Department of Forest Resources, University of Idaho, Moscow, ID 83844-1133
Our objective was to estimate winter carbon flux as a proportion of the annual flux for foliage, stems, soils and coarse woody debris within an inland Douglas-fir ecosystem. Through measurements distributed throughout the year we examined tissue-specific physiological response to climate and their contribution to winter and annual ecosystem respiration (Recosystem). Finally, we determined the relative error of assuming that 1) the last measurements prior to winter are representative of the winter fluxes, 2) temperature responses of maintenance respiration can be extrapolated throughout the winter and 3) no flux occurs over winter. Stem respiration (Rstem) was measured on eight to fifteen trees congruent with sapwood temperature (Fig. 2). Soil respiration (Rsoil) and temperature were measured at 25 locations, except in months of snow cover when it was measured as CO2 efflux from the snowpack. We have found that measurement of snowpack CO2 efflux provides results comparable to those based on gas diffusion when the snowpack [CO2] is at steady state (unpublished results). Coarse woody debris respiration (Rcwd) and temperature were measured on ten logs with the same apparatus as that used for soil respiration. All respiration measurements were made with a Li-Cor 6200. Photosynthesis between measurement dates were estimated by linear interpolation. Respiration during periods between measurements was estimated using tissue-specific Q10 equations (Fig. 3) developed either from 24-hour diel or annual measurements of respiration and temperature (Fig. 4). All temperatures used to force the respiration models were estimated from linear regression equations relating tissue-specific temperatures measured on site to average daily temperatures from a nearby weather station (Republic, WA). Wintertime carbon efflux and uptake varied among components due in part to variable temperature responses (Table 2). Average rates of winter Acanopy and Rfoliage were 0.1 µmols m¨² s¨¹ and 0.15 µmols m¨² s¨¹, respectively. Rstem dropped dramatically with the onset of freezing temperatures; no flux was measurable below sapwood temperatures of -5°C. Because of this drop, winter Rstem averaged 2.3 µmols m¨³ s¨¹. Winter efflux from coarse woody debris averaged 0.2 µmols m¨² s¨¹. The winter average Rsoil of 0.56 µmols m¨² s¨¹ was the highest of all components during this period. Table 3 presents these values expressed as a proportion of the total annual flux (Fig. 5) and Table 4 expresses them as percentages of the total respiratory efflux in winter and annually. If these fluxes are summed for the winter and their relative contributions to winter Recosystem are calculated, the uniqueness of each component is demonstrated (Table 4). For example, Rsoil dominated the winter flux, contributing 69% of Recosystem. Rstem, in contrast, contributed only 8% of winter Recosystem. When the same two fluxes are summed over the entire year and expressed as percentages of annual Recosystem, they exhibit inverse trends; Rsoil declines to 54% of Recosystem and stem respiration rises to 23%. Three common alternatives for estimating winter carbon flux can present error in ecosystem carbon budgets (Table 5Introduction
Although it is widely accepted that substantial carbon fluxes occur throughout winter in forests under maritime climates, less is known of forest carbon fluxes in cold continental winters. In temperate regions, winter metabolic rates are low in comparison to summer rates; however, studies suggest that the tissue components respond differently. For example, respiratory efflux from soil continues at substantial rates in areas of harsh and prolonged winter such as sub-alpine meadows and arctic tundra, due largely to the insulating effect of the snowpack on soil temperature. In contrast, the aboveground components of forests are not buffered against cold temperatures and as a result are presumed inactive. The difficulty of wintertime gas-exchange measurement combined with the expectation of low rates has led many researchers to assume negligible activity and therefore the winter contribution to whole ecosystem carbon budgets is rarely quantified.Methods and Materials
We measured the components of a carbon budget in a mature Douglas-fir (Pseudotsuga menziesii var glauca) forest in northeastern Washington (Table 1). All components were measured nine times throughout a 12-month period between July 1996 and June 1997. Net annual assimilation (Acanopy) was determined from interpolation between measured daily photosynthetic rates incorporating differences in canopy position and needle age class. Likewise, foliar respiration (Rfoliage) was measured from three canopy positions and on each foliar cohort at both maximum and minimum air temperature on dark acclimated foliage (Fig. 1). Photosynthesis measurements were made with a Li-Cor 6200 Portable Photosynthesis System. The light environment was quantified in the canopy and replicated on the ground with a Portable LED solid state lighting system. Results
Canopy net photosynthesis ranged from 0.00 µmols m¨² s¨¹ in February to 2.44 µmols m¨² s¨¹ in August. Variable temperature response of each respiratory flux was found, resulting in a range of Q10 values from 1.7 to 15.9 (Table 2). Q10 equations developed from the entire year's measurements accurately predicted both Rsoil and Rfoliage for each season. In contrast, accurate prediction of Rstem required three separate Q10 equations including one for the growing season (late May to mid-September), winter (mid-November through March) and non-winter maintenance periods in spring and fall. Rcwd was accurately predicted by one equation for all months excluding those of snow-cover, for which no suitable equation was found.
Discussion
Components differ in response to sub-zero temperatures and presence of ice and snow (Fig. 5). The large Q10 value for Rsoil is at least partially due to the depth at which soil temperature was measured (10 cm). This depth is below the region of highest respiratory production, and hence the observed temperature-respiration relationship was poorly coupled in comparison to aboveground components, many of which had Q10 values near 2. The high Q10 value for Rstem during the spring and fall may be due the low temperature range of measurement, and the very high value in winter may be a result of ice formation within the sapwood. Whatever the cause, our results support those of others who found increasing Q10 values associated with cooler seasons. Rfoilage was presumably well coupled to canopy air temperature and hence showed a more "typical" Q10.
Other studies have shown that Rsoil can comprise over 50% of Recosystem. At our site, this proportion increased in winter, likely due to the insulating effect of the snowpack, which allows microbial and root activity to persist. In contrast, the relatively greater exposure of aboveground tissues to cold temperatures may explain their low contribution to winter Recosystem. Positive Acanopy was measured at air temperatures as low as -1°C, allowing for enough carbon assimilation during winter to replace 23 % of winter respiratory losses.
We found that wintertime carbon fluxes represent a relatively small proportion of the annual carbon budget of this forest, with only 8% of annual Recosystem. The errors attending two common short-cuts are relatively minor, particularly for above ground tissues (Table 5
), but the extrapolation of fall fluxes significantly overestimates winter fluxes. Although the Acanopy results are preliminary, we suspect that canopy net photosynthesis will approximately balance ecosystem carbon losses over the winter. In future work we intend to model photosynthetic responses to light and temperature in order to test this hypothesis.