Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

1 Introduction

Carbon dioxide exchange between the ecosystem and the atmosphere would be a major component for carbon budget at boreal forests. In this chapter, net ecosystem exchange (NEE) of CO2 at a permafrost larch ecosystem will be discussed, based on a micrometeorological (tower flux) measurement. Movement of CO2 from ecosystem to atmosphere is customarily labeled as positive. The micrometeorological measurement can obtain NEE with a half-hourly time-resolution at an ecosystem scale. These temporal and spatial scales are advantages in carbon, water, and energy budget studies over ecological measurements.

Most studies of recent micrometeorological NEE measurements in boreal forest regions have been conducted in evergreen coniferous ecosystems and in broad-leaved mixed forests of aspen, birch, and other species (Jarvis et al. 2001); see, for example, the Boreal ecosystem-atmosphere study (BOREAS) in Canada (Sellers et al. 1997), the northern hemisphere climate-processes land-surface experiment (NOPEX) in Scandinavia (Halldin et al. 1999), and the EuroSiberian Carbonflux (Schulze et al. 1999) and the TCOS-Siberia (e.g., Shibistova et al. 2002; Röser et al. 2002) projects in Siberia.

On the other hand, deciduous conifer, i.e., larch, has only a few series of continuous micrometeorological measurements of NEE taken in northern Eurasia such as Russia (Dolman et al. 2004; Machimura et al. 2005; Nakai et al. 2005), Mongolia (Li et al. 2005), China (Wang et al. 2005), and Japan (Hirano et al. 2003; Wang et al. 2004), in spite of its growing locations distributed over extended areas. In these studies, the measurements conducted over the continuous permafrost are very limited. One could easily imagine reasons why so few studies carry out continuous measurements in the permafrost larch ecosystems. They are in regions of extremely cold and continental climate, and at geographically remote locations. Consequently, micrometeorological NEE data from permafrost larch forests of Siberia are expected to be highly valuable to the studies of global carbon budget.

Some pioneering micrometeorological measurements at a few Larix gmelinii forests in Yakutia in Northeastern Siberia were reported in the literatures. The ecosystem exchange of water vapor (Kelliher et al. 1997) and carbon dioxide (Hollinger et al. 1995, 1998) was assessed at a stand located at south of Yakutsk during midsummer. The GAME Siberia, CREST/WECNoF (Ohta 2005), and TCOS-Siberia projects had long-term measurements in a larch forest at Spasskaya Pad Experimental Site near Yakutsk in the continuous permafrost area. At this site, Ohta et al. (2001) reported year-round measurements of heat and water vapor fluxes. Also Ohta et al. (2008) measured the inter-annual variation of water balance (see also Chap. 13, this Vol.). Dolman et al. (2004) estimated seasonal variation in NEE. Machimura et al. (2005) estimated NEE of 80-130 gC m−2 for seasons from May to September by a micrometeorological measurement at Neleger near Yakutsk. However, the average tree size at these stands near Yakutsk was larger (e.g., heights >10 m) than the size typical of larch forests on Siberian permafrost. Generally, average tree heights in the canopy layer of old L. gmelinii stands do not exceed 10 m in the continuous permafrost regions of Central Siberia (Abaimov et al. 1998; Bondarev 1997; Kajimoto et al. 1999; see also Chap. 7, this Vol.).

As described above, recent information on NEE of larch forests in Siberia is still limited to stands with relatively large individuals near Yakutsk. To gain better understanding of NEE for typical permafrost larch ecosystems, there should be more sites for tower flux measurement in permafrost region of Siberian taiga. As one of such sites, a tower flux measurement was initiated in 2004 to observe energy, water, and CO2 above a mature larch stand near Tura, Central Siberia (Nakai et al. 2004). This site may warrant a long-term study since it has stand structure typical of old L. gmelinii forests of the region and was established on a plateau. Recently, enough metainformation was also obtained by intensive and extensive ecological studies at sites near the flux tower (Zyryanova and Shitova 1999; Zyryanova et al. 2000; Kajimoto et al. 1999, 2003; Abaimov et al. 2000; Osawa et al. 2000, 2003, 2004; Matsuura and Abaimov 2000; Tokuchi et al. 2004; Matsuura et al. 2005; see also related chapters, this Vol.).

