Abstract
M.Sc.
The catchment area of the Olifants River has, over a number of years, been
exposed to extensive mining, agriculture and urbanisation activities in the
Witbank-Middelburg and Phalaborwa regions which has largely contributed
to the deterioration of water quality. Of major concern is the influence
anthropogenic activities have on the aquatic ecosystem of the Olifants
River within the Kruger National Park and how me health of fish residing
In these waters is affected. To determine fish health, a biological
monitoring method, the Health Assessment Index (H41), was tested for the
first time In South Africa.
Four surveys were conducted at Mamba and Balule In the Olifants River,
Kruger National Park during 1994 (February, May, July and November).
An additional survey was conducted in February 1995 at Loskop Dam.
Samples of water and sediment were taken for analysis of metals and
physical and chemical water parameters. A maximum of 20 Clarias
gariepinus fish were sampled at each location. Evaluation of the fish was
done according to guidelines set in the HAI and parasite population
composition (prevalence, abundance, mean intensity) was determined.
Organ and tissue samples including gills, liver, muscle and skin were
analysed for the bioaccumulation of chromium, copper, Iron, manganese,
nickel, lead, strontium and zinc, using atomic absorption
spectrophotomeby.
Metal concentrations in the water at Mamba and Balule were within
guideline limits, whereas concentrations at Loskop Dam were above
guidelines. As reflected by the application of the Aquatic toxicity Index,
Loskop Dam presented with the poorest water quality followed by Mamba
then Balule. It was found that certain physical and chemical variables
namely fluoride, potassium, sulphate and total dissolved solids
concentrations at Mamba and Balule were relatively high, particularly
during drier months.
Metals accumulated in organs and tissues, with the highest concentrations
In the gills followed by the liver, skin and muscle. The discriminant
analysis, utilising metal bloaccumulation, discriminates between water
quality at Mamba and Balule, revealing a 100 % classification probability
for each survey.
Values obtained In the application of the HAI indicated that variables with
good predictor accuracy were plasma protein, all parasites, endoparasites,
liver, white blood cell counts, ectoparasites, skin, fins and gills. The discriminant function for the HAI generally Indicated variables similar to
those exhibiting high predictor accuracy. The discriminant function
showed relatively low classification probability for each survey. In case 1,
where separate endo- and ectoparasite variables were Included in the
determination of me discriminant function, probability for me entire study
ranged between 47.5 % and 84.2 %. In case 2, where endo- and
ectoparasite variables were given a refined score rating system, probability
ranged between 62.5 % and 100 %. The low classification probability
Indicates either the Importance of repetitive testing for this technique or a
total departure from it.
Results showed that fish populations with higher HAI values are found in
water of poorer quality (Mamba), while healthier fish populations i.e. with
lower RAI values are found In water of better quality (Balule). Parasite
data shows a similar tendency by indicating that ectoparasites are
abundant in water of higher quality, while endoparasftes increase in water
of poorer quality. The HAI reflects the condition of fish populations in
relation to their aquatic environment, therefore, the HA/ gives an
indication of water quality and should be used as a first level screening
tool. If complemented by a parasite survey, distinguishing between endoand
ectoparasites, the results from the HAI will be enhanced.