Monitoring and Predicting Proliferation of Invasive Aquatic Plants in Lake Victoria with MODIS and MERIS Data
Cheruiyot, E.K.1; Menenti, M.1; Gorte, B.G.H.1; Mito, C.O.2
1TU Delft, NETHERLANDS; 2University of Nairobi, KENYA
Availability of clean fresh water is one of the greatest environmental challenges worldwide. Lakes require regular monitoring in order to assess the quality of their water. In this work, we apply EO (Earth Observation) data to monitor the proliferation of aquatic plants in Lake Victoria, East Africa. We also seek to identify the core drivers of this proliferation by relating it to environmental factors like water quality (WQ) and hydrology. With a surface area of 68 800 km2, Lake Victoria is the largest of all African lakes and also the second largest freshwater body in the world. It has a huge socio-economic relevance, and an important economic resource for the riparian countries. It has however, since late 1980s, been faced with environmental challenges and human impacts which have perturbed the ecological balance affecting its biodiversity. Of great concern is the growth of invasive plants, especially the water hyacinth, which is associated with many adverse effects including obstruction to fishing, navigation and irrigation, interference with the aquatic biodiversity, water quality deterioration and a general risk to public health. The overall objective is to provide accurate and reliable information that would aid the management authorities in decision making towards weed control in the lake. This work is carried out within the ESA ALCANTARA Program and follows a successful TIGER 2 project, which provided a framework on mapping of aquatic vegetation using MODIS and MERIS data and proposed models for predicting vegetation proliferation. We refine these methods by first improving the accuracy of vegetation area estimation. We do this by defining a more accurate lake shoreline and also improving on the accuracy of endmember retrieval using higher spatial resolution imagery (Landsat 8 - TM and KOMPSAT-2). We then retrieve WQ indicators like total suspended matter (TSM), coloured dissolved organic matter (CDOM) and phytoplankton chlorophyll (Chl-a) from MERIS data. We also obtain hydrological data as well as industrial, domestic and agricultural discharge into the lake from a field campaign. We finally perform a regression analysis to establish statistical relationships between proliferation of aquatic plants and its drivers. From these relationships, we develop a model for predicting the proliferation of vegetation in the lake. Preliminary results indicate that vegetation proliferation responds optimally to the variations in the conditions of TSM, Chl-a and rainfall after a delay period of about two to three months with correlation coefficients R = 0.46, R = 0.57 and R = 0.67 respectively.