The objectives of this study were to estimate seasonal and annual carbon dioxide exchange and to evaluate whether the gmelin larch (Larix gmelinii) ecosystem functions as a carbon sink or source. For those purposes, measurement should be continued for many years. Measured data may be useful for understanding function of permafrost larch ecosystems. In this chapter, half-hourly, daily, and seasonal changes in NEE of a permafrost larch ecosystem are shown, based on the flux measurements at Tura in 2004 (Nakai et al. 2008).

2 Study Site for Micrometeorological Measurements

In 2002, a 105-year-old stand of larch (Larix gmelinii) forest (Carbon Flux Site in Fig. 1.3, this Vol.), east of the village of Tura, was selected for micrometeorological measurements. A 20-m high wooden tower was constructed in 2003 within the settlement of Tura, and this prefabricated structure was transported to the study site by a helicopter (Fig.10.1) in August 2003. The equipment was set up and flux measurements started in early June 2004 (Fig. 10.2).

Fig. 10.1
figure 1_10

A helicopter transporting the wooden tower to the study site. The tower was landing on the ground (Photo: Y. Matsuura)

Full size image
Fig. 10.2
figure 2_10

The tower equipped with all devices (Photo: Y. Nakai) (see Color Plates)

Full size image

The site is on a slightly sloping terrace of the Nyzhnyaya Tunguska River with a northern aspect. It consists of a nearly even-aged stand of 105-year-old larch trees that originated from an intensive forest fire in late 1890s. The prevailing winds are westerly (Nakai et al. 2005). According to tree census data from four permanent plots near the flux tower (Kajimoto et al. 2004, 2007; see Table 6.1, this Vol.), stem density of living larch trees is about 5,480 trees ha−1. Average diameter of the trees at breast height is 3.2 cm; mean tree height is 3.4 m. Individual tree crowns are slender and rarely overlap with one another (Fig. 10.3). Consequently, the stand has a very sparse canopy. Leaf biomass of the stand is 0.44 Mg ha−1, and leaf area index (LAI; calculated as projected needle area) is ca. 0.6 ha ha−1 (details see Table 6.2, this Vol.). The leaves of the larch trees are nearly exclusively short shoot leaves and are often distributed throughout the length of the stem.

Fig. 10.3
figure 3_10

Slender canopy of the larch trees in 105-year-old larch stand (Photo: Y. Nakai)

Full size image

Soil type is cryosol with permafrost table existing within the upper one meter of the soil profile (Matsuura et al. 2005). The parent material is old fluvial deposit of the Nyzhnyaya Tungusuka River. Soil texture of the surface-active layer is clay rich. Ground surface is densely covered with lichen and moss as is seen commonly at high latitude boreal forests. This cover forms a thick porous layer of 10-30 cm in depth above the mineral soil, which functions as a heat insulator above the mineral soil. Climate data of Tura from a station with long-term observations (Lydolph 1977) are shown in Fig.10.4. According to the statistics of 1968-1995 observations at the station, annual mean air temperature is −9°C and annual total precipitation is 360 mm. About 45% of the annual precipitation falls between June and August. More detailed climatic information will be described in the next section.

