Real-time measurement of total nitrogen for agricultural runoff based on multiparameter sensors and intelligent algorithms

Universal high-frequency monitoring methods of river water quality in China based on machine learning

The in-situ high-frequency monitoring of total nitrogen (TN) and total phosphorus (TP) in rivers is a challenge and key to instant water quality judgment and early warning. Based on the physical and chemical association between TN/TP and sensor-measurable predictors, we proposed a novel "indirect" measurement method for TN and TP in rivers. This method combines the timeliness of multi-sensor and the accuracy of intelligent algorithms, utilizing 188,629 data sets from 131 water monitoring stations across China. Under 5 algorithms and 4 predictor group scenarios, the results showed that: (1) extra tree regression (ETR) with 6 predictors exhibited the best precision, and the mean determination coefficient (R2) of TN and TP inversion across 131 stations reached 0.78 ± 0.25 and 0.79 ± 0.22 respectively; (2) among 6 potential predictors, the importance degrees of temperature, electrical conductivity, NH4-N, and turbidity were greater than that of pH and DO, and >80 % of stations exhibited acceptable prediction accuracy (R2 > 0.6) when the number of predictors (P) ranged from 4 to 6, which showed good tolerability to predictor variations; (3) the accurate classification rates of water quality standard (ACRws) of all stations based on TN and TP reached 90.41 ± 6.96 % and 92.33 ± 6.41 %; (4) in 9 regions/basins of China, this method showed universal application potential with no significant prediction difference. Compared with laboratory test, water quality automatic monitoring station, and remote sensing inversion, the proposed method offers high-frequency, high-precision, regional adaptability, low cost, and stable operation under rainy, cloudy, and nighttime conditions. The new method may provide important technological support for timely pollutant tracing, pre-warning, and emergency control for river pollution.

Water quality prediction based on sparse dataset using enhanced machine learning

Water quality in surface bodies remains a pressing issue worldwide. While some regions have rich water quality data, less attention is given to areas that lack sufficient data. Therefore, it is crucial to explore novel ways of managing source-oriented surface water pollution in scenarios with infrequent data collection such as weekly or monthly. Here we showed sparse-dataset-based prediction of water pollution using machine learning. We investigated the efficacy of a traditional Recurrent Neural Network alongside three Long Short-Term Memory (LSTM) models, integrated with the Load Estimator (LOADEST). The research was conducted at a river-lake confluence, an area with intricate hydrological patterns. We found that the Self-Attentive LSTM (SA-LSTM) model outperformed the other three machine learning models in predicting water quality, achieving Nash-Sutcliffe Efficiency (NSE) scores of 0.71 for CODMn and 0.57 for NH3N when utilizing LOADEST-augmented water quality data (referred to as the SA-LSTM-LOADEST model). The SA-LSTM-LOADEST model improved upon the standalone SA-LSTM model by reducing the Root Mean Square Error (RMSE) by 24.6% for CODMn and 21.3% for NH3N. Furthermore, the model maintained its predictive accuracy when data collection intervals were extended from weekly to monthly. Additionally, the SA-LSTM-LOADEST model demonstrated the capability to forecast pollution loads up to ten days in advance. This study shows promise for improving water quality modeling in regions with limited monitoring capabilities.

Identification of pollution source and prediction of water quality based on deep learning techniques

Semi-arid rivers are particularly vulnerable and responsive to the impacts of industrial contamination. Prompt identification and projection of pollutant dynamics are crucial in the accidental pollution incidents, therefore required the timely informed and effective management strategies. In this study, we collected water quality monitoring data from a typical semi-arid river. By water quality inter-correlation mapping, we identified the regularity and abnormal fluctuations of pollutant discharges. Combining the association rule method (Apriori) and characterized pollutants of different industries, we tracked major industrial pollution sources in the Dahei River Basin. Meanwhile, we deployed the integrated multivariate long and short-term memory network (LSTM) to forecast principal contaminants. Our findings revealed that (1) biological oxygen demand (BOD), chemical oxygen demand (COD), total nitrogen, total phosphorus, and ammonia nitrogen exhibited high inter-correlations in water quality mapping, with lead and cadmium also demonstrating a strong association; (2) The main point sources of contaminant were coking, metal mining, and smelting industries. The government should strengthen the regulation and control of these industries and prevent further pollution of the river; (3) We confirmed 4 key pollutants: COD, ammonia nitrogen, total nitrogen, and total phosphorus. Our study accurately predicted the future changes in this water quality index. The best results were obtained when the prediction period was 1 day. The prediction accuracies reached 85.85%, 47.15%, 85.66%, and 89.07%, respectively. In essence, this research developed effective water quality traceability and predictive analysis methods in semi-arid river basins. It provided an effective tool for water quality surveillance in semi-arid river basins and imparts a scientific scaffold for the environmental stewardship endeavors of pertinent authorities.

Controls for multi-temporal patterns of riverine nitrogen and phosphorus export to lake: Implications for catchment management by high-frequency observations

Intensifying human activity coupled with climate change increase the transport of excess riverine nitrogen (N) and phosphorus (P) loading from catchment to lake, leading to eutrophication and harmful algal blooms worldwide. To improve understanding of multi-temporal patterns of riverine N and P export and their hydro-biogeochemical controls over both episodic events and long-term trend, we analyzed and interpreted high-frequency data of total nitrogen (TN), ammonia-nitrogen (NH₄-N), and total phosphorus (TP) provided by an automatic water quality monitoring station in a typical agricultural catchment draining to Lake Chaohu, China. Mann-Kendall test revealed a significant decreasing trend of riverine N and P concentration most of the time during 2018-2020. At the sub-daily scale, intraday TN concentrations varied by more than 1 mg/L in 31.8% of the period. Monthly TN and TP concentrations were particularly high in December 2019, indicating combined effect of hydrologic (long dry antecedent period and subsequent intensive rainfall events) and anthropogenic controls (fertilization and agricultural drainage). Significantly higher TN concentrations in winter and TP concentrations in summer reflected coupled dominances of precipitation and temperature on hydrologic and biogeochemical processes. Rainfall events with very heavy intensity drove disproportionate N and P loads (more than 20% of the total export) in only 3.2% of the period. Moderate and very heavy events registered the highest TN and TP concentrations, respectively. Our results highlighted the importance of automatic water quality monitoring station to reveal dynamics of riverine N and P export, which may imply future nutrient loading abatement plans for lake-connected catchment.

Response Model for Urban Area Source Pollution and Water Environmental Quality in a River Network Region

With the development of cities, urban area source pollution has become more severe and a significant source of water pollution. To study the relationship between urban area source pollution and water environmental quality in a river network, this study uses a city in the Yangtze River Delta, China, as an example. The Storm Water Management Model (SWMM) model and the MIKE11 model were combined into a unified modeling framework and used to simulate dynamic changes in the water quality of a river network under light rain, moderate rain, and heavy rain. In the study period, the annual urban area source input loads of potassium permanganate (COD_Mn), total phosphorus (TP), and ammonia nitrogen were 29.8, 0.9, and 4.8 t, respectively. The influence of light rain on the water quality of the river network was lagging and temporary, and rainfall area pollution was the primary contributor. Under the scenario of moderate rain, overflow from a pipeline network compounded rainfall runoff, resulting in a longer duration of impact on the water quality in the river. Additionally, the water quality in the river course was worse under moderate rain than under light or heavy rain. Under the scenario of heavy rain, rain mainly served a dilutive function. This research can provide support for urban area source pollution control and management.