2024 AmeriFlux Annual Meeting Abstracts

click here to jump to the Data Session on Sept 4
click here to jump to the Remote Sensing Session on Sept 5
click here to jump to the Community Session on Sept 4
click here to jump to the General Flux Session on Sept 4

Download the booklet with the poster abstracts here.

Wednesday, September 4

Opening Session: Flux Data for Climate Insights

Invited: Rob Jackson, Stanford University

“FLUXNET-CH4, tropical wetlands, and potential methane tipping points”

Short Talk: Jackie Matthes, Harvard Forest, Harvard University

“Carbon & water cycling within a decade of precipitation extremes at Harvard Forest”

The northeastern U.S. has experienced a rapid rise in extreme precipitation events and total precipitation due to climate change. Despite higher precipitation overall, long-term near-surface soil moisture at Harvard Forest has decreased since 2010, a global pattern observed in other temperate forest regions. In this study, we used long-term data on ecosystem-atmosphere water and carbon exchange at Harvard Forest to deduce the impact of precipitation shifts on ecosystem water and carbon flux and the strength of land-atmosphere water coupling. The region’s well-drained alluvial and colluvial soils rapidly drain surplus moisture from large rain events, while the remaining moisture necessary to preserve local humidity is quickly lost to evapotranspiration unless frequently replenished by rainfall. Meteorological conditions in the nongrowing season have shifted to warmer, drier conditions with increased evaporation that sets the stage for more frequent summer soil moisture deficits. Since 2010, we have also observed a dampening of canopy light response curves, indicating lower rates of carbon uptake during the growing season. More frequent and intense dry conditions during key phenological windows, the intense delivery of rainfall during a shorter temporal window in the growing season, and rising summer temperatures and lower humidity have combined to decrease the ecosystem carbon uptake by photosynthesis and cause larger interannual variation in the strength of the net carbon sink at Harvard Forest over the past decade.


Session: In a new light – Approaches in data processing, modeling and upscaling

Invited: Alejandro Cueva, El Colegio de la Frontera Sur

“An assessment of uncertainties in measurements of carbon dioxide and energy fluxes across the MexFlux Network”

 

Short Talk: Ngoc Nguyen, UC Berkeley

“Unaccounted carbon losses across global drylands”

In drylands, rainfall rewets dry soils, causing large pulses of soil and net CO2 effluxes, yet the extent of these rain-induced carbon pulses remains unclear globally. Hence, we quantify the impact of such rewetting events on dryland carbon balance using 323 site-years of data across 34 global dryland eddy-covariance sites. We manually label 1857 rewetting events, and find that rewetting events contribute significantly to annual ecosystem respiration and net ecosystem productivity, on average 21.3% and 15.4% respectively. However, parametric and machine-learning eddy-covariance partitioning methods underestimate, on average 30% to 50%, ecosystem respiration and vegetation productivity during these events. As a result, we develop FLUXPULSE, an add-on tool to FLUXNET formatted data, to reduce underestimations of carbon fluxes during rewetting events by automatically detecting and bias-correcting rain-induced carbon pulses. Our research highlights the significant contribution of rewetting events to dryland carbon balance and suggests the application of FLUXPULSE in future eddy-covariance datasets to effectively capture the dynamics of carbon fluxes across global drylands.

 

Short Talk: Jake Searcy, University of Oregon

“High-Resolution CO2 Flux Prediction from Satellite Imagery with Footprint Aware Deep Learning”

