UBC Faculty Research and Publications

Spatial and temporal variability of CO2 concentration and flux in a boreal aspen forest Yang, P. C.; Black, T. Andrew; Neumann, Herman H.; Novak, M. D.; Blanken, P. D. Nov 30, 1999

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-Blacki_AGU_1999JD900295.pdf [ 736.36kB ]
JSON: 52383-1.0041907.json
JSON-LD: 52383-1.0041907-ld.json
RDF/XML (Pretty): 52383-1.0041907-rdf.xml
RDF/JSON: 52383-1.0041907-rdf.json
Turtle: 52383-1.0041907-turtle.txt
N-Triples: 52383-1.0041907-rdf-ntriples.txt
Original Record: 52383-1.0041907-source.json
Full Text

Full Text

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 104, NO. D22, PAGES 27,653-27,661, NOVEMBER 27, 1999 Spatial and temporal variability of COz concentration and flux in a boreal aspen forest P. C. Yang, • T. A. Black, • H. H. Neumann, 2 M.D. Novak, • and P. D. Blanken 3 Abstract. In conjunction with eddy covariance measurements of CO2 fluxes at the 39.5-m height over a 21.5-m-tall boreal aspen stand in northern Saskatchewan, CO2 concentration was measured at eight heights in order to calculate net ecosystem exchange. During both leafless and full-leaf peri- ods, daytime vertical CO2 concentration gradients above 9rn were weak (< 0.2 gmol mo1-1 m't), but were strong below this height. Little change in CO2 storage in the air column below 39.5 rn oc- curred during much of the daytime, while around sunrise and sunset CO2 storage changed mainly below 9 m. For the rest of the night, over 85% of the increase in CO2 storage occurred above 9 m. On some calm nights during the growing season, CO2 also accumulated below 9 rn resulting in a sudden upward CO2 flux at 39.5 rn following the resumption of mixing 2-3 hours after sunrise. A 1 O-day experiment was conducted to determine the spatial variability of CO2 flux in the trunk space. Two eddy covariance systems were mounted just above the understory about two tree heights apart. The correlation between CO2 fluxes were poor even under unstable (daytime) conditions, uggesting a relatively heterogeneous understory and soil. In contrast, thecorrelation between water vapor fluxes was high (r = 0.70) in unstable conditions. However, average daytime and nighttime CO2 fluxes over the 10 days agreed to within 5%. This suggests that partitioning net ecosystem exchange between overstory and understory on an hourly basis using a single-understory eddy covariance system is inadvisable; however, partitioning probably can be done quite reliably using 5-day average fluxes. 1. Introduction The eddy covariance technique has been widely and success- fully used in studies of CO2 exchange between various forests and the atmosphere [ .g., Verma et al., 1986; Wofsy et al., 1993; Hol- linger et al., 1994, 1998; Grace et al., 1995; Fan et al., 1995; Baldocchi and Vogel, 1996; Black et al., 1996; Jarvis et al., 1997]. Some researchers have also successfully used this tech- nique to study CO2 exchange above the forest floor (e.g., Baldoc- chi and Meyers [1991] in an oak-hickory forest in Tennessee, Lee et al. [1994] in a hemlock-Douglas-fir forest in coastal British Columbia, Black et al. [1996] in a boreal aspen forest, and Bal- docchi et al. [1997] in a boreal jack pine forest). The relatively small contributions of small eddies to turbulent transport above forests make it feasible to use the eddy covariance technique to measure CO2 fluxes above the forest, even using a closed-path infrared gas analyzer (IRGA) system with a relatively long sam- pling tube [Lee et al., 1999]. Large, intermittent eddy structures that penetrate through the overstory enable eddy covariance measurements to be made near the forest floor [Baldocchi and Meyers, 1991; Blanken et al., 1998]. Furthermore, cospectral analysis has shown that turbulent transfer mechanisms in the 1Faculty ofAgricultural Sciences, University ofBritish Columbia, Vancouver, Canada. 2Atmospheric Environment Service, Downsview, Ontario, Canada. 3Department of Geography and Environmental Studies, University of Colorado, Boulder. Copyright 1999 by the American Geophysical Union. Paper number 1999JD900295. 0148-0227/99/1999JD900295509.00 trunk space are similar to those above the forest at the Old Aspen (OA) site in this study [Blanken et al., 1998]. One of the important uses of the CO2 flux measured using the eddy covariance technique (eddy CO2 flux F½) is to estimate the net ecosystem exchange (NEE) between the atmosphere and the soil or vegetation [e.g., Wofsy et al., 1993; Black et al., 1996' Greco and Baldocchi, 1996; Jarvis et al., 1997; Hollinger et al., 1998]. This exchange, which is often referred to as the biotic flux, can be expressed as NEE = F½ + ASa/At, where Z•a/At is the change of the CO2 storage in the air column beneath the eddy flux sensors. The estimation of ASa/At requires half-hourly measure- ments of CO2 concentration at various heights between the eddy covariance instruments and the ground (referred to as the CO2 concentration profile). Several questions need to be answered in carrying Out this procedure. First, how does the distribution of CO2 beneath the eddy covariance sensors depend on time of day and atmospheric onditions? Second, how many sampling levels in the profile system are needed to properly estimate •a/•t? Third, how important are the half-hourly changes of CO2 storage in the air column in determining