Search

An in-situ tree interception experiment was conducted to determine the hydrological impact of a solitary standing Norway maple and small-leaved lime in an urban environment. During the two-year experiment, rainfall data was collected and divided into interception, throughfall and stemflow. With approximately 38 % of the gross precipitation intercepted by both trees, the interception storage was higher than for similar studies done in Mediterranean regions. A regression analyses for the Norway maple found rainfall duration, rainfall amount and the tree's leaf area index to be the most important variables influencing interception storage. The regression analysis and the tree interception models by Gash and Rutter, as well as an adapted version of the Water and Energy Transfer between Soil, Plants and Atmosphere model (WetSpa), were tested for their accuracy in modelling the measured interception storage. The models in general overestimated interception storage for small interception events (< interception storage) and underestimated interception storage for bigger interception events (> interception storage). The regression analysis wasn't stable throughout seasons, event sizes and trees making it unsuitable for generic use. The method of Gash slightly overperformed WetSpa and Rutter for all events throughout seasons and trees. However, WetSpa showed a better performance for rainfall events > 10 mm. A scenario analysis, featuring the construction of student houses on a university campus, demonstrated the potential of urban trees to retain rainfall water. Even though trees alone could not restore the natural hydrological balance, they could partly mitigate the increased runoff volume and peak discharge caused by sealing of the natural surface through decreasing the net rainfall that reaches the ground. This study highlights the role of solitary trees in an urban environment where natural hydrological processes are severely altered. [ABSTRACT FROM AUTHOR]

An essential factor in the propagation of drought from meteorological drought to groundwater drought is the delay between a precipitation event and the groundwater recharge reaching the groundwater table. This delay mainly occurs in the unsaturated soil zone of the hydrological cycle. However, studies of drought propagation in the hydrological cycle often neglect these lag effects caused by the vadose zone. Additionally, this effect is often not considered in groundwater modelling, analysis of climate change impact on groundwater, and the effective management and sustainability of future water resources.This research aims to estimate the spatially distributed vadose zone lag time in relation to drought propagation using the kinematic wave approximation of Richards' equation under the assumption of deep drainage occurring in response to rainfall infiltration, steady vertical flow (no flow barriers or artificial drainage), and isotropic and homogenous soil hydraulic properties. This theory is combined with the van Genuchten-Burdine and Brooks-Corey parametric models. This research is carried out in the Dijle and Demer catchments in central Belgium, which overly the vulnerable Brulandtkrijt and Central Campine groundwater systems. In this research, two approaches are implemented for groundwater recharge delay estimation for both van Genuchten-Burdine and Brooks-Corey parametric models. In the first approach, a detailed parametrization of soil and geological layers was obtained based on 168 borehole descriptions. In the second approach, lumped hydraulic and physical properties were used for a limited number of soil and geological layers. The results of both parametric models from both approaches were validated using the physical based flow model Hydrus-1D. As the result, detailed parametrization of soil and geological layers gives good validation result with a Nash-Sutcliffe Efficiency of 0.89 for Brooks-Corey and 0.80 for van Genuchten-Burdine.Under consideration of detailed parametrization of soil and geological layers, the estimated spatially distributed groundwater recharge delay in the vadose zone using the Brooks-Corey model ranges from 0 to 110 months, with a spatial mean of 10.3 months. Using the van Genuchten-Burdine model the delay ranges from 0 to 73 months with a spatial mean of 6.7 months. The thickness of the vadose zone is the main parameter influencing the spatial distribution of this delay. In regions with shallow groundwater mostly found in the northern part of the study area, groundwater recharge delay is less than 2 months whereas the delay is up to 110 months for the areas found in the southeast with deep water tables. This delay is combined with the groundwater recharge time series to compare the result with meteorological drought. This is done by comparing the time series of standardized precipitation index (SPI) with the delayed recharge. Finally, the result of the two models and approaches are compared. The Brooks-Corey model using a detailed parametrization of the geological layers results clearly shows groundwater drought in recharge occurred after meteorological drought.Keywords: Groundwater Drought, Groundwater Recharge Delay, Vadose zone, Kinematic wave approximation [ABSTRACT FROM AUTHOR]

In urban catchments prone to pluvial flooding, due to the short response times of these catchments, rainfall data with high spatial and temporal resolution are required to increase reliability of pluvial flood modelling. Whether an intense and localized rainfall event triggers a pluvial flood, highly depends on the precise location and rainfall values of the high intensity locations at the local scale.One of the objectives of the FloodCitiSense project, which aims to develop an early warning service for pluvial flooding in urban areas, is to explore the potential of alternative data sources for near real-time rainfall observation at ground level in urban areas via low-cost sensors and citizen science. The pilot cities within FloodCitiSense are Brussels, Rotterdam and Birmingham. In this abstract we focus on Brussels as case study.A professional network of 16 rain gauges (Flowbru.be) covers the Brussels Capital Region (162 km²) and provides open data in real-time. Although the spatial distribution of this network is rather dense, there are parts of the city with missing rainfall data. The specific topography of the Brussels Region moreover has an influence on the spatial distribution of rainfall over the region. Therefore we explore potential alternative (low-cost) sources to complement the professional rain gauge observations.A first alternative source is a network of relatively cheap commercial rain gauges (NetAtmo), which provide data in near real-time. We analysed the availability, quality and usefulness of the NetAtmo data by comparing with the professional network data. We collected one-year hourly rainfall measurements from several NetAtmo stations in Brussels and compare the rainfall data with the Flowbru professional rain gauge stations within two kilometres radius. The NetAtmo stations have on average 30 to 35% missing data and after removing the missing data they record slightly less amount of rainfall volume compared to the professional rain gauges. The Pearson correlation coefficients between the rainfall observations of the two rain gauge types on average are over 0.7. A second source is to deploy a network of low-cost sensors with the help of citizens (citizen science concept). The low-cost sensor measures the raindrop arrival rate by measuring by simple counting the number of drops hitting a piezo element. The piezo is placed on a 3D printed holder and the signal is processed by an Arduino datalogger. 20 low-cost sensors were installed, enabling to fill the gaps in the professional rain gauge network. Preliminary results are erratic, mainly due to difficult operation conditions (heat wave, water infiltration) as well as hard/software issues. Thirdly, we evaluate the intense rainfall quantification and pluvial flood occurrence via the FloodCitiSense App, which has been co-created with citizens and water actors. The data collection is done by enabling citizens to report via the MobileApp on rainfall intensity (5 categories) and pluvial flood occurrence (8 categories). This reporting is currently being tested.A preliminary conclusion is that the low-cost alternatives surely has potential to complement professional rainfall data collection, but that their remain still quite some challenges before turning them into operational products. [ABSTRACT FROM AUTHOR]