Huayang Cai
Institute of Estuarine and Coastal Research, School of Ocean Engineering and Technology, Sun Yat-sen University, Guangzhou, 510275, China Correspondence: Huayang Cai (caihy7@mail.sysu.edu.cn)
2022/09/23
A data-driven model to quantify the impact of river discharge on tide-river dynamics in river deltas. Version 1.0 - September 2022
Provided by Huayang Cai
Institute of Estuarine and Coastal Research, School of Ocean Engineering and Technology, Sun Yat-sen University
Email contacts: caihy7@mail.sysu.edu.cn
How to cite:
- Huayang Cai, Bo Li, Erwan Garel, Tongtiegang Zhao, Feng Liu, and Suying Ou (2022), A data-driven model to quantify the impact of river discharge on tide-river dynamics in the Yangtze River estuary, Journal of Hydrology, submitted
In this newly proposed version, we:
-
proposed a new definition of critical river discharge Qc defined as the value that leads to an apparent shift of tidal phases by approximately 100~180° with the increase of river discharge.
-
added an error estimation model.
How to use R_TIDE
-
Download and install R_TIDE toolbox Users can download the latest R_TIDE toolbox from Github: https://github.com/Huayangcai/R_TIDE-V1.0-Matlab-Toolbox.git
-
R_TIDE Demo
2.1. Harmonic analysis driven by river discharge
First of all, you need to load data provided by R_TIDE Toolbox (such as Data_Yangtze_river.mat). The demo can be executed using the main program labelled by R_demo_Yangtze.m.
The syntax of the main subroutine is illustrated below:
[nameu,fu,yout,st,ft,Eta,Phi,percent,si,cof]=R_tide(xin,Q,T,lat1,ray,synth,Qc,twin,sname,ipso,ipre);
Descriptions of the inputs:
xin1: hourly water level data used for harmonic analysis
Q1: hourly river discharge data used for harmonic analysis
T1: the corresponding time series of the input data in term of ‘datenum’
lat1: the latitude of the selected station
ray: Rayleigh criteria, the default value is 1, which indicates that the Rayleigh criteria is used to select tidal constituents. Otherwise, Rayleigh criteria is not used.
synth: signal noise ratio, the default value is 10. You can set it depending on your own time series.
Qc1: the critical discharge beyond which the tide is vanishing, the default value is the maximum of the Q variable. Generally, it can be set to be the value corresponding with a negligible tidal range last for more than 2 days.
twin: window spectrum, the default value is 366.
sname: the name of the selected station
ipso: the method you used for harmonic analysis. If ipso=1, it will invoke standard PSO (Particle Swarm Optimization) algorithm to optimize and save the optimized results. If ipso=2, it will directly invoke the optimized file derived from PSO last time. If ipso=3, it will invoke default Matlab FMINCON Function to optimize and save the optimized results. If ipso=4, it will directly invoke the optimized file derived from FMINCON function last time. So if it is the first time to use it, this parameter should be set as 1 or 3.
ipre: the method you used for prediction. If ipre=1, it will save the coefficients for prediction and then invoke the program, namely, Rtide_predict.m; If ipre=2, it will directly predict. Divide the time series into 2 parts, one of which is used to derive coefficient for prediction, the other is used for prediction. Users should set the length of time series in the file, namely, Rtide_pre.m. So if it is the first time to use it, this parameter should be set as 1.
tsnr: decide whether you’d like to add an error estimation. If tsnr=0, it will output the tidal properties. If tsnr=1, it will add an error estimation invoking the function 'Rtide_harmonic_witherr.m';
tau: the default value of travelling time of river discharge propagating to the studied tidal gauging station.
Descriptions of the outputs:
nameu: the name of the selected tidal constituents
fu: the frequency of the selected tidal constituents (/h)
yout: the reconstructed water level derived from R_TIDE, consisting of st and ft
st: the reconstructed water level derived by the residual water level model
ft: the reconstructed water level derived by the tidal-fluvial model
Eta: the time-dependent amplitude of each tidal constituent derived from R_TIDE
Phi: the time-dependent phase of each tidal constituent derived from R_TIDE
precent: the correlation coefficient between reconstructed and observed water levels
si: the Root Mean Square Error (RMSE) between reconstructed and observed water levels
cof: the regression coeffiicents adopted for each tidal bands derived from R_TIDE
tidecon: the tidal properties together with their errors;