Predicting high‐frequency variation in stream solute concentrations with water quality sensors and machine learning

Abstract

Stream solute monitoring has produced many insights into ecosystem and Earth system functions. Although new sensors have provided novel information about the fine-scale temporal variation of some stream water solutes, we lack adequate sensor technology to gain the same insights for many other solutes. We used two machine learning algorithms-Support Vector Machine and Random Forestto predict concentrations at 15-min resolution for 10 solutes, of which eight lack specific sensors. The algorithms were trained with data from intensive stream sensing and manual stream sampling (weekly) for four full years in a hydrologic reference stream within the Hubbard Brook Experimental Forest in New Hampshire, USA. The Random Forest algorithm was slightly better at predicting solute concentrations than the Support Vector Machine algorithm (Nash-Sutcliffe efficiencies ranged from 0.35 to 0.78 for Random Forest compared to 0.29 to 0.79 for Support Vector Machine). Solute predictions were most sensitive to the removal of fluorescent dissolved organic matter, pH and specific conductance as independent variables for both algorithms, and least sensitive to dissolved oxygen and turbidity. The predicted concentrations of calcium and monomeric aluminium were used to estimate catchment solute yield, which changed most dramatically for aluminium because it concentrates with stream discharge. These results show great promise for using a combined approach of stream sensing and intensive stream discrete sampling to build information about the highfrequency variation of solutes for which an appropriate sensor or proxy is not available.

Estimating metal loading into the sea from tidal rivers using total suspended solids and water quality models

“…Then, in the process of the transport of suspended solids by river flows, the metal loads can accumulate and exceed the carrying capacity of the suspended solids; under these conditions, some metals may settle into the sediments of the riverbeds. Due to the poor transport ability of sediments, in coastal rivers, metals enter the sea mainly in the freely dissolved and suspended forms (Donner et al, 2017;Green et al, 2020;Ruegner et al, 2019). Thus, when calculating metal loadings, the loadings of these two forms of metals are included and estimated.…”

Can machine learning accelerate process understanding and decision‐relevant predictions of river water quality?

“…Examples of recent efforts to incorporate DL approaches for these predictions include the use of an LSTM to predict DO concentrations in 506 pristine U.S. catchments achieving moderate accuracy (Zhi et al, 2021), and an LSTM paired with a CNN (that generated streamflow estimates) to predict total nitrogen, phosphorus, and organic carbon in a major Korean river basin (Baek et al, 2020). LSTM, RF, and hybrid ML models have also been used for short‐term predictions of a suite of water quality variables (e.g., DO, pH, conductivity, turbidity, nutrients, water quality indices) with high‐frequency monitoring; in some cases classical ML surrogate models are used as soft sensors to make predictions of variables that are difficult or laborious to measure directly such as those that require laboratory sample analysis (Bui et al, 2020; Green et al, 2021; Harrison et al, 2021; Liu et al, 2019; Lu & Ma, 2020; Paepae et al, 2021).…”

Geospatial Artificial Intelligence (GeoAI) in the Integrated Hydrological and Fluvial Systems Modeling: Review of Current Applications and Trends

“…Limitations: Solutes with atmospheric, episodic, or strong biotic and abiotic controls were much more poorly predicted than solutes least affected by ecosystem dynamics. [191] Many WQ parameters cannot easily be measured in situ and in real time for various reasons, such as high-cost sensors, low sampling rate, multiple processing stages, and the requirement of frequent cleaning and calibration. Therefore, a common practice is the estimation of a particular WQ parameter value based on other surrogate parameters, called soft sensors [181,183,184].…”

“…Overall, the GeoAI water quality prediction depends not only on the selected algorithms and settings but also on the WQ parameters, data size, and training data quality for the learning models [183,188,191].…”