You are here

Harvard Forest Data Archive

HF072

Net Carbon Exchange of a Young Upper-Slope Deciduous Forest at Harvard Forest LPH Tower since 2002

Related Publications

Data

Overview

  • Lead: Julian Hadley
  • Investigators: David Foster, Paul Kuzeja, Liza Nicoll, Jessica Schedlbauer, Mark VanScoy
  • Contact: Mark VanScoy
  • Start date: 2002
  • End date: 2010
  • Status: completed
  • Location: Prospect Hill Tract (Harvard Forest)
  • Latitude: +42.542
  • Longitude: -72.185
  • Elevation: 380 meter
  • Taxa: Acer rubrum (red maple), Betula alleghaniensis (yellow birch), Betula lenta (black birch), Betula papyrifera (white birch), Pinus resinosa (red pine), Pinus strobus (white pine), Quercus alba (white oak), Quercus rubra (red oak), Quercus velutina (black oak), Tsuga canadensis (eastern hemlock)
  • Release date: 2012
  • Revisions:
  • EML file: knb-lter-hfr.72.25
  • DOI: digital object identifier
  • Related links:
  • Study type: long-term measurement
  • Research topic: forest-atmosphere exchange
  • LTER core area: primary production, inorganic nutrients
  • Keywords: carbon dioxide, eddy covariance, heat flux, relative humidity, photosynthesis, respiration
  • Abstract:

    This data set contains sensible heat exchange, water vapor exchange and carbon exchange as well as environmental data for a deciduous forest dominated by red oak (Quercus rubra). It is 1.1 km WNW of the EMS tower where continuous eddy covariance measurements began in 1992 (see HF004). The High Deciduous site is about 385 m a.s.l., or 35 m higher in elevation than the EMS, which is situated in a relatively low area near a stream.

    The forest near this eddy covariance tower is broadly similar in species composition to the EMS site, but it is younger and shorter in stature. The site was cleared for pasture, but not deeply plowed or planted, in the 18th and19th centuries. Agriculture on the site was abandoned near the end of the 19th century. The forest within 200 to 300 m of the eddy covariance tower to the NW, W, SW, and S burned in an intense fire in 1957, which left few or no surviving trees.

  • Methods:

    Constraints on valid eddy covariance data caused by topography and turbulence

    The eddy covariance tower is on the west to northwest slope of Little Prospect Hill, which has a long SW-NE axis. The tower is about 150 m from the ridge top, on a slope with an 8 to 10% grade. When wind is from the E or S, lee-slope turbulence creates conditions unsuitable for valid eddy covariance measurements. This can be seen from unreasonably high apparent C fluxes observed when the compass bearing of wind direction is less than 215 degrees, or slightly S of SW. Therefore, all flux data are filtered to exclude wind directions less than 215 degrees, and fluxes during these periods are estimated from single or multiple regression models derived from other environmental data. The predictor variables in these models are soil and air temperature for nighttime fluxes, and photosynthetically active radiation (PAR) as well as temperatures and time of day for daytime fluxes.

    A friction velocity (u*) filter is also applied to eddy covariance data before calculating carbon exchange, because of evidence that CO2 is removed from the forest understory by nocturnal cold air subsidence during periods of low turbulence. This could lead to serious underestimation of ecosystem respiration from eddy covariance data during these periods. A careful analysis of nighttime C flux for 2002 and 2003 showed that a minimum friction velocity (u*) of 0.35 m/s was necessary to avoid underestimates of forest-to-atmosphere C flux at night. A minimum u* of 0.35 m/s is applied to both nighttime and daytime data. This relatively high u* threshold compared to the Harvard Forest Main Tower site (u* = 0.2 m/s) is likely due to the steeper slope on which the High Deciduous site is located.

    CO2 profile measurements at this site were begun in 2008. These measurements show a substantial storage of CO2 in the sub-canopy air space during the night hours, which is not subsequently detected as a release during the early morning hours, and occurs even at u* greater than 0.35 m/s. Therefore, it appears that there is a storage term, which must be added to the measured C flux in order to get a realistic estimate of net ecosystem C exchange. Until measurements of this nighttime CO2 storage term, or estimated values for it from modeling, can be added to the data, the flux data can not be used to calculate ecosystem respiration or any accurate 24-hour or long term carbon exchange. Daytime C flux measurements are a close approximation of daytime total carbon exchange, since measured daytime subcanopy CO2 storage (positive or negative) is very small.