the pattern of NEE during the day? The answers to these questions are important for obtaining reliable diurnal NEE patterns at forest CO2 flux measurement sites. A related question is how spatially variable is the CO2 flux above the understory? The answer to this question will help de- termine the feasibility of subtracting CO2 fluxes measured at one or more locations in the trunk space from those above the forest to provide an estimate of •he CO2 source or sink strength of the overstory. The flux measured at a single location above the un- derstory is representative of a much smaller area than that of a flux measured above the forest. For example, Blanken [1997], using the footprint model developed by Schuepp et al. [1990], found that at the OA site the upwind distance where the vegeta- tion makes the maximum contribution to the measured flux at the 27,653 27,654 YANG ET AL.' VARIABILITY OF COz FLUX IN AN ASPEN FOREST 39.5-m height was 100 and 300 m for typical daytime and night- time conditions, respectively. In contrast, the corresponding val- ues for the 4-m height (in the trunk space) were about 10 and 40 m. Using a Lagrangian random walk model, Baldocchi [1997] found that flux footprints beneath the overstory were even more contracted when horizontal wind velocity fluctuations were con- sidered. This means that the validity of subtracting the two fluxes is questionable when there is marked horizontal variability of the vertical fluxes in the trunk space. One way of assessing this is to determine the horizontal variability of eddy fluxes in the trunk space. Therefore the specific objectives of this paper are (1) to learn how the vertical distribution of CO2 below the eddy covariance sensors depends on time of day and atmospheric conditions, (2) to determine how the accuracy of the estimation of ASa/At de- pends on the number of sampling heights in the concentration profile system, (3) to assess the relative importance of ASa/At in the process of CO2 exchange between the atmosphere and the for- est, and (4) to assess the horizontal variability of COz flux in the lower trunk space above the understory. 2. Experimental Methods 2.1 Site Description The OA forest site (53.63øN, 106.20øW) is located in the southern part of Prince Albert National Park, Saskatchewan, Can- ada. The site is an even-aged stand of trembling aspen (Populus tremuloides Michx.) with scattered balsam poplar (Populus bal- samifera L.). In 1994, the stand was 70 years old, mean canopy height was 21.5 m, diameter at the 1.3-m height was 20 cm, and stem density was 830 stems per hectare. The trunk space with almost no branches extended up to about 15 m. The understory was mainly composed of 2-m-tall hazelnut (Corylus cornuta Marsh.) with occasional clumps of alder (Alnus crispa (Ait.) Pursch) and sparse shrubs (e.g., prickly rose, Rosa acicularis Lindl.). The soil, an Orthic Gray Luvisol, has a surface organic layer about 8-10 cm thick above a silty-clay textured subsoil. The topography is relatively level, and the fetch is at least 3 km in all directions. More detailed site descriptions can be found in the work of Black et al. [1996] and Chen et al. [1999]. 2.2 COz Flux and Concentration-Profile Measurements During late fall of 1993 and much of 1994, COz concentra- tions were measured at 0.8, 2.3, 9.5, 15.7, 18.8, 21.9, 25.0, and 34.2 m above the ground. With an eight-input manifold, a 200 L rain -• rotary pump drew air down a 9.3-ram inner diameter (ID) Dekoron tube from each height (25 L min '• per height) o a data logging hut. The tubes were heated to prevent condensation us- ing a 22-gauge bare nichrome wire passing inside. Using eight solenoid valves in the hut, air from each level was sequentially pumped using a diaphragm pump through an IRGA (model 6262, LI-COR Inc., Lincoln, Nebraska). A similar system, but with un- heated tubes, was used during July through October 1996 [Chen et al., 1999]. COz concentrations at all eight levels were meas- ured for approximately 1 min twice every half hour, and the sys- tem was calibrated automatically every 6 hours. Changes in COz storage in the air column beneath the eddy covariance sensors were calculated from these CO2 concentration profile data. The rate of change for a given half hour was estimated by calculating the difference between the mean air COz storage in the previous and following half hours. Half-hourly CO2 fluxes were measured using the eddy covari- ance method at the 39.5-m height above the ground on a 37-m walk-up scaffold tower (main tower) in 1993, 1994, and 1996. Fluctuations in the wind vector components were measured using a three-dimensional sonic anemometer-thermometer: a model DAT-310 with model TR-61B probe, Kaijo-Denki Co., Tokyo, Japan (20-cm path length) in 1993 and 1994 and a model 1012R2A, Gill Instruments, Lymington, England (15-cm path length) in 1996. Fluctuations in CO2 and water vapor concentra- tion were measured using the closed-path method with a tem- perature-controlled LI-COR 6262 IRGA. Air was drawn through a heated tube (6 m long by 3.2 mm ID Bev-a-line in 1993 and 1994 and 4.7 m long by 4 mm ID Dekoron in 1996) at 6.5 L min -• in 1993 and 1994 and at 10 L min '• in 1996 insuring turbulent flow. Using a similar eddy covariance system, CO2 fluxes above the understory were measured continuously at the 4- m height on a 6-m scaffold tower 