Search for: All records

Hybrid Knowledge‐Guided Machine Learning (KGML) models, which are deep learning models that utilize scientific theory and process‐based model simulations, have shown improved performance over their process‐based counterparts for the simulation of water temperature and hydrodynamics. We highlight the modular compositional learning (MCL) methodology as a novel design choice for the development of hybrid KGML models in which the model is decomposed into modular sub‐components that can be process‐based models and/or deep learning models. We develop a hybrid MCL model that integrates a deep learning model into a modularized, process‐based model. To achieve this, we first train individual deep learning models with the output of the process‐based models. In a second step, we fine‐tune one deep learning model with observed field data. In this study, we replaced process‐based calculations of vertical diffusive transport with deep learning. Finally, this fine‐tuned deep learning model is integrated into the process‐based model, creating the hybrid MCL model with improved overall projections for water temperature dynamics compared to the original process‐based model. We further compare the performance of the hybrid MCL model with the process‐based model and two alternative deep learning models and highlight how the hybrid MCL model has the best performance for projecting water temperature, Schmidt stability, buoyancy frequency, and depths of different isotherms. Modular compositional learning can be applied to existing modularized, process‐based model structures to make the projections more robust and improve model performance by letting deep learning estimate uncertain process calculations.

Airborne electromagnetic surveys collected in December 2011 and November 2018 and three soil sampling transects were used to analyze the spatial heterogeneity of shallow (<4 m) soil properties in lower Taylor Valley (TV), East Antarctica. Soil resistivities from 2011 to 2018 ranged from ∼33 Ωm to ∼3,500 Ωm with 200 Ωm assigned as an upper boundary for brine‐saturated sediments. Elevations below ∼50 m above sea level (masl) typically exhibit the lowest resistivities with resistivity increasing at high elevations on steeper slopes. Soil water content was empirically estimated from electrical resistivities using Archie's Law and range from ∼<1% to ∼68% by volume. An increase in silt‐ and clay‐sized particles at low elevations increases soil porosity but decreases hydraulic conductivity, promoting greater residence times of soil water at low elevations near Lake Fryxell. Soil resistivity variability between 2011 and 2018 shows soils at different stages of soil freeze‐thaw cycles, which are caused predominantly by solar warming of soils as opposed to air temperature. This study furthers the understanding of the hydrogeologic structure of the shallow subsurface in TV and identifies locations of soils that are potentially prone to greater rates of thaw and resulting ecosystem homogenization of soil properties from projected increases in hydrological connectivity across the region over the coming decades.

In limnological studies of temperate lakes, most studies of carbon dioxide (CO₂) and methane (CH₄) emissions have focused on summer measurements of gas fluxes despite the importance of shoulder seasons to annual emissions. This is especially pertinent to dimictic, small lakes that maintain anoxic conditions and turnover quickly in the spring and fall. We examined CO₂and CH₄dynamics from January to October 2020 in a small humic lake in northern Wisconsin, United States through a combination of discrete sampling and high frequency buoy and eddy covariance data collection. Eddy covariance flux towers were installed on buoys at the center of the lake while it was still frozen to continually measure CO₂and CH₄across seasons. Despite evidence for only partial turnover during the spring, there was still a notable 19‐day pulse of CH₄emissions after lake ice melted with an average daytime flux rate of 8–30 nmol CH₄m⁻²s⁻¹. Our estimate of CH₄emissions during the spring pulse was 16 mmol CH₄m⁻²compared to 22 mmol CH₄m⁻²during the stratified period from June to August. We did not observe a linear accumulation of gases under‐ice in our sampling period during the late winter, suggesting the complexity of this dynamic period and the emphasis for direct measurements throughout the ice‐covered period. The results of our study help to better understand the magnitude and timing of greenhouse gas emissions in a region expected to experience warmer winters with decreased ice duration.

Millions of lakes worldwide are distributed at latitudes or elevations resulting in the formation of lake ice during winter. Lake ice affects the transfer of energy, heat, light, and material between lakes and their surroundings creating an environment dramatically different from open‐water conditions. While this fundamental restructuring leads to distinct gradients in ions, dissolved gases, and nutrients throughout the water column, surprisingly little is known about the resulting effects on ecosystem processes and food webs, highlighting the lack of a general limnological framework that characterizes the structure and function of lakes under a gradient of ice cover. Drawing from the literature and three novel case studies, we present the Lake Ice Continuum Concept (LICC) as a model for understanding how key aspects of the physical, chemical, and ecological structure and function of lakes vary along a continuum of winter climate conditions mediated by ice and snow cover. We examine key differences in energy, redox, and ecological community structure and describe how they vary in response to shifts in physical mixing dynamics and light availability for lakes with ice and snow cover, lakes with clear ice alone, and lakes lacking winter ice altogether. Global change is driving ice covered lakes toward not only warmer annual average temperatures but also reduced, intermittent or no ice cover. The LICC highlights the wide range of responses of lakes to ongoing climate‐driven changes in ice cover and serves as a reminder of the need to understand the role of winter in the annual aquatic cycle.