    Eddy covariance system and environmental sensors

    Eddy covariance data are collected with a Campbell Scientific CSAT-3 sonic anemometer and a Licor LI-6262 closed path CO2/H2O analyzer. Data are recorded at 5 hz and the flow rate is 5.7 l/min. A 600 s running mean filter is used in covariance calculations. CO2 fluxes are not spectrally corrected, as tests during the first year showed that the appropriate smoothing time (0.2 s) did not significantly change the flux values. H2O fluxes are spectrally corrected using a by smoothing the sonic temperature data by 1.6s, calculating a degraded heat flux value after this procedure and multiplying the original calculated H2O flux by the ratio of the raw to the smoothed heat flux. This typically increases the calculated H2O flux by 15 to 20%.

    The system was run side-by-side with the eddy covariance system at the EMS site for three days during September 2001. C flux and sensible heat fluxes measured were nearly identical except for a 7% lower vertical wind measurement by the e.c. system used at the High Deciduous site, which was reflected in net C fluxes lower by a similar percentage. The CO2/H2O analyzer for the High Deciduous site is calibrated weekly using CO2-free nitrogen and a commercial CO2 standard of about 490 ppm. Two such standards have been used since the inter-calibration with the e.c. system at the EMS site, and the two standards were cross-calibrated to ensure consistency.

    Above canopy (AC) air temperature data is from a user-constructed shaded and passively ventilated thermocouple for data from August 2002 through mid-July 2003, and thereafter from a Campbell Scientific passively ventilated air temperature/relative humidity sensor.

    Soil temperature is an average of 4 thermocouples randomly placed at 10 cm depth within 30 m of the tower base.

    Understory air temperatures at 20 cm and 1 m above ground (measured only after Day 308 in 2003) are from bare thermocouples that are shielded from direct solar illumination during the mid-day period, but not at all times. They can only be assumed to give a fairly accurate estimate of air temperature at night, and were installed to detect nocturnal air temperature inversions connected with advective removal of CO2 from the forest understory.

    Net radiation data are only available for the site during summer 2002 and beginning again in November 2004. Data from an aspirated high precision thermometer are available for about 1 month during 2002, and will be posted at a later date.

    Further information on the data

    As of this point (February 2005) it has not been possible to measure subcanopy CO2 storage at this site. Therefore, only FCO2 data are reported, and not NEE. However, integration of the FCO2 values over time should approximate actual NEE, since subcanopy storage of CO2 is transient. The relatively short stature of the canopy at the High Deciduous site (about 16 m compared to 25 m at the EMS site) and the high minimum u* for valid data also reduce the magnitude of the storage term.

    Ecosystem respiration ( R ) is estimated using a statistical model for nighttime FCO2, with soil temperature and/or air temperature as the predictor variables. Gross Ecosystem Exchange (GEE) is calculated by subtracting R from FCO2. Since FCO2 may be slightly different from NEE due to subcanopy CO2 storage, GEE is approximate, but when integrated over time it should closely approximate the true value.

    Similarly to carbon flux, sensible heat, latent heat and water vapor fluxes are also only accepted as valid if the compass bearing of wind direction is 215 degrees or greater. However, no u* threshold is applied to sensible heat, latent heat or water fluxes.

  • Use:

    This dataset is released to the public under Creative Commons license CC BY (Attribution). Please keep the designated contact person informed of any plans to use the dataset. Consultation or collaboration with the original investigators is strongly encouraged. Publications and data products that make use of the dataset must include proper acknowledgement.

  • Citation:

    Hadley J. 2012. Net Carbon Exchange of a Young Upper-Slope Deciduous Forest at Harvard Forest LPH Tower since 2002. Harvard Forest Data Archive: HF072.

Detailed Metadata

hf072-01: eddy flux (2002 to 2004)