40 m (about wo tree heights) away from the main tower in 1993 and 1994. Corrections were made for the effect of fluctuations in air density on the CO2 fluxes [Webb et al., 1980]. A two-dimensional coordinate rotation was applied to make the average vertical and lateral wind velocity components equal to zero above the forest [Tanner and Thurtell, 1969], and a one-dimensional rotation was applied to bring the lateral velocity component to zero above the understory [Baldocchi and Hutchison, 1987]. A detailed description of the eddy covariance systems and analytical procedures can be found in the work of Black et al. [1996] and Chen et al. [1999]. Com- parisons of the three eddy covariance systems, with the sensors mounted above the forest, made at Camp Borden, Ontario, Can- ada [Lee et al., 1996] and at the site [Yang, 1998] indicated that CO2 fluxes agreed to within 7%. During the 1 O-day period, August 12-22, 1994, an experiment was conducted to determine the degree of spatial variability in the CO2 fluxes above the hazelnut understory. Eddy fluxes of CO2 at the 5.9-m height on the main tower were made using a Kaijo- Denki DAT-310 sonic anemometer already operating at this height and an IRGA identical to that on the 6-m tower. These fluxes were compared with those being measured at the 4-m height on the 6-m tower. Because the ground level at the main tower was 0.3-0.5 m lower than that at the 6-m tower, the eleva- tion difference between the two measurement heights was only 1.3-1.6 m, which probably had little effect on the CO2 flux com- parison. To compare CO2 concentrations at the 4-m height at the two towers, CO2 concentrations on the main tower were calcu- lated using a rectangular hyperbolic fit to the concentrations measured at the eight heights. 3. Results and Discussion 3.1 Diurnal Courses of CO2 Concentration and the Estimation of ASa/At Figure 1 shows the ensemble-averaged diurnal courses of CO2 concentration during the leafless and full-leaf periods at the OA site in 1994. Nighttime and daytime CO2 concentrations from the 9.5-m height to the 39.5-m height were very similar (magnitudes of vertical concentration gradients <0.2 gmol mol '• m'l), showing that considerable turbulent mixing extended below the bottom of the aspen canopy. Mixing within aspen canopy on calm nights was clearly indicated by reduced vertical air temperature and CO2 concentration gradients between the 15- and 21-m heights, which often extended down to 9 m. This mixing was likely due to cold YANG ET AL.' VARIABILITY OF COz FLUX IN AN ASPEN FOREST 27,655 37O 366 362 'E 8 5oo õ O0 400 (a) - (b) 0 4 20 8 12 16 24 Local Time (CST) Figure 1. Ensemble-averaged CO2 concentrations (mole fraction i  gmol mol ' moist air) at eight heights (0.8, 2.3, 9.5, 15.7, 18.8, 21.9, 25.0, and 34.2 m) at the Old Aspen site in 1994 for (a) the leafless (February 4 to April 10) pe- riod and (b) the full-leaf (June 1 to August 31) period. The CO2 concentration at an additional height (0.5 m, cir- cles) during the summer was measured using a different gas analyzer. The symbols are squares (0.8 m), inverted triangles (2.3 m), triangles (9.5 m), left pointing triangles (15.7 m), right pointing triangles (18.8 m), pluses (21.9 m), asterisks (25.0 m), and crosses (34.2 m). air shedding from foliage as a result of radiative cooling [Yang, 1998]. Nighttime CO2 concentrations below 9 m were signifi- cantly higher than those above due to the strong temperature in- version in this layer, which suppressed the upward turbulent transport of CO2. This inversion was the result of the radiative cooling of the hazelnut understory due to the openness of the as- pen canopy. Concentrations below 9 m changed most rapidly near sunrise and sunset, and during the full-leaf period the mag- nitude of these changes was 20-50 gmol mol ' h 'l (Figure lb). In the evening the rapid increase in concentration usually ceased at about 22:00 LT (CST), and for the rest of the night, concentra- tions below 9 m remained quite constant. In contrast, concentra- tions between the 9- and 39.5-m heights increased only slightly during the 5 hours prior to 22:00 LT but increased steadily at 2-3 gmol mol ' h 'l during the rest of the night. The common occur- rence at the OA site of relatively constant CO2 concentration be- low 9 m after 22:00 LT indicates a quasi steady state condition with the rate of upward turbulent diffusion through the aspen canopy being approximately equal to the effiux of respiratory CO2 from the soil. This is likely because (1) steady nighttime soil temperature sulted in a relatively constant CO2 effiux and (2) the friction velocity u, at the 39.5-m height during the period 22:00-05:00 LT was also relatively constant (e.g., for the full-leaf period in 1994 the ensemble average for this period varied from 0.27 to 0.29 m To determine how many CO2 concentration sampling heights are necessary and to determine the best choices for this site, the change in CO2 storage in the 0- to 39.5-m air column was calculated using all combinations of the eight heights taken 1, 2, 3, 4, 5, 6, 7, or 8 at a time. AS,,/At was calculated using i:I where M is the molecular weight of moist air (: 29 g mol'l), Pi is the density of the air in layer i, and Ci is the CO2 concentration (ILl. mol mo1-1 wet air) in layer i. The thickness of each layer (Azi), except for the bottom and top