Determination of sediment quality in the Nyl River system , Limpopo Province , South Africa

The importance o� wetland management and conservation is becoming more and more prevalent in the world today. It is thus important to determine baseline contamination values �or wetlands to assist in making in�ormed management decisions. Sediment from the Nyl River flood plain in the dry Limpopo Province was analysed using sequential extraction and ICP-MS to determine baseline metal concentrations, and bioavailability thereof. Eight heavy metal (Cu, Cd, Cr, Al, As, Zn, Mn, Pb) concentrations were determined and compared to sediment quality guideline values to assess sediment quality. Fractionation o� the elements was also noted to assess the bioavailability o� the metals. The results indicated that the sediment is o� a �air quality in comparison to the sediment quality guideline values. They also indicate that the metals will only become available in the presence o� strong reducing agents as most o� the metal concentrations were recorded in the �th and 5th �ractions obtained from the Tessier sequential extraction of the sediment samples. The study concluded that the sediment is of a fair quality and that it poses little potential threat to the system.


Introduction
From the myriad results reported in the literature it is possible to conclude that metals have a high to�icity and worldwide distribution in the aquatic environment.They are also known to accumulate in sediments (Klavins et al., 1998) Data concerning environmental e��ects o� chemicals clearly indicate the accelerated and negative e��ects o� the dispersal o� metals and metalloids in the environment by anthropogenic activities, and the changes made to global chemical cycles (Mester et al., 1998).The study o� sediments in wetlands is important as wetlands act as natural filters for water in a system and thus act as a sink for contaminated suspended particles in the water column.Sediments also provide an indication o� potential contamination on a temporal scale.The analysis o� water indicates the contamination status at present whereas sediment can provide in�ormation on the systems' contamination history (Shine, 2004).Wetlands can also act as a source o� increased contaminant levels in a water body during periods of increased water flow by remobilising the settled particles resulting in the re-suspension o� the contaminants into the water.These sediments are transported downstream and affect the ecosystems of the river downstream as well as flooded wetlands (Ulbrich et al., 1997).The mobilisation of sediments by water flow allows contaminants to penetrate deep into wetlands by flood waters (Ulbrich et al., 1997).It is thus important to assess the sediment quality throughout the system.
Contaminants bind to the sediment particles (Buykx et al., 2002).These contaminants can be either metal compounds, or originate �rom chemical compounds released into the system via a number o� anthropogenic activities.
These metal compounds may be present in several geochemical phases that act as reservoirs or sinks o� trace metals in the environment (Li et al., 1995).These phases include the broad categories: exchangeable, specifically adsorbed, carbonate, Fe-Mn oxides, organic matter and mineral lattice (Li et al., 1995).It is thus recognised that the quantification of the chemical forms o� metals in the sediment is essential �or estimating the mobility and bioavailability of metals in the environment (Leschber et al., 1985;Li et al., 1995).
The study area falls within the Waterberg catchment area.It �ollows the course o� both the Klein and Groot Nyl Rivers, �rom their sources to their confluence, and then the course of the Nyl River to Moorddrift Dam near Mokopane (Potgietersrus).The Nyl River flows through or is impacted on by the towns of Modimolle (Nylstroom) and Moogkopong (Naboomspruit).The Nyl River flows in a north-easterly direction from Modimolle in the west to Mokopane in the east.At Mokopane the river changes course northwards and is renamed the Makgalakwena River.The Makgalakwena River then flows into the Limpopo River.The Nyl River and its flood plain are subjected to various potential impacts via anthropogenic activities, such as mining and �arming, as well as the associated problems caused by �ormal and informal settlements.According to the Mookgopong (Anon, 2003a) and Mokopane (Anon, 2003b) tourism bureau the farming activities that take place include both agriculture and livestock.Crops such as maize, groundnuts, tobacco, citrus, cotton, millet, wheat, rice and sunflowers are planted in the fertile soil of the Waterberg catchment area.Cattle are also farmed extensively in the rich grazing provided by the flood plain.Around the sources o� the Klein Nyl River and the Groot Nyl River game �arming takes place on a large scale with a large number o� privately owned game farms (Anon, 2003c).The mining of tin, chrome and fluorspar also takes place in the area between Mookgopong and Mokopane.The Waterberg catchment is characterised by its richness in minerals with 545 known mineral deposits (Anon, 2003d).
The Nyl River flood plain is South Africa's largest ephemeral wetland, being 24 250 ha in size, and 500 ha of this wetland is situated in a nature reserve and is classified as a RAM-SAR site.The flood plain is unique in the fact that it is the only place in South Africa where the wild rice Oryza longistaminata grows (Gibbs et al., 1991).The nature reserve also provides valuable breeding ground �or the endangered Roan antelope (Hippotragus equinus) and numerous waterfowl species.In all there are 23 Red Data bird species found in the wetland (Tarboton, 1987).

