Both water and land resources are becoming increasingly stressed through the action or inaction of humans. Foreign substances can alter the physical and chemical parameters of both the water column and sediment. Copper, while an essential micronutrient to both plants and animals, can be detrimental to organisms at both high and low concentrations. Sediments are sensitive indicators for contaminants in aquatic environments. AAS and XRF have been used extensively to determine trace metal levels in sediments samples. Both Portsmouth Hard and Langstone Harbour are coastal mudflats, but with differing levels of development and exploitation. The results reported by AAS and XRF did not allow for a determination of which, if either, of the sites had a greater concentration of Cu, although statistical analysis did uphold the hypotheses that there is no significant difference between the two techniques. All reported concentration of Cu fell below probable effect level (PELs) for marine sediment.
Get your grade
or your money back
using our Essay Writing Service!
Both water and land resources are increasingly becoming stressed through the action or inaction of humans (Eletta, 2007). Marine ecosystems are threatened by major anthropogenic disturbances and, although degradation of the marine environment could have global consequences, the significance remains poorly understood (Lohrer, Thrush, & Gibbs, 2004). Water pollution has become a serious concern. Foreign substances can alter the physical and chemical parameters of both the water column and sediment. While some substances can be considered a source of nutrients for microorganisms, others are toxic to marine ecosystems (Eletta, 2007).
Marine sediments are commonly used as a gauge in environmental monitoring programs because they accumulate pollutants at concentrations above those in the water. This accumulation can be enhanced in enclosed and semi-enclosed areas, where the exchange of water with open sea is reduced (Bakan & Ã-zkog, 2007). Cu is ubiquitous in the environment, with 50 ppm in the Earth's crust and 0.25 ppb in ocean water (Chester, 1990). Variations in concentrations of Cu occur due to both naturally occurring and anthropogenic sources. Cu is an essential micronutrient to both plants and animals (Bakan & Ã-zkog, 2007). As such, organisms have mechanisms to deal with the specific Cu levels in their environment. The amount of Cu required for normal metabolism is small, at both high and low concentrations can be detrimental.
AAS and XRF, along with other mass analysis techniques such as ICP-MS, have been used extensively to determine trace metal levels in water and sediments samples (Eletta, 2007; Atgina, El-Aghab, ZararsÄ±zc, Kocatas, Parlakd, & Tuncela, 2000).
The aim of this report is to investigate variation of Cu concentrations in sediments, between Portsmouth Hard and Langstone Harbour, as well at comparing the techniques used to determine these concentrations. Additionally, nutrients in overlaying waters are compared for the two sites.
1.1. Study areas
Portsmouth Hard (site A)
Portsmouth Hard is a tidal coastal mudflat in busy commercial harbour. The sample site (fig.1) is located adjacent to a concrete slipway, close to a railway station, Gosport ferry and a landing area for small fishing boats. This site has a potentially a high risk of contamination from a variety of sources, principally engine hydrocarbon waste and heavy metal contamination (Fones, 2010).
Langstone Harbour (site B)
Langstone Harbour is a coastal mudflat in semi-enclosed harbour, which lies between Hayling Island and Portsmouth. The sample site (fig. 1) is near a slipway and close to large number of recreational boats (Fones, 2010). The harbour is covered by a wide range of designations, including SSSI and Ramsar. In addition to being an important environmental site, Langstone continues to be active an harbour for both commercial and recreational vessels, with 263 marine aggregates vessels passing through the harbour in 2009 (Langstone Harbour Board, 2009). This site has a potential risk of hydrocarbons and trace metals from the recreational boats and aggregate ships.
Figure 1: Location of sampling sites A and B. Maps adapted from Google Maps (Google, 2009).
2. Materials and Method
2.1 Sample collection
The samples were collected from Portsmouth Hard on the 11th March 2010 and from Langstone Harbour on the 12th March 2010. Sediment samples were collected at low tide from a 5m sampling area that was chosen using judgemental sampling. At each site six replicate sediment samples were collected from the upper 2cm of sediment, using a plastic spatula, and transferred into plastic bags. The bags were labelled, placed into a coolbox and transferred back to the University where they were stored at 4Â°C. A 125mL plastic bottle was filled with distilled water to act as the analytical blank sample (Fones, 2010).
Always on Time
Marked to Standard
2.2 Sediment analysis
Sediment samples were analysed using a Perkin Elmer 1100B flame atomic absorption spectrometer (AAS), using an air-acetylene flame, and a Philips PW1480 X-ray fluorescence spectrometer (XRF). AAS calibration was carried out using standard solutions.