layers, was approximated by halv- ing the difference between the sampling height above and below. The coefficients of determination (r 2)resulting from the linear e- gression of changes in storage using all heights on the values cal- culated using all sampling height combinations are plotted against number of sampling heights used during the leafless and full-leaf periods (Figure 2). The plots clearly show that one height is in- adequate to estimate AS``/At regardless of which height is chosen. Furthermore, the plots show three relatively distinct groups of data: (1) there is at least one height above and either two heights or only the 2.3-m height below the base of the aspen canopy (15- m height), (2) there is no height below 9.5 m, and (3) the only height below 15 m is the 0.8-m height. The first group has the highest r 2 values, and of these the highest always includes the 2.3-m height for any number of heights for leafless and full-leaf periods. The latter is also shown in Table 1, which lists the maximum r 2 value for each number of sampling heights. The other two groups in Figure 2 cannot be used to adequately esti- mate AS``/At no matter how many heights are used. More effec- tive mixing within the stand prior to leaf emergence probably accounts for the higher r 2 values for the first and second groups during the leafless period. These results suggest that AS``/At can be reasonably estimated by using at least one height above and either two heights or only the 2.3-m height below the base of the aspen canopy. If the choice were restricted to two heights, the best combination would be 2.3 and 15.7 m with a standard error of estimate (syx) of 0.87 lu, mol m '2 s '• during the full-leaf period (Table 1). If the choice were three heights, the best combination would be 2.3, 9.5, and 25.0 m with an syx of 0.58 gmol m '2 s 'l. In this case, the difference inr 2 values caused by using a different above-canopy height, namely, 21.9 or 34.2 m, was only about 2%. 27,656 YANG ET AL.: VARIABILITY OF CO2 FLUX IN AN ASPEN FOREST 1 0.8 0.6 0.4- (a) leafless (b) full-leaf 6 I I I I I I I I I I I I I I 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 8 Number of sampling heights Figure 2. Coefficients of determination (r z) from the linear regression of COz storage change inthe 0- to 39.5-m air column calculated using all eight sampling heights on values calculated using all other combinations of heights ver- sus number of sampling heights for the same periods as those in Figure 1. Circles are combinations with at least one height above and either two heights or only the 2.3-m height below the base of the aspen canopy (15 m), crosses are those with no heights below 9.5 m, and squares are those with 0.8 m being the only height below 15 m. The solid line connects the means of combinations in the first group, the dotted line connects the means in the sec- ond group, and the dash-dotted line connects the means in the third group. 3.2 Diurnal Changes in AS./At and Its Effect on Net Ecosystem Exchange Figure 3 shows the ensemble-averaged diurnal pattern of &S,/At for the 0- to 39.5-m air column, the eddy COz flux Fc at the 39.5-m height, and NEE for the full-leaf period in 1994 at the OA site. During this period, sunrise occurred between 5:00 and 6:00 LT, and sunset occurred between 20:30 and 22:00 LT. Be- tween 8:00 and 9:00 LT, AS,/At accounted for 50-60% of the NEE. During this hour and the following hour, 65 and 60%, re- spectively, of AS,/At was due to the decrease in the COz concen- tration of the air beneath the 9-m height. Between late morning and late afternoon, AS,/At made a negligible contribution. to NEE. During the early evening with a sharp drop in photosyn- thesis, there was a marked increase in AS,/At as respiration con- tinued while air and soil temperatures were high. The large in- crease in COz concentration in the air beneath the 9-m height de- scribed in section 3.1 accounted for 65-90% of AS,/At between 17:00 and 22:00 LT. After 22:00 LT, the upward eddy flux of COz as well as the rate of COz accumulation became relatively constant, averaging 3-4 and 1 gmol m 'z s '•, respectively. Over 85% of the latter occurred in the layer between the 9- and 39.5-m heights as a result of the relatively unchanging COz concentration beneath the 9-m height as discussed in section 3.1. Similar pat- terns of the daily courses of these fluxes have been observed by other workers (e.g., Fan et al. [1995] in a boreal ichen woodland Table 1. Statistics ofthe Linear Regression between COz Storage Change Calculated Using All Eight Sampling Heights (AS,/At) and Calculated Storage Change Using the Best Combination for a Given Number (n) of Sampling Heights (AS,/At)' Number of Heights, Syx, Best Combination of n a r z pmol m 'z s '• Sampling Heights, m Leafless 1 0.881 0.822 0.144 9.5 2 0.910 0.948 0.078 2.3, 15.7 3 0.993 0.969 0.060 2.3, 9.5, 25.0 4 0.974 0.994 0.026 0.8, 2.3, 9.5, 25.0 5 0.997 0.998 0.014 0.8, 2.3, 9.5, 18.8, 34.2 6 0.998 1.000 0.006 0.8, 2.3, 9.5, 15.7, 25.0, 34.2 7 1.000 1.000 0.004 0.8, 2.3, 9.5, 15.7, 18.8, 25.0, 34.2 Full-Leaf 1 0.592 0.599 1.862 9.5 2 0.766 0.913 0.866 2.3, 15.7 3 0.947 0.962 0.576 2.3, 9.5, 25.0 4 0.950 0.988 0.326 0.8, 2.3, 9.5, 25.0 5 0.986 0.995 0.205 0.8, 2.3, 9.5, 18.8, 34.2 6 0.994 0.999 0.085 0.8, 2.3, 9.5, 15.7, 25.0, 34.2 7 0.999 1.000 0.048 0.8, 2.3, 9.5, 15.7, 18.8, 25.0, 34.2 Regression is AS./At =a(AS./At)' + b. The absolute values ofb are less than 0.004 •tmol m -z S '1. $yx is the standard error of estimate. Allvalues ofcoefficient of determination r z are shown i Figure 2. YANG ET AL.' VARIABILITY OF CO2 FLUX IN AN ASPEN FOREST 27,657 -10 I I I I I 5 , , . /":•• --5 i .." . .