Application of neural network and time series modeling to study the suitability of drain water quality for irrigation: a case study from Egypt

Prediction of mine water quality by the Seq2Seq model based on attention mechanism

In recent years, as China's industrialization level has advanced, the issue of environmental pollution, particularly mine water pollution, has become increasingly severe. Water quality prediction is a fundamental aspect of water resource protection and a critical approach to addressing the water resource crisis. For improvement in water quality prediction, this research first analyzes the characteristics of mine water quality changes and provides a brief overview of water quality prediction. Subsequently, the Long Short-Term Memory and Sequence to Sequence (Seq2Seq) models, derived from Artificial Neural Networks, are introduced. The Seq2Seq water quality prediction model is implemented, incorporating the attention mechanism. Experimental validation confirms the effectiveness of the proposed model. The results demonstrate that the attention mechanism-based Seq2Seq model accurately predicts parameters such as pH value, Dissolved Oxygen, ammonia nitrogen, and Chemical Oxygen Demand, exhibiting a high degree of consistency with actual results. They play a vital role in assessing the health of the water and its ability to support aquatic life. The change of these indicators can reflect the degree and type of water pollution. Moreover, the Seq2Seq + attention model stands out with the lowest predicted Root Mean Square Error of 0.309. Notably, in comparison to the traditional Seq2Seq model, the incorporation of attention mechanisms in the Seq2Seq model results in a substantial 2.94 reduction in Mean Absolute Error. This research on the Seq2Seq water quality prediction model with attention mechanism provides valuable insights and references for future endeavors in water quality prediction.

Identification of influential weather parameters and seasonal drought prediction in Bangladesh using machine learning algorithm

Droughts pose a severe environmental risk in countries that rely heavily on agriculture, resulting in heightened levels of concern regarding food security and livelihood enhancement. Bangladesh is highly susceptible to environmental hazards, with droughts further exacerbating the precarious situation for its 170 million inhabitants. Therefore, we are endeavouring to highlight the identification of the relative importance of climatic attributes and the estimation of the seasonal intensity and frequency of droughts in Bangladesh. With a period of forty years (1981-2020) of weather data, sophisticated machine learning (ML) methods were employed to classify 35 agroclimatic regions into dry or wet conditions using nine weather parameters, as determined by the Standardized Precipitation Evapotranspiration Index (SPEI). Out of 24 ML algorithms, the four best ML methods, ranger, bagEarth, support vector machine, and random forest (RF) have been identified for the prediction of multi-scale drought indices. The RF classifier and the Boruta algorithms shows that water balance, precipitation, maximum and minimum temperature have a higher influence on drought intensity and occurrence across Bangladesh. The trend of spatio-temporal analysis indicates, drought intensity has decreased over time, but return time has increased. There was significant variation in changing the spatial nature of drought intensity. Spatially, the drought intensity shifted from the northern to central and southern zones of Bangladesh, which had an adverse impact on crop production and the livelihood of rural and urban households. So, this precise study has important implications for the understanding of drought prediction and how to best mitigate its impacts. Additionally, the study emphasizes the need for better collaboration between relevant stakeholders, such as policymakers, researchers, communities, and local actors, to develop effective adaptation strategies and increase monitoring of weather conditions for the meticulous management of droughts in Bangladesh.

Prediction of water distribution uniformity of sprinkler irrigation system based on machine learning algorithms