For the samples analysed by AAS the sediment was homogenised in the plastic bag. Approximately 30mL of the sediment/water slurry was added to a 50mL centrifuge tube. The tubes from each group were placed into a Fisher Scientific Accuspin 1 centrifuge, along with a tube containing 30ml of distilled water to be used to check for contamination from the vessels, and spun for 10 minutes at 3,500rpm. The supernatant was poured off leaving the sediment. The centrifuged sediment was emptied into a PTFE beaker and dried in the oven for a week.
The dried samples were homogenised and 1.0g weighed out and placed into a 15mL centrifuge tube. Approximately 0.5g of certified sediment reference material, HR-1, was also placed into a 15mL centrifuge tube. A final 15mL centrifuge tube was labelled ARBlank. Acid digest was carried out by adding 2mL of Aqua Regia (three parts concentrated hydrochloric acid and one part of concentrated nitric acid) to each tube.
Acid digest samples were diluted to bring them within the calibration range. Each samples had 8mL of Milli-Q water added, 1mL of this solution was put into a fresh tube and a further 9mL of Milli-Q water added. The diluted samples were analysed for Cu using flame ASS.
Initial preparations for the XRF samples were carried out in a similar manner. The dried samples were homogenised and pressed into powder pellets, using a press that exerts a pressure of 4000psi, before being analysed for Cu using XRF. All values were blank corrected.
2.3 Nutrient analysis
The nutrient samples were analysed using a QuAAtro Auto Analyser. The water samples for nutrient analysis were filtered, using 0.7Î¼m glass fibre filters, at the time of collection. For each replicate 20mL of water was filtered into a plastic pot and 100Î¼L of Mercury II chloride added to fix the sample.
2.4 Statistical analysis
Statistical analysis was carried out using the PASW package. The samples were assessed for normality using visual analysis of histograms and the Kolmogorov-Smirnov test, before undergoing statistical analysis.
3. Results and Discussion
3.1 Comparison of Cu concentrations between the two sites
Sources of Cu in the marine environment are varied. Natural inputs include minerals in the rock that make up sediments, biological particles and hydrothermal systems. Anthropogenic inputs can be directly into the water, or leached after deposition on land, as well as from dust from the atmosphere. Historically Cu has been a key biocide in antifouling marine paints. Preservatives, such as Chromated Copper Arsenate (CCA) are used extensively on vessels constructed of wood. Studies have shown that after submersion the components of CCA, Cu, Cr and As, are lost from the wood (Brown, Eaton, & Thorp, 2001). Both sites are subject to anthropogenic disturbance, with the greater development at Portsmouth Hard suggesting that it would be more likely to have elevated levels of heavy metals.
The data from both Portsmouth Hard and Langstone Harbour was found to be normally distributed, with Kolmogorov -Smirnova results of z = 0.200 n = 11 and z = 0.200 n = 7 respectively.
For XRF the mean concentration of Cu was 36.00Î¼g/g Â±3.915 at Portsmouth Hard and 49.57Î¼g/g Â±3.387 at Langstone Harbour. The independent samples T-test disproved the null hypotheses, that there is no significant difference in concentrations of Cu between Langstone Harbour and the Hard for XRF. The Levene's test showed that the significance of F (2.068) was greater than 0.05 (p = 0.170) so consulting the equal variance estimate, t (16) = -2.412, p = 0.028. The two-tail significance indicates that the probability is less than 0.05 and therefore the mean difference of 13.571 is significantly different from zero.
Comparatively, the independent samples T-test for AAS data was not shown to be significant. The mean concentration of Cu was 31.80Î¼g/g Â±4.499 at Portsmouth Hard and 30.86Î¼g/g Â±2.955 at Langstone Harbour. The Levene's test showed that the significance of F (4.812) was less 0.05 (p = 0.044) so consulting the Equal variances not assumed estimates, t (15) = 0.175, p = 0.863. The mean difference of 0.943 is not significantly different from zero. In this case, the null hypotheses, that there is no statistically significant difference in concentrations of Cu between Langstone Harbour and the Hard for AAS, was accepted.
This Essay is
a Student's Work
This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.Examples of our work
The results from the two techniques do not allow for a determination of which, if either, of the sites has the greater concentration of Cu. The possible reasons for the discrepancy in results between the two techniques are discussed below.
3.2 Comparison of AAS and XRF trace metal techniques
A key decision in experimental design is the choice of analytical technique. For sediment geochemistry factors affecting this choice including the range of chemical elements to be determined, the acceptable level of uncertainty and the cost (Phedorin, et al., 2000).
Both techniques employed in this investigation have associated strengths and weaknesses. ISO 11466, extraction of trace elements soluble in aqua regia, suggests that aqua regia will not totally dissolve most soils and similar materials. It goes to on to state that as such elements extractable in aqua regia cannot be described as 'totals'. Additionally the standard suggests that these elements cannot be regarded as the bio-available fraction, an important factor when considering potential toxicity, as the extraction procedure is too vigorous to represent biological process. The process used in this investigation also differs from that given in the standard.
