A5049969849- Hydrological Forecasting Using AI
- Hydrology and Watershed Management Studies
- Flood Risk Assessment and Management
- Neural Networks and Applications
- Hydrology and Drought Analysis
- Hydrology and Sediment Transport Processes
- Energy Load and Power Forecasting
- Meteorological Phenomena and Simulations
- Plant Water Relations and Carbon Dynamics
- Water resources management and optimization
- Groundwater flow and contamination studies
- Computational Physics and Python Applications
- Soil and Unsaturated Flow
- Soil erosion and sediment transport
- Dam Engineering and Safety
- Environmental Monitoring and Data Management
- Air Quality Monitoring and Forecasting
- Soil Moisture and Remote Sensing
- Model Reduction and Neural Networks
- Energy and Environment Impacts
- Geotechnical Engineering and Analysis
- Oil, Gas, and Environmental Issues
- Statistical and Computational Modeling
- Research Data Management Practices
- Simulation Techniques and Applications
University of Nottingham
2006-2017
University of Greenwich
1999-2003
University College Cork
1999
University of Leeds
1997-1998
This paper traces two decades of neural network rainfall-runoff and streamflow modelling, collectively termed ‘river forecasting’. The field is now firmly established the research community involved has much to offer hydrological science. First, however, it will be necessary converge on more objective consistent protocols for: selecting treating inputs prior model development; extracting physically meaningful insights from each proposed solution; improving transparency in benchmarking...
The forecasting power of neural network (NN) and autoregressive moving average (ARMA) models are compared. Modelling experiments were based on a 3-year period continuous river flow data for two contrasting catchments: the Upper River Wye Ouse. Model performance was assessed using global storm-specific quantitative evaluation procedures. NN ARMA solutions provided similar results, although naïve predictions yielded poorer estimates. annual then grouped into set distinct hydrological event...
Abstract The internal behaviour of an artificial neural network rainfall—runoff model is examined and it demonstrated that specific architectural features can be interpreted with respect to the quasi-physical dynamics a parsimonious water balance model. Neural solutions were developed for daily discharge series simulated by conceptual given observed precipitation totals evaporation rates Test River basin in southern England. outputs associated each hidden node, produced from output node...
"Panta Rhei – Everything Flows" is the science plan for International Association of Hydrological Sciences scientific decade 2013–2023. It founded on need improved understanding mutual, two-way interactions occurring at interface hydrology and society, their role in influencing future hydrologic system change. calls strategic research effort focused delivery coupled, socio-hydrologic models. In this paper we explore synthesize opportunities challenges that socio-hydrology presents...
Abstract. This paper evaluates six published data fusion strategies for hydrological forecasting based on two contrasting catchments: the River Ouse and Upper Wye. The input level discharge estimates each river comprised a mixed set of single model forecasts. Data was performed using: arithmetic-averaging, probabilistic method in which best from last time step is used to generate current forecast, different neural network operations soft computing methodologies. results this investigation...
Abstract. Two recent studies have suggested that neural network modelling offers no worthwhile improvements in comparison to the application of weighted linear transfer functions for capturing non-linear nature hydrological relationships. The potential an artificial perform simple transformations under controlled conditions is examined this paper. Eight models were developed: four full or partial emulations a recognised rainfall-runoff model; solutions developed on identical set inputs and...
Abstract Several studies have observed that neural network models will often produce phase-shift errors or timing lags in their output results. This paper investigates a potential solution to the error problem through application of procedure first applied sunspot prediction. was two hydrological forecasting for River Ouse, northern England, using neuro-evolution toolbox. Models were optimised on combination root mean squared and correction factor. The this produced improvements up about six...
Multi-step ahead inflow forecasting has a critical role to play in reservoir operation and management Taiwan during typhoons as statutory legislation requires minimum of 3-h warning be issued before any releases are made. However, the complex spatial temporal heterogeneity typhoon rainfall, coupled with remote mountainous physiographic context, makes development real-time rainfall-runoff models that can accurately predict several hours time challenging. Consequently, there is an urgent,...
This paper deals with the application of an innovative method for combining estimated outputs from a number rainfall-runoff models using gene expression programming (GEP) to perform symbolic regression. The GEP multimodel combination uses synchronous simulated river flows four conventional produce set combined flow estimates different catchments. selected combinations are linear perturbation model (LPM), linearly varying gain factor (LVGFM), soil moisture accounting and routing (SMAR) model,...
Four design tool procedures are examined to create improved neural network architectures for forecasting runoff from a small catchment. Different algorithms used remove nodes and connections so as produce an optimised model, thereby reducing computational expense without loss in performance. The results also highlight issues selecting analytical methods compare outputs different procedures.
This paper presents a Colombian-based study on hydrological modelling metrics, arguing that redundancies and overlap in statistical assessment can be resolved using principal component analysis. Numerous scores for optimal operator water level models developed at 20 monitoring stations, producing daily, weekly ten-day forecasts, are first reduced to set of five composite orthogonal metrics not interdependent. Each is next replaced by single surrogate measure, selected from several original...