Fig. 10.4
figure 4_10

Monthly climate variation of Tura. Error bar with monthly mean air temperature indicates mean daily difference. Data from Lydolph (1977)

Full size image

Carbon dioxide, water vapor, heat, and momentum fluxes were measured at the top of the 20-m tall tower from June to early September, using the eddy covariance technique. Here, momentum flux can be calculated as covariance between horizontal wind and vertical wind. A three-dimensional sonic anemo-thermometer and a fast-response open-path infrared CO2/H2O gas analyzer (IRGA) were set at the top of the tower (see Table 10.1 for measurement sensors). Voltage signals from these eddy sensors were recorded at 10 Hz with a data logger (CR-5000, Campbell). The correction for angle of attack errors (Nakai et al. 2006) was applied to raw sonic wind data. Quality-control procedures designed by FFPRI FluxNet (Ohtani et al. 2005) were applied to the raw data as follows. First, graphs of the raw eddy time-series data plotted for an interval of every 30-min were visually checked. The noise-bearing data due to wetting on the sensor surfaces were excluded from further flux calculations. Second, within each half-hour observation, spikes and abnormal values out of plausible physical ranges, absolute variance, discontinuity, and stationarity were checked, using the procedures of Vickers and Mahrt (1997) and Foken and Wichura (1996). Coordinate axes for the wind field were rotated twice, so that the mean lateral and vertical velocities were zero (McMillen 1988). Humidity correction to sonic temperature and linear trends removal in scalar components (sonic temperature, water vapor, and CO2 mixing ratio) were conducted. Water vapor and CO2 fluxes were corrected for the air density effects using heat and water vapor fluctuation (WPL correction proposed by Webb et al. 1980). After the flux calculations, half-hourly fluxes with absolute angles of principal wind flow against a horizon >10°, measurable precipitation, or friction velocity <0.2 m s−1 were excluded.

Table 10.1 List of measurement items, heights, sensors for the micrometeorological observation
Full size table

NEE between forest and the atmosphere can be computed as the sum of eddy covariance flux and storage change in CO2 below the eddy flux measurement height (20 m in this case). The method by Hollinger et al. (1994) was employed for estimating the storage term since profile of CO2 concentration was not measured at the site. However, the storage change in CO2 should be verified by comparing it with profile measurements, because the estimates of storage changes here are uncertain, especially in half-hourly resolution.

Energy balance closure was examined to evaluate the surface flux measurements. The sum of turbulent energy fluxes is frequently less by 20% or more than the available radiative energy in many field measurements (e.g., Wilson et al. 2002). In this study, energy closure for quality-assured dataset was expressed by linear regressions as shown in Fig.10.5, where H is sensible heat flux, LE is latent heat flux, Rn is net radiation, and G is ground heat flux. Left hand side of Fig.10.5 (half-hourly time-resolution) revealed imbalance of total turbulent energy flux (H+LE) to some 40% of total radiative energy flux (Rn-G). On the other hand, right hand side of Fig.10.5 (daily time-resolution) shows much better agreement of total turbulent energy flux and total radiative energy flux. Such imbalance in half-hourly data might be due to the storage terms in sensible and latent heat fluxes, and to ignoring the latent heat of ice melting in the soil. Careful and detailed analysis of energy fluxes would still be needed for further discussion of energy balance closure.

Fig. 10.5
figure 5_10

Energy balance closure for observed energy fluxes. H sensible heat flux; LE latent heat flux; Rn net radiation; G ground heat flux; Total turbulent energy fluxes (H+LE) are plotted against total radiative energy fluxes (Rn-G)

Full size image

Micrometeorological variables were also measured at the top of the tower: air temperature, relative humidity, wind speed and direction, precipitation, air pressure, short-wave, long-wave, and photosynthetic active radiation for both downward- and upward-direction. Soil temperature, moisture, and soil heat flux had been measured since late July 2004 (see Table 10.1 for measurement sensors with height or depth). For all micrometeorological variables, a half-hourly average was calculated and recorded with a data-logger (CR-10X, Campbell). Electricity for all devices was supplied with six solar panels (110 W maximum for each) through deep cycle batteries. A whole integrated measurement system (CFX-N1, Climatec, Inc.) had been operating steadily.

