Forest Evaporation Tree Life Form and Nitrogen

Among boreal zone vegetation, forest E is unique in its variable relation with E (Table 3). According to three of these studies that are comparable, the differences in £/Eeq are associated with the three tree life-forms found in the boreal zone, namely, deciduous broad-leaved, deciduous needle-leaved, and evergreen needle-leaved (Table 6). During summer, the former two life-forms mostly bear fully grown leaves (see Black et al., 1996). However, life-form is not the only cause of difference in E among these three forests. Leaf area index was identically small in the two sparse needle-leaved forests and greater by more than a factor of 3 in the relatively closed-canopy, broad-leaved forest. Because of this, half of E emanated from the soil of the Pinus forest and from the under-storey vegetation (leaf area index = 0.5) and soil of the Larix forest. The corresponding value was 22% for the understorey (leaf area index = 3.3) of the Populus forest, including only 5% from the soil.

The large contribution of soil evaporation in the two Siberian forests meant E was closely coupled to rainfall frequency. Rapid surface drying occurs in the boreal zone during summer because of the generally high D. Thereafter, during fine weather, E is greatly reduced from the turbulence-driven, wet-surface, energy-limited rate to a much lower rate limited by diffusion through the dry surface layer. An illustrative example comes from measurements made during and after the 12-mm fall of rain on 12 July 1996 [two 1.5-h long showers that ended at 1130 hours (1.5 mm) and 1900 hours (10.5 mm)] at the central Siberian Pinus sylvestris forest site. The soil in that forest was mostly (65%) covered with a 30-mm-thick carpet of lichen (the rest being forest floor beneath the tree crowns) that had a surface area index of 6 and (dry) bio-

TABLE 6 The Ratio of Forest Evaporation (E) to the Equilibrium Rate (Ee<l) and Overstorey (Tree) Leaf Nitrogen Content and Area Density (Leaf Area, Expressed on a One-Sided Basis, Produced per kg of Carbon Assimilated) (Normalized Values in Parentheses) during Summer in Three Life Forms of Boreal Forest

TABLE 6 The Ratio of Forest Evaporation (E) to the Equilibrium Rate (Ee<l) and Overstorey (Tree) Leaf Nitrogen Content and Area Density (Leaf Area, Expressed on a One-Sided Basis, Produced per kg of Carbon Assimilated) (Normalized Values in Parentheses) during Summer in Three Life Forms of Boreal Forest

Overstorey Leaf

Overstorey Leaf


Nitrogen Content

Area Density



(mg g-1)

(m2 kg-')


Deciduous broad-leaved forest (full leaf)


30 (1)

20.0 (1)

Black et til., 1996;

Dang et al, 1997

Deciduous needle-leaved forest (full leaf)


16 (0.53)

7.2 (0.36)

Kelliher et al., 1997;

Vygodskaya etal, 1997

Evergreen needle-leaved forest


10 (0.33)

1.8 (0.09)

Kelliher et al., 1998; Wirth, 1998

(personal communication).

The three tree species compared, with their leaf area index and age in years, respectively in parentheses, are Poptiltis trermtkndes (5.1, 70), Larix grnelinii (1.5, 130), and Pinns sylvestris (1.5, 215), respectively.

The three tree species compared, with their leaf area index and age in years, respectively in parentheses, are Poptiltis trermtkndes (5.1, 70), Larix grnelinii (1.5, 130), and Pinns sylvestris (1.5, 215), respectively.

mass density of 0.8 kg m-2. Throughfall, captured in live 150-mm diameter lysimeters at ground level, indicated that the tree and lichen canopies intercepted 5.1 (42%) and 1.5 mm of rain, respectively. The lichen canopy thus intercepted 22% of the tree canopy throughfall, a percentage identical to that of feather moss covering the floor of a boreal Picea mariana forest in northern Canada (Price el al., 1997), and 78% or only 5.4 mm actually reached the sandy soil. At 0900 hours on 13 July, the lichen water content was 1.7 g g-1, which is within the range of values obtained for boreal forest mosses in equilibrium over distilled water (1.5-2.5 gg Busby and Whitfield, 1978). For Busby and Whitfield, these values were only 10-20% of the saturated values (determined by immersion followed by several minutes of drainage before weighing), suggesting that 80-90% of the water was retained as capillary films on (i.e., not within) their moss samples. Our field evaporation data support this conclusion because on the fine day after the rain, the lysimeters averaged a weight loss equivalent to 1.6 mm day ', suggesting that soil evaporation contributed only 0.1 mm (see also Skre et al., 1983). On the following two days, soil evaporation was 1.2 and 0.6 mm day ', respectively, and it remained at about 0.5 mm day-1 for the next nine fine days until the next rainfall (Kelliher et al., 1998). Rainfall intercepted by the lichen canopy was thus "gone" within one day. Nearly a quarter of the rainfall that reached the soil was evaporated away on the second day and, as stated above, the asymptotic lower limit of soil evaporation was reached by the third day.