  1. year: year
  2. doy: day of the year with hours and minutes converted to a decimal fraction of a day (unit: nominalDay )
  3. co2: carbon dioxide concentration of the air drawn into the eddy covariance system at 21 m above ground or 4.5 m above the average tree canopy top (unit: dimensionless / missing value: NA)
  4. h2o: water vapor concentration of the air drawn into the eddy covariance system at 21 m above ground or 4.5 m above the average tree canopy top (unit: dimensionless / missing value: NA)
  5. u: wind speed measured by the sonic anemometer at 21 m or 4.5 m above the average tree canopy top (unit: metersPerSecond / missing value: NA)
  6. ustar: friction velocity measured by the sonic anemometer. Friction velocity is the square of momentum flux from the atmosphere above the sonic to the air layers below the sonic, and is a measure of atmospheric turbulence (unit: metersPerSecond / missing value: NA)
  7. wind.dir: compass direction in degrees of the average wind vector, with 0 and 360 degrees indicating geographic north (unit: degree / missing value: NA)
  8. h: sensible heat flux from the forest to the atmosphere, calculated by the sonic anemometer from the covariance of air temperature and the vertical component of wind velocity (unit: wattPerMeterSquared / missing value: NA)
  9. le: flux of latent heat (heat used in evaporating water) from the forest to the atmosphere, calculated by multiplying FH2O (see definition below) by the heat of evaporation of water (unit: wattPerMeterSquared / missing value: NA)
  10. fco2: measured carbon dioxide flux from forest to atmosphere. Includes all data collected, some of which are invalid (see below) (unit: micromolePerMeterSquaredPerSecond / missing value: NA)
  11. fco2.valid: FCO2 data after removal of data that are invalid due to wind direction or low atmospheric turbulence (see Methods description) (unit: micromolePerMeterSquaredPerSecond / missing value: NA)
  12. nee.value: best estimate of FCO2, using either valid measurement from the column to the left, or a model estimate (unit: micromolePerMeterSquaredPerSecond / missing value: NA)
  13. nee.data: binary variable identifying whether measured FCO2 or an estimate was used
    • 1: data point
    • 0: estimate
    • NA: NA
  14. r.estimate: estimated ecosystem respiration. This is equal to the measured FCO2 at night, if the wind direction is greater than 215 and ustar is greater than 0.35 m/s. Under all other circumstances it is an estimate of CO2 production by the ecosystem, based on a statistical model that uses soil and air temperatures and valid nighttime FCO2 values to predict FCO2 under other circumstances (unit: micromolePerMeterSquaredPerSecond / missing value: NA)
  15. gee.estimate: estimate of gross carbon fixation by the forest, calculated difference between NEE and R (unit: micromolePerMeterSquaredPerSecond / missing value: NA)
  16. fh2o: measured water vapor flux from forest to atmosphere. Includes all data collected, some of which are invalid (see below) (unit: millimolePerMeterSquaredPerSecond / missing value: NA)
  17. fh2o.valid: FH2O data after removal of data that are invalid due to wind direction (see Methods description). Low turbulence data are not removed as the H2O flux during the foliated season for the trees is from the canopy, which therefore does not act as a barrier to movement of H2O, as it does for the large amount of CO2 produced by soil and forest-floor litter. When trees do not have foliage, the leafless canopy is not a significant barrier to evaporation of water from the soil, which occurs primarily when the ground and air near the ground are warm, generating upward movement of air. (unit: millimolePerMeterSquaredPerSecond / missing value: NA)
  18. sonic.tair: air temperature estimated from the speed of sound measured by the sonic anemometer. This estimate is based on air density which is directly related to the speed of sound. The sonic air temperature estimate can differ from actual air temperature by a few degrees due to the presence of water vapor, which lowers air density. (unit: celsius / missing value: NA)
  19. tair.above.canopy: air temperature measured at 23 m height, or about 6.5 m above the average tree canopy top (unit: celsius / missing value: NA)
  20. rh.above.canopy: relative humidity measured at 23 m height (same location as Tair.above.canopy) (unit: dimensionless / missing value: NA)
  21. vpd.above.canopy: water vapor pressure deficit (equals saturation water vapor pressure at Tair.above.canopy minus actual water vapor pressure) (unit: kilopascal / missing value: NA)
  22. tsoil.10cm: soil temperature measured at 10 cm depth. The average of 3 to 4 values at randomly located points within 15 m of the flux tower base. (unit: celsius / missing value: NA)
  23. par: photosynthetically active radiation measured at 22 m height, or about 5.5 m above the average tree canopy top (unit: micromolePerMeterSquaredPerSecond / missing value: NA)
  24. net.radiation: net radiant energy exchange, all wavelengths, measured by a sensor about 27 m above ground, or 10.5 m above the average tree canopy top (unit: wattPerMeterSquared / missing value: NA)
  25. understory.tair.20cm: air temperature 20 cm above the soil surface. The average of 2 values at randomly located points within 15 m of the tower. This data is not accurate if the sensors receive direct sunlight. The data is used to identify low-turbulence periods at night or near dawn or dusk when air temperature near the soil is significantly lower than above the canopy, contributing to cold air drainage and invalidating FCO2 and FH2O measurements. (unit: celsius / missing value: NA)
  26. understory.tair.1m.: air temperature 1 m above the soil surface. The average of 2 values at randomly located points within 15 m of the tower. Also not accurate if the sensors receive direct sunlight. (unit: celsius / missing value: NA)

hf072-02: eddy flux (2005 to 2010)