-", _ ' ! .................. _ - I I I I I -200 4 8 12 16 20 24 Local Time (CST) Figure 3. Ensemble-averaged COz flux Fc measured using the eddy covariance system at the 39.5-m height (squares), the rate of change in COz storage (Z•a/Z•) in the air column beneath the eddy covariance system (circles), and net ecosystem exchange (NEE) (Fc + Z•a/At, dash-dotted line) during the full-leaf period in 1994 at the Old Aspen (OA) site. The vertical bars indicate one standard eviation. and Jarvis et al. [1997] in a boreal black spruce forest). In the latter study, bSa/At accounted for about 20% of the nighttime eddy flux measured at the 26-m height, which is slightly less than the corresponding fraction (35%) in this study. Jarvis et al. [1997] also found a significant proportion of bSa/bt accumulated above and in the upper part of the canopy; however, CO•. also tended to increase below the 6-m height. On very calm nights, CO•. continued to accumulate after mid- night in the air beneath the 9-m height while eddy fluxes above the forest were smaller than usual. An example of this (July 16- 17, 1996) is shown in Figure 4. In this case, COz concentration below the 9-m height continued to increase somewhat sporadi- cally until well after sunrise (8:00 LT). At the 0.5-m height, COz concentration reached 860 gmol mol 'l around sunrise. Around 9:00 LT, the stable boundary layer broke up, resulting in a strong mixing of CO2 throughout the 39.5-m air column (clearly seen in Figure 4a). At this time, there was a strong upward flux (approximately 26 gmol m '2 s -1) at the 39.5-m height (Figure 4b). 55O I 5OO 450 F 400' 350 (a) I I I I I •, 20- E 0 - E • 0 x = -2:0 ' 0 r,.) 18 I I I I I' (b) j •r• - '• ..... '._- ................... • ...... .,.,., ,. .,,, ...,', ..... 4,•, •.._.•. I I I I I 22 2 6 10 14 18 Local Time (CST) Figure 4. (a) Diurnal courses of the average COz concentration (gmol mol 'l moist air) of the air between the 0- and 9-m heights (circles), the 9- and 20-m heights ( quares), the 20- and 30-m heights (inverse triangles), and the 30- and 39.5-m heights (triangles) on July 16-17, 1996, at the OA site. (b) The corresponding courses ofFc at the 39.5- m height (solid line) and the 2-hour unning mean of Z•a/At (dash-dotted line). The strong upward COz flux at 9:00-9:30 LT (arrow in Figure 4b) corresponds with the uniform high COz concentration as a result of mixing (arrow in Figure 4a). 27,658 YANG ET AL.: VARIABILITY OF CO2 FLUX IN AN ASPEN FOREST 20 • • I • • o 1:1 line (a) unstable & neutral o oo o o 10 -10 -20 10 0 v v (b) stable v v r 2 = 0.22 -10 v v 1'1 line v v v v ,• r 2 = 0.08 I v I I I I I -2-020 -15 -10 -5 0 5 10 15 6-m scaffold tower (gmol m '2 S '1) Figure 5. Comparison of the half-hourly values of net hazelnut understory CO2 exchange (NHE) measured using eddy covariance systems atthe 6-m scaffold and main towers (40 m apart) in the trunk space during August 12-22, 1994, in (a) unstable and neutral conditions and (b) stable conditions. The slopes and intercepts of the regressions (dash-dotted lines) are 0.682 and 0.947 gmol m '2 s '• for Figure 5a and 0.308 and 1.581 gmol m '2 s 'l for Figure 5b. Although the accuracy of this measurement was probably not high, it clearly showed the sudden upward discharge of CO2 stored in the air column corresponding to the unusually large drop in CO2 storage (-13 •mol m '2 s '•) (compare Figure 3). About 20 such events were observed during the 1994 and 1996 growing seasons. This phenomenon of loss of CO2 to the atmos- phere in the early morning was commonly observed in a tropical rain forest by Grace et al. [1995]. In that case, it was almost he only time during the day that CO2 was lost to the atmosphere above the eddy covariance sensors because during the nighttime stability above the dense canopy appeared to trap respired CO2. Hollinger et al. [1998] also reported that the CO2 storage in the air column accounted for the low nighttime eddy fluxes over a Siberian larch forest. In contrast to Grace et al. [1995] and Hol- linger et al. [1998], on many calm nights in this study, Fc+ •a/• did not appear to entirely account for the likely rate of respired CO2. This "missing" CO2 is probably accounted for by mass flow [Lee, 1998] or CO2 storage in the air-filled pores of the soil [Chen et al., 1999]. 3.3 Horizontal Variability of the Fluxes within the Trunk Space Figure 5 compares net hazelnut understory CO2 exchange (NHE, which is the sum of Fc at the 4-m height and Z•u.