Simulation Modeling of Runoff,Groundwater Flow and Sediment Transport into Kastoria Lake,Greece

Kastoria Lake is located in northwestern Greece. The whole basin of the lake is about 253 km². For the computation of the surface water volume inflowing into the lake from the main streams of the sub-basins located around Kastoria Lake, a rainfall-runoff sub-model is used. A quasi-three-dimensional simulation model of the Kastoria basin aquifer is also realized, in order to estimate the groundwater contribution to the volumetric budget of the lake and the whole basin as well. For the computation of sediment load inflowing into the lake from the main streams of the sub-basins, the rainfall-runoff sub-model is combinedwith a soil erosion sub-model and a sediment transport sub-model for streams. A GIS was developed in the hydrologic basin with all data needed for parameter identification and model application. The data base was enriched by a series of on site measurements of water discharge made in all main streams for one whole hydrologic year. By means of the resulting mathematical sediment model, those sub-basins, which deliver most sediment load to the lake, are identified. On the basis of this identification, a series of control measures, for the reduction of sediment inflowing into the lake, at certain places of the above mentioned sub-basins is proposed.     

Machine Learning Utilization for Bed Load Transport in Gravel-Bed Rivers

Three data-driven techniques, namely artificial neural networks, adaptive-network-based fuzzy inference system, and symbolic regression based on genetic programming, are employed for the prediction of bed load transport rates in gravel-bed steep mountainous streams and rivers in Idaho (U.S.A.), and the potential of several input variables is investigated. The input combinations that were tested are based, mainly, on unit stream power, stream power, and shear stress, and exhibited similarly good performance, with respect to the machine learning technique used, accentuating the importance of the regression model. The derived models are robust, generalize very well in unseen data, and generated results superior to those of some of the widely used bed load formulae, without the need to set a threshold for the initiation of motion, and consequently avoid predicting erroneous zero transport rates.     

Numerical Simulation of Sediment Transport Following a Dam Break

For the computation of the morphological changes on reservoir beds, in scale model experiments, a hydrodynamic model is combined with the sediment continuity equation. The hydrodynamic model is based upon the equations of mass and momentum conservation. The sediment transport capacity appearing in the sediment continuity equation is calculated by using the relationship proposed by Smart. For the solution of the above system of three partial differential equations, the explicit numerical schemes of Lax and Wendroff and MacCormack are tested. The whole mathematical model is applied to the experimental case of sediment release from a reservoir after a dam break.     

A Machine Learning Approach for the Mean Flow Velocity Prediction in Alluvial Channels

Water Resources Management - In natural alluvial channels, the determination of the flow resistance constitutes a problem with additional complexity compared to rigid bed channels, due to the bed...     

Computation of Drought Index SPI with Alternative Distribution Functions

The Standardized Precipitation Index (SPI) is widely used as drought meteorological index, to identify the duration and/or severity of a drought. The SPI is usually computed by fitting the gamma probability distribution to the observed precipitation data. In this work, the possibility to calculate SPI by fitting to the precipitation data the normal and the log-normal probability distributions was studied. For this purpose, 19 time series of monthly precipitation of 76?years were used, and the assumption that the gamma probability distribution would provide better representation of the precipitation data than log-normal and normal distributions, at various time scales (1, 3, 6, 12 and 24?months) was tested. It is concluded that for SPI of 12 or 24?months, the log-normal or the normal probability distribution can be used for simplicity, instead of gamma, producing almost the same results.