Using remote sensing to assess how intensive agriculture impacts the turbidity of a fluvial lake floodplain

Abstract

Lake St-Pierre is set in the largest floodplain in the province of Quebec, Canada, and is a rich ecosystem of great ecological importance. However, Lake St-Pierre has seen its ecological integrity deteriorate in recent decades, largely due to the development of agriculture in and around its floodplain. This study uses a simple turbidity retrieval model (NIR and RED reflectances from Sentinel-2) to quantify the impact of land use on water turbidity within the lake floodplain during the 2019 and 2020 spring flood. Using a linear mixed effect model, we assessed how land use (wet meadows, cultivated grasslands, soybean, corn fields) impacts turbidity retrieved from Sentinel-2. Water turbidity was found to increase with the level of agricultural perturbations. During the severe and long 2019 flood, the turbidity was 5% higher over cultivated grassland fields, 35% higher over soybean fields and 70% higher over corn fields, compared to wet meadows. For 2020, a comparatively drier year, the corresponding impacts were 9% for cultivated grassland fields, 65% for soybean fields and 93% for corn fields. However, land use only explained 4% (2019) and 5% (2020) of the water turbidity spatiotemporal variance, which was instead mostly driven by the sediment load coming from upstream watersheds. Our results indicate that, despite complex processes driving water mass movements in the floodplain, intensive farming practices lead to higher water turbidity compared to natural lands. This study provides evidence that intensive agriculture impacts the water quality in a critical moment for the long-term health of the Lake St-Pierre.