• a/at for the 0-4 m air column) obtained uring the 10-day period at the two tower locations. Half-hourly rates of change in CO2 storage during the daytime were usually less than 10% of Fc and about 20-30% at night. Over this period the mean daytime and night- time values of u, at the 39.5-m height were 0.40 and 0.26 m s 'l, respectively. The corresponding values at the 4-m height were 0.06 and 0.04 m s 'l [Blanken t al., 1998]. Generally, agreement between NHE values at these two locations was poor; however, there was better agreement in neutral and unstable conditions (mainly daytime) or when mixing was more vigorous ( tandard deviation of vertical velocity ow at the 4-m height was greater than 0.12 m s -•) than under stable conditions (mainly nighttime) when mixing was weak (Figure 5 and Table 2). In the latter case, values often differed greatly, to the degree that at times they were of opposite sign. Generally, there was slightly better agreement between NHE values at the two locations than there was between the corresponding values of Fc. On calm nights, negative CO2 exchange frequently occurred at both locations simultaneously, suggesting that these negative fluxes were not the result of meas- urement error. Negative fluxes over a relatively large area and the persistence of significant differences between CO2 concentra- tions at the two towers for 6 or more hours suggested the occur- rence of CO2 advection in the trunk space. Advection flux expressed on a ground surface area basis was calculated to be as large as 2 •mol m '2 s -1, corresponding to a sustained horizontal concentration difference of3 •mol mol 'l between the two towers and a wind speed of 0.5 m s -• at the 4-m height. These high ad- vective fluxes were consistent with the above spatial variability in NHE. This and the considerable short-term horizontal variability in the CO2 concentration at the 4-m height (Figure 6) indicated that the forest floor was not as uniform as it appears to be as con- cluded by Fan et al. [1995]. The slightly better correlation be- tween NHE at the two locations (higher r 2) for flow parallel to the line connecting the two towers than that for perpendicular flow also indicated the patchiness of the forest floor and understory with regard to CO2 exchange (Table 2). Nonuniformity was probably a result of patchiness of the hazelnut, variability in the depth of the surface organic layer, stone content in the glacial till derived mineral soil, and topographic variability. These results showed that half-hour CO2 exchange rates measured at one loca- tion were usually not representative of a large area of forest floor or understory. Over the 1 O-day period, however, the average val- ues of NHE at the two locations were almost identical: 2.3 and 2.2 !.tmol m '2 s 'l on the 6-m and main towers, respectively (Table 3). Furthermore, the daytime means at the two locations were very similar as was the case for the nighttime means. YANG ET AL.: VARIABILITY OF COz FLUX IN AN ASPEN FOREST 27,659 Table 2. Coefficients of Determination (r z)Using Different Stratification Criteria in the Correlation between Fluxes Measured at Two Locations 40 m Apart in the Trunk Space Eddy Flux Experiment, August 12-22, 1994 Daytime Wind Direction at 4 m Relative to a Line t7w at the 4-m u, at the 4-m Time of Day Connecting the Towers Height Height Daytime n = 236 Nighttime Parallel Perpendicular crw>0.12 m s -1 crw<0.12 m s '1 u,>0.06 m s 4 u,<0.06 m s 4 n = 196 n = 32 n = 25 n = 98 n = 385 n = 155 n = 345 NHE, gmol m -z S -1 0.113 H, W m 'z 0.424 AE, W m 'z 0.594 0.059 0.152 a 0.115 b 0.290 0.066 0.100 0.069 0.007 0.357 0.022 0.455 0.041 0.507 0.119 0.018 0.730 0.700 0.328 0.496 0.659 0.521 See also Figures 5 and 7. aFor nighttime, n = 33. bFor nighttime, n = 17. Figure 7 compares half-hour sensible (H) and latent (/IE) heat fluxes in the trunk space at the two towers. Unlike COz, storage changes accounted for a very small proportion of these fluxes. Similar to NHE, the sensible heat fluxes measured at the two lo- cations agreed better under neutral and unstable conditions or when u, at the 4-m height exceeded 0.06 m s -• than they did un- der stable conditions (Figures 7a and 7b and Table 2). Over the 1 O-day period the average values of H on the 6-m and main tow- ers were 1.6 and 0.9 W m 'z, respectively (Table 3), which is close agreement considering that the accuracy of the eddy covariance measurement of H and/IE was not better than -I-5 W m 'z. Half- hour latent heat fluxes at the two towers were much better corre- lated than NHE and H in unstable and neutral conditions were (Figure 7c and Table 2). In contrast to NHE and H, the diurnal patterns of/IE were very similar at the two locations on 7 of the 10 days [Yang, 1998]. Average values of AE over the 1 O-day riod on the 6-m and main towers were 18 and 21 W m 'z, respec- tively (Table 3), which indicates close agreement. Analysis also showed that correlation between/IE at the two locations was very similar for both parallel and perpendicular wind directions (Table 2), indicating that the forest floor and understory were more ho- mogeneous with regard to/IE than NHE. For sensible heat flux, r z was higher for parallel f ow (Table 2); however, this difference is questionable because sensible heat fluxes above the understory were generally low [Blanken et al., 1998]. These comparisons uggest hat significant short-term (e.g., half hourly) horizontal variability exists for COz flux in the trunk space all day although it is slightly less during the daytime. Short-term horizontal variability in sensible heat flux is less than that for NHE but still very significant. Latent heat flux shows the least short-term horizontal variability especially when mixing is sufficient, which indicates the feasibility of partitioning evapora- 520• • 480 440 • • o A A A O A O O 360 1'1 line 32O 320 360 400 440 480 520 6-m scaffold tower (gmol mo1-1) Figure 6. Comparison of the half-hourly COz concentrations (i  gmol mol ' moist air) measured at the 4-m height on the 6-m scaffold tower with those calculated for the 4-m height on the main tower using a rectangular hyperbolic fit to the concentrations measured at eight heights during August 12-22, 1994. The triangles are nighttime data, and the circles are daytime data. The dash-dotted line is the daytime regression li e (C6-m = 0.971Cmain + 15 gmol mol 4, r z = 0.88, Syx = 8.8 }.tmol mol'l), and the dotted line is the nighttime regression line (C6. m = 0.695Cmaia + 123 }.tmol mol '1, r z = 0.45, S. vx = 19.9 gmol mol-1). 