Search for: All records

ABSTRACT Machine‐learning models have been surprisingly successful at predicting stream solute concentrations, even for solutes without dedicated sensors. It would be extremely valuable if these models could predict solute concentrations in streams beyond the one in which they were trained. We assessed the generalisability of random forest models by training them in one or more streams and testing them in another. Models were made using grab sample and sensor data from 10 New Hampshire streams and rivers. As observed in previous studies, models trained in one stream were capable of accurately predicting solute concentrations in that stream. However, models trained on one stream produced inaccurate predictions of solute concentrations in other streams, with the exception of solutes measured by dedicated sensors (i.e., nitrate and dissolved organic carbon). Using data from multiple watersheds improved model results, but model performance was still worse than using the mean of the training dataset (Nash–Sutcliffe Efficiency < 0). Our results demonstrate that machine‐learning models thus far reliably predict solute concentrations only where trained, as differences in solute concentration patterns and sensor‐solute relationships limit their broader applicability. 

Abstract The Lamprey River Hydrological Observatory (LRHO) is a lowland coastal watershed in southeastern New Hampshire (USA). The LRHO offers a platform to investigate the effects of suburbanization and changing seasonality on watershed hydrology, biogeochemistry, and nutrient export to an estuarine ecosystem. The LRHO utilizes a nested‐watershed design to examine headwater stream and main‐stem river dynamics distributed across a mixed land‐use environment. Data sets from the LRHO now comprise over 20 years of weekly grab sample data as well as 7 years of high‐frequency sensor data. Collectively these data sets include measures of discharge, dissolved organic matter, nutrients, cations and anions, greenhouse gases, and other physio‐chemical properties. Here we share information on the setting and motivating questions of the LRHO and data availability. 

Abstract Synthesis research in ecology and environmental science improves understanding, advances theory, identifies research priorities, and supports management strategies by linking data, ideas, and tools. Accelerating environmental challenges increases the need to focus synthesis science on the most pressing questions. To leverage input from the broader research community, we convened a virtual workshop with participants from many countries and disciplines to examine how and where synthesis can address key questions and themes in ecology and environmental science in the coming decade. Seven priority research topics emerged: (1) diversity, equity, inclusion, and justice (DEIJ), (2) human and natural systems, (3) actionable and use‐inspired science, (4) scale, (5) generality, (6) complexity and resilience, and (7) predictability. Additionally, two issues regarding the general practice of synthesis emerged: the need for increased participant diversity and inclusive research practices; and increased and improved data flow, access, and skill‐building. These topics and practices provide a strategic vision for future synthesis in ecology and environmental science.