is a concept based on the realization by the scientific community that the total content of an element in a soil or plant tissue is not a reliable indicator of nutritional availability, plant requirements, or element uptake. As a result, numerous and diverse techniques (including partial extraction geochemistry) have been developed to estimate the degree to which chemical elements in soils, including trace elements, are available to plants and animals (Iskander & Kirkham 2001). When partial extractions are used to transfer this concept to reality a number of problems exist: it is very difficult, if not impossible for a single solution to adequately mimic the transfer of nutrients, micronutrients and non-nutrients across the root zone interface of a plant from a soil pore water, for a range of species and for a range of soil types. At best ‘bioavailability’ when estimated by a partial extraction is ‘operationally defined’, which also means that no two extractants will provide identical measurements.
In the present case, for an element such as Cu, similarities in the continental distribution patterns can be demonstrated for the partial digestion (Fig. 13) and the total analysis (Fig. 14), although on a one-to-one basis there are differences. For example, the highest MMI® Cu value (30.5 ppm) is on the Finniss River south of Darwin in the Northern Territory. The second highest value for Cu by total analysis (113 ppm) is in Northern Queensland. In both cases the alternative technique has a high (>75th percentile), but not an outlier, value at the site. The likely explanation is that, due to the high proportion of mine tailings in the Finniss River catchment, the catchment outlet sample here has a very high percentage ‘bioavailability’ or lability.
Point distribution of MMI® Cu in NGSA catchment outlet sediments. A small circle symbol (○) is assigned for data between the minimum and the 25th percentile, a small dot (.) for data between the 25th percentile and the median, a larger dot (●) for data between the median and the 75th percentile and a plus sign (+) for data between the 75th percentile and the maximum. Lower outliers, if present, are assigned a large circle symbol (○). Upper outliers, if present, are assigned a large solid square symbol (■). For more detail, please see Caritat & Cooper (2011a).
Point distribution of Total Cu in NGSA catchment outlet sediments. For symbol explanation, see Figure 13.
By contrast, Figures 15 and 16 show the distribution of MMI® Fe and total Fe, respectively. Clearly where Fe is in highest overall abundance (e.g. the Pilbara) as measured by total analysis, it is relatively ’unavailable’ to plants as measured by MMI® extraction. This contrasts to, for example, the eastern seaboard of Australia where high rainfall and weathering rates ensure a good supply of ‘soluble’ Fe, despite the underlying lithology having a lower overall Fe content. Manganese shows a similar large difference between total and bioavailable methods.
Point distribution of MMI® Fe in NGSA catchment outlet sediments. For symbol explanation, see Figure 13.
Point distribution of Total Fe in NGSA catchment outlet sediments. For symbol explanation, see Figure 13.
In the case of agriculture, soil scientists have developed partial extraction techniques including methods for P (Colwell 1967a,b), and the DTPA test for Zn, Fe, Mn and Cu (Lindsay & Norvell 1978) which are now relatively standard soil testing procedures for plant-available nutrients. It is interesting to compare the results of the present study to these methods. Table 3 shows the range of concentrations extracted from the NGSA catchment outlet sediments by the MMI® technique. Average ’bioavailable’ concentrations range from below the ppm level for Mo, Ni, Se, and Zn to hundreds of ppm for Ca, Mg and K. Tiller (1983) lists extracted concentrations in Australian agricultural soils ranging from <0.1 to 45 ppm for Cu, 0.3 to 2560 ppm for Mn and <0.05 to 47 ppm for Zn, using the DTPA method of Lindsay & Norvell (1978). Values presented here (Table 3) based on catchment outlet sediments cover a similar range, indicating that the extraction efficiency and the estimation of ’bioavailability’ based on the MMI® extraction compare favourably with those from the DTPA method. Comparison of data from nearly 100 agricultural soils after MMI® extraction for P with the Colwell (1967a, b) method suggests that the two procedures correlate reasonably well (r2 = 0.60) and have similar extraction efficiencies (Mann, unpublished). Tests with MMI® on these same soils for Ca, Mg and K against standard agricultural soil laboratory procedures confirm similar findings with r2 values of 0.96, 0.97 and 0.95, respectively.