Key shortcomings of XRF include matrix and particle-size effects. These result in variations in the fluorescent intensities of the excited elements due to both the chemical composition and granulation of the sample (Stallard, Apitz, & Dooley, 1995). This can limit the accuracy and precision of XRF analysis, a factor that was possibly not considered carefully enough when the power pellets were formed.
To make a comparison between two analytical techniques it is important to assess the relative accuracy of the methods against the known certified values of a reference sample (Pilotto, Goff, & Weatherburn, 1998). The data from both the XRF standard, XRF HR-1, and the reference sample extracted in Aqua Regia, HR-1, was found to be normally distributed (Kolmogorov -Smirnova z = 0.200 n = 17 and z = 0.200 n = 6).
The mean for the concentration of Cu in the reference samples, XRF HR-1, was found to be 100.17Î¼g/g Â± 2.44, while the concentration of the known standard is reported in the company literature as 79.9Î¼g/g Â±11.4 (National Water Research Institute, 2006). In this case, the technique appears to have over estimated the concentration of Cu, although the null hypotheses, that there is no significant difference between the mean concentration of Cu measured in the reference samples and the known concentration of the reference standard, is upheld by the independent sample T-test, t (5) = 8.301, p < 0.05.
The mean for the concentration of Cu in the reference samples extracted in Aqua Regia, HR-1, is 53.82Î¼g/g Â± 3.23. In this case the techniques appears to have underestimated the concentration of Cu, although similarly the null hypotheses, that there is no significant difference between the mean concentration of Cu measured in the reference samples and the known concentration of the reference standard extracted in aqua regia, is upheld by the independent sample T-test, t (16) = -8.064, p <0.05. Means, plus the percentage error from the certified value can be found in table 1.
Table 1. Comparison of acid digest (AAS) and XRF analysis of HR-1 certified reference material.
Percentage Error from the Certified Value
Percentage Error from the Certified Value
100.17Î¼g/g Â± 2.44
53.82Î¼g/g Â± 3.23
79.9Î¼g/g Â± 11.4
The reference material used in this investigation was from a freshwater, not a marine source, although this does not appear to hold a significant bearing on the results and was employed by Stallard, Apitz, & Dooley (1995) when investigating the use of XRF for analysis of metals in marine sediments. The level of Cu in the reference sample, compared to that in the samples to be analysed, appears to be a more important consideration. In this case the lowest concentration of Cu found by XRF was 18Î¼g/g, with the highest being 64Î¼g/g, for AAS the lowest concentration was 11Î¼g/g and with the highest being 54Î¼g/g, all of which are notably lower than the Cu concentration given for the certified reference sample.
3.3 Comparison of the two trace metal techniques at each site
Having concluded that there is no signification difference between the two trace metal techniques, when looking at the certified reference sample, the below calculations investigate whether this conclusion is upheld within the field data.
The mean difference in Cu concentrations, measured by XRF and AAS, is 10.941. The paired samples T-test (t (16) = 4.737, P < 0.05) shows that mean difference is not significantly different from zero. Therefore the null hypotheses, that the mean difference in the total Cu concentrations measured by XRF and AAS is not significantly different from zero, can be accepted.
The mean difference in the total Cu concentrations measured by XRF and AAS at Portsmouth Hard is 5.500. The paired samples T-test shows that mean difference in Cu concentrations measured by XRF and AAS is not significantly different from zero (t (9) = 2.661, P < 0.05). Therefore the null hypotheses, that the mean difference in the total Cu concentrations at Portsmouth Hard, measured by XRF and AAS, is not significantly different from zero, can be accepted. The mean difference in the total Cu concentrations measured by XRF and AAS at Langston Harbour is 18.714. The paired samples T-test for Langstone Harbour shows that mean difference in totally Cu concentrations measured by XRF and AAS is not significantly different from zero (t (6) = 6.507, P < 0.05). Therefore the null hypotheses, that mean difference in the total Cu concentrations at Langstone Harbour, measured by XRF and AAS, is not significantly different from zero, can also be accepted.
These results are consistent with the findings from the above section and with other studies that have found no statistical difference between the trace metal measurements made with the two techniques (Atgina, El-Aghab, ZararsÄ±zc, Kocatas, Parlakd, & Tuncela, 2000; Radu & Diamond, 2009).
3.4 Comparison of nutrient concentrations between Langstone Harbour and the Hard for NO3, PO4 and Si
Not all nutrient data (table 2) for NO3, PO4 and Si was found to be normally distributed, with Kolmogorov-Smirnova results for Portsmouth Hard of n = 10 z = 0.110, z = 0.018 and 0.003 respectively, and for Langstone Harbour of n= 9, z = 0.004, z= 0.200 and z= 0.143 (results greater than z = 0.05 representin normal distribution). The removal of outliers resulted in normal distribution for PO4 (n = 9 z = 0.200) and Si (n= 9 z = 0.200) at Portsmouth Hard. Both outliers are from replicate 6 and it is possible that this sample was contaminated, either during collection or sample preparation. The NO3 data for Langstone Harbour is not normally distributed, this is possibly due to discrepancies in sampling, with some samples being collected from the harbour side of the sand spit and some from the open water side (fig. 1).