27,660 YANG ET AL.' VARIABILITY OF CO2 FLUX IN AN ASPEN FOREST Table 3. Comparison of the Means of Scalar Fluxes Obtained on the 6-m Scaffold Tower and the Main Tower in the Trunk Space Eddy Flux Experiment, August 12-22, 1994 Daytime Flux Nighttime Flux 24-Hour Flux 6-m Scaffold Main 6-m Scaffold Main 6-m Scaffold Main NHE, gmol m '2 s '1 2.0 1.9 2.8 2.7 2.3 2.2 H, W m '2 3.1 2.6 -0.9 -1.9 1.6 0.9 AE, W m -2 30.0 34.0 0.5 0.5 18.1 20.8 tion between overstory and understory on a 1-2 hour basis during the daytime. For the 1 O-day period, averaging over 5 consecutive days was required to obtain agreement in the CO2 fluxes at the two locations within 10%. This suggests the feasibility of parti- tioning CO2 uptake between overstory and understory using aver- ages over 5-day periods. In an experiment designed to examine the spatial variability of turbulent fluxes above a uniform even- aged 14 m tall loblolly pine stand (Duke Forest, North Carolina) involving 7 towers, Katul et al. [1999] found that F½ was also spatially heterogeneous. In contrast o our study, they found that ZE was as spatially heterogeneous a F½, while H was relatively homogeneous. 4. Conclusions 1. During the leafless and full-leaf perind,q in a 21.5-m tall aspen forest, vertical COz concentration gradients in the 9- to 39.5-m layer were small during the daytime and nighttime. Be- low the 9-m height there were significant vertical gradients and marked diurnal variations particularly around sunrise and sunset. 2. Changes in COz storage in the air column beneath the above-forest eddy covariance system (39.5 m) between late , morning and late afternoon were small compared to eddy fluxes F½ at the 39.5-m height. Storage changes were mainly due to the loss or gain of CO2 below the 9-m height near sunrise and sunset, when it accounted for up to 60% of net ecosystem exchange, and to the gain of CO2 between the 9- and 39.5-m heights during the rest of the night. On some calm nights, CO2 accumulated below the 9-m height, which usually resulted in a sudden strong upward eddy flux of CO2 shortly after sunrise. 3. CO2 profile analysis showed that a minimum of three heights was required to estimate the rate in change of CO2 storage in the 0- to 39.5-m air column within 5% of values obtained us- ing concentrations measured at eight heights. The three heights had to include one height above and two heights below the base of the aspen canopy. 4. The correlation between fluxes measured at two locations beneath the aspen canopy about two tree heights apart increased in the following order: F½, H, and/IE. The low correlation in the case of Fc, especially in stable conditions, indicates significant short-term (e.g., half hour) horizontal variability. However, when averaged over a period of 5 days, daytime and nighttime CO2 fluxes at the two locations agreed to within 10%, suggesting that partitioning net ecosystem exchange between the overstory and 60 20 40 10 20 E -20 g -?40 o ,- 150 E 100 50 0 2O 4O 6O -10 -20 -3-030 -20 -10 0 10 20 40 20 ! (d) stable v 2E v v v v Vr 2 = 0.09 -5-050 0 50 100 150 2) 0 20 40 6-m scaffold tower (W m -2) Figure 7. Same as Figure 5 except for sensible (H) and latent (AE) heat fluxes. The solid lines are one-to-one lines. The slopes and intercepts of the regressions (dash-dotted lines) are: (a) 0.741, 0.724 W m 'z, (b) 0.280, -1.495 W m 'z, (c) 0.684, 16.943 W m 'z, and (d) 0.363, 0.867 W m '2. YANG ET AL.: VARIABILITY OF CO2 FLUX IN AN ASPEN FOREST 27,661 understory/soil can be done using averages over periods at least 5 days long, but not on a half-hourly basis. Acknowledgment. The authors from UBC gratefully acknowledge funding provided by a 4-year Natural Science and Engineering Research Council (NSERC) Collaborative Special Project Grant in support of BO- REAS, an NSERC Operating Grant (TAB), and grants from the AES/NSERC Joint Science Subvention Program and the Canadian Forest Service. We greatly appreciate the support given by numerous individu- als. On-site assistance was provided by John Deary, Tom Hertzog, and Monica Eberle. Field operations were made possible by the support of Prince Albert National Park personnel, especially Mary Dahlman, Paula Pacholek, and Murray Heap. Janusz Olejnik and Marian Breazu assisted with instrument construction and electronics. Gerry den Hartog, Zoran Nesic, Xuhui Lee, Ralf Staebler, Alan Barr, Joe Eley, Rick Ketler, Uwe Gramann, Craig Russell, Aisheng Wu, hobel Simpson, Grant Edwards, and Jose Fuentes provided various forms of field assistance. Siguo Chen and Wenjun Chen assisted us both in the field and with data analysis. We greatly appreciate the helpful comments and suggestions made by three anonymous reviewers. References Baldocchi, D. D., Flux footprints