scholarly journals Improved higher lead time river flow forecasts using sequential neural network with error updating

2014 ◽  
Vol 62 (1) ◽  
pp. 60-74 ◽  
Author(s):  
Om Prakash ◽  
K.P. Sudheer ◽  
K. Srinivasan
Keyword(s):  
Neural Network ◽  
Lead Time ◽  
River Flow ◽  
Forecast Accuracy ◽  
Flood Management ◽  
Residual Variance ◽  
Case Examples ◽  
River Flow Forecasting ◽  
Forecast Lead Time ◽  
Artificial Neural Network Ann

Abstract This paper presents a novel framework to use artificial neural network (ANN) for accurate forecasting of river flows at higher lead times. The proposed model, termed as sequential ANN (SANN), is based on the heuristic that a mechanism that provides an accurate representation of physical condition of the basin at the time of forecast, in terms of input information to ANNs at higher lead time, helps improve the forecast accuracy. In SANN, a series of ANNs are connected sequentially to extend the lead time of forecast, each of them taking a forecast value from an immediate preceding network as input. The output of each network is modified by adding an expected value of error so that the residual variance of the forecast series is minimized. The applicability of SANN in hydrological forecasting is illustrated through three case examples: a hypothetical time series, daily river flow forecasting of Kentucky River, USA and hourly river flow forecasting of Kolar River, India. The results demonstrate that SANN is capable of providing accurate forecasts up to 8 steps ahead. A very close fit (>94% efficiency) was obtained between computed and observed flows up to 1 hour in advance for all the cases, and the deterioration in fit was not significant as the forecast lead time increased (92% at 8 steps ahead). The results show that SANN performs much better than traditional ANN models in extending the forecast lead time, suggesting that it can be effectively employed in developing flood management measures.

scholarly journals Improving reliability of river flow forecasting using neural networks, wavelets and self-organising maps

2012 ◽  
Vol 15 (2) ◽  
pp. 486-502 ◽  
Author(s):  
Mukesh K. Tiwari ◽  
Ki-Young Song ◽  
Chandranath Chatterjee ◽  
Madan M. Gupta
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Lead Time ◽  
River Flow ◽  
Simulation Studies ◽  
River Flow Forecasting ◽  
Testing Dataset ◽  
Crucial Problem ◽  
Daily Discharge ◽  
Selection Of

Neural network (NN) models have gained much attention for river flow forecasting because of their ability to map complex non-linearities. However, the selection of appropriate length of training datasets is crucial and the uncertainty in predictions of the trained NNs with new datasets is a crucial problem. In this study, self-organising maps (SOM) are used to classify the datasets homogeneously and the performance of four types of NN models developed for daily discharge predictions – namely traditional NN, wavelet-based NN (WNN), bootstrap-based NN (BNN) and wavelet-bootstrap-based NN (WBNN) – is analysed for their applicability cluster-wise. SOM classified the training datasets into three clusters (i.e. cluster I, II and III) and the trained SOM is then used to assign testing datasets into these three clusters. Simulation studies show that the WBNN model performs better for the entire testing dataset as well as for values in clusters I and III; for cluster II the performance of BNN model is better compared with others for a 1-day lead time forecasting. Overall, it is found that the proposed methodology can enhance the accuracy and reliability of river flow forecasting.

A comparison between neural-network forecasting techniques-case study: river flow forecasting

1999 ◽  
Vol 10 (2) ◽  
pp. 402-409 ◽  
Author(s):  
A.F. Atiya ◽  
S.M. El-Shoura ◽  
S.I. Shaheen ◽  
M.S. El-Sherif
Keyword(s):  
Neural Network ◽  
River Flow ◽  
River Flow Forecasting ◽  
Forecasting Techniques

River flow forecasting for multiple time periods

1988 ◽  
Vol 15 (1) ◽  
pp. 58-65 ◽  
Author(s):  
Donald H. Burn
Keyword(s):  
Kalman Filtering ◽  
River Flow ◽  
Forecast Accuracy ◽  
Forecasting Model ◽  
Lead Times ◽  
Multiple Time ◽  
Filtering Algorithm ◽  
River Flow Forecasting ◽  
Forecast Precision ◽  
Time Periods

The performance of a river flow forecasting model employing a Kalman filtering algorithm was evaluated for increasing forecast lead times. The expected decrease in forecast accuracy was quantified and a decrease in forecast precision was noted for increased lead times. The merits of external estimates of meteorological inputs to the model were evaluated through an examination of different forecasting options. It was revealed that even noisy estimates of meteorological events improved the flow forecasts. Key words: forecasting, Kalman filter, real time, precipitation, snowmelt.

scholarly journals A Comparison of Neural-Network and Surrogate-Severe Probabilistic Convective Hazard Guidance Derived from a Convection-Allowing Model

2020 ◽  
Vol 35 (5) ◽  
pp. 1981-2000
Author(s):  
Ryan A. Sobash ◽  
Glen S. Romine ◽  
Craig S. Schwartz
Keyword(s):  
Neural Network ◽  
United States ◽  
Lead Time ◽  
Eastern United States ◽  
Spatial Smoothing ◽  
Large Magnitude ◽  
Feed Forward Neural Network ◽  
Operational Forecasting ◽  
Forecast Lead Time ◽  
Probability Forecasts

AbstractA feed-forward neural network (NN) was trained to produce gridded probabilistic convective hazard predictions over the contiguous United States. Input fields to the NN included 174 predictors, derived from 38 variables output by 497 convection-allowing model forecasts, with observed severe storm reports used for training and verification. These NN probability forecasts (NNPFs) were compared to surrogate-severe probability forecasts (SSPFs), generated by smoothing a field of surrogate reports derived with updraft helicity (UH). NNPFs and SSPFs were produced each forecast hour on an 80-km grid, with forecasts valid for the occurrence of any severe weather report within 40 or 120 km, and 2 h, of each 80-km grid box. NNPFs were superior to SSPFs, producing statistically significant improvements in forecast reliability and resolution. Additionally, NNPFs retained more large magnitude probabilities (>50%) compared to SSPFs since NNPFs did not use spatial smoothing, improving forecast sharpness. NNPFs were most skillful relative to SSPFs when predicting hazards on larger scales (e.g., 120 vs 40 km) and in situations where using UH was detrimental to forecast skill. These included model spinup, nocturnal periods, and regions and environments where supercells were less common, such as the western and eastern United States and high-shear, low-CAPE regimes. NNPFs trained with fewer predictors were more skillful than SSPFs, but not as skillful as the full-predictor NNPFs, with predictor importance being a function of forecast lead time. Placing NNPF skill in the context of existing baselines is a first step toward integrating machine learning–based forecasts into the operational forecasting process.

Sign in / Sign up

Export Citation Format

Share Document