Improving carbon dioxide (CO2) flux measurement and prediction is becoming increasingly essential, particularly in ecosystems monitored with eddy-covariance (EC) flux towers. One significant challenge in enhancing these measurements and predictions is the highly variable footprint of EC towers. This variability is influenced by factors such as atmospheric conditions and site characteristics, leading to variation in the tower’s measurement area by orders of magnitude, which can significantly affect the accuracy and representativeness of the flux data. As satellite imagery continues to advance in resolution, it becomes possible to account for the measurement footprint of EC towers, which are generally much larger than individual pixels. This talk will present a new deep learning framework developed to explicitly utilize time varying footprints and site heterogeneity to provide high temporal and spatial resolution CO2 flux predictions. Our methodology utilizes a dual-arm neural network architecture. One arm of the network leverages all 11 spectral bands from Landsat 8/9 satellites at 30m2 resolution combined with temperature, relative humidity, and incoming solar radiation data from AmeriFlux towers to predict half-hourly CO2 fluxes. The other arm of the network uses local atmospheric conditions and tower height to generate an attention map that accurately models the measurement footprint, and is used to weight each pixel’s prediction from the first arm to predict the overall flux measured at the tower. Both arms are trained simultaneously on data from 214 Ameriflux sites providing new high resolution flux predictions and a data driven model of the tower’s footprint. The last year of each site’s data is withheld for validation, and monthly flux predictions show good agreement R2 =0.77. Validation on withheld sites yields R2 =0.63. One example of upscaling to landscape and regional fluxes is done utilizing estimates from Parameter-elevation Relationships on Independent Slopes Model (PRISM) as inputs to produce a detailed map of CO2 fluxes across Oregon.

 

Short Talk: Arman Ahmadi, UC Berkeley

“Elucidating Water Availability’s Impact on Thermal Optimality in Terrestrial Ecosystems Using AmeriFlux Data and Machine Learning”

Ecosystem productivity significantly contributes to global carbon sequestration. Air temperature directly affects the leaf-scale photosynthetic rate and ecosystem productivity. Ecosystem productivity exhibits a tipping point behavior versus air temperature: it rises with increasing temperatures and diminishes after reaching a maximum. Extreme temperatures due to global warming can trigger this productivity tipping point. However, along with temperature, the availability and spatiotemporal distribution of soil and atmospheric moisture affect the thermal behavior of terrestrial ecosystems. This study uses interpretable machine learning through two model-agnostic methods, partial dependence plot (PDP) and accumulated local effects (ALE), to analyze the thermal optimality of terrestrial ecosystems and evaluate the role of water availability in regulating thermal behavior and productivity. Machine learning models are trained to simulate Jarvis’s empirical photosynthesis model. Through this multi-input data-driven approach, this study moves beyond unimodal photosynthesis-temperature response curves and quantitatively visualizes the three-dimensional temperature-moisture-productivity space for different ecosystems. Our study analyzes 112,683 daily data samples of biotic and abiotic variables from 108 AmeriFlux sites. Our findings uncover the critical role of water availability in shaping the terrestrial ecosystems’ thermal behavior. Arid ecosystems reach their productivity tipping point at lower temperatures, while wet ecosystems have higher temperature tipping points. We quantify the interplay of water availability and temperature in driving terrestrial ecosystems’ productivity, where water availability emerges as the controlling factor in arid ecosystems. We also found a statistically significant increase in air temperature trends in most study sites. This warming trend can cause different ecosystems to reach their productivity tipping point and lead to unprecedented ecological consequences.


Thursday, September 5

Session: Remote Sensing

Invited: Elsa Ordway, UCLA

Pangea NASA Scoping Study

 

Invited: Marcy Litvak, University of New Mexico & Dave Moore, University of Arizona & Russ Scott, USDA-ARS

ARID NASA Scoping Study

 

Short Talk: Emma Reich, Northern Arizona University

“Decoupling of transpiration, gross primary productivity, and solar-induced fluorescence at the ecosystem scale”