Materials and methods
The �� sampling sites were chosen to be evenly distributed throughout the system, thereby providing representative sample points in the Nyl River, the Nyl River flood plain and some of the tributaries with more permanent flow throughout the season.Figure 1 indicates a schematic representation of the study area and sampling points.These sites were selected so that possible point sources of pollution could be identified.
It is generally recognised that in�ormation about the physico-chemical �orms o� elements is necessary to understand their environmental behaviour such as mobility and bioavailability (Tack and Verloo, 1995).Sediment samples were collected from the 18 localities throughout the system and subjected to a 5-point sequential extraction.The sequential extraction of metals �rom solid media is a common tool used in the analysis o� environmental geochemistry (Sutherland and Tack, 2003).This process uses di��erent reducing and o�idising agents to remove each �raction o� the bound metals in the sample.This allows �or the evaluation o� the total metal content available �or uptake by organisms or bioavailability.
The 5 �ractions e�tracted are listed in Table � with a brie� e�planation o� the remobilisation o� metal ions �rom each �raction.
Sediment samples were collected quarterly from March �00� to July �00� �rom the upper 50 mm o� the substrate and placed in 350 mℓ plastic honey jars and frozen.Frozen samples were returned to the laboratory �or �urther analysis.
In the laboratory the sediment samples underwent a process of sequential extraction.It was decided to use a 5-fraction extraction, which identifies the non-residual metal concentrations among the three basic operationally-defined host fractions (Ngiam and Lim, 2000).The process followed was modified from the process set out by Tessier et al. (1979).The process involves subjecting the sediment samples to chemicals of decreasing pH and increasing o�idising strength, to remove the operationally defined host fractions corresponding to the e�changeable, carbonate, reducible and organic/sulphide phases (Ngiam and Lim, 2000).Sediment samples were dried in an oven at 60ºC.Approximately 1 g of dry sample was placed in a 50 mℓ nalgene polyethylene centrifuge tube before it underwent extraction.Figure 2 gives a brief outline of the process followed during the e�traction process.Two replicates were prepared �or each e�traction.
All samples were then analysed for metal content using standard inductively coupled plasma mass spectrophotometry (ICP-MS) techniques.ICP-MS analyses were carried out on an X-series thermo elemental quadrupole-based ICP-MS.Yttrium was used as an internal standard to correct �or high dissolved solids arising �rom matri� e��ects.The metals chosen �or the study were selected �rom a list o� metals that were above acceptable levels in a scan done on water �rom the system.
For statistical analysis metal concentrations in sediment fractions were analysed using one-way ANOVA with season and site as independent factors.Significance level was taken as P<0.05.Organic matter such as detritus, living organisms and coatings on mineral particles bind trace metals through comple�ation and bioaccumulation processes.
Trace metals may be released under o�idising conditions during the degradation o� these substances.

Fraction 5 Residual or inert �raction
Residual �ractions are the trace metals bonded strongly to the crystal structure o� the minerals comprising the sediment.
These metals are not likely to be remobilised under normal environmental conditions

Aluminium
The increased bioavailability o� aluminium in sediments comes about by the remobilisation o� sediment particles by increased water flow and agitation in conjunction with decrease in pH (Buykx et al., 2002).This increased bioavailability can have various physiological e��ects on the organisms in the system.would indicate that the aluminium in the sediment is not very bioavailable and will only be released in the presence o� a strong reducing agent or i� the pH in the system were to decrease.
No significant differences were found between the mean aluminium concentrations for Fractions 1, 2, 4 and 5, when the fractions �or each month were compared to each other.The comparison between mean aluminium concentrations in Fraction 3 for March and July were, however, significantly different (P<0.05).
The results indicate that the majority of the aluminium is concentrated in the 5 th or inert �raction.This �urther implies that the majority of the aluminium extracted from the sediment is �rom a lattice or detrital origin and can be regarded as coming from a natural source (Jain, 2004) and that aluminium toxicity is not a threat to the organisms in the system.

