Energy balance closure at FLUXNET sites
Introduction
The global proliferation of long-term eddy covariance sites measuring carbon and energy fluxes (FLUXNET) provides a unique contribution to the study of the environmental, biological and climatological controls of net surface exchange between vegetation and the atmosphere (Baldocchi et al., 2001). Independent methods of evaluating the reliability of the eddy covariance measurements at FLUXNET sites are highly desirable. One method of independently evaluating scalar flux estimates from eddy covariance is energy balance closure. Energy balance closure, a formulation of the first law of thermodynamics, requires that the sum of the estimated latent (LE) and sensible (H) heat flux be equivalent to all other energy sinks and sourceswhere Rn is the net radiation, G the heat flux into the soil substrate, S the rate of change of heat storage (air and biomass) between the soil surface and the level of the eddy covariance instrumentation, and Q the sum of all additional energy sources and sinks. Typically, Q is neglected as a small term, and an imbalance between the remaining independently measured terms on the left- and right-hand sides of Eq. (1) may indicate inaccurate estimates of scalar fluxes.
Energy balance closure is directly relevant to the evaluation of latent and sensible heat fluxes and not to other scalar fluxes (e.g. CO2 fluxes, the primary constituent for FLUXNET). However, at least four considerations make an analysis of energy balance closure at FLUXNET sites relevant and functional. First, although the source–sink distributions for water, heat and CO2 are different from each other in an ecosystem, the atmospheric transport mechanisms within and above the canopy, which are measured by eddy covariance, are similar for all scalars. Second, the computation of all scalar fluxes using the eddy covariance technique is founded on similar theoretical assumptions in the conservation equations and their Reynolds decomposition (Paw U et al., 2000). Thirdly, at FLUXNET sites all the terms in Eq. (1) are typically measured and recorded continuously at a high temporal resolution, providing a data set for independently evaluating eddy covariance fluxes across multiple time scales. Finally, an independent evaluation of energy balance closure across FLUXNET sites is useful for its own purposes, especially in view of the close biophysical coupling between carbon, energy and water fluxes (Collatz et al., 1991, Baldocchi and Meyers, 1998).
Historically, energy balance closure has been accepted as an important test of eddy covariance data (Anderson et al., 1984, Verma et al., 1986, Mahrt, 1998), and a number of individual sites within the FLUXNET network report energy balance closure as a standard procedure (e.g. Hollinger et al., 1999, Anthoni et al., 1999, Aubinet et al., 2000; Goldstein et al., 2000; Wilson and Baldocchi, 2000, Schmid et al., 2000). From these studies and many others, a general concern has developed within the micrometeorological community because surface energy fluxes (LE+H) are frequently (but not always) underestimated by about 10–30% relative to estimates of available energy (Rn−G−S). The imbalance is often present, though often to a lesser extent, even over flat, homogeneous surfaces and short vegetation (Stannard et al., 1994, Mahrt, 1998, Twine et al., 2000), which are presumably ideal conditions for the eddy covariance method.
An energy imbalance has implications on how energy flux measurements should be interpreted and how these estimates should be compared with model simulations (Liu et al., 1999, Twine et al., 2000, Culf et al., 2002). Within the FLUXNET community, the consequence of a possible widespread energy imbalance on the interpretation of CO2 fluxes is somewhat less certain and is dependent on the mechanisms creating the imbalance. General hypotheses have been suggested to account for the lack of energy balance closure, including: (1) sampling errors associated with different measurement source areas for the terms in Eq. (1), (2) a systematic bias in instrumentation, (3) neglected energy sinks, (4) the loss of low and/or high frequency contributions to the turbulent flux and (5) neglected advection of scalars. If the energy imbalance results from different measurement source areas for the terms in Eq. (1), or bias errors in net radiation, or neglected energy sinks and sources, there may be no reason to suspect that the measured CO2 fluxes are systematically inaccurate. However, if the mechanisms creating the imbalance result from inappropriate assumptions made from the scalar conservation equation, such as neglecting mean advection (Paw U et al., 2000), or result from low or high frequency loss of flux, the estimated CO2 flux may contain similar errors in relation to the actual terrestrial exchange. An overall evaluation of site characteristics and meteorological conditions associated with the energy imbalance may provide circumstantial evidence for its cause, and also suggest whether similar errors are likely in the CO2 flux estimates.
