To apply an AVS normalization, a good understanding of the amount and geographical distribution of AVS in sediments is needed. AVS concentrations are not yet measured on a routine basis in sediment monitoring programs but some data have been collected on the AVS distribution in different countries within the context of metal risk assessments where the SEM-AVS concept has been used. An overview of AVS ranges that are typically encountered in sediment is given per country in Figure 3 as Box-Whisker plots. A box and whisker diagram, or boxplot, provides a graphical summary of a set of data based on the quartiles of that data set. The ‘box,’ or rectangle, in Figure 3 contains 50% of the data, and the extremes of that box are the 25th percentile and 75th percentile. Each ‘whisker’ represents the remaining 25% of the data and the extremities of these whiskers are the minimum and maximum values of the data.
The largest database available is for the Flanders region of Belgium. This database, which is representative for EU low midland rivers, contains 200 sediments sampled over a depth of 0-10 cm. The 50th percentile of the AVS distribution yields an AVS value of 8.7 µmol/g dry wt. (Vangheluwe et al., 2005). The lowest AVS concentration in the Flanders dataset is 0.045 µmol/g dry wt. The 10th percentile is 0.77 µmol/g dry wt. The latter value has been used in the different ongoing metal risk assessments to be used as a generic default correction value for low midland rivers (the Netherlands, Germany, and possibly Northern France) when site-specific measurements are lacking.
In 2008, three additional countries were sampled: Finland, United Kingdom, and Spain (Vangheluwe et al., 2008). Although the intention was to sample in the spring season (April-May) when AVS concentrations are expected to be the lowest, this was only possible for Finland. In the United Kingdom, sampling was conducted in June-September and Spain in October. In Finland, a total of 25 samples were taken (13 lakes, 12 rivers). Analysis of the AVS concentrations gives a 10th percentile of 1 µmol/g dry wt. and a 50th percentile of 11 µmol/ g dry wt. The lowest concentration measured was 0.3 µmol/g dry wt. For the United Kingdom, 16 sediments from 16 different rivers were sampled. Analysis of the AVS concentrations gives a 10th percentile of 0.31 µmol/g dry wt. and a 50th percentile of 7.95 µmol/ g dry wt. The lowest concentration measured was 0.071 µmol/g dry wt. For Spain, 20 samples of the river Ebro were sampled. Analysis of the AVS concentrations gives a 10th percentile of 3.68 µmol/g dry wt. and a 50th percentile of 13.5 µmol/ g dry wt. The lowest AVS concentration measured was 1.7 µmol/g dry wt.
Burton et al. (2007) investigated AVS concentrations for 84 sites in wadable streams of 10 countries and nine ecoregions of Europe. The results showed AVS concentrations ranging from 0.004 µmol/g dry wt. to 44 µmol/g dry wt. with a median value of 0.1 µmol/g dry wt. (sample depth 0-5 cm) and an average value of 2.5 µmol/g dry wt. It should be noted that sediments in this program were collected in head streams resulting in very low AVS and SEM levels.
Sediment Sampling Recommendations
The use of AVS is sometimes criticized because of the dynamic behavior of AVS in natural systems. AVS concentrations have shown temporal and spatial (horizontal and vertical) variations depending on sediment type and hydrological conditions (Poot et al., 2007). Most often the AVS concentration increases with increasing sediment depth (even over small sample distances 0-10 cm,) and is linked to the redox gradient present in the sediment (Van Den Berg et al., 1998; Van Den Berg et al., 2001a, 2001b). In addition, there seems to be a strong seasonal component where AVS concentrations tend to be the higher at the end of the summer and during fall and lower in winter and spring (Howard and Evans, 1993; van den Hoop et al., 1997; Grabowski et al., 2001). However, it should be noted that transient nature of AVS in this regard maybe overstated as most of the studies relate to uncontaminated sediments where the oxidation of iron sulfide in sediments cannot be taken as indicative of the oxidation of the other metal sulfide complexes, which are more stable [e.g., copper, zinc, cadmium sulfide (Peterson, 1996) and nickel sulfide (Buykx et al., 2000)].
Despite the higher stability of nickel sulfide complexes, both temporal and spatial variations are important to be considered when collecting SEM-AVS data. As AVS concentrations have the tendency to be lower in spring and winter than in summer, it is recommended to sample sediments late winter/early spring. Furthermore, it is recommended to sample the top layer (0-5 cm) as AVS concentrations are lower than those found in the deeper layers (> 10 cm) (Van Den Berg et al., 1998; De Lange et al., 2008). As such, realistic worst case exposure conditions are guaranteed to conduct bioassays.
Acid volatile sulfide is defined operationally as those sulfides that are readily extracted by the cold extraction of sediment in approximately 1 molar HCl acid (Allen et al., 1993). Another term that is used in conjunction with AVS is SEM. SEM (Simultaneously Extracted Metal) can be defined as the metal, which is simultaneously extracted under the same conditions under which the AVS content is determined. AVS is a complex and variable fraction represented by a variety of reduced sulfur components, although often dominated by relatively labile iron and manganese monosulfides (Morse, 2004). In the past multiple techniques, employing different chemical reagents and methodologies have been used to extract AVS from sediment. As the efficiency of sulfide extraction from minerals including pyrite varies amongst methods different results can be expected from these measurements. In part to promote a greater intercomparability of AVS and SEM results extraction with 1 molar HCl has been proposed (Allen, 1991). Nevertheless, it has been shown by Hammerschmidt and Burton (2010) that caution is needed to interpret results that lack well described quality control procedures which may lead to a high variability among laboratories. This was even more apparent in cases where sediments with low AVS concentrations (< 0.5 µmol/ g dry wt.) were used. However, recently Brumbaugh et al. (2011) has shown, in an interlaboratory comparison of measurements of Acid Volatile Sulfide and Simultaneously extracted nickel in spiked sediments, that measurements of AVS and SEM-AVS can be reproducible among different laboratories when performed with structured analytical guidelines.