| dc.description.abstract | Lake Kyoga, a shallow lake surrounded by numerous swamps and one of the great African lakes in Uganda is facing an increasing pressure from human activities yet there is limited information on water quality of the lake. Unlike Lake Victoria, little study has been carried out to monitor the water quality of the lake. Lake surface temperature (LST) and water clarity are important indicators of water quality, thus highly influence the lake ecosystem. In-situ measurements have been carried out to study these water quality parameters. Even though in situ observations usually have high accuracy, their availability is limited to a few sites, with even fewer sites having a continuous, reliable, long-term record of observations. MODIS sensor which has been used to study water quality is being operated beyond its life span, a new sensor VIIRS was launched to provide data continuity of its predecessors the MODIS and AVHRR. This study therefore aimed at comparing the VIIRS estimations to those of MODIS to determine whether the new launched sensor will be able to provide predictions with accuracy comparable to its predecessors.
Satellite derived Lake Surface Temperature for both VIIRS and MODIS were compared to in situ LST. Out of the 33 available points predetermined by in-situ measurements, 15 were used because the rest were obscured by cloud cover thereby making them unusable. The statistical analysis performed showed that VIIRS LST modelled was almost comparable to that of MODIS. Both sensors exhibited the same trend patterns when their mean annual estimates for years 2013-2018 was plotted.
Two Secchi disk depth algorithms were explored, Mueller and Lee et al. The algorithms were applied on MODIS and VIIRS imagery and the output compared to in situ SDD. Out of the 33 in situ points, only 10 points were used because the rest were obscured by cloud cover. The results from the statistical analysis showed that Mueller was the best SDD retrieval algorithm for VIIRS and that of MODIS was Lee et al.
It was concluded that VIIRS can actually be relied upon to carry out accurate estimation of LST in succession to its predecessors. It was recommended that the best SDD algorithm retrieval for Lake Kyoga using VIIRS imagery should further be validated and more in situ measurements should be incorporated. | en_US |
