Future increases in the intensity of hurricanes and El Niño periods predicted by climate change models have focused attention on their role in stimulating harmful algal blooms (HABs). A series of hurricanes that recently impacted Florida (USA) provided a unique opportunity to explore the relationships between hurricanes, El Niño and HABs in two Florida estuaries subject to repeated intense ecosystem disruptive HABs, the Indian River Lagoon and the St. Lucie Estuary. The roles that hurricanes and El Niño play in contributing to HAB events are examined in the context of key structural and functional features of each estuary and their watersheds, including morphology, water residence time and hydrology, such as the influence of Lake Okeechobee discharges into the St. Lucie Estuary. The most direct impact was the increase in rainfall associated with hurricanes and El Niño, resulting in enhanced nutrient loads which drive HABs in the Indian River Lagoon and Lake Okeechobee. Major HABs in Lake Okeechobee also present an indirect threat of freshwater HAB blooms in the St. Lucie Estuary via mandated discharges from the lake into the estuary during high rainfall periods. Conversely, during the absence of HABs in Lake Okeechobee, short water residence times produced by discharges into the St. Lucie Estuary can result in lower bloom intensities.

Impedimetric biosensors for measuring small molecules based on weak/transient interactions between bioreceptors and target analytes are a challenge for detection electronics, particularly in field studies or in the analysis of complex matrices. Protein–ligand binding sensors have enormous potential for biosensing, but achieving accuracy in complex solutions is a major challenge. There is a need for simple post hoc analytical tools that are not computationally expensive, yet provide near real time feedback on data derived from impedance spectra. Here, we show the use of a simple, open source support vector machine learning algorithm for analyzing impedimetric data in lieu of using equivalent circuit analysis. We demonstrate two different protein-based biosensors to show that the tool can be used for various applications. We conclude with a mobile phone-based demonstration focused on the measurement of acetone, an important biomarker related to the onset of diabetic ketoacidosis. In all conditions tested, the open source classifier was capable of performing as well as, or better, than the equivalent circuit analysis for characterizing weak/transient interactions between a model ligand (acetone) and a small chemosensory protein derived from the tsetse fly. In addition, the tool has a low computational requirement, facilitating use for mobile acquisition systems such as mobile phones. The protocol is deployed through Jupyter notebook (an open source computing environment available for mobile phone, tablet or computer use) and the code was written in Python. For each of the applications, we provide step-by-step instructions in English, Spanish, Mandarin and Portuguese to facilitate widespread use. All codes were based on scikit-learn, an open source software machine learning library in the Python language, and were processed in Jupyter notebook, an open-source web application for Python. The tool can easily be integrated with the mobile biosensor equipment for rapid detection, facilitating use by a broad range of impedimetric biosensor users. This post hoc analysis tool can serve as a launchpad for the convergence of nanobiosensors in planetary health monitoring applications based on mobile phone hardware.

In this study, we explored the use of statistical machine learning and long-term groundwater nitrate monitoring data to estimate vadose zone and saturated zone lag times in an irrigated alluvial agricultural setting. Unlike most previous statistical machine learning studies that sought to predict groundwater nitrate concentrations within aquifers, the focus of this study was to leverage available groundwater nitrate concentrations and other environmental variables to determine mean regional vertical velocities (transport rates) of water and solutes in the vadose zone and saturated zone (3.50 and 3.75 m yr⁻¹, respectively). The statistical machine learning results are consistent with two primary recharge processes in this western Nebraska aquifer, namely (1) diffuse recharge from irrigation and precipitation across the landscape and (2) focused recharge from leaking irrigation conveyance canals. The vadose zone mean velocity yielded a mean recharge rate (0.46 m yr⁻¹) consistent with previous estimates from groundwater age dating in shallow wells (0.38 m yr⁻¹). The saturated zone mean velocity yielded a recharge rate (1.31 m yr⁻¹) that was more consistent with focused recharge from leaky irrigation canals, as indicated by previous results of groundwater age dating in intermediate-depth wells (1.22 m yr⁻¹). Collectively, the statistical machine learning model results are consistent with previous observations of relatively high water fluxes and short transit times for water and nitrate in the primarily oxic aquifer. Partial dependence plots from the model indicate a sharp threshold in which high groundwater nitrate concentrations are mostly associated with total travel times of 7 years or less, possibly reflecting some combination of recent management practices and a tendency for nitrate concentrations to be higher in diffuse infiltration recharge than in canal leakage water. Limitations to the machine learning approach include the non-uniqueness of different transport rate combinations when comparing model performance and highlight the need to corroborate statistical model results with a robust conceptual model and complementary information such as groundwater age.

Due to strong abiotic forcing, dissolved oxygen (DO) in shallow tidal creeks often disobeys the conventional explanation of general aquatic DO cycling as biologically regulated. In the present work, we seek to quantify the relative importance of abiotic (hydrologic and climatic), and biotic (primary productivity as represented by chlorophyll‐a) descriptors of tidal creek DO. By fitting multiple linear regression models of DO to hourly chlorophyll‐a, water quality, hydrology, and weather data collected in a tidal creek of a Chesapeake Bay marsh (Maryland, USA), temporal shifts (summer‐early winter) in the relative importance of tidal creek DO descriptors were uncovered. Moreover, this analysis identified an alternative approach to evaluating tidal stage as a driver of DO by dividing stage into two DO‐relevant variables: stage above and below bankfull depth. Within the hydrologic variable class, stage below bankfull depth dominated as an important descriptor, thus highlighting the role of pore water drainage and mixing as influential processes forcing tidal creek DO. Study findings suggest that tidal creek DO dynamics are explained by a balance of hydrologic, climatic, and biotic descriptors during warmer seasons due to many of these variables (i.e., chlorophyll‐a, water temperature) acting as tracers of estuarine‐marsh water mixing; conversely, in early winter months when estuarine and marsh waters differ less distinctly, hydrologic variables increase in relative importance as descriptors of tidal creek DO. These findings underline important distinctions in the underlying mechanisms dictating DO variability in shallow tidal marsh‐creek environments relative to open water estuarine systems.