within and over forest canopy, Bound- ary Layer Meteorol., 85, 273-292, 1997. Baldocchi, D. D., and B. A. Hutchison, Turbulence in an almond orchard: Vertical variations in turbulent statistics, Boundary Layer Meteorol., 40, 127-146, 1987. Baldocchi, D. D., and T. P. Meyers, Trace gas exchange above the forest floor of a deciduous forest, 1, Evaporation and COz flux, J. Geophys. Res., 96, 7271-7285, 1991. Baldocchi, D. D., and C. Vogel, A comparative study of water vapor, en- ergy and COz flux densities above and below a temperate broadleaf and a boreal pine forest, Tree Physiol., 16, 5-16, 1996. Baldocchi, D. D., C. Vogel, and B. Hall, Seasonal variation of carbon di- oxide exchange rates above and below a boreal jack pine forest, Agric. For. Meteorol., 83, 147-170, 1997. Black, T. A., et al., Annual cycle of water vapor and carbon dioxide above a boreal aspen forest, Global Change Biol., 2,219-229, 1996. Blanken, P. D., Evaporation within and above a boreal aspen forest, Ph.D. thesis, Univ. of B.C., Vancouver, Canada, 1997. Blanken, P. D., T. A. Black, H. H. Neumann, G. den Hartog, P. C. Yang, Z. Nesic, R. Staebler, W. Chen, and M.D. Novak, Turbulent flux measurements above and below the overstory of a boreal aspen forest, Boundary Layer Meteorol. , 89, 109-140, 1998. Chen, W. J., T. A. Black, P. C. Yang, A. G. Barr, H. H. Neumann, Z. Ne- sic, M.D. Novak, J. Eley, R. J. Ketler, and R. Cuenca, Effects of cli- matic variability on the annual carbon sequestration by a boreal aspen forest, Global Change Biol., 5, 41-53, 1999. Fan, S. -M., M. L. Goulden, J. W. Munger, B.C. Daube, P.S. Bakwin, S. C. Wofsy, J. S. Amthor, D. R. Fitzjarrald, K. E. Moore, and T. R. Moore, Environmental controls on the photosynthesis and respiration of a boreal lichen woodland: A growing season of whole-ecosystem exchange measurements by eddy correlation, Oecologia, 102, 443- 452, 1995. Grace, J., et al., Carbon dioxide uptake by an undisturbed tropical rain forest in southwest Amazonia, 1992 to 1993, Science, 270, 778-780, 1995. Greco, S., and D. D. Baldocchi, Seasonal variations of COz and water va- por exchange rates over a temperate deciduous forest, Global Change Biol., 2, 183-197, 1996. Hollinger, D. Y., F. N. Kelliher, J. N. Byers, J. E. Hunt, T. M. McSeveny, and P. L. Weir, Carbon dioxide exchange between an undisturbed old- growth temperate forest and the atmosphere, Ecology, 75, 134-150, 1994. Hollinger, D. Y., et al., Forest-atmosphere carbon dioxide exchange in eastern Siberia, Agric. For. Meteorol., 90, 291-306, 1998. Jarvis, P. G., J. M. Massheder, S. E. Hale, J. B. Moncrieff, M. Rayment, and S. L. Scott, Seasonal variation of carbon dioxide, water vapour and energy exchanges of a boreal black spruce forest, J. Geophys. Res., 102, 28,953-28,966, 1997. Katul, G., et al., Spatial variability of turbulent fluxes in the roughness sublayer of an even-aged pine forest, Boundary Layer Meteorol., in press, 1999. Lee, X., On micrometeorological observations of surface-air exchange over tall vegetation, Agric. For. Meteorol., 91, 39-49, 1998. Lee, X., T. A. Black, and M.D. Novak, Comparison of flux measure- ments with open- and closed-path gas analyzers above an agricultural field and a forest floor, Boundary Layer Meteorol., 67, 195-202, 1994. Lee, X., T. A. Black, G. den Hartog, H. H. Neumann, Z. Nesic, and J. Olejnik, Carbon dioxide exchange and nocturnal processes over a mixed deciduous forest, Agric. For. Meteorol., 81, 13-29, 1996. Lee, X., J. D. Fuentes, R. M. Staebler, and H. H. Neumann, Long-term observation of the atmospheric exchange of COz with a temperate de- ciduous forest in southern Ontario, Canada, J. Geophys. Res., in press, 1999 Schuepp, P. H., M. Y. Leclerc, J. I. MacPerson, and R. L. Desjardins, Footprint prediction of scalar fluxes from analytical solutions of the diffusion equation, Boundary Layer Meteorol., 50, 355-373, 1990. Tanner, C. B., and G. W. Thurtell, Anemoclinometer measurements of Reynolds tress and heat transport in the atmospheric boundary layer, Res. Dev. Tech. œep. ECOM-66-G22F, Univ. of Wis., Madison, 1969. Verma, S. B., D. D. Baldocchi, D. E. Anderson, D. R. Matt, and R. E. Clement, Eddy fluxes of COz, water vapor and sensible heat over a deciduous forest, Boundary Layer Meteorol., $6, 71-97, 1986. Webb, E. K., G.I. Pearman, and R. Leuning, Correction of flux meas- urements for density effects due to heat and water vapor transfer, Q. J. œ. Meteorol. Soc., 106, 85-100, 1980. Wofsy, S.C., M. L. Goulden, J. W. Munger, S. -M. Fan, P.S. Bakwin, B. C. Daube, S. L. Bassow, and F. A. Bazzaz, Net exchange of COz in a mid-latitude forest, Science, 260, 1314-1317, 1993. Yang, P. C., Carbon dioxide flux within and above a boreal aspen forest, Ph.D. thesis, Univ. of B.C., Vancouver, Canada, 1998. T. A. Black, M.D. Novak, and P. C. Yang, Faculty of Agricultural Sciences, University of British Columbia, 139-2357 Main Mall, Vancou- ver, British Columbia, Canada V6T 1ZA. (ablack@interchange.ubc.ca; novk@ interchange.ubc.ca; henggan @interchange.ubc.ca.) P. D. Blanken, University of Colorado, Boulder, CO 80302. (blanken @ spot.colorado.edu.) H.H. Neumann, Atmospheric Environment Service, Downsview, On- tario, Canada M3H 5T4. (hneumann@ec.gc.ca.) (Received September 15, 1998; revised April 21, 1999; accepted April 28, 1999.)


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items