Lower limit of detection (LLD), minimum, median, average and maximum concentrations determined for 11 nutrients by MMI® extraction of 1190 NGSA catchment outlet sediment samples (0–10 cm, <2 mm)
Tiller (1983) also provides an extensive list of total concentrations of micronutrients (Cu, Mn, Zn, Mo) in Australian soils, categorized by soil groups. However, he points out that in some cases a negative correlation exists between total content and micronutrient deficiency in plants. For example, Zn deficiency in black earths and grey, brown and red clays occurs despite these soils having total Zn at the upper end of the range. Another similar example occurs for Mo in lateritic podzols. The only soil group for which total concentrations give reliable results are podzols and deep sands which have very low total micronutrient concentrations (Cu 2–19 ppm, Mn 40–530 ppm, Zn 1.5–164 ppm and Mo 1.8–7.2 ppm) in relation to the amounts being required for crop production. It is for this reason that assessment of soil status for plant growth in Australia now relies heavily on ‘bioavailability’ and partial extraction methods. Tiller (1983) goes on to suggest that organic ligands, similar to those used in the DTPA method of Lindsay & Norvell (1978) and in MMI®, are present in natural soils and complement inorganic complexes for these micronutrients. Concentrations of acetic acid and polybasic acids of up to 10-3 M and free amino acid concentrations of 10-4 M in bulk incubated soil solutions have been suggested. Hodgson et al. (1965) show that Cu and Zn in soil solutions are present mainly as metal-organic complexes; however metal-inorganic complexes are more likely in arid regions due to the higher concentrations of chloride and sulphate ions and the probable lower concentrations of organic complexes (in tropical Australia, conversely, the presence of organic acids in rain is well documented, e.g. Gillett et al. 1990).
Trace elements have played a vital role in the development of Australian agriculture and few countries have derived so much benefit from their use (Williams & Raupach 1983). Most important are deficiencies in Cu, Mo and Zn; the widespread occurrence of their deficiency can be attributed to prolonged weathering and leaching in predominantly geologically stable landscapes. Large areas in the lower ranges (lightest shade) in the distribution patterns of Figures 9 and 10 for Zn and Mo are in accord with this fact. Manganese deficiency in oats grown on rendzina soils of the Mt Gambier district of South Australia was one of the first reported examples of trace element deficiency in Australia (Samuel & Piper 1928). Magnesium and K deficiencies are less common. The continental distribution of MMI® Mg (Fig. 5) shows relatively high ‘bioavailable’ concentrations throughout the arable lands of the eastern states. Calcium deficiencies have been reported in eastern Australia on acid soils and have been responsible for problems in pastures and apple orchards. Low ’bioavailable’ Ca is only evident along the immediate coastline in eastern Australia (Fig. 4).
Iron oxides have a profound effect on the availability of P in Australian soils (Hingston et al. 1974; Taylor et al. 1983). In general, the surfaces of Fe oxides and oxyhydroxides below a pH of 8 are positively charged; anions such as phosphate (PO43-) are strongly adsorbed and tightly held. As pointed out by Taylor et al. (1983), this can have positive and negative effects. Large amounts of phosphate fertilizer, usually superphosphate, need to be applied to Australian soils to have an impact on plant growth. In some cases eutrophication of rivers and lakes may be minimized or at least delayed by the amount of P bound to soils and immobilized, provided the soil particles remain in situ.
