Talk Abstracts—2018 AmeriFlux PI Meeting

Linking variation in intrinsic water-use efficiency to isohydricity: a comparison at multiple spatiotemporal scales
*Koong Yi, Justin T. Maxwell, Matthew K. Wenzel, D. Tyler Roman, Peter E. Sauer, Richard P. Phillips, and Kimberly A. Novick
*, Indiana University Bloomington
Species-specific responses of plant intrinsic water-use efficiency (iWUE) to multiple environmental drivers associated with climate change, including soil moisture (θ), vapor pressure deficit (D), and atmospheric CO2 concentration (Ca), are poorly understood. We present how the iWUE and growth of canopy-dominant deciduous trees growing at Morgan-Monroe State Forest, which span a gradient of isohydric to anisohydric water- use strategies, respond to key environmental drivers (θ, D, and Ca). iWUE was calculated for individual tree species using leaf-level gas exchange and tree-ring δ13C in wood measurements, and for the whole forest using the eddy covariance method. We show that iWUE of the isohydric species was generally more sensitive to environmental change than the anisohydric species, and increased significantly with rising D during the periods of water stress. At longer timescales, the influence of Ca was pronounced for isohydric tulip poplar but not for others. Our results imply that (1) trees’ physiological response to changing environmental drivers can be interpreted differently depending on the observational scale, and (2) care should be taken in interpreting observed or modeled trends in iWUE that do not explicitly account for the influence of D.

Why (and how) to determine soil water retention curves for Ameriflux sites
*Benjamin N. Sulman, Jing Yan and Teamrat A. Ghezzehei
*, Oak Ridge National Laboratory
Nearly all Ameriflux sites include continuous measurements of volumetric soil water content (VWC) among their meteorological and supporting datasets. These measurements are invaluable for identifying drought periods and interpreting seasonal and interannual variations in CO2 and water vapor fluxes related to droughts and other variations in hydrology. However, variations in soil physical and hydrological characteristics among sites make VWC difficult to compare across sites. Biological and hydrological impacts of soil water variation are often more tightly related to soil water potential (Ψs) than to VWC. Determining soil water retention curves, which allow VWC to be transformed into Ψs, facilitate more robust interpretation of ecohydrological variations and allow more natural comparisons across sites. As one example, we show that variations in evapotranspiration and photosynthesis at the Morgan Monroe State Forest Ameriflux site are better explained by variations in Ψs than by variations in VWC. As a second example, we demonstrate how Ψs can provide more physically relevant estimates of how soil moisture affects microbial respiration across soils with different structural properties. Finally, we summarize methods for determining soil water retention curves for Ameriflux sites using data that are already available for most sites.

From leaf to canopy to global: using flux data to test and improve terrestrial biosphere models across scales
Gordon B. Bonan, National Center for Atmospheric Research, 
The coupling between the biosphere and atmosphere in climate models occurs with the near- instantaneous fluxes of momentum, energy, and mass over the diurnal cycle as mediated by plant physiology, the microclimate of plant canopies, soils, and boundary layer processes. The central paradigm of land surface models has been to represent plant canopies as a “big leaf” without vertical structure, though with separate source fluxes for sunlit and shaded leaves and for soil. A particular challenge is how to represent turbulent processes in plant canopies. A further challenge, largely ignored in models, is that Monin-Obukhov similarity theory fails in the roughness sublayer extending to twice the canopy height or more. Stomatal conductance and soil moisture stress are additional regulators of canopy transpiration, but most models have an ad-hoc parameterization of soil moisture stress. Parameterization of soil moisture stress from more fundamental principles of plant hydrodynamics is a key research frontier. Here, results are shown for a multilayer canopy model with a roughness sublayer parameterization and plant hydrodynamics to test if this theory provides a tractable parameterization of surface fluxes extending from the ground through the canopy and the roughness sublayer. Leaf gas exchange measurements, eddy covariance flux data, and vertical profile data are used to test the model. The multilayer canopy improves simulations while also advancing the theoretical basis for surface flux parameterizations.

