Linking Spatio-Temporal Water Quality Dynamics to Fish Kill Events in Lake Victoria, Kenya Using Remote Sensing

Fish kill events have been increasingly reported in Lake Victoria, posing a significant threat to cage aquaculture, a rapidly growing sector supporting food security and local livelihoods in the region. These events are often associated with environmental stressors such as eutrophication, algal blooms, and upwelling; however, the mechanisms driving fish mortality remain poorly understood. In particular, environmental conditions preceding and during fish kill events are rarely captured comprehensively.

This study aims to analyze spatio-temporal variations in water quality and assess their relationship with fish kill events in the Kenyan portion of Lake Victoria using remote sensing techniques. The specific objectives are to (i) map spatial and temporal patterns of key water quality indicators and (ii) identify environmental conditions associated with fish kill occurrences. The study focuses on nearshore areas of Kisumu and Homabay counties, where a severe fish kill event in November 2022 resulted in mortality rates of up to 12.5% and significant economic losses.

Multi-source satellite datasets are being utilized, including Sentinel-2 (turbidity), Sentinel-3 OLCI (chlorophyll-a), MODIS (chlorophyll-a and surface temperature), and Landsat (historical trends). Time series analysis, spatial mapping, and event-based approaches are applied, including lag analysis to evaluate whether changes in chlorophyll-a, turbidity, and temperature precede fish kill events.

Data collection and processing are currently ongoing, with preliminary observations indicating notable spatial and temporal variability in chlorophyll-a and turbidity levels in affected areas. Final analysis will focus on identifying statistically significant relationships between water quality dynamics and fish mortality events. This study is expected to provide new insights into the environmental drivers of fish kills in Lake Victoria and demonstrate the potential of remote sensing for early warning and management of aquaculture risks.