Genetic programming based monthly groundwater level forecast models with uncertainty quantification

Wavelet gated multiformer for groundwater time series forecasting

Developing accurate models for groundwater control is paramount for planning and managing life-sustaining resources (water) from aquifer reservoirs. Significant progress has been made toward designing and employing deep-forecasting models to tackle the challenge of multivariate time-series forecasting. However, most models were initially taught only to optimize natural language processing and computer vision tasks. We propose the Wavelet Gated Multiformer, which combines the strength of a vanilla Transformer with the Wavelet Crossformer that employs inner wavelet cross-correlation blocks. The self-attention mechanism (Transformer) computes the relationship between inner time-series points, while the cross-correlation finds trending periodicity patterns. The multi-headed encoder is channeled through a mixing gate (linear combination) of sub-encoders (Transformer and Wavelet Crossformer) that output trending signatures to the decoder. This process improved the model's predictive capabilities, reducing Mean Absolute Error by 31.26 % compared to the second-best performing transformer-like models evaluated. We have also used the Multifractal Detrended Cross-Correlation Heatmaps (MF-DCCHM) to extract cyclical trends from pairs of stations across multifractal regimes by denoising the pair of signals with Daubechies wavelets. Our dataset was obtained from a network of eight wells for groundwater monitoring in Brazilian aquifers, six rainfall stations, eleven river flow stations, and three weather stations with atmospheric pressure, temperature, and humidity sensors.

A catchment scale assessment of water balance components: a case study of Chittar catchment in South India

The detailed analyses of the water balance components (WBCs) of the catchment help assess the available water resources, especially in the arid climate regions for their sustainable management and development. This paper mainly used the Soil and Water Assessment Tool (SWAT) model to analyze the variation in the WBCs considering the change in the Land Use and Land Cover (LULC) and meteorological variables. For this purpose, the model used the inputs of LULC and meteorological variables between 2001 and 2020 at 5 years and daily time intervals, respectively, from the Chittar river catchment. The developed models were calibrated using SWAT-CUP split-up procedure (pre-calibration and post-calibration). The model was found to be good in calibration and validation, yielding the coefficient of determination (R²) of 0.94 and 0.81, respectively. Furthermore, WBCs of the catchment were estimated for the near future (2021-2030) at the monthly and annual scales. For this endeavor, LULC was forecasted for the years 2021 and 2026 using Cellular Automata (CA)-Artificial Neural Network (ANN), and for the same period, meteorological variables were also forecasted using the smoothing moving average method from the historical data.

Long-Term Groundwater Level Prediction Model Based on Hybrid KNN-RF Technique

Reliable seasonal prediction of groundwater levels is not always possible when the quality and the amount of available on-site groundwater data are limited. In the present work, a hybrid K-Nearest Neighbor-Random Forest (KNN-RF) is used for the prediction of variations in groundwater levels (L) of an aquifer with the groundwater relatively close to the surface (<10 m) is proposed. First, the time-series smoothing methods are applied to improve the quality of groundwater data. Then, the ensemble K-Nearest Neighbor-Random Forest (KNN-RF) model is treated using hydro-climatic data for the prediction of variations in the levels of the groundwater tables up to three months ahead. Climatic and groundwater data collected from eastern Rwanda were used for validation of the model on a rolling window basis. Potential predictors were: the observed daily mean temperature (T), precipitation (P), and daily maximum solar radiation (S). Previous day’s precipitation P (t − 1), solar radiation S (t), temperature T (t), and groundwater level L (t) showed the highest variation in the fluctuations of the groundwater tables. The KNN-RF model presents its results in an intelligible manner. Experimental results have confirmed the high performance of the proposed model in terms of root mean square error (RMSE), mean absolute error (MAE), Nash–Sutcliffe (NSE), and coefficient of determination (R2).

