tree-branch of Hymenaea courbaril L. have been performed applying a dynamic cuvette system (see Section 2.2.4) from 25 to 31 October 1999 (LBA-EUSTACH 2). Leaf scale O3 resistances
,O3
rL have been determined from incoming and outgoing O3 mixing ratios (see Section 3.6 Eqs. (12) and (13)). Fig. 15 (a) exhibits the diel course of median
,O3
rL together with that of
(
2 3)
3 /
,O s HO O
s r D D
r = , the leaf scale stomatal resistance (see Eq. (13)), which was also determined by the branch cuvette measurements. There is no significant difference between
,O3
rL and
,O3
rs from 1100 LT to 1800 LT, indicating that daytime O3 deposition is entirely controlled by stomatal aperture. The identity of the two resistances during that time period is also supporting the usual assumption, that the O3 mesophyll resistance
,O3
rm (Eq. (13)) is negligible (see also Gut et al., 2002b). The mean values, averaged for daylight hours (0600 LT to 1800 LT) are 640 s m-1 for
,O3
rL and 770 s m-1 for rs,O3, where any significant differences would arise from the measurements during the
early morning hours. Contrastingly huge during night. Especially after sunset (1800 LT), in the first half of the night, when fast O3 depletion resulted in huge storage corrections of O3 fluxes at canopy level, a strong divergence of
,O3
rL and
,O3
rs is visible.
However there is no significant reduction of the leaf resistance during that time, it ranges between 1000 and 3000 s m-1 during the whole night (mean: 1440 s m-1, 1900 LT to 0500 LT).
Although the difference to
,O3
rs is smaller in the second half of the night, there is still distinct additional O3 deposition in parallel to any stomatal uptake. Within the branch cuvette, this can be attributed to O3 deposition to the leaves’ cuticles and to bark surfaces.
This non-stomatal resistance does not show a
systematic diel variation, such as found for coniferous forests by Rondon et al. (1993).
They obtained a reduction of the non-stomatal resistances during daylight hours concluding a corresponding dependence on air temperature or solar radiation.
To elucidate the behavior of the non-stomatal O3 deposition, the measured leaf scale O3 resistances (rL,O3) are compared with those calculated by the leaf resistance model of Baldocchi et al. (1988), rLm,O3 (see Section 3.6, Eq. (13)). Still keeping the proved assumption
,O3
rm ≈ 0 s m-1, rLm,O3 was calculated with three different hypothetical values of O3 cuticula resistances (
0000 0400 0800 1200 1600 2000 2400
102
0000 0400 0800 1200 1600 2000 2400
102
Fig. 15. (a): Diel variation of measured leaf scale O3 stomatal resistance (median: dashed line; inter quartile range IQR: gray hatched area) and O3 leaf resistance (median: solid line; IQR: light gray shaded area), both obtained by branch cuvette measurements at Hymenaea courbaril (IBAMA site). (b): Diel variation of measured and modeled O3 leaf resistance for three different cuticula resistances (dashed line: 4000 s m-1; dashed-dotted line: 2000 s m-1; dotted line: 1000 s m-1).
Fig. 15 (b). For
,O3
rct = 4000 s m-1, the diel variation of the modeled leaf resistance (rLm,O3) agrees quite well with the measured one (rL,O3). The resulting mean values of rLm,O3 are 570 s m-1 (0600 LT to 1800 LT) and 1550 s m-1 (1900 LT to 0500 LT). These values compare well to the mean
,O3
rL data (640 s m-1 and 1440 s m-1, respectively; see above). In a greenhouse experiment, Gut et al. (2002b) used the same type of dynamic cuvettes to investigate O3 deposition to another young tropical deciduous tree (Pouteria glomerata), which is typical for the Amazon region. In contrast to the findings here, they obtained no significant O3 deposition for night time, which is indicative for a very large cuticula resistance. This was supported by a statistically
insignificant difference between
,O3
rL and
,O3
rs
for daytime hours. According to the survey of non-stomatal O3 deposition velocities from leaf level experiments by Kerstiens and Lendzian (1989) rather similar values for a variety of other plant species are known. The present results from Hymenaea courbaril L (mean: vdL of 7.0 × 10-2 cm s-1) rank in the upper third of the reported range. These results show the potential for considerable nocturnal O3 deposition if O3 mixing ratios are high.
Nevertheless, to explain the fast O3 depletion during the first half of the night, additional sinks, most likely chemical reactions (besides (5)) are necessary.
