Author(s): Dorota Miroslaw-Swiatek; Mateusz Grygoruk; Sylwia Szporak-Wasilewska; Tomasz Okruszko
Linked Author(s):
Keywords: No Keywords
Abstract: In a typical lowland agricultural landscape of Europe, dominated by small (+/- 5 m wide and +/- 1 m deep) rivers and adjacent wetlands, processes of the response of these systems to management should be studied to avoid their degradation, yet allowing for successful and sustainable agriculture. One of the purposes of river maintenance works is to reduce the hydraulic roughness of river channels, by mowing vegetation. This action allows for lowering water levels during the peak flows of floods. To reduce negative impacts on the river biodiversity, vegetation mowing can be performed in sections, leaving plants (and even whole patches of macrophytes) in the channel (Bal et al., 2011). The affected river channel level regulates the hydraulic head of adjacent groundwater and affects the status of ecosystems in the valley. In practical modeling applications, assessment of the impact of maintenance measures on adjacent wetlands uses a simplified approach of combining the groundwater flow model with the river flow model through an internal boundary condition value defined as the position of the water table in the river (Grodzka-Lukaszewska et al., 2022). In this study, we proposed an artificial neural networks model (ANN) as a tool in the prediction of the water level, and the effect of vegetation mowing on river flow conditions was taken into account by the flow reduction factor (k). The research was made for the river reach (6 km) of the river Upper Biebrza (NE Poland, Fig. 1). The maximum channel width is 8 m and the maximum depth does not exceed 2.5 m. The river is characterized by a slight water slope not exceeding 0.00024 and is very intensively overgrown. Vegetation starts in May and continues until the end of November, reaching its maximum extent in September. This is illustrated by a bankful flow value of 0.96 m3/s during the non-vegetation period and 0.37 m3/s in September. Water levels were measured with automatic pressure transducers installed in piezometers located in the stream bed at four cross-sections located along the reach in km 5+946 (RI), 3+556 (RII), 2+548 (RIV), 0+531 (RV) (Fig. 1). The loggers were in continuous operation between 27 June 2018 and 10 June 2021.
Year: 2024