Because of the large geographic and biological variability at the contributing sites, FLUXNET offers a unique data set to compare energy balance closure for different vegetation types, terrain features and climates. Previous studies have focused on energy balance closure at multiple sites (Stannard et al., 1994, Barr et al., 1994, Aubinet et al., 2000, Twine et al., 2000), but FLUXNET encompasses a larger quantity of sites and with extensive variability in site characteristics. Although diverse instrumentation and processing methods in FLUXNET increase the possible sources affecting energy balance closure, instrumentation differences also present an opportunity to examine the effect of available techniques, especially potential differences between open and closed path infrared gas analyzers (IRGAs). Because FLUXNET sites typically measure continuously over annual cycles, an evaluation of closure over different seasons is possible, which has not generally been reported in other synthesis studies of energy balance closure (Stannard et al., 1994, Mahrt, 1998, Twine et al., 2000).
The goals of this study are: (1) to evaluate energy balance closure at all FLUXNET sites, and (2) to evaluate the potential source of differences in the imbalance within and between sites. Based on these considerations, (3) we will also discuss the relationship between energy balance closure and the consequences for interpreting CO2 fluxes.
Section snippets
Materials and methods
Methodology and review papers for many of the FLUXNET eddy covariance sites are published in Aubinet et al. (2000) and Baldocchi et al. (2001). Eddy covariance and the supporting environmental and meteorological data were contributed by individual investigators to the FLUXNET database at Oak Ridge National Laboratory’s Data Archive Center. The data set contains annual files of half-hourly flux and meteorological data at eddy covariance stations across Europe and North America (Table 1) and
Overall energy balance closure
Regression coefficients of LE+H against Rn−G−S, using OLSs on all the half-hour data for each of the 50 site-years, are shown in Table 2. The slope was less than 1 for all site-years, ranging from 0.53 to 0.99, with a mean of 0.79±0.01. The intercept ranged from −32.9 to 36.9 W m−2, with a mean of 3.7±2.0 W m−2. There were more positive (31) than negative intercepts (19). The mean coefficient of determination (r2) was 0.86, ranging from 0.64 to 0.96.
Using the RMA approach, which accounted for
Discussion
Although the intercepts of the OLS regressions for the site-years were typically non-zero, the OLS slope (for positive and negative intercepts) was always less than 1. The EBR was less than 1 for over 80% of the site-years, further suggesting a general lack of closure. The general lack of complete energy balance closure at FLUXNET sites (linear regression slope and EBR typically less than 1) is consistent with historical evidence that the energy balance is often not closed using the eddy
Conclusions
- 1.
There was a general lack of energy balance closure at FLUXNET sites, with the scalar fluxes of sensible and LE being underestimated and/or available energy being overestimated. The mean imbalance was in the order of 20%.
- 2.
Energy balance closure is typically poor during nocturnal periods, especially when the turbulent mixing is weak. On average, energy balance closure was better in the afternoon than in the morning, possibly suggesting the underestimation of storage terms, which are usually larger
Acknowledgements
This work is the result of the FLUXNET (NASA) Workshop in Marconi, CA, USA (June 2000). It is also a contribution to the AmeriFlux program (US Department of Energy’s Terrestrial Carbon Program). We thank Bruce Hicks, Marc Pitchford and Jose Fuentes for reviewing and improving this manuscript.
References (70)
- et al.
Carbon and water vapor exchange of an open-canopied ponderosa pine ecosystem
Agric. For. Meteorol.
(1999) - et al.
Variation in net radiation over heterogeneous surfaces: measurements and simulations in a juniper-sagebrush ecosystem
Agric. For. Meteorol.
(2000) - et al.
On using eco-physiological, micrometeorological and biogeochemical theory to evaluate carbon dioxide, water vapor and trace gas fluxes over vegetation: a perspective
Agric. For. Meteorol.
