A method to improve the stability and accuracy of ANN- and SVM-based time series models for long-term groundwater level predictions

2016 ◽  
Vol 90 ◽  
pp. 144-155 ◽  
Author(s):  
Heesung Yoon ◽  
Yunjung Hyun ◽  
Kyoochul Ha ◽  
Kang-Kun Lee ◽  
Gyoo-Bum Kim
Keyword(s):  
Time Series ◽  
Groundwater Level ◽  
Time Series Models ◽  
The Stability ◽  
Stability And Accuracy

scholarly journals Analysis of long-term observations of the groundwater level In an aseismic region

2019 ◽  
pp. 47-67
Author(s):  
A. A. Lyubushin ◽  
O. S. Kazantseva ◽  
A. B. Manukin
Keyword(s):  
Time Series ◽  
Groundwater Level ◽  
Atmospheric Pressure ◽  
Active Regions ◽  
Sampling Interval ◽  
Pressure Effects ◽  
Adaptive Compensation ◽  
Groundwater Level Fluctuations ◽  
And Cluster Analysis

The results of the analysis of continuous precise time series of atmospheric pressure and groundwater level fluctuations in a well drilled to a depth of 400 m in the territory of Moscow are presented. The observations are remarkable in terms of their duration of more than 22 years (from February 2, 1993 to April 4, 2015) and by the sampling interval of 10 min. These long observations are suitable for exploring the stationarity of the properties of hydrogeological time series in a seismically quiet region, which is important from the methodological standpoint for interpreting the similar observations in seismically active regions aimed at earthquake prediction. Factor and cluster analysis applied to the sequence of multivariate vectors ofthe statistical properties of groundwater level time series in the successive 10-day windows after adaptive compensation for atmospheric pressure effects distinguish five different statistically significant states of the time series with the transitions between them. An attempt to geophysically interpret the revealed states is made. Two significant periods – 46 and 275 days – are established by spectral analysis of the sequence of the transitions times between the clusters.

Rescue of groundwater level time series: how to identify and treat errors

2021 ◽  
Author(s):  
Jānis Bikše ◽  
Inga Retike ◽  
Andis Kalvāns ◽  
Aija Dēliņa ◽  
Alise Babre ◽  
...  
Keyword(s):  
Time Series ◽  
Groundwater Level ◽  
Monitoring Network ◽  
Observation Data ◽  
Automatic Data ◽  
Mean Values ◽  
Temporal Prediction ◽  
National Monitoring ◽  
Groundwater Drought

<p>Groundwater level time series are the basis for various groundwater-related studies. The most valuable are long term, gapless and evenly spatially distributed datasets. However, most historical datasets have been acquired during a long-term period by various operators and database maintainers, using different data collection methods (manual measurements or automatic data loggers) and usually contain gaps and errors, that can originate both from measurement process and data processing. The easiest way is to eliminate the time series with obvious errors from further analysis, but then most of the valuable dataset may be lost, decreasing spatial and time coverage. Some gaps can be easily replaced by traditional methods (e.g. by mean values), but filling longer observation gaps (missing months, years) is complicated and often leads to false results. Thus, an effort should be made to retain as much as possible actual observation data.</p><p>In this study we present (1) most typical data errors found in long-term groundwater level monitoring datasets, (2) provide techniques to visually identify such errors and finally, (3) propose best ways of how to treat such errors. The approach also includes confidence levels for identification and decision-making process. The aim of the study was to pre-treat groundwater level time series obtained from the national monitoring network in Latvia for further use in groundwater drought modelling studies.</p><p>This research is funded by the Latvian Council of Science, project “Spatial and temporal prediction of groundwater drought with mixed models for multilayer sedimentary basin under climate change”, project No. lzp-2019/1-0165.</p>

scholarly journals Improving Inflation Forecasts in the Medium to Long Term

2013 ◽  
pp. 1-6 ◽  
Author(s):  
Saeed Zaman
Keyword(s):  
Time Series ◽  
Forecast Accuracy ◽  
Time Series Models ◽  
Powerful Technique ◽  
Inflation Forecasts ◽  
Statistical Time

A simple but powerful technique for incorporating a changing underlying inflation trend into standard statistical time series models can improve forecast accuracy significantly—about 20 percent to 30 percent, two to three years out.