FLUXNET methane synthesis activity: Objectives, Observations, and Future Directions
*Gavin McNicol, Sara Knox, Rob Jackson, Ben Poulter, Zhen Zhang, Etienne Fluet-Chouinard, Gustaf Hugelius, Philippe Bousquet, Pep Canadell, Marielle Saunois, Dario Papale, Trevor Keenan, Housen Chu, Dennis Baldocchi, Dave Campbell, Samuel Chamberlain, Ankur Desai, Eugenie Euskirchen, Thomas Friborg, Mathias Goeckede, Manuel Helbig, Ken Krauss, Annalea Lohila, Bhaskar Mitra, Asko Noormets, Shuli Niu, Olli Peltola, Benjamin Runkle, Torsten Sachs, Karina Schäfer, Narasinha Shurpali, Oliver Sonnentag, Angela Tang, Timo Vesala, Lisamarie Windham-Myers, and Donatella Zona
*, Stanford University
To help reconcile bottom-up and top-down estimates of CH4 emissions, a better quantification of CH4 sources and sinks is needed. Direct observations of ecosystem-scale CH4 emissions with high measurement frequency offer a means of constraining regional and global CH4 budgets. Unlike the well-coordinated effort for synthesizing CO2 observations, which resulted in international data networks such as FLUXNET, no such data synthesis and initiative exists for CH4. With the increasing number of eddy covariance sites measuring CH4 emissions worldwide, the time is now right for such an effort.

Here we present a recent FLUXNET coordination network for CH4 organized by the Global Carbon Project in collaboration with other initiatives and regional flux networks. This activity aims to develop a FLUXNET-type analysis and synthesis for CH4 emissions. We present the objectives of this initiative, provides an overview of the current coverage of eddy covariance CH4 flux measurements worldwide, presents initial results comparing CH4 fluxes across ~50 sites, and discuss future research directions and needs. Preliminary observations suggest similar trends in CH4 emissions across regions and ecosystem types as reported in previous syntheses of primarily chamber-based measurements, however, the present database indicates considerably lower CH4 emissions for tundra and boreal wetlands. In contrast with earlier CH4 syntheses, temperature is found to be the strongest predictor of annual CH4 flux across wetland sites. Through this analysis and synthesis of flux tower CH4 measurements, this activity ultimately aims to better characterize CH4 emissions from terrestrial ecosystems and reconcile differences between bottom-up and top-down CH4 budgets.

New Constraints on the Methane Budget for the US Corn Belt and Upper Midwest: Initial Results from the GEM Project
*M. Julian Deventer1, Xiang Li1, Kelly Wells1, Xueying Yu1, Zichong Chen1,2, Tyler Roman3, Randall Kolka3, Jeffrey D. Wood1,4, Tim Griffis1, Dylan B. Millet1
1University of Minnesota – Twin Cities; 2Johns Hopkins University; 3US Forest Service – Northern Research Station Grand Rapids; 4University of Missouri
The US Corn Belt and Upper Midwest is critical to the US methane (CH4) budget as a global hotspot for crop and animal agriculture, with >40% crop cover and >700 million livestock. It also includes some of the most wetland-rich areas in the coterminous US, along with major urban and anthropogenic emissions. The Greenhouse Emissions in the Midwest (GEM) project combines an array of aircraft-based, tall-tower, in-situ agricultural, and wetland measurements with remote sensing and modeling tools in a multi-scale network to better understand the drivers of methane emissions in this key region.
This presentation will highlight some initial results from the GEM project, including (i) regional CH4 budget constraints inferred from tall-tower gradient measurements and inverse modeling; (ii) bottom-up (tracer release and Lagrangian dispersion) versus top-down (airborne mass balance) facility-scale flux measurements at large dairy and cattle operations, and (iii) results from ongoing ecosystem scale CH4 eddy flux measurements at a natural peatland site. For the latter we will explore the uncertainty in annual wetland CH4 flux budgets through an analysis of current CH4 instrumentation, post-processing, correction and gap-filling routines.