Long-term carbon and water flux studies generally assume a consistent, tight relationship between plant carbon gain and water use. However, certain environmental conditions such as heatwaves have been shown to lead to a decoupling between gross primary productivity (GPP), transpiration (T), and solar-induced fluorescence (SIF). SIF could be a promising intermediary to investigate decoupling between carbon and water relations, as it is a proxy for photosynthesis, but also decouples from GPP under stressful conditions. Because decoupling alters ecosystem carbon and water cycles that are typically tightly correlated, it is of interest to determine the environmental conditions and the ecosystem types under which these decoupling events occur. However, at the ecosystem scale, it is not yet clear in what ecosystem types and under what conditions T, GPP, and SIF decouple. To better quantify the conditions under which T, GPP, and SIF decouple, we compiled paired eddy covariance and SIF data from flux tower sites in Europe, Asia, and North America using collaborations established with the FLUXNET Secondment Program. We used a stochastic antecedent model in a Bayesian framework to test when SIF, T, and GPP are linked to concurrent and antecedent drivers and their interactions (e.g., climate variables) to identify net sensitivities to environmental drivers (e.g., under what conditions T decouples from SIF and GPP, such as heatwaves). By using this framework to disentangle the effects and timescales of influence of individual climate variables on SIF, T, and GPP, we found evidence of decoupling at the ecosystem scale related to high temperatures. Furthermore, we determined the temperatures at which decoupling events are likely to occur across a range of different ecosystem types.

 

Short Talk: Yanghui Kang, UC Berkeley

“Assessing global impacts of CO2 fertilization on photosynthesis through machine learning upscaling of eddy covariance fluxes”

Elevated atmospheric CO2 has enhanced global photosynthesis, a process known as the CO2 fertilization effect, which substantially mitigates climate change. However, significant uncertainties and discrepancies exist in its quantification from various satellite data and process-based models, hindering a robust assessment of terrestrial carbon dynamics. Here, we upscale eddy covariance fluxes with machine learning and multi-source remote sensing data to provide a comprehensive quantification of global Gross Primary Productivity (GPP), i.e. ecosystem photosynthesis, over the past four decades. We incorporate CO2 fertilization using both a data-driven approach and a hybrid modeling approach that combines machine learning with optimality theories. Our findings reveal a widespread positive effect of elevated CO2 on photosynthesis, which is underestimated by conventional upscaled and satellite-based data products due to a neglect of direct CO2 effects on photosynthetic light use efficiency. Furthermore, our estimates of GPP spatial and temporal dynamics demonstrate improved consistency with terrestrial biosphere models from the TRENDY ensemble. Our dataset, named CEDAR-GPP, provides global monthly GPP estimates at 0.05-degree resolution from 1982 to 2020. Our work reconciles the long-standing discrepancies of long-term photosynthesis changes between satellite-derived and terrestrial biosphere model simulations, offering important insights and robust benchmarks for assessing the global carbon cycle.

 

Short Talk: Rubaya Pervin, Indiana University

“DryFlux 2.0: Incorporating high frequency soil moisture to improve upscaled dryland carbon and water fluxes”

Upscaling dryland in situ carbon and water fluxes are important to better understand the role drylands play in global carbon cycle variability and their sensitivity to future changes in climate and hydrological cycling. However, global scale upscaled in situ flux GPP products have been shown to perform poorly in drylands because of the complexity of plant responses to limited and variable water availability. The DryFlux product developed by Barnes et al. (2021) addresses this issue by taking dryland specific ecohydrological responses into account. However, Dryflux did not incorporate soil moisture data. In this study, we develop an improved upscaled flux product for drylands – DryFlux 2.0) – that incorporates soil moisture data at different depths into the machine learning algorithm and addresses antecedent conditions. Furthermore, we adjust the model to operate on a daily time step to align with the temporal resolution of our flux tower data, therefore better capturing the dynamic nature of dryland fluxes. We run the models in loops across varying time windows for the purpose of optimization procedures to understand how previous weather conditions impact current ecosystem functioning. This new approach to upscaling dryland fluxes will help identify where and when better representation of shallow and deep soil moisture pools can significantly improve upscaled GPP and ET flux model predictions in North American darlands. Our goal is to use the new version of DryFlux to better understand the impact of the ongoing megadrought in North American drylands. Beyond this study, DryFlux 2.0 can be used for a wide range of applications, including benchmarking process-based models that have been shown to underestimate the magnitude and variability of ecosystem fluxes in drylands, as well as addressing broader questions about dryland ecosystem carbon and water cycle responses to climate and environmental change.

