Wavelet and artificial neural network analyses of tide forecasting and supplement of tides around Taiwan and South China Sea

Chen, Bang-Fuh; Wang, Han-Der; Chu, Chih-Chun

doi:10.1016/j.oceaneng.2007.04.003

Cited by 54 publications

(21 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Frequency domain analysis provides information about frequency localization that reflects the amplitude of different frequency signals; however, it is applicable only to stationary time series. A Fourier transform provides frequency domain representation for the temporal processing, but it offers no information about the local variations in time and it requires the signals to be stationary [ Chen et al ., ]. In the 1980s, French scientist Morlet proposed a wavelet analysis method that incorporates both time and frequency multiresolving functions [ Morlet , ; Lau and Weng , ; Torrence and Compo , ].…”

Section: Theory and Methodsmentioning

confidence: 99%

Temporal and spatial characteristics of pan evaporation trends and their attribution to meteorological drivers in the Three‐River Source Region, China

Wang

Zhao

et al. 2015

JGR Atmospheres

View full text Add to dashboard Cite

Pan evaporation is an important indicator of atmospheric evaporative demand, and its long-term variation is of much concern in studies of climate change. Based on data from 33 meteorological stations from 1962 to 2012, this work considered the temporal and spatial trends of pan evaporation and the meteorological variables that affect them in the Three-River Source Region (TRSR) of southwestern China. Pan evaporation in the TRSR has decreased significantly since 1988 with an obvious abrupt change from 1993 to 2003. Furthermore, a 27 year period of oscillation has existed over the past 51 years. Pan evaporation reflects the combined effects of four meteorological variables: net radiation (R n ), wind speed (u 2 ), actual vapor pressure (e a ), and air temperature (T a ). Based on this research, a number of conclusions were drawn. (1) The pace of climate change increased after 1980 and pan evaporation decreased at a rate of À13.3 mm/a 2 from 1980 to 2012, which is much faster than the rate of À1.2 mm/a 2 from 1962 to 1979. (2) For the decrease of pan evaporation from 1980 to 2012, the quantifying contributions of R n , u 2 , e a , and T a were À8.7, À6.4, À1.8, and +3.6 mm/a 2 , respectively. Thus, it was established for the TRSR that "global dimming" was the main reason, and "wind stilling" was a close second to global dimming for the decrease in pan evaporation. (3) Different regions of the TRSR are affected differently by the effects of the meteorological variables. Low-elevation regions in the TRSR are more susceptible to the effects of net radiation and wind speed, whereas high-elevation regions are affected more by actual vapor pressure and air temperature.

show abstract

Section: Theory and Methodsmentioning

confidence: 99%

Temporal and spatial characteristics of pan evaporation trends and their attribution to meteorological drivers in the Three‐River Source Region, China

Wang

Zhao

et al. 2015

JGR Atmospheres

View full text Add to dashboard Cite

show abstract

“…Thus, an important consideration in applying B-P learning is how well the network generalizes (Chen et al 2007) …”

Section: Artificial Neural Networkmentioning

confidence: 99%

A comparison between artificial neural network method and nonlinear regression method to estimate the missing hydrometric data

Bahrami

Kavianpour

Abdi

et al. 2010

Journal of Hydroinformatics

View full text Add to dashboard Cite

Missing values are a common problem faced in the analysis of hydrometric data. The need for complete hydrological data, especially hydrometric data for planning, development and designing hydraulic structures, has become increasingly important. Reasonably estimating these missing values is significant for the complete analysis and modeling of the hydrological cycle. The major objective of this paper is to estimate the missing annual maximum hydrometric data by using According to the coefficient of determination (R 2 ) and the root mean squared value of error (RMSE), it was concluded that ANN provides a better estimation of the missing data.

show abstract

“…Though Kalman filtering requires fewer amounts of data, its prediction are for short-term duration. Numerical models, such as the finite difference method, require accurate boundary conditions and geometric information 5) . Although including more number of constituents in the harmonic analysis improves the accuracy, it leads to the problem of growing memory and calculation time.…”

Section: Introductionmentioning

confidence: 99%

“…The results showed that the ANN model is as powerful as HM (60) when one-year tidal data are used and with 2-hour lead time. ANN and wavelet analyses were combined to extend the predictions for 5-year duration and to improve the prediction quality 5) . Feed forward neural network with Resilient Back Propagation (RBP) learning algorithm was used to predict tide levels and supplement missing data with quicker computation 16) .…”

Section: Introductionmentioning

confidence: 99%

Tidal-Level Forecasting Using Artificial Neural Networks Along the West Coast of India

2014

View full text Add to dashboard Cite

Knowledge of tide level is essential for safe navigation of ships in harbor, disposal and movement of sediments, environmental observations, explorations, and many more coastal and ocean engineering applications. Traditional methods such as harmonic analysis, least mean squares method, and hydrodynamic models have disadvantages in that they require excessive data, are time consuming, and are tedious to carry out. Artificial Neural Network (ANN) has been widely applied in the coastal engineering field in the last two decades for solving various problems related to time series forecasting of waves and tides; predicting sea-bed liquefaction and scour depth; and estimating design parameters of coastal engineering structures. Its ability to learn highly complex interrelationship based on provided data sets with the help of a learning algorithm, along with built-in error tolerance and less amount of data requirement, makes it a powerful modeling tool in the research community. In the present study, an attempt was made to predict tides at Karwar, located at the west coast of India, using a type of network called Non-linear Auto Regressive eXogenous input (NARX). It has an advantage in that the generated output is fed back to the network creating a loop. Conceptually, it differs in the fact that it uses the target given to it also as an input. Predictions were carried out for various durations using the weekly and monthly data sets. It was found that at Karwar, one year s prediction can be successfully carried out using one month data as an input with correlation coefficient ('r') greater than 0.97. The developed model was further applied to predict tides at New Mangalore Port Trust, Panambore, along the west coast of India, which is 260 kms south of Karwar. Results obtained were encouraging with 'r' value greater than 0.96.

show abstract

Wavelet and artificial neural network analyses of tide forecasting and supplement of tides around Taiwan and South China Sea

Cited by 54 publications

References 6 publications

Temporal and spatial characteristics of pan evaporation trends and their attribution to meteorological drivers in the Three‐River Source Region, China

Temporal and spatial characteristics of pan evaporation trends and their attribution to meteorological drivers in the Three‐River Source Region, China

A comparison between artificial neural network method and nonlinear regression method to estimate the missing hydrometric data

Tidal-Level Forecasting Using Artificial Neural Networks Along the West Coast of India

Contact Info

Product

Resources

About