The coefficients of uniformity Christiansen's uniformity coefficient (CU) and distribution uniformity (DU) are an important parameter for designing irrigation systems, and are an accurate measure for water lose. In this study, three machine learning algorithms Random forest (RF), extreme gradient boosting (XGB) and random forest-extreme gradient boosting (XGB-RF) were developed to predict the water distribution uniformity based on operating pressure, heights of sprinkler, discharge, nozzle diameter, wind speed, humidity, highest and lowest temperature for three different impact sprinklers (KA-4, FOX and 2520) for square and triangular system layout based on four scenarios (input combinations). The main findings were; the highest CU value was 86.7% in the square system of 2520 sprinkler under 200 kPa, 0.5 m height and 0.855 m3/h (Nozzle 2.5 mm). Meanwhile, in the triangular system, it was 87.3% under the same pressure and discharge and 1 m height. For applied machine learning, the highest values of R2 were 0.796, 0.825 and 0.929 in RF, XGB and XGB-RF respectively in the first scenario for CU. Moreover, for the DU, the highest values of R2 were 0.701, 0.479 and 0.826 in RF, XGB and XGB-RF respectively in the first scenario. The obtained results revealed that the sprinkler height had the lowest impact on modeling of the water distribution uniformity.

Prediction of irrigation water suitability using geospatial computing approach: a case study of Agartala city, India

An increase in population expansion, urban sprawling environment, and climate change has resulted in increased food demand, water scarcity, environmental pollution, and mismanagement of water resources. Groundwater, i.e., one of the most precious and mined natural resources is used to address a variety of environmental demands. Among all, irrigation is one of the leading consumers of groundwater. Various natural heterogeneities and anthropogenic activities have impacted the groundwater quality. As a result, monitoring groundwater quality and determining its suitability are critical for the sustainable long-term management of groundwater resources. In this study, groundwater samples from 35 different sampling stations were collected and tested for various parameters associated with irrigation water quality. Hybrid MCDM (fuzzy-AHP) method was used to determine the groundwater suitability for irrigation purposes. The suitability map obtained using spatial overlay analysis was classified into low, moderate, and high irrigation water suitability zones. Along with suitability analysis, various regression-based machine learning models such as multiple linear regression (MLR), random forest (RF), and artificial neural network (ANN) were used and compared to predict irrigation water suitability. Results depicted that the ANN model with the highest R2 value of 0.990 and RMSE value near to zero (0) has outperformed all other models. The present methodology could be found useful to predict irrigation water suitability in the region where regular sampling and analysis are quite challenging.

A Methodology Based on Machine Learning and Soft Computing to Design More Sustainable Agriculture Systems

Advances in new technologies are allowing any field of real life to benefit from using these ones. Among of them, we can highlight the IoT ecosystem making available large amounts of information, cloud computing allowing large computational capacities, and Machine Learning techniques together with the Soft Computing framework to incorporate intelligence. They constitute a powerful set of tools that allow us to define Decision Support Systems that improve decisions in a wide range of real-life problems. In this paper, we focus on the agricultural sector and the issue of sustainability. We propose a methodology that, starting from times series data provided by the IoT ecosystem, a preprocessing and modelling of the data based on machine learning techniques is carried out within the framework of Soft Computing. The obtained model will be able to carry out inferences in a given prediction horizon that allow the development of Decision Support Systems that can help the farmer. By way of illustration, the proposed methodology is applied to the specific problem of early frost prediction. With some specific scenarios validated by expert farmers in an agricultural cooperative, the benefits of the methodology are illustrated. The evaluation and validation show the effectiveness of the proposal.

Comparing linear and non-linear data-driven approaches in monthly river flow prediction, based on the models SARIMA, LSSVM, ANFIS, and GMDH

