<abstract xmlns="http://www.w3.org/1999/xhtml"><p>The ice cover on lakes is one of the most influential factors in the lakes’ winter aquatic ecosystem. The paper presents a method for predicting ice coverage of lakes by means of multilayer perceptrons. This approach is based on historical data on the ice coverage of lakes taking Lake Onega as an example. The daily time series of ice coverage of Lake Onega for 2004–2017 was collected by means of satellite data analysis of snow and ice cover of the Northern Hemisphere. Input signals parameters for the multilayer perceptrons aimed at predicting ice coverage of lakes are based on the correlation analysis of this time series. The results of training of multilayer perceptrons showed that perceptrons with architectures of 3-2-1 within the Freeze-up phase (arithmetic mean of the mean square errors for training epoch 
<inline-formula><alternatives><inline-graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="graphic/j_eko-2017-0019_eq_001.png"></inline-graphic><math xmlns:mml="http://www.w3.org/1998/Math/MathML"><mrow><mover><mrow><mi>MSE</mi></mrow><mo>¯</mo></mover><mo>=</mo><mn>0.0155</mn></mrow></math><tex-math>$\overline {MSE} = 0.0155$</tex-math></alternatives></inline-formula>
) and 3-6-1 within the Break-up phase (
<inline-formula><alternatives><inline-graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="graphic/j_eko-2017-0019_eq_002.png"></inline-graphic><math xmlns:mml="http://www.w3.org/1998/Math/MathML"><mrow><mover><mrow><mi>MSE</mi></mrow><mo>¯</mo></mover><mo>=</mo><mn>0.0105</mn></mrow></math><tex-math>$\overline {MSE} = 0.0105$</tex-math></alternatives></inline-formula>
) have the least mean-squared error for the last training epoch. Tests within the holdout samples prove that multilayer perceptrons give more adequate and reliable prediction of the ice coverage of Lake Onega (mean-squared prediction error <italic>MSPE</italic> = 0.0076) comparing with statistical methods such as linear regression, moving average and autoregressive analyses of the first and second order.</p></abstract>

The ice cover on lakes is one of the most influential factors in the lakes’ winter aquatic ecosystem. The paper presents a method for predicting ice coverage of lakes by means of multilayer perceptrons. This approach is based on historical data on the ice coverage of lakes taking Lake Onega as an example. The daily time series of ice coverage of Lake Onega for 2004–2017 was collected by means of satellite data analysis of snow and ice cover of the Northern Hemisphere. Input signals parameters for the multilayer perceptrons aimed at predicting ice coverage of lakes are based on the correlation analysis of this time series. The results of training of multilayer perceptrons showed that perceptrons with architectures of 3-2-1 within the Freeze-up phase (arithmetic mean of the mean square errors for training epoch 
MSE¯=0.0155$\overline {MSE} = 0.0155$
) and 3-6-1 within the Break-up phase (
MSE¯=0.0105$\overline {MSE} = 0.0105$
) have the least mean-squared error for the last training epoch. Tests within the holdout samples prove that multilayer perceptrons give more adequate and reliable prediction of the ice coverage of Lake Onega (mean-squared prediction error MSPE = 0.0076) comparing with statistical methods such as linear regression, moving average and autoregressive analyses of the first and second order.

Selection of Parameters and Architecture of Multilayer Perceptrons for Predicting Ice Coverage of Lakes

Northern Water Problems Institute, Karelian Research Centre, Russian Academy of Sciences

Ekológia (Bratislava)

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 3.0 License.

The ice cover on lakes is one of the most influential factors in the lakes’ winter aquatic ecosystem. The paper presents a method for predicting ice coverage...

Selection of Parameters and Architecture of Multilayer Perceptrons for Predicting Ice Coverage of Lakes

Published Online: Sep 23, 2017

Page range: 226 - 234

DOI: https://doi.org/10.1515/eko-2017-0019

Keywords
Freeze-up phase, Break-up phase, Lake Onega, training sample, statistical methods

© 2017 Nikolaevich Vyacheslav Baklagin, published by De Gruyter Open

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 3.0 License.

Selection of Parameters and Architecture of Multilayer Perceptrons for Predicting Ice Coverage of Lakes

Published Online: Sep 23, 2017

Page range: 226 - 234

DOI: https://doi.org/10.1515/eko-2017-0019

KeywordsFreeze-up phase, Break-up phase, Lake Onega, training sample, statistical methods

© 2017 Nikolaevich Vyacheslav Baklagin, published by De Gruyter Open

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 3.0 License.

Keywords
Freeze-up phase, Break-up phase, Lake Onega, training sample, statistical methods