Robust clustering-based hybrid technique enabling reliable reservoir water quality prediction with uncertainty quantification and spatial analysis.

Machine learning methodology has recently been considered a smart and reliable way to monitor water quality parameters in aquatic environments like reservoirs and lakes. This study employs both individual and hybrid-based techniques to boost the accuracy of dissolved oxygen (DO) and chlorophyll-a (Chl-a) predictions in the Wadi Dayqah Dam located in Oman. At first, an AAQ-RINKO device (CTD+ sensor) was used to collect water quality parameters from different locations and varying depths in the reservoir. Second, the dataset is segmented into homogeneous clusters based on DO and Chl-a parameters by leveraging an optimized K-means algorithm, facilitating precise estimations. Third, ten sophisticated variational-individual data-driven models, namely generalized regression neural network (GRNN), random forest (RF), gaussian process regression (GPR), decision tree (DT), least-squares boosting (LSB), bayesian ridge (BR), support vector regression (SVR), K-nearest neighbors (KNN), multilayer perceptron (MLP), and group method of data handling (GMDH) are employed to estimate DO and Chl-a concentrations. Fourth, to improve prediction accuracy, bayesian model averaging (BMA), entropy weighted (EW), and a new enhanced clustering-based hybrid technique called Entropy-ORNESS are employed to combine model outputs. The Entropy-ORNESS method incorporates a Genetic Algorithm (GA) to determine optimal weights and then combine them with EW weights. Finally, the inclusion of bootstrapping techniques introduces a stochastic assessment of model uncertainty, resulting in a robust estimator model. In the validation phase, the Entropy-ORNESS technique outperforms the independent models among the three fusion-based methods, yielding R2 values of 0.92 and 0.89 for DO and Chl-a clusters, respectively. The proposed hybrid-based methodology demonstrates reduced uncertainty compared to single data-driven models and two combination frameworks, with uncertainty levels of 0.24% and 1.16% for cluster 1 of DO and cluster 2 of Chl-a parameters. As a highlight point, the spatial analysis of DO and Chl-a concentrations exhibit similar pattern variations across varying depths of the dam according to the comparison of field measurements with the best hybrid technique, in which DO concentration values notably decrease during warmer seasons. These findings collectively underscore the potential of the upgraded weighted-based hybrid approach to provide more accurate estimations of DO and Chl-a concentrations in dynamic aquatic environments.

Soil nutrients and plant uptake parameters as related to greenhouse gas emissions.

Fertilizer and water management practices have short- and long-term effects on soil chemical and physical properties and, in turn, greenhouse gas (GHG) emissions. The goal of this 4-yr field study was to establish the relationships between soil properties, agronomic practices, and GHG (CO2 and N2 O) emissions under different fertilizer and water table management practices. There were two fertilizer treatments: inorganic fertilizer (IF) and a mix of solid cattle manure and inorganic fertilizer (SCM), combined with tile drainage(DR) and controlled drainage with subirrigation(CDS). The cropping system was a maize (Zea mays L.)-soybean [Glycine max (L.) Merr.] rotation. Nitrogen in biomass (BMN) and N in grain (GRN) were measured and used to calculate other plant N parameters. Nitrous oxide and CO2 fluxes were collected weekly, and their respective cumulative emissions were calculated. The results show that soil organic matter (OM), soil total carbon (C), and soil total nitrogen (N) were greater in SCM than IF by 23.7, 35.2, and 24.4%, respectively. Water table management did not significantly affect soil N and C. Increased CO2 emissions were witnessed under higher soil OM, soil total C, and total N. Plant N uptake parameters were negatively correlated with N2 O and CO2 emissions. Higher plant N uptake can reduce environmental pollution by limiting N2 O and CO2 emissions.

Machine learning for predicting greenhouse gas emissions from agricultural soils.