 

Short Talk: Keenan Ganz, University of Washington

“Overstory and understory leaves warm faster than air in evergreen needleleaf forests”

The limited homeothermy hypothesis states that leaves maintain their temperature within an optimal range for photosynthesis by increasing transpiration during warm conditions. Under limited homeothermy, plants may offset thermal stress caused by climate change. If this hypothesis is true, we should observe: 1) leaf temperature increasing more slowly than air temperature and 2) leaves cooler than air during warm conditions. We tested these predictions with an energy balance model for evergreen needleleaf forest sites in the National Ecological Observatory Network. Our model divides a forest canopy into vertical strata and estimates leaf temperature in each stratum from measurements of microclimate, water and carbon fluxes, and canopy structure. The model let us partition opposing forcings on canopy temperature: warming from absorbed radiation and cooling from transpiration. Our results do not support limited homeothermy. In all forests and strata, leaf temperature increased faster than air and periods with leaves cooler than air were rare. In such cases, cooling was due to emitted radiation, not transpiration. Temperature forcing due to absorbed radiation was larger than forcing due to transpiration in all sites. We attribute these results to the needle-like shape of leaves in our study sites. This leaf shape increases boundary layer conductance and causes heat gain from surrounding air to overpower heat loss from transpiration when leaves are cooler than air. Our results indicate that needleleaf forests cannot avert thermal stress in a warming world. Thermal limits on photosynthesis and non-linear increases in respiration with temperature may weaken the role of evergreen forests as a global carbon sink. A manuscript from this work is in review at Agricultural and Forest Meteorology.

 

Short Talk: Pramit Kumar Deb Burman, Indian Institute of Tropical Meteorology

“Spaceborne SIF signal detects the imprint of monsoon on the photosynthetic carbon uptake by a subtropical forest in India”

It is imperative to monitor the long-term variability and environmental drivers of the terrestrial carbon cycle as these ecosystems play a crucial role in climate mitigation due to their offering of a large carbon storage. The plant photosynthetic carbon uptake, defined as the gross primary productivity (GPP), is estimated most accurately using the ground-based eddy covariance (EC) technique. Almost 22% of the large Indian landmass is forested, stretching from the tropics to the subtropics located in different climates. A unique feature of their carbon uptake strength and variability is the impact of monsoon, a planetary-scale event signature of this region. However, due to the missing EC measurements, the carbon sequestration nor the impact of monsoon and associated processes on it are well-understood. This region, however, is densely populated and remains vulnerable to rapid land use, land cover change and adversities of climate variability. The space-based estimation of solar-induced chlorophyll fluorescence (SIF) offers a complementary means; however, it also remains largely unexplored in India. In this work, we check the efficacy of space-based SIF measurements at 740 nm by GOME-2 and at 757 nm and 771 nm by OCO-2 and GOSAT in tracking the carbon uptake features of a broadleaf deciduous forest in the Kaziranga National Park in India, in conjunction with EC measurements. We find the GOSAT SIF products to outperform the others in the process. Nevertheless, linear relationships exist between all the SIF products and GPP of this forest, which improve with temporal downscaling; at finer temporal resolutions, the microclimatic variations introduce nonlinearity. However, this linearity between SIF and GPP is broken during the monsoon period when cloudiness and vapor pressure deficit deviate significantly from their year-round values. We also report the SIF use efficiency of this ecosystem and establish SIF as a potential proxy for forest monitoring.