Comparative Study of Time Series Models, Support Vector Machines, and GMDH in Forecasting Long-Term Evapotranspiration Rates in Northern Iran

2020 ◽  
Vol 146 (6) ◽  
pp. 04020010 ◽  
Author(s):  
Afshin Ashrafzadeh ◽  
Ozgur Kişi ◽  
Pouya Aghelpour ◽  
Seyed Mostafa Biazar ◽  
Mohammadreza Askarizad Masouleh
Keyword(s):  
Time Series ◽  
Support Vector Machines ◽  
Comparative Study ◽  
Support Vector ◽  
Time Series Models ◽  
Northern Iran ◽  
Vector Machines

Time series study for groundwater level evaluation in monitoring well

2021 ◽  
Vol 13 (2) ◽  
Author(s):  
Rubens Oliveira da Cunha Júnior ◽  
João Victor Mariano da Silva
Keyword(s):  
Time Series ◽  
Groundwater Level ◽  
Sedimentary Basin ◽  
Precipitation Data ◽  
Environmental Aspects ◽  
Aquifer System ◽  
Groundwater Levels ◽  
Monitoring Well ◽  
Series Study

Climate and hydrogeological conditions of the Brazilian semi-arid demand sustainable and efficient water solutions. Groundwater monitoring programs are tools to subsidize the decision-making in this sense. In Ceará state, the monitoring of Araripe sedimentary basin aquifers is important for the development of the region. In this scenario, the present work aimed to study the groundwater level through an exploratory analysis of time series. The study area covered the eastern portion of the Araripe sedimentary basin, in the municipality of Milagres, in Ceará state. As the object of this study, it was obtained the time series of monthly average groundwater levels in a monitoring well of RIMAS/CPRM and installed in the Middle Aquifer System. Graphical and numerical methods were applied for the identification and description of time series main characteristics. Precipitation data in the study area were used to evaluate the system recharge. Results were discussed according to the environmental aspects of the study area. As a result, it was possible the identification and description of time series patterns such as trend and seasonality through the applied methods. It is also highlighted the sharp drawdown of groundwater levels in long term in the time series, reflecting the quantitative state of the aquifer system, as well as the groundwater recharge during the rainy season of the region, evidenced by the study of time series seasonality together with the precipitation data..

The ILRS Contribution to ITRF2020

2021 ◽  
Author(s):  
Vincenza Luceri ◽  
Erricos C. Pavlis ◽  
Antonio Basoni ◽  
David Sarrocco ◽  
Magdalena Kuzmicz-Cieslak ◽  
...  
Keyword(s):  
Time Series ◽  
Tracking System ◽  
A Priori ◽  
Systematic Errors ◽  
Novel Approach ◽  
Priori Information ◽  
The Stability ◽  
The Individual ◽  
Stability And Consistency

<p>The International Laser Ranging Service (ILRS) contribution to ITRF2020 has been prepared after the re-analysis of the data from 1993 to 2020, based on an improved modeling of the data and a novel approach that ensures the results are free of systematic errors in the underlying data. This reanalysis incorporates an improved “target signature” model (CoM) that allows better separation of true systematic error of each tracking system from the errors in the model describing the target’s signature. The new approach was developed after the completion of ITRF2014, the ILRS Analysis Standing Committee (ASC) devoting almost entirely its efforts on this task. The robust estimation of persistent systematic errors at the millimeter level permitted the adoption of a consistent set of long-term mean corrections for data collected in past years, which are now applied a priori (information provided by the stations from their own engineering investigations are still taken into consideration). The reanalysis used these corrections, leading to improved results for the TRF attributes, reflected in the resulting new time series of the TRF origin and especially in the scale. Seven official ILRS Analysis Centers computed time series of weekly solutions, according to the guidelines defined by the ILRS ASC. These series were combined by the ILRS Combination Center to obtain the official ILRS product contribution to ITRF2020.</p><p>The presentation will provide an overview of the analysis procedures and models, and it will demonstrate the level of improvement with respect to the previous ILRS product series; the stability and consistency of the solution are discussed for the individual AC contributions and the combined SLR time series.</p>

scholarly journals Influence of the range of data on the performance of ANN- and SVM- based time series models for reproducing groundwater level observations

2019 ◽  
Author(s):  
Heesung Yoon ◽  
Yongcheol Kim ◽  
Soo-Hyoung Lee ◽  
Kyoochul Ha
Keyword(s):  
Time Series ◽  
Groundwater Level ◽  
Model Building ◽  
Model Performance ◽  
Support Vector ◽  
Time Series Models ◽  
Ann Model ◽  
Data Set ◽  
Groundwater Level Fluctuations ◽  
Numerical Tests

In the present study, we designed time series models for predicting groundwater level fluctuations using an artificial neural network (ANN) and a support vector machine (SVM). To estimate the model sensitivity to the range of data set for the model building, numerical tests were conducted using hourly measured groundwater level data at a coastal aquifer of Jeju Island in South Korea. The model performance of the two models is similar and acceptable when the range of input variable lies within the data set for the model building. However, when the range of input variables is beyond it, both the models showed abnormal prediction results: an oscillation for the ANN model and a constant value for SVM. The result of the numerical tests indicates that it is necessary to obtain various types of input and output variables and assign them to the model building process for the success of design time series models of groundwater level prediction.

Sign in / Sign up

Export Citation Format

Share Document