Machine learning (ML) models are increasingly used to study complex environmental phenomena with high variability in time and space. In this study, the potential of exploiting three categories of ML regression models, including classical regression, shallow learning and deep learning for predicting soil greenhouse gas (GHG) emissions from an agricultural field was explored. Carbon dioxide (CO2) and nitrous oxide (N2O) fluxes, as well as various environmental, agronomic and soil data were measured at the site over a five-year period in Quebec, Canada. The rigorous analysis, which included statistical comparison and cross-validation for the prediction of CO2 and N2O fluxes, confirmed that the LSTM model performed the best among the considered ML models with the highest R coefficient and the lowest root mean squared error (RMSE) values (R = 0.87 and RMSE = 30.3 mg·m-2·hr-1 for CO2 flux prediction and R = 0.86 and RMSE = 0.19 mg·m-2·hr-1 for N2O flux prediction). The predictive performances of LSTM were more accurate than those simulated in a previous study conducted by a biophysical-based Root Zone Water Quality Model (RZWQM2). The classical regression models (namely RF, SVM and LASSO) satisfactorily simulated cyclical and seasonal variations of CO2 fluxes (R = 0.75, 0.71 and 0.68, respectively); however, they failed to reasonably predict the peak values of N2O fluxes (R < 0.25). Shallow ML was found to be less effective in predicting GHG fluxes than other considered ML models (R < 0.7 for CO2 flux and R < 0.3 for estimating N2O fluxes) and was the most sensitive to hyperparameter tuning. Based on this comprehensive comparison study, it was elicited that the LSTM model can be employed successfully in simulating GHG emissions from agricultural soils, providing a new perspective on the application of machine learning modeling for predicting GHG emissions to the environment.

Assessing climate change impacts on greenhouse gas emissions, N losses in drainage and crop production in a subsurface drained field.

Future climate change-driven alterations in precipitation patterns, increases in temperature, and rises in atmospheric carbon dioxide concentration ([CO2]atm) are expected to alter agricultural productivity and environmental quality, while high latitude countries like Canada are likely to face more challenges from global climate change. However, potential climate change impact on GHG emissions from tile-drained fields is poorly documented. Accordingly, climate change impacts on GHG emissions, N losses to drainage and crop production in a subsurface-drained field in Southern Quebec, Canada were assessed using calibrated and validated RZWQM2 model. The RZWQM2 model was run for a historical period (1971-2000) and for a future period (2038 to 2070) using data generated from 11 different GCM-RCMs (global climate models coupled with regional climate models). Under the projected warmer and higher rainfall conditions mean drainage flow was predicted to increase by 17%, and the N losses through subsurface drains increase by 47%. Despite the negative effect of warming temperature on crop yield, soybean yield was predicted to increase by 31% due to increased photosynthesis rates and improved crop water use efficiency (WUE) under elevated [CO2]atm, while corn yield was reduced by 7% even with elevated [CO2]atm because of a shorter life cycle from seedling to maturity resulted from higher temperature. The N2O emissions would be enhanced by 21% due to greater denitrification and mineralization, while CO2 emissions would increase by 16% because of more crop biomass accumulation, higher crop residue decomposition, and greater soil microbial activities. Soil organic carbon storage was predicted to decrease 22% faster in the future, which would result in higher global warming potential in turn. This study demonstrates the potential of exacerbating GHG emissions and water quality problems and reduced corn yield under climate change impact in subsurface drained fields in southern Quebec.

Water table management and fertilizer application impacts on CO2, N2O and CH4 fluxes in a corn agro-ecosystem.

Water table management with controlled drainage and subsurface-irrigation (SI) has been identified as a Beneficial Management Practice (BMP) to reduce nitrate leaching in drainage water. It has also been shown to increase crop yields during dry periods of the growing season, by providing water to the crop root zone, via upward flux or capillary rise. However, by retaining nitrates in anoxic conditions within the soil profile, SI could potentially increase greenhouse gas (GHG) fluxes, particularly N2O through denitrification. This process may be further exacerbated by high precipitation and mineral N-fertilizer applications very early in the growing season. In order to investigate the effects of water table management (WTM) with nitrogen fertilization on GHG fluxes from corn (Zea mays) agro-ecosystems, we conducted a research study on a commercial farm in south-western Quebec, Canada. Water table management treatments were: free drainage (FD) and controlled drainage with subsurface-irrigation. GHG samples were taken using field-deployed, vented non-steady state gas chambers to quantify soil CO2, N2O and CH4 fluxes weekly. Our results indicate that fertilizer application timing coinciding with intense (≥24 mm) precipitation events and high temperatures (>25 °C) triggered pulses of N2O fluxes, accounting for up to 60% of cumulative N2O fluxes. Our results also suggest that splitting bulk fertilizer applications may be an effective mitigation strategy, reducing N2O fluxes by 50% in our study. In both seasons, pulse GHG fluxes mostly occurred in the early vegetative stages of the corn, prior to activation of the subsurface-irrigation. Our results suggest that proper timing of WTM mindful of seasonal climatic conditions has the potential to reduce GHG emissions.