 

Short Talk: Sean Burns, University of Colorado

“Using a pair of GPS sensors to measure vegetation optical depth (VOD) and relate it to the evaporation of intercepted rainfall in a subalpine forest”

Between summer of 2022 and summer 2023, we deployed a pair of Global Navigation Satellite System (GNSS) receivers at the Niwot Ridge Subalpine Forest AmeriFlux site (US-NR1) to estimate the vegetation optical depth (VOD) and determine the timing of the evaporation of intercepted water over the diel cycle. We compared the changes in canopy water mass inferred by the VOD measurements and concurrent tree sway measurements and found that they both suggested canopy evaporation lasted around 10-14 hours following a precipitation event. We also compared the VOD and tree sway data with the US-NR1 long-term above-canopy (21.5 m) and subcanopy (2.5 m) eddy-covariance water vapor flux measurements and highlight regions of agreement and disagreement between these independent measurements as-related to the timing and amount of canopy evaporation.


Community Session: More than the sum of its parts

Invited: Dennis Baldocchi, UC Berkeley

“Ameriflux 2024: Where the Whole Exceeds the Sum of the Parts”

We discuss an overview of where Ameriflux has come from and where it is going, based on our ability to measure fluxes, share data and address pressing problems facing society and ecosystems in a changing world.

 

Short Talk: Stefan Arndt, University of Melbourne & Peter Isaac, TERN Australia

“TERN OzFlux – the flux network in Australia and New Zealand”

This talk will describe the history and future of OzFlux, the flux network in Australia and New Zealand. OzFlux started out in the early 2000s as a network of a few individual sites, including the longest running sites of Tumbarumba and Howard Springs. OzFlux was (and still is) a community of practice where individual site PIs share ideas, exchange, analyse and submit data and try to improve the overall flux science. This changed in 2009 with the introduction of the Terrestrial Ecosystem Research Network (TERN), a government initiative to support research infrastructure. TERN provided a stable operational basis for many site PIs and enabled the establishment of a collaborative network of sites in important ecosystems in Australia. Researchers from New Zealand are part of OzFlux as a network, even if they do not benefit from TERN funding. TERN OzFlux is now expanding its operations and changes its operational model. We will move away from a site PI focused model where data are collected by an individual site and uploaded and processed every few months to a model where data are streamed constantly to a central server and processed more rapidly. Multiple sites will be serviced by full-time technicians, site set-up and communications are streamlined, allowing for network wide monitoring of functionality of sensors and a more rapid detection and repair of system faults. The changes to our operational model should lead to more consistent data streams, greater data coverage and also the inclusion of more non-TERN funded flux sites.

 

Short Talk: Jen Diehl, Northern Arizona University

“The Great Thermal Bake–Off: A Hands-On Workshop for Enhancing Temperature Measurement Precision and Standardization for Improved Flux Interpretation and Application”

Advances in thermal infrared remote (TIR) sensing are rapidly revolutionizing our understanding of ecosystem functioning by providing detailed measurements of surface temperatures and energy fluxes. When coupled with eddy covariance measurements, TIR data enhances our understanding of ecosystem-scale carbon and water exchanges, bridging the gap between ground-based data and satellite observations. However, the adoption of thermal infrared-based temperature measurement has been slow due to challenges in accuracy, reliability, and lack of standardized practices. In August 2024, forty scientists from around the world gathered near Flagstaff, Arizona, to participate in a workshop sponsored by FLUXNET and AmeriFlux. Our workshop addressed these challenges by conducting a ‘thermal camera bake-off’ to compare and evaluate different thermal sensors and establish standardized methods for TIR measurements, including calibration techniques and data processing protocols. This rigorous comparison was essential for identifying the strengths and weaknesses of various thermal sensors, leading to the development of best practices for their use in ecological research. The primary goal was to improve the standardization and reliability of TIR measurements, fostering cross-disciplinary collaboration and promoting best practices. By creating a platform for scientists to share knowledge and techniques, the workshop facilitated the exchange of ideas and innovations, ultimately enhancing the quality of TIR data collected globally. This initiative laid the foundation for a network of high-quality TIR measurements, contributing significantly to the advancement of ecosystem science. By addressing these critical issues, we enabled more precise and reliable temperature measurements, which are crucial for a better understanding of the impacts of climate change on ecosystems. The outcomes of this workshop will be pivotal in driving future research and ensuring the robustness of TIR data in ecological studies.