An upscaling framework for methane emissions in an ombrotrophic peat bog in Ohio
*C. Rey-Sanchez1, G.M. Davies2, J. Slater2, Y Hao2,3, R. Grau-Andres2, V. Rich3, G. Bohrer1
1Department of Civil and Environmental Engineering and Geodetic Science, The Ohio State University, Columbus, Ohio, USA. 2School of Environment and Natural Resources, The Ohio State University, Columbus, Ohio, USA. 3Microbiology Department, The Ohio State University, Columbus, Ohio, USA
Peatlands store up to a third of Earth’s soil carbon pool but they also emit methane (CH4), a powerful greenhouse gas. Understanding the fluxes of CH4 from peatlands is important for predicting Earth’s greenhouse gas budget, and climate feedbacks under changing conditions. The objective of this study was to develop an upscaling framework that integrates the heterogeneity of an ombrotrophic peat bog into estimates of total CH4 emissions. To achieve this, we studied Flatiron Bog, near Akron, Ohio. The bog is a traditional kettle lake bog with an open water subsection surrounded by other three concentric subsections: a floating mat, a tamarack tree (Larix laricina) woodland, and a lower-bush mosaic subsection dominated by blueberry (Vaccinium spp.). Through the peak growing seasons of 2017 and 2018 (June-Oct and May-Oct, respectively), we used chambers to measure plant and surface CH4 emissions at each of the subsections in the bog. There was a large spatial and temporal variability in CH4 emissions within the bog. Emissions were higher in the Tamarack subsection than in other areas such as the open water, or the outer blueberry subzones. Temporal variations were mostly explained by changes in water table depth. We found small fluxes of CH4 through plant tissue, which were sometimes negative, indicating that these plants could be a slight sink of CH4. The information from this study can be used to improve estimates of peat bog CH4 emissions that include within- site land cover heterogeneity.

Carbon and water exchanges in managed forests in North American cold region
M. Altaf Arain, McMaster University, Hamilton, Ontario, Canada,
Forest ecosystems play a major role in the atmospheric CO2 sequestration and providing sustainable water resources and healthy environments. In North America, a large portion of these forests are located in northern cold regions and has traditionally been managed for timber production. In Canada alone, 230 Mha of forests (~58% of total cover area) are managed for wood production. Many of these managed forests are in different stages of their development and their response to climate variability, extreme events such as drought and heat stresses and management regimes is not fully understood. In this study, eddy covariance flux measurements in an age sequence (78-, 43-, and 17-years old as of 2017) of conifer (white pine) plantations and a managed deciduous (>90-yr old oak) forest located in a temperate-boreal transition zone in southern Ontario, Canada are examined to determine the impact climate variability and extreme events. While in conifer stands, simultaneous occurrence of heat and drought stress in the early growing season was a major factor to cause a large decrease in annual net ecosystem productivity (NEP), in deciduous forest drought stress early in the growing season was a major factor for reduced NEP. In both forests, severity of heat and drought stress impacts was highly dependent on the timing of these events. This study helps to assess the vulnerability of managed forests to climate change and extreme weather events in cold regions of North America.

Influence of concurrence of extreme drought and heat events on carbon and energy fluxes in dominant ecosystems in the Pacific Northwest region
*H. Kwon, W. Creason, B. E. Law, C. J. Still, and C. Hanson
*, Oregon State University
Temperature in June 2015 was the highest recorded in the Pacific Northwest, USA and strongly coupled with low soil moisture. June is usually the best month for growth in the region. We examined the impact of the June 2015 climate extremes on carbon and energy fluxes at a high- desert sagebrush, semi-arid young and mature ponderosa pine forests, and a mesic Douglas-fir forest compared to an average climate year (2014). We assessed if the ecosystems recover from extreme climate stress within the growing season. In sagebrush, the carry-over effect of precipitation mitigated the immediate impact of extreme climate, leading to a 25–40% increase in net ecosystem production (NEP) and gross primary production (GPP) and a 65% increase in latent heat flux. The drought and heat lowered NEP by 35–65% and GPP by 15–33% in ponderosa pine and Douglas-fir. A greater increase in latent heat flux was observed in Douglas- fir than in ponderosa pine driven by increased evaporation. The decline in NEP was correlated with vapor pressure deficit. NEP recovered in October (the beginning of the rainy season) in ponderosa pine but partially recovered in Douglas fir, resulting in the largest seasonal reduction in carbon fluxes among sites (64–128 g C m-2 season-1). Our results suggest that the responses of carbon and energy fluxes to climate extremes differ depending on site- and species-specific characteristics. Given the likelihood of future drought and heat extremes, identifying these anomalous ecological responses to anomalous climate is critical to improve predictions of physiological thresholds and tolerance of different tree species.

