Using Landsat satellite imagery to monitor the spatial and temporal dynamics of aquatic weed extent in Lakes Chivero and Manyame, located in an urban catchment of Zimbabwe

Abstract

This study quantified the spatial and temporal variation of aquatic weeds in two lakes in an urban catchment of Zimbabwe using the automatic water extraction index (AWEI) and normalised difference vegetation index (NDVI) derived from Landsat satellite data from 1986 to 2020. Extent of aquatic weeds estimated using AWEI in Lake Chivero increased from less than 1 km2 (4%) in 1986 to 7 km2 (27%) in 2020. NDVI-based aquatic weed estimation gave the least spatial extent in the first few years. Similarly, in Lake Manyame aquatic weeds occupied ~62 ha (<1% in 1986) before reaching a peak extent of 60 km2 (~70%) in 1995, based on AWEI estimates. NDVI-derived aquatic weed extent ranged from less than 2 km² in 1997 to a maximum of 56.12 km² in 1994. Although AWEI and NDVI estimated similar extents, NDVI had higher estimates than AWEI. A non-significant positive trend in aquatic weed extent was detected for Lake Manyame based on AWEI (Mann-Kendal tau = 0.139, s = 69, p = 0.27) and NDVI (Mann-Kendal tau = 0.129, s = 64, p = 0.307). In Lake Chivero, a non-significant negative trend was observed in aquatic weed extent based on NDVI (Mann- Kendal tau = −0.06, s = −30, p = 0.6382), while a positive trend was detected using AWEI (tau = 0.0036, s = 18, p = 0.7827). Results of the regression analysis indicate that phosphorus (R² = 0.7957, p = 0.00122) and nitrogen (R² = 0.8992, p = 0.0011) significantly explained variations in aquatic weed infestation in Lake Chivero. These results suggest that phosphorus and nitrogen enrichment are key drivers of aquatic weed proliferation in the two lakes. Thus, sustainable management of water resources in the catchment hinges on reducing the amount of nutrients released into the lakes from sewage treatment plants and croplands.