 

Short Talk: Tonantzin Terrazas, Instituto de Ecología, UNAM

Mexflux Update

 

Short Talk: Dario Papale, University of Tuscia

ICOS Update

 

Short Talk: Kyle Delwiche, UC Berkeley

FLUXNET Coop Update


Friday, September 6

Session: Fluxes everywhere, all at once (General Flux Science session)

Invited: Jennifer Watts, Woodwell Climate Research Center

“Insights from 30+ Years of Arctic-Boreal Flux Observations: Key Discoveries and Future Directions”

 

Short Talk: Sonia Wharton, Lawrence Livermore National Laboratory

“Doppler lidar in forests for advancing our understanding of canopy-ABL processes”

The AmeriFlux community can benefit from recent advances in Doppler lidar technology. Here, we present two field campaigns funded by the U.S. DOE’s Wind Energy Technology Office that are studying the wind resource over tall canopies for improving wind power forecasts in forested areas. Likewise, these lidars offer a wealth of information about the atmospheric boundary layer above the forest and subsequent shear and turbulence events which can penetrate the canopy and ultimately affect our interpretation of ecosystem mass and energy fluxes via enhanced canopy coupling. The field sites include a 50-m tall conifer forest (Wind River) and a 20-m tall deciduous forest in the Appalachian Mountains (Mountain Lake Biological Station). Both sites are operated by NEON; Wind River was also an AmeriFlux site from 1998-2016. A vertical-profiling lidar was placed directly on top of the NEON tower and took measurements of wind velocity, direction and turbulence up to 220 m above ground level at both towers. At Wind River, a scanning lidar was placed in a nearby clearing and programmed to scan the wind field over the forest canopy, including overlapping its scans with the profiling lidar on top of the tower. The scanning lidar captured spatially-variable terrain induced flows across the surrounding mountain-valley terrain. Both lidars captured wind jets and periods of intermittent turbulence over the forest canopy. How and when these turbulence events penetrate the tall forest are studied using NEON’s eddy covariance and tower profile measurements. At MLBS, we are attempting to use multiple scanning lidars to build “virtual meteorological towers” over the forest canopy. This allows for high-resolution observations of the wind components without making assumptions about flow homogeneity (which lidar traditionally rely on). Our talk will discuss the logistics for deploying Doppler lidar in forests and give examples of how the AmeriFlux community could use these technologies.

 

Short Talk: Christopher Still, Oregon State University

“Impacts of the June 2021 Heat Dome event on trees and forests of the Pacific Northwest, USA”

Background

Many areas of the Pacific Northwest (PNW) experienced record high air temperatures during an extreme heat wave (“Heat Dome”) in late June of 2021. Notably, sunlit leaf temperatures reached 45-50 ºC for multiple hours, exceeding known damage thresholds. There is an increasing recognition that heat waves can lead to widespread growth reductions and forest vulnerabilities. As the frequency of coincident heat waves atop drought is increasing, understanding the environmental drivers, ecological consequences, and biophysical and physiological mechanisms underlying heat wave damage incurred by forests is essential. In this work we analyze ecosystem responses to the Heat Dome using AmeriFlux data from four contrasting forest sites: US-Me6, CA-Ca3, US-xWR, and US-xAB.

