Water quality has become a global concern due to ever increasing population and developmental
activities that are polluting water resources. Botswana’s water resources are threatened by various
pollution sources such as agricultural runoff, industrial and domestic effluents. This study was
carried out to assess the water quality of Dikgathong Dam in Botswana using physicochemical
analysis of water quality parameters and remote sensing techniques. The study first assessed
landuse patterns before construction (2010) and after construction (2015) to establish dominant
landuse in order to select water quality parameters related to the principal landuse in the catchment.
Images for 2010 and 2015 were acquired from Landsat and were classified using the supervised
classification through the Maximum Likelihood algorithm. Results showed that forest and shrubs
were the dominant landuse covering 73.7 % of the total area, followed by settlements (21.1 %) and
agricultural fields (2.76 %). Chl_a, COD, EC, TP, TN, TSS, NO3 and NO2 were selected for testing
and analysis based on their relationship with forest, settlements and agricultural fields. For
assessment of water quality, ten points were sampled in the dam from 15th January to 07th April
2016. Temperature, pH, EC, COD, TDS, TSS, turbidity, chloride, nitrates, sodium, potassium,
calcium, magnesium, sulphates, phosphates, total phosphorus, alkalinity, hardness and algae were
tested and analysed according to standard methods. Only COD, turbidity and TSS exceeded the
limits set by Environmental Protection Agency (EPA) surface water standards of 2001, making
Dikgathong Dam slightly polluted. One way ANOVA showed significant variations (p<0.05)
between water quality values in all sampling points only for NO3, SO4, pH, algae and Na. Five
different groups of sites were identified from ten sites using cluster analysis. The principal
component analysis identified ten parameters (COD, EC, turbidity, TSS, Ca, Mg, NO3, SO4, total
hardness and alkalinity) based on similarities of water quality characteristics. The Water Utilities
Corporation, which is responsible for the dam, can therefore monitor water quality at five points
focusing mainly on ten parametersfound to be principal. This study also investigated the likelihood
of integrating remote sensing and in-situ measurements to assess the water quality status of the
dam. Quasi analytical algorithms and MODIS data were used to quantify Chl_a and TSS
concentrations in the dam. Values for Chl_a were between 1.74 and 24.4 mg/m3
, while TSS ranged
from 2.34 mg/l to 59.2 mg/l. Based on chlorophyll concentrations the dam can be classified asboth oligotrophic and mesotrophic as per the EPA 2001 standard. The QAA and MODIS can
therefore be deployed as a mechanism for near real time monitoring of water quality in Botswana
reservoirs. Spearman’s correlation was used to test whether satellite retrieved water quality
parameters relate to in-situ measurements. Strong positive significant correlation was observed
between chl_a and turbidity (r=0.794 and 0.830), TSS (r = 0.819 and 0.770), SO4 COD (r=0.781
and 0.769).), SO4 (r= 0.851 and 0.646) and alkalinity (r= 0.847). Moderate positive and nonsignificant
relationship is observed for temp (r= 0.055), pH (r= 0.587), EC (r= 0.409), TDS
(r=0.348), Na (r= 0.406) and Cl (r= 0.394). Strong positive and significant correlation was
observed between remote sensing retrieved TSS and in-situ measured TSS (r= 0.733) and turbidity
(r= 0.867). This study concludes that there is strong positive correlation between parameters
retrieved through remote sensing and in-situ measurements and therefore can be used in
monitoring and assessment of the water quality in the lake at any point in time.