Chromium
Chromium is a relatively scarce metal that occurs in several states.The most to�ic o� these states is the chromium VI or he�avalent state.Fytianos and Lourantou ( 2004) observed in their study of sediment from Lake Volvi and Koronia in Northern Greece, that chromium is primarily distributed in the reducible (Fe/Mn oxide), residual and oxidisable fractions.They found that metals bound to these di��erent �ractions have di��erent potentials for remobilisation and for uptake by biota (Fytianos and Lourantou, 2004).
Figure 4  The highest chromium concentrations recorded all �ell within the Sediment Quality Guideline (SQG) range of 81 to 370 mg/kg (EPA, 1999).The majority of the chromium concentrations were below the lower limit o� �� mg/kg or ERL (effect range low) (EPA, 1999).This would indicate that chromium is not a potential threat to the system.

Manganese
Manganese is an essential element (Health and Welfare Canada, 1980) that is a functional component in nitrate assimilation and is used as a catalyst in many enzymatic systems in both plants and animals (DWAF, 1996).Manganese is readily o�idisable and settles out o� the water column as MnO � (DWAF, 1996).
Figure 5 indicates the manganese concentrations recorded during the sampling period in the system.The results indicate that the third �raction had the highest concentrations o� manganese during the sampling period.No significant di��erences were recorded in manganese concentrations between corresponding �ractions o� each sampling month.This would indicate that manganese could be released into the system i� the pH o� the water in the system were to decrease.

Zinc
Zinc is an essential micronutrient �or all organisms and �orms the active site �or various metalloenzymes (DWAF, 1996).Figure 6 illustrates the zinc concentrations in the system during the sampling period.
There were no significant differences between Fractions 4 and 5 in the mean zinc concentrations during the di��erent sampling months.In Fraction 3 significant differences (P<0.The increased levels of zinc measured in July �00� samples were predominantly �ound at the sites close to the source o� the river and this would thus indicate that increased levels were �rom a natural source.

Copper
Copper is a common environmental metal and is essential in cellular metabolism but at high concentrations it can be highly toxic to fish (Grosell et al., 1997).Copper is generally remobilised with acid-base ion e�change or o�idation mechanism (Gomez Ariza et al., 2000).The site is, however, situated on a �arm about 50 m �rom the source o� the Klein Nyl River so it can be assumed the levels come �rom natural sources.
A comparison between the corresponding �ractions �rom each sampling month indicated that Fraction 3 was the only �raction that e�hibited significant differences.Significant differences in mean copper concentration were between November �00� and July 2002, and March 2002 and July �00�.

Arsenic
Arsenic is a highly toxic metalloid element (Rodrigues et al., 2003;Pizzaro et al., 2003).It is widely distributed as a trace element in rocks and soils and is mainly mobilised by microbial activities (Garcia-Sanchez and Alvarez-Ayuso, 2003).
Figure 8 indicates the arsenic concentrations in the system during the sampling period.The graphs indicate that most o� the arsenic is concentrated in the residual or inert fraction (Fraction 5).
Fractions 1 and 5 were the only �ractions to show significant differences between sampling months.Concentrations in Fraction 1 were significantly different for August 2001 and July 2002.Fraction 5 showed a significant difference between November �00� and March 2002.

Cadmium
Cadmium is a non-essential trace element that enters the  , 1996).Cadmium adsorbs strongly to sediments and organic matter (Sanders et al., 1999).Cadmium has a range of negative physiological e��ects on a organism, such as decreased growth rates and negative e��ects on embryonic development (Newman and McIntosh, 1991) Figure 9 indicates the cadmium concentrations recorded during the sampling period.
The ma�imum cadmium concentration o� �.5 mg/kg recorded during sampling �ell within the SQG and thus cadmium poses little potential threat to the organisms in the system.No di��erences were recorded between corresponding �ractions �or each month with the e�ception of Fraction 4, where significant differences were �ound between November 2001 and March 2002 and November �00� and July �00�.

Lead
Lead is a non-essential trace element (Ewers and Schlipkoter, 1991).The toxicity o� lead is dependent on the li�e stage o� the organism, and the presence o� organic material (Hellawell, 1986).Decreases in water pH can increase the bioavailability o� lead in the system (Hellawell, 1986).