Results

The extreme heat elicited multi-scale responses in trees and forests of the PNW, including widespread foliar scorch from the coast to the Cascades. The heat also led to unusual early pulses of litterfall and marked reductions in canopy greenness within weeks of the event at US-xAB and US-xWR. At all forest sites daily mean fluxes responded strongly. At two of the sites (US-xWR, and US-xAB), ET increased markedly during the event compared to before and after. At US-xAB, the hottest night of the event led to extraordinary nighttime ET (peaking at ~150 W m-2) leading to pronounced canopy cooling (up to 10 K relative to canopy-top air temperature) and very large negative sensible heat fluxes (peaking around -250 W m-2). Heat impacts lingered as shown by cumulative fluxes, particularly at US-Me6 and CA-Ca3, where ET trajectories notably flattened and water fluxes at Me6 were the lowest on record for the remainder of the season. NEE was on average positive, i.e., the forests were C sources throughout the event. Cumulative annual C uptake was also affected: CA-Ca3 was absorbing more CO2 at the time of the event than in any previous year, but afterwards its NEE trajectory tapered off sharply. The forest at US-Me6 reversed from being a C sink to a source by late in the growing season. At both sites the change was driven by heat impacts on GPP.

 

Short Talk: Carlos Wang, UC Berkeley

“The reign of surface energy balance under advection”

Orbiting around the non-closure problem in eddy covariance, a new generation of high-resolution thermal imagery has revealed that advection may be more common than previously expected. To investigate this, we deployed multiple tower arrays and profile measurements at an irrigated alfalfa site to measure: 1) horizontal heat and moisture advection; 2) vertical heat advection; and 3) vertical heat flux divergence. Over the five analysis periods, advective fluxes modified latent heat fluxes (LE) due to local and non-local processes. Locally, advected moisture from upwind humidified the atmosphere, creating two competing processes: either the increased stomatal opening enhanced downwind LE, or the lowered atmospheric demand suppressed downwind LE. Here, we found that stomatal regulation played a dominant role, resulting in an overall enhancement. However, the enhancement was also influenced by non-local processes, as spectral analysis revealed that low frequency (i.e., large) eddies contributed high heat and moisture via advection. Additionally, thermal remote sensing observations from ECOSTRESS and Landsat 8/9 showed that these large eddies were generated over the upwind surface, indicating the enhancement of LE through non-local transport of heat and moisture. Results from profile measurements suggested that vertical heat advection and heat flux divergence were present. However, accurate estimates of this remains challenging due to difficulties in measuring rapid changes in vertical wind speed and temperature. Lastly, by accounting for different previously neglected terms, we found that the energy balance closure was improved by ~21% (r2 = 0.97, p<0.001). Our findings here showed that spatiotemporal heterogeneity induced advection, which modified the surface energy balance. Additionally, this improved understanding on advection provides valuable insights for remote sensing evapotranspiration models, which often treat pixels as isolated columns rather than also considering the lateral effects of heat and moisture.

 

Short Talk: Lianhong Gu, Oak Ridge National Laboratory

“Misunderstanding and mismeasurements of heat transfer are a key contributor to the observed Earth surface energy imbalance”

The eddy covariance (EC) approach is our most advanced technique for measuring mass and energy exchanges at the Earth surface. Yet the EC approach cannot close the Earth surface’s energy budget. This problem has puzzled Earth system scientists for many decades. It casts doubt on the reliability of data used to validate Earth system models, and questions whether our current understanding of energy processes is complete. Heat transfer is a key energy exchange process at the Earth surface. The current EC theoretic framework and guidance for heat transfer measurements have been heavily influenced by studies of heat transfer in solid states. Yet heat transfer in liquid and gas states differs drastically from that in solid state as the former is simultaneously coupled with mass transfer whereas mass transfer does not occur in the latter. The lack of appreciation for this crucial difference has led to simplistic equating of heat energy with enthalpy exchange, formation of ill-conceived concepts and theories, and misguided measurements of turbulent heat flux. Starting from first principles of physical fluid mechanics and thermodynamics, we systematically derive fundamental equations of coupled mass and energy transfer at the Earth surface that simultaneously conserve mass and total energy. We demonstrate that the conventional sensible heat calculations miss several substantial terms while miscalculate others. We suggest that both measurements and modeling of thermal energy exchange at the Earth surface will need to be revised to better understand and predict energy processes and temperature dynamics in a critical part of our biosphere.