Assessment of lake Bunyonyi's water clarity status using Satellite Derived Turbidity.

Abstract

Lake Bunyonyi is the principal source of drinking water in Kigezi, as well as a habitat for faunal and foral biodiversity and a source of income. As a result, monitoring of its quality is of concern. Traditional water quality monitoring approaches involve using in-situ methods that are generally tiresome, costly and are limited in time and space due to the lake’s expanse. Previous studies on the use of satellite imagery to monitor water quality on Uganda lakes have been carried out on large lakes i.e. Lake Victoria and Lake Kyoga. However this research consequently evaluates the use of Sentinel 2 imagery as an option to improve monitoring of turbidity of small lakes in Uganda. To do this, an excursion was carried out on Lake Bunyonyi to collect this in-situ turbidity data. At each sampling point, turbidity was determined by use of a turbidimeter. Turbidity measurements were determined monthly at nine sampling stations for a full year resulting into 108 samples. The mean annual In-Situ data at each station were then compared with their respective mean annual satellite- derived data from both Necchad and Dogliotti retrieval algorithms. The choice of these algorithms was driven by their global applicability. From the comparison, Necchad algorithm had r = 0.824, RMSE=0.25, Bias= 0.19, mean deviation = 0.188 , whereas Dogliotti algorithm had r = 0.808, RMSE = 0.62 , Bias=0.60, mean deviation= 0.597 .The results demonstrated that Necchad algorithm has the best approximation of In Situ turbidity of the lake hence suitable to be used in this region. On classification of the turbidity concentration maps from the best algorithm, it was shown that the water in middle of the lake was more transparent as compared to that closer to the shore. This was more so for water close to the highly industrious areas along the lake’s shore line. Thus the results generally demonstrated that water clarity for small lakes can be observed from Sentinel 2 imagery.