0000 0400 0800 1200 1600 2000 2400
-15
0000 0400 0800 1200 1600 2000 2400
-15
Fig. 16. Diel variation of the mean (median) up-scaled leaf level O3 flux (median: solid line; inter quartile range:
gray shaded area) and mean (median) canopy scale O3 fluxes at RBJ tower site during (a) LBA-EUSTACH 2 (full squares: O3 flux; open circles: storage corrected O3 flux) and (b) LBA-EUSTACH 1 (open circles: storage corrected O3 flux). Details see text.
Finally the leaf scale results of O3 deposition are dicussed versus those obtained at the canopy scale. A direct comparison is not feasible, since both data sets have not gathered simultaneously. Moreover, there are general difficulties transferring results from measurements on single plants to the entity of forest canopies by so-called bottom-up scaling (e.g. Baldocchi et al., 1991). Especially here, where results of branch cuvette measurements, which have been performed on one single tree species, are compared with those obtained for the entire canopy of the RBJ rain forest with its large biodiversity – this seems to be, at a first glance, a hopeless undertaking. Therefore any comparison between leaf and canopy scale results is more qualitatively, i.e. it is not focusing on absolute values, the emphasize is rather on the relative behavior of mean diel variation. For this purpose, the O3 fluxes, obtained from the cuvette measurements, were scaled up just by multiplication with the total LAI of 5.6 observed at RBJ. This rather crude scaling probably overestimates the real canopy O3 deposition, because it is inherently assumed by this approach, that the O3 deposition determined at crown level is vertically constant within the entire canopy. The diel variation of the up-scaled O3 fluxes (25 to 31 October) are shown in Fig. 16 (a) together with the diel variation of measured (eddy covariance) O3 fluxes, as well as the storage corrected canopy scale O3 fluxes measured between 21 September and 20 October (see also Fig.
9 (b)). There is a marked difference between the up-scaled leaf level results and the eddy
covariance O3 fluxes, especially during the first half of day-light hours. Compared to that, the up-scaled leaf level O3 flux shows a delayed increase and a more moderate slope during this time period. This different behavior nearly vanishes, if the storage corrected O3 flux is compared with the up-scaled O3 flux.
Keeping in mind, that for cuvette based flux measurements no storage contribution has to be considered, the agreement between fluxes around noon is excellent. Nevertheless, one have to bear in mind that canopy and leaf scale measurements represent slightly different time periods (21 September to 20 October vs. 25 to 31 October, respectively). Therefore, the development growth state of the deciduous part of the RBJ canopy in early October might have been different from that of Hymenaea courbaril L. end of October. Secondly, during the time period of eddy covariance measurements, there were much higher O3 mixing ratios than end of October. This may have two consequences: (i) cancellation of any major difference between canopy and leaf scale O3 fluxes during day-light hours due to the expected overestimation of the latter (see above), and (ii) the considerable higher nocturnal canopy scale O3 fluxes, especially during the first half of the night. Mean nocturnal leaf resistances (
,O3
rL ), determined from the cuvette measurements, would allow relative high O3 deposition fluxes, if higher O3
mixing ratios would have been experienced end of October.
In this context, it is interesting to oppose up-scaled leaf level O3 fluxes from end of October, to storage corrected canopy scale O3
fluxes from LBA-EUSTACH 1 (May). As shown in Fig. 16 (b), there is an exceptionally good agreement between the diel variations of both quantities. At a first glance this is surprising, at least for the fact, the ambient O3 mixing ratios in May have been somewhat lower than end of October (see Fig. 13).
Obviously, the higher uptake capacity of the leaves (higher O3 deposition velocity) in May have balanced the lower mixing ratios as well as possible differences which arise from up-scaling. But besides the coincidentally quantitative agreement of canopy scale and up-scaled leaf level fluxes, the more or less identical diel variation is most striking. This may be most likely due to (i) SHD regimes in May 1999 and end of October 1999 have been rather comparable (Fig. 13), and (ii) the developmental state of the Hymenaea courbaril L end of October, characterized by the presence of mostly mature leaves (Rottenberger et al., 2005) might correspond more to the developmental state of the RBJ canopy during the wet season than to that during the period 21 September to 20 October.
Finally, it should be stated generally, that during daytime the diel variation of O3 deposition to the entire RBJ rainforest canopy seems to be very well characterized by the results of the leaf level measurements at a single deciduous tree. In addition the results suggest under dry season conditions the storage corrected eddy covariance flux (in
contrast to the measured flux itself) to be a good measure for the actual O3 uptake of the RBJ canopy during daytime.
4.6 Deforestation Impact on Ozone