Recovery of subalpine forest ecosystems following insect and fire disturbance
*John M. Frank, Mario Bretfeld, William J. Massman, Brent E. Ewers, Kate A. Dwire, Paula J. Fornwalt, Laurie S. Huckaby, John L. Korfmacher, José F. Negrón, Michael G. Ryan, Daniel Beverly, Andrew Parsekian, David E. Reed
*, USDA Forest Service
As forest disturbance has become more widespread and frequent, the subsequent recovery of these ecosystems has major ecological, economical, and sociological implications. Here, we present ongoing work at two AmeriFlux sites in southeastern Wyoming where insect disturbance and forest fire have dramatically changed the landscapes and altered ecosystem processes. At the GLEES site (US-GLE), in the decade following a spruce beetle outbreak, summertime carbon fluxes have returned to magnitudes similar to before. Yet, these fluxes are relatively suppressed in early spring, concurrent with a large snowpack that enshrouds small trees and understory plants, and late summer when understory plants senesce. As an explanation for this, repeated measurements of surviving trees and understory plants within the eddy covariance flux footprint show that over the decade mid-sized trees have grown considerably and vegetation in plots with high overstory mortality has flourished. In June 2018, the Chimney Park site (US-CPk), which a decade earlier had a mountain pine beetle outbreak, burned with mixed severity within the Badger Creek Fire. While much of the footprint of the tower burned with zero or low severity, the eddy covariance system was destroyed. In August/September 2018 a new flux scaffold was installed in an area of high-severity, stand-replacing fire ~100 m away from the US-CPk tower. Here, recovery will be quantified as this area progresses from a landscape driven by the surface energy balance and soil physical properties towards an ecosystem with understory plant recruitment and tree establishment.

Estimating evapotranspiration from weather station data
*Angela Rigden and Guido Salvucci
*, Harvard University
Evapotranspiration is a key component of the terrestrial surface energy and water balance. However, measurements of evapotranspiration are spatially and temporally sparse, limiting observation-based regional studies and trend analyses. Modeling evapotranspiration has also proved difficult due to, for example, land surface heterogeneities, complex feedbacks between the land and the atmosphere, and the need for numerous site-specific calibrated parameters. Here, we develop a model to estimate evapotranspiration based on the surface energy balance, Monin-Obukhov similarity theory, diffusion equations, and a new optimization principle, which determines the surface conductance to water vapor transport from vertical profiles of relative humidity. The key advantage of this approach is that biophysical and hydrological surface parameters are not required as inputs to estimate evapotranspiration. In developing the model, we synthesize eddy covariance measurements from the AmeriFlux network to calibrate sub-models of downwelling longwave radiation and ground heat flux, and to explore the dependence of thermal roughness length on land cover type and atmospheric states. Once formulated, we evaluate the evapotranspiration model at over 60 AmeriFlux sites, finding good agreement between observations and estimates across times scales, climates, vegetation types, and water/energy limitations. Since the evapotranspiration model relies primarily on meteorological data, it can also be applied at common weather stations, which are more spatially and temporally abundant. The ability to estimate evapotranspiration at weather stations has facilitated analyses on trend detection and attribution, as well as evapotranspiration partitioning.

The role of tree hydraulic traits in response to soil water availability
*Bohrer Gil; Mirfenderesgi, Golnazalsadat; Matheny, Ashley M.
*, Ohio State University

Hydraulic performance of plants is governed at the tissue level by traits that define the properties of conductive tissues, such as xylem, and control structures, such as stomata guard cells. At a higher level, the effects of these tissue-level traits integrate to emergent whole-tree traits that govern the apparent relationships between the tree hydrodynamics and soil and atmosphere environmental forcing. These emergent tree-level traits could be further upscaled across trees of similar structure to the canopy and forest plot scales. Provided a mechanistic description of plant hydrodynamics, these traits could be characterized using the parameters of the formulations describing the water flow through the integrated tree conductive system. Such traits include the characteristics of percent-loss of conductivity (PLC) curve, and the stomatal response curve to internal and external forcing, and root properties. The diversity of such traits produces a wide range of response strategies to both short-term variation of soil moisture and vapor-pressure-deficit (VPD), and to long-term changes to climate and hydrological cycles which affect water availability.