Beyond the leaf area index and fractional understorey E differences, overstorey leaf nitrogen contents (N) are also significantly different among the three life-forms of forest. Recalling Eq. (2), for simplicity, we consider N in light of its proportionality to the maximum value of gs (gs (mm s-') = 0.9 N (mg g-') for vegetation; Schulze et al., 1994). For the Populus, Larix, and Finns forests, we thus estimate gs = 27, 14, and 9 mm s-1, respectively. Reasonably corroborating the Siberian estimates are os values derived from rnicrorneteorlogical measurements using the Penman —Monteith equation (Monteith and Unsworth, 1990): 10 and 8 mm s-1 for the Larix and Pitius forests, respectively (Kelliher et al,, 1997; Schulze et al,, 1999). The proportionality between gs and N originates at the scale of a leaf with the maximum value of stomatal conductance (gst ) (Schulze et al., 1994). The evergreen needle-leaved boreal forest data of Roberntz and Stockfors (1998) are illustrative here. Fertilized Picea abies trees growing in northern Sweden (64°N) gave values of gst and N, both of which were 30% larger than for their unfertilized counterparts.

Although gs in Eq. (2) does not constantly equal gs , it is an important parameter determining the capacity of vegetation for evaporation governed by atmospheric demand or D0 in Eq. (2). Because E cannot always meet this demand, it is also instructive to examine the relation between gs and D0. The Lohammer function fits most field data of this relation and it is simply written as 8s = Ss„n, [ 1 /i 1 + CD0/D5q)}], where D50 is the value of D0 when Ss = &smi/2- For the Siberian Larix and Pinus forest data, Schulze et al. (1999) found that a single relation emerged and we find that the Lohammer function fits it sufficiently with D50 =10 ramol mol-1 and a slightly different gs =11 mm s~1.

There are thus a number of similarities and differences between the deciduous and evergreen needle-leaved Siberian forests that can affect £/£eq during summer. The two forests had the same sparse overstorey leaf area indices (1.5), response of gs to D0, and understorey contributions to £. On the other hand, in terms of the vegetation, overstorey N and gs were greater for Larix than for Pinus, in agreement with the wider ranging study of Kloeppel et al. (1998). Leaf area density (m2 of leaf area produced per kg of carbon assimilated) is larger by a factor of 4 for Larix than Pinus, but the Pinus leaves live up to 6 years in the Zotino stand. The evergreen Pinus leaves also contain higher concentrations of secondary compounds such as lignin that deter herbivores (Kloeppel et al., 1998) and waxes which, along with stomatal closure, minimize winter desiccation when soil water is frozen and unavailable. These additional constituents, not used in carbon assimilation and thus not determining stomatal conductance (Wong et al, 1979), which is a component of gs, may also effectively dilute N (Kloeppel et al, 1998). Another relevant climatological difference is the warmer and drier summer conditions, leading to a generally higher D0, in the eastern Siberian Larix forest (Table 1). Higher gs and D0 in the Larix forest, and Eq. (2), thus explain the higher £/£eq there compared to the central Siberian Pinus forest.

