Forecasting surface water-level fluctuations of a small glacial lake in Poland using a wavelet-based artificial intelligence method

Abstract

Lake waters are a significant source of drinking water and contribute to the local economy (e.g. enabling irrigation, offering opportunities for tourism, waterways for transport, and meeting utility water demands); therefore, the ability to accurately forecast lake water levels is important. However, given the significant lack of research with respect to forecasting water levels in small lakes (i.e. 0.05 km 2 \ area \ 10 km 2 ), the present study sought to address this knowledge gap by testing a pair of hypotheses: (1) it is possible to forecast water levels in small surface lakes using artificial neural networks (ANN), and (2) better water-level forecasts will be obtained when the wavelet transform (WT) is used as an input data preprocessing tool. Based on an analysis of a case study in Lake Biskupinskie (1.16 km 2 ) in Poland and based on a range of model performance statistics (e.g. mean absolute error, root mean square error, mean squared error, coefficient of determination, mean absolute percentage error), both hypotheses were confirmed for monthly forecasting of lake water levels. ANNs provided good forecasting results, and WT pre-processing of input data led to even better forecasts. Additionally, it was found that meteorological variables did not have a significant impact in forecasting water-level fluctuations. In light of the results and the limited scope of the present study, proposed future research directions and problems to be resolved are discussed.

Simulating Power Generation from Photovoltaics in the Polish Power System Based on Ground Meteorological Measurements—First Tests Based on Transmission System Operator Data

“…They belong to a group of artificial intelligence information processing paradigms that has gained immense popularity in recent years. Simulation or forecasting models based on ANNs have been successfully applied in various areas of cognitive and application research, including water-demand forecasting [10], lake water-level forecasting [11], renewables integration studies [12], in the area of color image identification and reconstruction [13], multi-core optic fibers [14], wind speed prediction [15], or, most importantly from this paper's perspective, in the areas of direct and global radiation prediction [16] and PV energy yield forecasting [17].…”

A systematic review on machine learning algorithms used for forecasting lake‐water level fluctuations

“…Thus, apart from the information on past water‐level data, the studies have additionally utilized evaporation, rainfall, and inflow or outflow as input data samples to the models (Guldal and Tongal in 2010 15 used water levels with evaporation, rainfall, discharges, and incoming runoff; Talebizadeh and Moridnejad in 2011 23 used water levels with evaporation, precipitation, and inflow; Kakahaji et al 2013 used water levels with inflow, evaporation, and rainfall; Buyukyildiz et al in 2014 16 used water level with outflow, evaporation, rainfall, and inflow; Young et al in 2015 38 used water levels with rainfall, outflow, and inflow discharge; Liang et al in 2018 45 used water levels with rainfall, and water inflow). In addition to these research findings, that is, apart from the information on past water‐level data, the works have additionally used some meteorological variables 18,19,21,24,27,29,33,34,42,43,46 . In the aforementioned works, the meteorological variables used are wind, relative humidity, pressure, evapotranspiration, temperature, evaporation, and rainfall.…”

“…[100][101][102][103][104][105][106] In the case of the problem of forecasting water levels in lakes, the same findings are concluded. The example research works for this are listed in Table 1, where Altunkaynak in 2014 30 46 evaluated the effectiveness of a hybrid wavelet with ANN algorithm for forecast prediction of water level variations inside smaller glacial lakes located in Poland and claimed that this wavelet analysis can act as a pre-processing tool to give better forecasting results. The cause for the better performance of the hybrid WA integrated with ML algorithms is that the wavelet analysis tends to extract only the much-needed details pertaining to raw data input in which it removes the redundant details.…”

“…Herein, the outcomes of the study specified that the wavelet with ANN and wavelet with ANFIS hybrid algorithms outpaced the simple MLP and ANFIS algorithms; In the work of 40 by Shafaei and Kisi in the year 2016, they coupled the WA with both SVM and ABFIS algorithms, then they revealed that the hybrid algorithms provided good performance in predicting water level variations in Van lake located in Turkey; Another research work of 42 by Yadav and Eliza in the year of 2017 used the hybrid version of WA and SVM algorithm in forecasting water level variations in Loktak lake located in India and revealed that the wavelet analysis influenced the effectiveness and performance of the SVM algorithm. Piasecki et al in the year 2018 46 evaluated the effectiveness of a hybrid wavelet with ANN algorithm for forecast prediction of water level variations inside smaller glacial lakes located in Poland and claimed that this wavelet analysis can act as a pre‐processing tool to give better forecasting results. The cause for the better performance of the hybrid WA integrated with ML algorithms is that the wavelet analysis tends to extract only the much‐needed details pertaining to raw data input in which it removes the redundant details 107 …”

Hybrid Models for Weather Parameter Forecasting

“…Turgay et al applied the wavelet neural network to predict precipitation [17]. Piasecki et al predicted water level using wavelet ANN [18]. Khan et al studied the comparison among wavelet ANN and ANN models [19].…”