We observed and characterize different types of hydraulic “behaviour” using a large array of sap-flow sensors at the footprint of the UMBS flux towers. We used a set of virtual sensitivity experiments to determine the role of different hydraulic trait combinations that govern trees’ vulnerability to limitations in soil water availability. We use a quantitative hydrodynamic modeling framework which allows studying the influence of each suits of plant hydraulic traits independently, and assesses how the different trait groups interact with each other to form viable hydraulic strategies in response to reduced soil moisture availability.

We demonstrate that coordination between traits leads to distinct whole-tree hydraulic strategies. We demonstrate how the relationship between plant behavior at the iso/anisohydric continuum and hydraulic safety margin, which has been previously hypothesized and observed by others, emerges from this trait coordination and their tree-level outcomes. Finally, our results suggest that hydrodynamic models represent an exciting new possibility to define and study plant traits and hydrodynamics.

The effect of land-cover conversions on surface temperature in semi-arid ecosystems at the Southwestern United States
*Tomer Duman, Cheng-Wei Huang, and Marcy Litvak
*, University of New Mexico
Land-cover conversions are expected to increase in the coming decades, especially in drought sensitive regions, such as the Southwestern United States (SW), which are particularly sensitive to climate-driven mortality events that have already triggered large-scale structural changes across the region. One unknown is whether or not such structural changes are sufficient to alter local energy and water vapor exchange with the atmosphere, given that vegetation density and structure directly regulates surface reflectivity, roughness, and eco-physiological functioning. We used an 11-year continuous record of turbulent fluxes, aerodynamic and radiation measurements from seven sites in the New Mexico Elevation Gradient to explore expected changes in surface temperature associated with land-cover conversions in the SW. We used the energy balance equation to analytically separate the contribution of surface albedo and emissivity, soil heat flux, aerodynamic conductance, vegetation water conductance as well as their interactions, to the predicted increase or decrease of surface temperature following land-cover change. The results show that expected land cover conversions mostly trigger an increase in surface temperature. For example, an estimated 3o°C increase on average following the conversion of piñon-juniper woodlands to juniper savanna. The largest contribution to this predicted increase in temperature was due to decrease in eco-physiological functioning embedded in lower vegetation water conductance. However, aerodynamic changes following most land cover conversions had a compensating cooling effect on the system. Surprisingly, the contribution of changes in albedo to surface temperature was an order of magnitude smaller than these other components, though it was not negligible.

Testing conceptual and mechanistic soil hydrology model simulations against observed water and carbon fluxes across semi-arid sites in the southwestern US
*Natasha MacBean, Russ Scott, Joel Biederman, Thomas Kolb, Sabina Dore, Nicolas Vuichard, Agnes Ducharne, Philippe Peylin, and David J.P. Moore
*, Indiana University
Climate change may have a significant adversarial impact on water-limited semi-arid ecosystems, given that they will likely experience more intense droughts with an increasing contrast between wet and dry seasons. Predicting future trends in moisture availability in these ecosystems, with concurrent impacts on vegetation phenology and carbon fluxes, relies on the ability of land surface models (LSMs) to adequately capture the correct ecosystem-level processes. Yet, LSMs have rarely been evaluated against semi-arid ecosystem site-level data. Here, we evaluated time series of soil moisture and evapotranspiration simulated by the ORCHIDEE land surface model against observations from five Ameriflux sites across semi-arid grass, shrub and forest sites in Arizona, USA. We tested two hydrological schemes of differing complexity: (1) a simple conceptual 2-layer bucket scheme; and (2) a 11-layer mechanistic scheme based on Fokker-Planck equations. We found that the more complex 11-layer mechanistic hydrology scheme is much better able to capture the highly dynamic nature of semi-arid ecosystem soil moisture variability. The improved soil moisture results in a better match between simulated and observed evapotranspiration (ET). Simulated runoff and drainage are also more realistic with the mechanistic model. However, it is likely the ratio between evaporation and transpiration is biased high. Including a bare soil evaporation resistance term serves to lower the ratio, but also results in a negative impact on modeled ET, suggesting model structural deficits remain. In spite of this issue, we demonstrate the mechanistic hydrology model better captures observed photosynthetic flux response to periods of moisture stress.