Ensemble modelling framework for groundwater level prediction in urban areas of India

India is facing the worst water crisis in its history and major Indian cities which accommodate about 50% of its population will be among highly groundwater stressed cities by 2020. In past few decades, the urban groundwater resources declined significantly due to over exploitation, urbanization, population growth and climate change. To understand the role of these variables on groundwater level fluctuation, we developed a machine learning based modelling approach considering singular spectrum analysis (SSA), mutual information theory (MI), genetic algorithm (GA), artificial neural network (ANN) and support vector machine (SVM). The developed approach was used to predict the groundwater levels in Bengaluru, a densely populated city with declining groundwater water resources. The input data which consist of groundwater levels, rainfall, temperature, NOI, SOI, NIÑO3 and monthly population growth rate were pre-processed using mutual information theory, genetic algorithm and lag analysis. Later, the optimized input sets were used in ANN and SVM to predict monthly groundwater level fluctuations. The results suggest that the machine learning based approach with data pre-processing predict groundwater levels accurately (R > 85%). It is also evident from the results that the pre-processing techniques enhance the prediction accuracy and results were improved for 66% of the monitored wells. Analysis of various input parameters suggest, inclusion of population growth rate is positively correlated with decrease in groundwater levels. The developed approach in this study for urban groundwater prediction can be useful particularly in cities where lack of pipeline/sewage/drainage lines leakage data hinders physical based modelling.

State-of-art of genetic programming applications in water-resources systems analysis

Evolutionary algorithms (EAs) have become competitive solvers of a wide variety of water-resources optimization problems. Genetic programming (GP) has become a leading EA since its inception in 1985. This paper reviews the state-of-the-art of GP and its applications in water-resources systems analysis. A comprehensive knowledge about GP's theory and modeling approach is essential for its successful application in water-resources systems analysis. This review presents variants of GP that have been proven useful in various applications to water resources problems. Several examples of applications of GP in water-resources systems analysis are herein presented. This review reveals GP's capability and superiority compared to other conventional methods, which makes it suitable for solving a wide variety of water-related problems including rainfall-runoff modeling, streamflow sediment prediction, flood prediction and routing, evaporation and evapotranspiration forecasting, reservoir operation, groundwater modeling, water quality modeling, water demand forecasting, and water distribution systems.

Development of genetic programming-based model for predicting oyster norovirus outbreak risks

Oyster norovirus outbreaks pose increasing risks to human health and seafood industry worldwide but exact causes of the outbreaks are rarely identified, making it highly unlikely to reduce the risks. This paper presents a genetic programming (GP) based approach to identifying the primary cause of oyster norovirus outbreaks and predicting oyster norovirus outbreaks in order to reduce the risks. In terms of the primary cause, it was found that oyster norovirus outbreaks were controlled by cumulative effects of antecedent environmental conditions characterized by low solar radiation, low water temperature, low gage height (the height of water above a gage datum), low salinity, heavy rainfall, and strong offshore wind. The six environmental variables were determined by using Random Forest (RF) and Binary Logistic Regression (BLR) methods within the framework of the GP approach. In terms of predicting norovirus outbreaks, a risk-based GP model was developed using the six environmental variables and various combinations of the variables with different time lags. The results of local and global sensitivity analyses showed that gage height, temperature, and solar radiation were by far the three most important environmental predictors for oyster norovirus outbreaks, though other variables were also important. Specifically, very low temperature and gage height significantly increased the risk of norovirus outbreaks while high solar radiation markedly reduced the risk, suggesting that low temperature and gage height were associated with the norovirus source while solar radiation was the primary sink of norovirus. The GP model was utilized to hindcast daily risks of oyster norovirus outbreaks along the Northern Gulf of Mexico coast. The daily hindcasting results indicated that the GP model was capable of hindcasting all historical oyster norovirus outbreaks from January 2002 to June 2014 in the Gulf of Mexico with only two false positive outbreaks for the 12.5-year period. The performance of the GP model was characterized with the area under the Receiver Operating Characteristic curve of 0.86, the true positive rate (sensitivity) of 78.53% and the true negative rate (specificity) of 88.82%, respectively, demonstrating the efficacy of the GP model. The findings and results offered new insights into the oyster norovirus outbreaks in terms of source, sink, cause, and predictors. The GP model provided an efficient and effective tool for predicting potential oyster norovirus outbreaks and implementing management interventions to prevent or at least reduce norovirus risks to both the human health and the seafood industry.