Iron is ubiquitous in the Australian landscape, and its mobility and redox state have a profound effect on the regolith and the geochemical behaviour of many elements. In actively weathering profiles, Fe is released into solution as Fe2+ at the weathering interface, and subsequently oxidized to Fe oxide/oxyhydroxide, commonly goethite (FeOOH). The oxidation and immobilisation of soluble Fe2+ according to the reaction:
has been referred to as ferrolysis (oxidation and hydrolysis of Fe). Importantly, whenever and wherever oxidation of one mole of Fe occurs by this process, two moles of acid (H+) are produced. The acidity produces seepages with pH values below 2.5 on the western side of the Yilgarn Craton (Mann 1983), and has a very important effect on mobility of most metals and some nutrients. High concentrations of Cu and Pb (i.e. in the ppm range) occur in salt lake brines on the Yilgarn Block in Western Australia (Mann 1983). In Figure 6 much of the east coast of Australia is shown to have high concentrations of mobile Fe in catchment outlet sediments. Values of up to 555 ppm extractable Fe are evident. These coastal environments have acid sulphate soils (ASS) which pose environmental problems (e.g. Bush et al. 2004). The catchment outlet sediments with high mobile Fe concentrations have low pH (Caritat et al. 2011a). The low pH of these coastal sediments was related in that paper to the Prescott Index, PI (Prescott 1950):
where Pr is the average monthly precipitation and Ev is the monthly evaporation from a free water surface. High PI is associated with higher rainfall, less evaporation and overall greater leaching of soils. It is here suggested that because of its high concentrations in most rock types, its propensity to produce acid upon oxidation and the distribution of mobile or ‘bioavailable’ Fe in relation to low pH, that Fe is a primary causative agent of low soil pH.
The MMI® technique was originally designed and is widely used for mineral exploration on a prospect scale (Mann et al. 1998) where the primary aim is to locate mineralization more or less vertically beneath (in situ regolith). Although the technique also has the ability to detect anomalous geochemistry in stable exotic overburden such as till (e.g. Cross Lake, Mann 2010), practitioners are advised to avoid sampling active aeolian or alluvial landforms. It is interesting that the technique has been used successfully here with catchment outlet sediments sampled mainly on floodplain landforms. The aim in the case of the NGSA strategy is, however, quite different to that of locating buried mineralization: it is to provide an integrated representative geochemical record of the rocks and soils within a catchment. Transported sediments will provide information from remote (i.e. transported) sources within catchments, whether they be related to lithology, mineralization, environmental processes or agricultural practices. Perhaps even more unexpected is the fact that the combination of two techniques, catchment outlet sediment sampling and MMI® partial extraction, has worked so well on a continental scale and on samples collected at an extremely low density. It has not only worked well for the 11 nutrients studied in detail here but for at least another 40 elements, whose Australia-wide distributions have been documented in a preliminary fashion elsewhere (Caritat & Cooper 2011a; Caritat et al. 2011b). Mapping of bioavailable forms of elements other than nutrients, such as As, Cd and Hg, which may have adverse environmental consequences (e.g. Albanese 2008), is potentially useful for both determination of baseline levels and identifying significant sources.
Low density geochemical mapping is very dependent on having reliable, representative sampling locations. It is evident that at times of peak flooding useful geochemical information has been carried via fine sediments to the catchment outlet sampling locations, and that these locations have provided representative material for regional geochemistry. It is also apparent that MMI® extraction followed by ICP-MS analysis has provided useful alternative geochemical information to total analysis and is a measure of ’bioavailability’ for many elements.
The National Geochemical Survey of Australia has been conducted cost-effectively over an area >6 million km2 by applying a sample density of 1 site per 5200 km2. As a result of implementing the MMI® technique to the NGSA catchment outlet sediment samples, useful partial extraction geochemical information has been obtained and interpreted on a continental scale for the first time.
Acknowledgments
The NGSA project was part of the Australian Government’s Onshore Energy Security Program 2006–2011, from which funding support is gratefully acknowledged. NGSA was lead and managed by Geoscience Australia and carried out in collaboration with the geological surveys of every State and the Northern Territory under National Geoscience Agreements. The authors acknowledge and thank all landowners for granting access to the sampling sites and all those who took part in sample collection. The sample preparation and analysis team at Geoscience Australia is thanked for its contributions, as is analytical staff at SGS Perth laboratories. The authors would also like to thank two anonymous reviewers for their helpful comments. PdC published with permission from the Chief Executive Officer, Geoscience Australia.
- © 2012 AAG/Geological Society of London
References
Bioavailability of nutrients in Australia from Mobile Metal Ion® analysis of catchment outlet sediment samples: continental-scale patterns and processes