Flashy, patchy and coupled: refined dryland carbon flux predictions developed from the North American Southwest reveal altered global dryland dynamics
*Mallory L. Barnes, Russell L. Scott, David. J.P. Moore, Guillermo Ponce Campos, Joel A. Biederman, Natasha MacBean, David. D. Breshears
*, Indiana University
Terrestrial plants play a crucial role in regional and global carbon cycling. Recent studies have highlighted the dominant role of semi-arid ecosystems in both the trend and interannual variability in the terrestrial carbon sink. Despite the importance of water-limited systems, current earth system models and remote-sensing-driven estimates of vegetation production do not adequately capture dryland carbon dynamics. Potential explanations for this phenomenon include spatial heterogeneity, tightly coupled water and carbon cycles, and rapid carbon uptake in response to precipitation pulses in drylands. We present a new product that fuses flux estimates with remotely sensed observations in a machine learning algorithm to estimate dryland GPP. Our machine learning model was driven by a dense network of eddy covariance sites spanning dryland ecosystem types and climate spaces and included a multiscalar drought index as a predictor. Improved characterization of intra- and inter- annual drought impacts on carbon uptake in our model resulted in more accurate predictions of inter-annual and seasonal variability in dryland carbon uptake. Our model trained in the Southwest provides more realistic estimates of drought impacts on carbon cycling in global drylands, underscoring the process-based underpinnings of the drought terms in our model. We anticipate our data-driven product to be a starting point for a more sensitive accounting of drought impacts on the global carbon cycle and an improved benchmark for earth system models in drylands. More generally, our results demonstrate that the well-established link between hydrologic and carbon cycles is likely crucial for accurate regional and global carbon modeling.

The infinite flux tower project: Introducing the Chequamegon Heterogeneous Ecosystem Energy-balance Study Enabled by a High-density Extensive Array of Detectors (CHEESEHEAD19)
Ankur Desai, University of Wisconsin-Madison,
For decades, long-term eddy-covariance observations of surface-atmosphere exchange have led to to better understanding of ecosystem functioning, climate feedbacks, and the carbon cycle. However, long-standing biases in surface energy balance closure and location bias limit eddy-covariance data from accurately sampling the regional scale of land surface processes and the corresponding atmospheric responses. These two challenges hamper our ability to better benchmark Earth system models with eddy-covariance data. Here, we introduce the Chequamegon Heterogeneous Ecosystem Energy-balance Study Enabled by a High-density Extensive Array of Detectors (CHEESEHEAD19), an June-October 2019 NSF-supported field experiment designed to specifically address energy- balance and location bias challenges by deploying a large number (~20) of flux towers and related surface in situ temote sensing and airborne atmospheric sampling within a single model grid cell (10×10 km). We will present initial model results demonstrating how spatial heterogeneity in surface fluxes can generate long-wave eddies that contribute to a large fraction (~40%) of energy imbalance and how CHEESEHEAD19 will test for this. Further, we demonstrate that several advanced flux computation methods across time, space, and frequency domains allow us to correct systematic bias in flux towers. Applying these approaches to scaling algorithms can lead to improved representation of the domain mean flux and its spatial distribution with an efficient number of towers. We are seeking collaborators who would like to complement our experiment.

The Indianapolis Flux Experiment (INFLUX): A city-sized, flow-through flux chamber
*Nikolay Balashov, Kenneth J. Davis and the INFLUX Science Team.
*, The Pennsylvania State University
The Indianapolis Flux Experiment (INFLUX) is urban test bed experiment aimed at testing, developing and improving our ability to quantify greenhouse gas (GHG) emissions from urban centers. Airborne- and tower-based GHG measurements collected around the city since 2011 have been used to estimate CO2 and CH4 emissions from the city using a variety of atmospheric mass balance methods. Micrometeorological measurements including Doppler lidar and eddy covariance flux towers provide critical information regarding the urban boundary layer and its development, essential information for urban-scale mass balance flux estimates. Recent work has shown methodological convergence on urban CO2 emissions to the level of 8%, multi-year evaluation of the CO2 and CH4 balance of the city, and the importance of understanding background conditions and the urban biosphere. This presentation will provide an overview of methods and results from INFLUX, and a discussion of future needs for urban GHG emissions studies.