  1. year: year
  2. doy: day of the year with hours and minutes converted to a decimal fraction of a day (unit: nominalDay )
  3. day.night: Indicates whether data was collected during a 30-minute period when average PAR was greater than 5 µmol m-2 s-1 (Day) or less than 5 µmol m-2 s-1 (Night).
    • Day: day
    • Night: night
  4. co2: carbon dioxide concentration of the air drawn into the eddy covariance system at 21 m above ground or 4.5 m above the average tree canopy top (unit: dimensionless / missing value: NA)
  5. h2o: water vapor concentration of the air drawn into the eddy covariance system at 21 m above ground or 4.5 m above the average tree canopy top (unit: dimensionless / missing value: NA)
  6. u: wind speed measured by the sonic anemometer at 21 m or 4.5 m above the average tree canopy top (unit: metersPerSecond / missing value: NA)
  7. ustar: friction velocity measured by the sonic anemometer. Friction velocity is the square of momentum flux from the atmosphere above the sonic to the air layers below the sonic, and is a measure of atmospheric turbulence (unit: metersPerSecond / missing value: NA)
  8. wind.dir: compass direction in degrees of the average wind vector, with 0 and 360 degrees indicating geographic north (unit: degree / missing value: NA)
  9. h: sensible heat flux from the forest to the atmosphere, calculated by the sonic anemometer from the covariance of air temperature and the vertical component of wind velocity (unit: wattPerMeterSquared / missing value: NA)
  10. le: flux of latent heat (heat used in evaporating water) from the forest to the atmosphere, calculated by multiplying FH2O (see definition below) by the heat of evaporation of water (unit: wattPerMeterSquared / missing value: NA)
  11. fco2.reported: measured carbon dioxide flux from forest to atmosphere. Includes all data collected, some of which are invalid (see below) (unit: micromolePerMeterSquaredPerSecond / missing value: NA)
  12. fco2.valid: FCO2 data after removal of data that are invalid due to wind direction or low atmospheric turbulence (see Methods description) (unit: micromolePerMeterSquaredPerSecond / missing value: NA)
  13. fco2.value: best estimate of FCO2, using either valid measurement from the column to the left, or a model estimate (unit: micromolePerMeterSquaredPerSecond / missing value: NA)
  14. fco2.data: binary variable identifying whether measured FCO2 or an estimate was used
    • 0: estimate
    • 1: measured FCO2
  15. fh2o: measured water vapor flux from forest to atmosphere. Includes all data collected, some of which are invalid (see below) (unit: millimolePerMeterSquaredPerSecond / missing value: NA)
  16. fh2o.valid: FH2O data after removal of data that are invalid due to wind direction (see Methods description). Low turbulence data are not removed as the H2O flux during the foliated season for the trees is from the canopy, which therefore does not act as a barrier to movement of H2O, as it does for the large amount of CO2 produced by soil and forest-floor litter. When trees do not have foliage, the leafless canopy is not a significant barrier to evaporation of water from the soil, which occurs primarily when the ground and air near the ground are warm, generating upward movement of air. (unit: millimolePerMeterSquaredPerSecond / missing value: NA)
  17. sonic.tair: air temperature estimated from the speed of sound measured by the sonic anemometer. This estimate is based on air density which is directly related to the speed of sound. The sonic air temperature estimate can differ from actual air temperature by a few degrees due to the presence of water vapor, which lowers air density. (unit: celsius / missing value: NA)
  18. tair.above.canopy: air temperature measured at 23 m height, or about 6.5 m above the average tree canopy top (unit: celsius / missing value: NA)
  19. rh.above.canopy: relative humidity measured at 23 m height (same location as Tair.above.canopy) (unit: dimensionless / missing value: NA)
  20. vpd.above.canopy: water vapor pressure deficit (equals saturation water vapor pressure at Tair.above.canopy minus actual water vapor pressure) (unit: kilopascal / missing value: NA)
  21. tsoil.10cm: soil temperature measured at 10 cm depth. The average of 3 to 4 values at randomly located points within 15 m of the flux tower base. (unit: celsius / missing value: NA)
  22. par: photosynthetically active radiation measured at 22 m height, or about 5.5 m above the average tree canopy top (unit: micromolePerMeterSquaredPerSecond / missing value: NA)
  23. net.radiation: net radiant energy exchange, all wavelengths, measured by a sensor about 27 m above ground, or 10.5 m above the average tree canopy top (unit: wattPerMeterSquared / missing value: NA)
  24. understory.tair.20cm: air temperature 20 cm above the soil surface. The average of 2 values at randomly located points within 15 m of the tower. This data is not accurate if the sensors receive direct sunlight. The data is used to identify low-turbulence periods at night or near dawn or dusk when air temperature near the soil is significantly lower than above the canopy, contributing to cold air drainage and invalidating FCO2 and FH2O measurements. (unit: celsius / missing value: NA)
  25. understory.tair.1m.: air temperature 1 m above the soil surface. The average of 2 values at randomly located points within 15 m of the tower. Also not accurate if the sensors receive direct sunlight. (unit: celsius / missing value: NA)