Mitigating greenhouse gas emissions in subsurface-drained field using RZWQM2.

Greenhouse gas (GHG) emissions from agricultural soils are affected by various environmental factors and agronomic practices. The impact of inorganic nitrogen (N) fertilization rates and timing, and water table management practices on N2O and CO2 emissions were investigated to propose mitigation and adaptation efforts based on simulated results founded on field data. Drawing on 2012-2015 data measured on a subsurface-drained corn (Zea mays L.) field in Southern Quebec, the Root Zone Water Quality Model 2 (RZWQM2) was calibrated and validated for the estimation of N2O and CO2 emissions under free drainage (FD) and controlled drainage with sub-irrigation (CD-SI). Long term simulation from 1971 to 2000 suggested that the optimal N fertilization should be in the range of 125 to 175 kg N ha-1 to obtain higher NUE (nitrogen use efficiency, 7-14%) and lower N2O emission (8-22%), compared to 200 kg N ha-1 for corn-soybean rotation (CS). While remaining crop yields, splitting N application would potentially decrease total N2O emissions by 11.0%. Due to higher soil moisture and lower soil O2 under CD-SI, CO2 emissions declined by 6% while N2O emissions increased by 21% compared to FD. The CS system reduced CO2 and N2O emissions by 18.8% and 20.7%, respectively, when compared with continuous corn production. This study concludes that RZWQM2 model is capable of predicting GHG emissions, and GHG emissions from agriculture can be mitigated using agronomic management.

Comparison of cyanobacterial microcystin synthetase (mcy) E gene transcript levels, mcy E gene copies, and biomass as indicators of microcystin risk under laboratory and field conditions.

Increased incidences of mixed assemblages of microcystin-producing and nonproducing cyanobacterial strains in freshwater bodies necessitate development of reliable proxies for cyanotoxin risk assessment. Detection of microcystin biosynthetic genes in water blooms of cyanobacteria is generally indicative of the presence of potentially toxic cyanobacterial strains. Although much effort has been devoted toward elucidating the microcystin biosynthesis mechanisms in many cyanobacteria genera, little is known about the impacts of co-occurring cyanobacteria on cellular growth, mcy gene expression, or mcy gene copy distribution. The present study utilized conventional microscopy, qPCR assays, and enzyme-linked immunosorbent assay to study how competition between microcystin-producing Microcystis aeruginosa CPCC 299 and Planktothrix agardhii NIVA-CYA 126 impacts mcyE gene expression, mcyE gene copies, and microcystin concentration under controlled laboratory conditions. Furthermore, analyses of environmental water samples from the Missisquoi Bay, Quebec, enabled us to determine how the various potential toxigenic cyanobacterial biomass proxies correlated with cellular microcystin concentrations in a freshwater lake. Results from our laboratory study indicated significant downregulation of mcyE gene expression in mixed cultures of M. aeruginosa plus P. agardhii on most sampling days in agreement with depressed growth recorded in the mixed cultures, suggesting that interaction between the two species probably resulted in suppressed growth and mcyE gene expression in the mixed cultures. Furthermore, although mcyE gene copies and McyE transcripts were detected in all laboratory and field samples with measureable microcystin levels, only mcyE gene copies showed significant positive correlations (R(2) > 0.7) with microcystin concentrations, while McyE transcript levels did not. These results suggest that mcyE gene copies are better indicators of potential risks from microcystins than McyE transcript levels or conventional biomass proxies, especially in water bodies comprising mixed assemblages of toxic and nontoxic cyanobacteria.

Modelling phosphate adsorption to the soil: application of the non-ideal competitive adsorption model.

Phosphorus (P) transport in subsurface runoff has increased despite the limited mobility of P in soils. This study investigated the ability of the non-ideal competitive adsorption (NICA) model to describe phosphate (PO(4)) adsorption for soils in southern Quebec (Canada). We measured the surface charge and PO(4) adsorption capacity for 11 agricultural soils. Using the experimental data and a nonlinear fitting function, we derived the NICA model parameters. We found that the NICA model described accurately the surface charge of these soils with a mean R(2)>0.99, and described the adsorption data with a mean R(2)=0.96. We also found that the variable surface charge was distributed over the two binding sites with the low pH sites demonstrating a stronger binding energy for hydroxyl and PO(4) ions. We established that the NICA model is able to describe P adsorption for the soils considered in this study.