As the data for NO3 was not normally distributed a non-parametric Mann-Whitney U test was carried out. The supported the null hypotheses, that, under these circumstances, there is no significant difference in concentrations of NO3 between Langstone Harbour and the Hard, z = -3.674, p = <0.05.
For PO4 an independent samples T-test was carried out and disproved the null hypotheses, that there is no significant difference in concentrations of PO4 between Langstone Harbour and the Hard. The Levene's test for equal variance showed that the significance of F (0.521) was greater than 0.05 (p = 0.481) so consulting the equal variance estimates, t (16) = 2.878, p = 0.011, the two-tail significance indicates that the probability is less than 0.05 and therefore the mean difference of 0.207 is significantly different from zero.
Table 2. Nutrient data for Portsmouth Hard and Langstone Harbour. *outlying data, removed for statistical analysisFor Si the independent samples T-test also disproved the null hypotheses, that there is no significant difference in concentrations of Si between Langstone Harbour and the Hard. The Levene's test for equal variance showed that the significance of F (2.782) was greater than 0.05 (p = 0.115) so consulting the equal variance estimates, t (16) = -3.075, p = 0.008, the two-tail significance indicates that the probability is less than 0.05 and therefore the mean difference of 2.048 is significantly different from zero.
Nutrient Data for Portsmouth Hard - Âµmol/L
Nutrient Data Langstone Harbour - Âµmol/L
3.5 Additional sources of error
Short term variability in coastal environments can greatly affect trace metal concentrations in surface sediments. For example, increased runoff can cause higher concentrations of metals leached by weathering, while surface conditions affect the resuspension of particulate matter (Laslett, 1995).
Errors can be introduced during sampling processing. An important consideration for trace metal analysis is contact with metal equipment, which may potentially contaminate the samples. Another source of error can result from changes in composition due to temperature changes. It has been suggested that warming samples, from the ocean bottom to room temperature, may shift the ion exchange equilibrium and cause a release of ions from the sediment into the pore water (Bufflap & Allen, 1995).
Errors can also be introduced during processing. For example, when samples are centrifuged some fine particles can remain suspended in the water, especially if the sediment is disturbed while pouring extracted water from the tube. Ankley et.al. (1991), as cited by Bufflap & Allen (1995), reported that pore water samples showed significant losses (35-63%) of trace metals when filtered. While some of this loss could be due to the formation of precipitates or sorption by the glass fibre filters, it is possible that some of this is due to sediment particles remaining suspended in the water.
Although it is not possible to say with any certainty which site has the greater concentration of Cu, all concentrations reported fell below the probable effect level (PELs) of 108Î¼g/g for marine sediment (Canadian Council of Ministers of the Environment, 1999). While these figures are not officially recognised in the UK, they are widely used as baseline to evaluate the degree to which adverse biological effects are likely to occur (Pilotto, Goff, & Weatherburn, 1998; Bakan & Ã-zkog, 2007). The occurrence of adverse biological effects cannot be predicted solely from concentration data. The likelihood of adverse effects occurring in response to Cu exposure depends on the sensitivity of individual species, as well as a variety of physicochemical, biological and geochemical factors, which combine to affect bioavailability (Canadian Council of Ministers of the Environment, 1999).
The reasons for the discrepancy between comparisons of Cu concentrations, at the two sites, for two techniques, is not clear at this time. It is possible that errors introduced, either during sampling, processing or though analytical calibration, have affected the results. Further investigation, with improved methods and quality control measures, could provide a clearer outcome.
Both acid digest (AAS) and XRF techniques have been widely employed in the study of trace metals in sediment (Atgina, El-Aghab, ZararsÄ±zc, Kocatas, Parlakd, & Tuncela, 2000; Bakan & Ã-zkog, 2007, Bilinski, FranÄiÅ¡koviÄ‡-Bilinski, NeÄemer, HanÅ¾el, Szalontai, & Kovács, 2010), although the use of portable XRF appears to be favoured in some cases as it has a relatively fast throughput and is non-destructive (Stallard, Apitz, & Dooley, 1995; Radu & Diamond, 2009).
The most appropriate technique depends largely on the aim of the investigation. For example, if the data required is the total concentration of metal it possible that XRF would be the better technique, considering the inability of aqua regia to digest the whole sediment and XRF's that reports only the total element available in the fully oxidised state. Although, statistical, in this case, the data does not favour one technique over the other.