Equation (5) states how stomata connect evaporation, carbon assimilation, and nutrient cycles. Distilled in the CANVEG model, the coordination of low leaf N, leaf area, photosynthetic capacity, os> and £/£a] was demonstrated for the boreal forest by Baldocchi et al (2000). Experimentally, the fertilized Picea abies trees of Roberntz and Stockfors (1998) obtained a maximum value of A (AmjJ(Wbng et al, 1979) that was, like £st and N, 30% larger than for their unfertilized counterparts. Furthermore, over a four-year period, stemwood growth rate of the fertilized trees was greater by a factor of 4.3 (Linder et al, 1996). For co-occurring Larix and Pinus trees, including two central Siberian sites, leaf carbon isotope discrimination was 1 -3%o greater for Larix (Kloeppel et al, 1998). Kloeppel et al interpreted their data as indicating a greater water loss per unit of carbon assimilated for Larix because of its higher stomatal conductance.

The deciduous broad-leaved forest evaporates at the equilibrium rate during summer, like the boreal zone's other vegetation. As stated earlier, overstorey N and leaf area index in the Populus forest are greater by as much as a factor of 3 than in the needle-leaved Siberian forests. This is not coincidental. The higher N of Populus leaves reflects its earlier successional position following a disturbance (i.e., fire) that releases a significant quantity of nitrogen. As time goes on following disturbance, the later successional coniferous forest leaves "lose" much of this N in woody tissue (Schulze et al, 1999), although subsequent nonlethal fires may release more nitrogen. Photosynthetic capacity is proportional to N (Schulze et al, 1994), so Populus forest is relatively fast-growing, including a large leaf area index that intercepts virtually all of the above-forest irradiance (recall how soil E was only 5% of the total in the Populus forest).

Higher N also means gs is much larger in the Populus forest than in the needle-leaved Siberian forests. Using Eq. (2) to further explore the effects of life-form on E will be partly dependent on wind speed because the larger Populus leaves may be isolated from the effects of D during relatively calm conditions when essentially D D0. This is incorporated into Eq. (4) but, because of the generally similar and tall heights of the three forests (the Larix, Pinus, and Populus trees were up to 20, 22, and 21m tall, respectively), it will generally not significantly affect the nature of our comparison and so we shall continue with Eq. (2). Thus, if the Populus forest Dw in the Lohammer function is similar to the Larix and Pinus forest value of 10 mmol mol-1 and the D0 climate is similar to that in Siberia, although we know it varies there, the higher E/ELX] of the Populus forest could be explained solely by the higher N and gs ^ In northern Canada, leaf carbon isotope discrimination values were around 3%o greater for Populus than for Pinus, indicating a much more conservative water use efficiency for the latter (Brooks et al, 1997). Brooks et al obtained consistent results in years with above- and below-average precipitation, suggesting that environmental changes did not alter the life-form ranking. More generally, in a global study of leaf carbon isotope discrimination, Lloyd and Farquhar (1994) showed that deciduous species use water less efficiently than evergreen conifers.

In terms of deciduous versus evergreen habit, a boreal tree's life-form also reflects air bubble production inside its conduits during the freezing of xylem sap because dissolved air in the sap is insoluble in ice (i.e., winter embolism; Sperry, 1995). Larger xylem conduits tend to be more vulnerable to cavitation by freezing and thawing than smaller ones. Populus can avoid this problem because of its deciduous habit, whereas Larix is an exception, although its habit is reflected in its higher N compared to Pinus, as discussed earlier. For Populus tremuloides trees in Alaska (ca. 65°N, 148°W), vessel diameter ranged from 5 to 50 ¿u,m with around half being between 20 and 30 /xm (Sperry et al, 1994). For the xylem of conifers, lumen diameters in tracheids (i.e., the radial hole) are generally around an order of magnitude smaller and every few millimeters, the sap has to pass through the extremely fine pores of bordered pit membranes (Whitehead and Jarvis, 1981; Vysot-skaya and Vaganov (1991) report distributions of radial tracheid diameters for co-occurring central Siberian Pinus and Larix trees that are similar to the range of values given by Sperry et al, but one cell-wall thickness was included with the tracheid lumens.) In accordance with Poisueille's equation, increased conduit size should lead to increased conducting efficiency of the xylem. Sperry et al found that the branch xylem hydraulic conductance was nearly an order of magnitude larger for Populus tremuloides than for Larix laricina. This would also confer an evaporative advantage on Populus, compared to Larix and Pinus, which would be enhanced by its relatively higher leaf area index to yield a greater forest E relative to Eeq.

0 0

Post a comment