River flow variations directly affect the hydro-climatological, environmental, and ecological characteristics of a region. Therefore, an accurate prediction of river flow can critically be important for water managers and planners. The present study aims to compare different data-driven models in predicting monthly flow. Two river catchments located in the Guilan province in Iran, where rivers play an essential role in agricultural productions (mainly rice), are studied. The monthly river flow dataset was provided by Guilan Regional Water Authority during 1986-2015. The models are derived from two different numerical types of stochastic and machine learning (ML) models. The stochastic model is seasonal autoregressive integrated moving average (SARIMA), and the MLs are least square support vector machine (LSSVM), adaptive neuro-fuzzy inference system (ANFIS), and group method of data handling (GMDH). The inputs were selected by autocorrelation and partial autocorrelation functions (ACF and PACF) from the flow rates of the previous months. The data was divided into 75% of training and 25% of testing phases, and then the mentioned models were implemented. Predictions were evaluated by the criteria of root mean square error (RMSE), normalized RMSE (NRMSE), and Nash Sutcliff (NS) coefficient. According to the calculated values of different criteria during the test phase, RMSE = 1.138 cms, NRMSE = 0.109, and NS = 0.826, it can be concluded that the SARIMA model was superior to its ML competitors. Among the ML models, GMDH had the best performance (by RMSE = 1.290 cms, NRMSE = 0.124, and NS = 0.777) because it has more optimization parameters and sample space for network make-up. The models were also evaluated in hydrological drought conditions of both rivers. It was resulted that the rivers' flow can be well predicted in drought conditions by using these models, especially the SARIMA stochastic model. According to the NRMSE values (ranged between 0.1 and 0.2), the accuracy of predictions is evaluated in the appropriate range, and the present study shows promising results of the current approaches. Consequently, a comparison between the performance of linear stochastic models and complex black-box MLs, reveals that linear stochastic models are more suitable for the current region's monthly river flow prediction.

A Hybrid Model for Water Quality Prediction Based on an Artificial Neural Network, Wavelet Transform, and Long Short-Term Memory

Clean water is an indispensable essential resource on which humans and other living beings depend. Therefore, the establishment of a water quality prediction model to predict future water quality conditions has a significant social and economic value. In this study, a model based on an artificial neural network (ANN), discrete wavelet transform (DWT), and long short-term memory (LSTM) was constructed to predict the water quality of the Jinjiang River. Firstly, a multi-layer perceptron neural network was used to process the missing values based on the time series in the water quality dataset used in this research. Secondly, the Daubechies 5 (Db5) wavelet was used to divide the water quality data into low-frequency signals and high-frequency signals. Then, the signals were used as the input of LSTM, and LSTM was used for training, testing, and prediction. Finally, the prediction results were compared with the nonlinear auto regression (NAR) neural network model, the ANN-LSTM model, the ARIMA model, multi-layer perceptron neural networks, the LSTM model, and the CNN-LSTM model. The outcome indicated that the ANN-WT-LSTM model proposed in this study performed better than previous models in many evaluation indices. Therefore, the research methods of this study can provide technical support and practical reference for water quality monitoring and the management of the Jinjiang River and other basins.

Using Machine Learning Models to Predict Hydroponically Grown Lettuce Yield

Prediction of crop yield is an essential task for maximizing the global food supply, particularly in developing countries. This study investigated lettuce yield (fresh weight) prediction using four machine learning (ML) models, namely, support vector regressor (SVR), extreme gradient boosting (XGB), random forest (RF), and deep neural network (DNN). It was cultivated in three hydroponics systems (i.e., suspended nutrient film technique system, pyramidal aeroponic system, and tower aeroponic system), which interacted with three different magnetic unit strengths under a controlled greenhouse environment during the growing season in 2018 and 2019. Three scenarios consisting of the combinations of input variables (i.e., leaf number, water consumption, dry weight, stem length, and stem diameter) were assessed. The XGB model with scenario 3 (all input variables) yielded the lowest root mean square error (RMSE) of 8.88 g followed by SVR with the same scenario that achieved 9.55 g, and the highest result was by RF with scenario 1 (i.e., leaf number and water consumption) that achieved 12.89 g. All model scenarios having Scatter Index (SI) (i.e., RMSE divided by the average values of the observed yield) values less than 0.1 were classified as excellent in predicting fresh lettuce yield. Based on all of the performance statistics, the two best models were SVR with scenario 3 and DNN with scenario 2 (i.e., leaf number, water consumption, and dry weight). However, DNN with scenario 2 requiring less input variables is preferred. The potential of the DNN model to predict fresh lettuce yield is promising, and it can be applied on a large scale as a rapid tool for decision-makers to manage crop yield.