(1998) - et al.
On measuring and modeling energy fluxes above the floor of a homogeneous and heterogeneous conifer forest
Agric. For. Meteorol.
(2000) - et al.
Regulation of stomatal conductance and transpiration: a physiological model of canopy processes
Agric. For. Meteorol.
(1991) - et al.
The energy and water budgets of an area of patterned woodland in the Sahel
Agric. For. Meteorol.
(1993) - et al.
Bowen ratio, eddy correlation, and portable chamber measurements of sensible and latent heat flux over irrigated spring wheat
Agric. For. Meteorol.
(1991) - et al.
Gap filling strategies for long term energy flux data sets
Agric. For. Meteorol.
(2001) - et al.
Effects of climate variability on the carbon dioxide, water, and sensible heat fluxes above a ponderosa pine plantation in the Sierra Nevada (CA)
Agricultural and Forest Meteorology
(2000) - et al.
Inconsistencies in net radiation estimates from use of several models of instruments in a desert environment
Agric. For. Meteorol.
(1998)
Variability in soil heat flux from a mesquite dune site
Agric. For. Meteorol.
Validation of eddy flux measurements above the understorey of a pine forest
Agric. For. Meteorol.
A simple method for estimating frequency response corrections for eddy covariance systems
Agric. For. Meteorol.
Soil surface heat flux: some general questions and comments on measurements
Agric. For. Meteorol.
Experimental design for flux measurements: matching scales of observations and fluxes
Agric. For. Meteorol.
Measurements of CO2 and energy fluxes over a mixed hardwood forest in the mid-western United States
Agric. For. Meteorol.
Correcting eddy-covariance flux underestimates over a grassland
Agric. For. Meteorol.
Calibration of soil heat flux sensors
Agric. For. Meteorol.
Thermal properties for vineyard (EFEDA-I) and savanna (HAPEX-Sahel) sites
Agric. For. Meteorol.
Seasonal and interannual variability of energy fluxes over a broadleaved temperate deciduous forest in North America
Agric. For. Meteorol.
Factors controlling evaporation and energy balance partitioning beneath a deciduous forest over an annual cycle
Agric. For. Meteorol.
A comparison of methods for determining forest evapotranspiration and its components: sap flow, soil water budget, eddy covariance and catchment water balance
Agric. For. Meteorol.
Eddy correlation measurements of CO2, latent heat and sensible heat fluxes over a crop surface
Bound.-Lay. Meteorol.
Heat storage in a young eucalypt forest
Agric. For. Meteorol.
Estimates of the annual net carbon and water exchange of European forests: the EUROFLUX methodology
Adv. Ecol. Res.
A comparison of Bowen ratio and eddy-correlation sensible and latent-heat flux measurements above deciduous forest
Bound.-Lay. Meteorol.
Long-term carbon dioxide fluxes from a very tall tower in a northern forest
J. Atmos. Ocean. Technol.
Turbulent flux measurements above and below the overstory of a boreal aspen forest
Bound.-Lay. Meteorol.
Evaluation of the accuracy with which dry deposition can be measured with current micrometeorological techniques
J. Clim. Appl. Meteorol.
Calibration, comparison, and correction of net radiation instruments used during FIFE
J. Geophys. Res.
Cited by (1928)
Multimodel ensemble estimation of Landsat-like global terrestrial latent heat flux using a generalized deep CNN-LSTM integration algorithm
2024, Agricultural and Forest MeteorologyModelling soil moisture and daily actual evapotranspiration: Integrating remote sensing surface energy balance and 1D Richards equation
2024, International Journal of Applied Earth Observation and GeoinformationSpatial-temporal patterns of land surface evapotranspiration from global products
2024, Remote Sensing of EnvironmentA coupled hourly water-carbon flux model at plot and field scales for water-saving irrigated rice paddy
2024, Agricultural Water ManagementDrought effects on evapotranspiration and energy exchange over a rain-fed maize cropland in the Chinese Loess Plateau
2024, Agricultural Water Management