Conclusions
The results clearly indicate that the metals do not appear to pose an environmental problem in the system.Figures 11 A to D indicate a stacked graph o� the percentage makeup o� the di��erent fractions for the different metals.The figure indicates that the majority of the metals are partitioned into Fractions 3, 4 and 5 are generally not bioavailable.This would signi�y that metals �rom the sediments pose little to no potential threat to the organisms in the system.This would also imply that most o� the metals in the system are �rom a natural source.
The results also indicate that all metal concentrations �ell within the lower end o� the Sediment Quality Guideline Range.These will thus have little or no e��ect on the organisms in the system.Zinc was the only metal that had concentrations greater than the guideline value.This is however little cause �or concern as they were recorded in the residual or inert fraction (Fraction 5).This would imply that they are natural concentrations in the sediment.

Figure 1 Figure 2 Figure
Figure 1Schematic of sampling points in the Nyl River System (inlay provincial and national positioning(Anon, 2007)).
Figures 3 A to D indicate the various aluminium concentrations, �or each �raction, observed at each locality during the di�-�erent sampling periods.The concentrations recorded indicate that the majority of the aluminium in the sediment is partitioned into the � th and 5 th �ractions, that is, they are bound to the �ractions containing organic matter and the residual or inert �raction.This indicates the chromium concentrations found in the samples collected during the di��erent sampling months.The graphs illustrate that the chromium is partitioned in the different fraction during the different months.Chromium is predominantly found in the oxidisable Fraction 4 and inert Fraction 5.During the August 2001, November 2001 and March 2002 sampling periods the observed partitioning of chromium concentrations is primarily in Fractions 4 and 5.During the July sampling the majority of the chromium is �ound in the 5 th �raction.In the comparison between the di��erent �ractions and the months sampled, Fractions 4 and 5 indicated no signifi-Available on website http://www.wrc.org.zaISSN 0378-4738 = Water SA Vol.33 No. 5 October 2007 ISSN 1816-7950 = Water SA (on-line) cant differences (P>0.05) in mean chromium concentrations between August 2001, November 2001, March 2002 and July 2002.Fraction 3 indicated a significant difference in mean chromium concentration between March 2002 and July 2002.Fraction 1 indicated significant differences in mean chromium concentrations between August 2001 and July 2002 but no significant differences in mean chromium concentrations between August 2001 and March 2002 and July �00�, and November �00� and March 2002 and July 2002.No significant differences in mean chromium concentrations were recorded between August 2001 and November 2001.Fraction 2 indicated significant differences between August 2001 and November 2001, August 2001 and March 2002, November 2001 and March 2002 and November �00� and July �00�.

Figure 4
Figure 4(a-d) Chromium concentrations (mg/kg) in the different fractions during the different sampling months.A: August 2001, B: November 2001, C: March 2002, D: July 2002 05) in the mean zinc concentrations were found between August 2001 and July 2002, August 2001 and March 2002, November �00� and July �00� and March 2002 and July 2002.Fraction 2 indicated significant di��erences in the mean zinc concentrations between August 2001 and March 2002 and Fraction 1 between August 2001 and July 2002.
observed in the system during the sampling period.Copper concentrations recorded all

Figure 6
Figure 6(a-d) Zinc concentrations (mg/kg) in the different fractions during the different sampling months.A: August 2001, B: November 2001, C: March 2002, D: July 2002

Figure 8
Figure 8(a-d) Arsenic concentrations (mg/kg) in the different fractions during the different sampling months.A: August 2001, B: November 2001, C: March 2002, D: July 2002 indicates lead concentrations determined in the system during the sampling period.The results indicate that the majority of the lead is partitioned in Fractions 3, � and 5.No significant differences were noted in mean lead concentrations between corresponding �ractions �or Fraction 2 and 3 from August 2001 to July 2002.In Fraction 1 significant differences were recorded between August 2001 and March 2001, August 2001 and July 2002, November 2001 and March 2001 and November 2001 and July 2002.In Fraction 5 significant differences were recorded between August 2001 and November 2001, March 2002 and July 2002,

Figure
Figure 10(a-d) Lead concentrations (mg/kg) in the different fractions during the different sampling months.A: August 2001, B: November 2001, C: March 2002, D: July 2002