At Tura station of the Russian Federal Service of Hydro-Meteorology (located in a residential area of Tura; see Fig. 1.3, this Vol.), a year-round measurement was initiated for standard meteorological variables such as air temperature, humidity, solar radiation, wind speed, and precipitation. This provided additional meteorological data (Fig. 10.6) to those from the tower flux measurements. The year-round meteorological data help estimation of fluxes in early spring and late autumn when access to the tower site was difficult owing to instability of river ice, and provided basic data for annual NEE estimation (Nakai et al. 2007).

Fig. 10.6
figure 6_10

Principal meteorological variations at Tura station of Russian Federal Service of Hydro-Meteorology from June 2004 to September 2005 (located in the residential area of Tura; see Fig. 1.3, this Vol.)

Full size image

3 Meteorological Condition and Features of the Measurement Site

Monthly mean air temperature and total precipitation in the summer months (June, July, August) at the study site in 2004 are compared with the 30-year averages (1960-1989) recorded at the station of Russian Federal Service of Hydro-Meteorology in Tura (Table 10.2): monthly solar radiation in the summer 2004 is compared with climatic average values from Lydolph (1977). Mean monthly air temperature for 2004 was higher in June, but lower in July and August. Monthly precipitation was approximately half in June, but about the same in July and August. The temperature of the mineral soil surface under the lichen- or moss-litter layer, being affected by the permafrost, was nearly 0°C in June, and showed a maximum in July, decreasing in August, but was higher than that in June.

Table 10.2 Meteorological characteristics of summer months at Tura, compared to long-term climatic mean values
Full size table

Daily climate variables are shown in Fig. 10.7. More than three quarters of the days in June were sunny (Fig. 10.7a). There were only several days of occasional showers. In both July and August, a majority of the days had rain, which was weaker and longer than in June. Total precipitation during the whole period was 125.5 mm (41 days with rainfall). In August, daily photosynthetic active radiation (PAR in Fig.10.7b) was considerably less than that in June and July, as the long-term average of monthly total solar radiation reveals (Table 10.2). Average air temperature was 13.5°C throughout the measurement period (Fig. 10.7d). Daily maximum of vapor pressure deficit (VPDmax in Fig.10.7b) was generally higher in June and early July than in the later period.

Fig. 10.7
figure 7_10

Daily time-series of (a) precipitation (Pr), (b) daily maximum of vapor pressure deficit (VPDmax), photosynthetic active radiation (PAR), (c) net ecosystem production (NEP = −NEE), and (d) air temperature, and soil temperature (at 0.05 m depth of the mineral soil layer)

Full size image

Variations of the daily sums of photosynthetic active radiation (PAR) above the canopy (at 20 m height) and at the ground during the growing season are shown in Fig. 10.8. Day-to-day variations in PAR are large enough to obscure the seasonal change, though monthly PAR was the highest in June and lowest in August among the three months in summer. PAR at the ground was reduced to 55-70% of the global PAR above the canopy, as shown by PAR ratio (see dots in Fig. 10.8). This indicates that light interception by the larch canopy is not likely to limit the growth of ground vegetation. Dense ground vegetation under a sparse larch canopy in the permafrost forests might play a significant role in NEE, though seasonal variation in NEE during early growing season appears synchronized with larch needle development (for details see Sect. 10.4). Partitioning of CO2 exchange into tree canopy and ground vegetation should be carried out to understand ecosystem function and its relationship to ecosystem structure.

Fig. 10.8
figure 8_10

Variation in daily photosynthetic active radiation (PAR) above (at 20 m height above ground) and below (at 1.1 m above the ground) the larch canopy

Full size image

4 Intensity and Seasonal Variations in Net Ecosystem Exchange and Larch Tree Phenology

Average diurnal cycles of NEE, PAR, and air temperature are indicated for each of 10-day periods in Fig. 10.9. Slightly negative NEE values of −1 µmol m−2 s−1 at its minimum were observed during the daytime in early June. This indicates very weak net CO2 uptake, though the open-path Infrared gas analyzer has a problem of underestimating NEE under a cold condition (e.g., Grelle and Burba 2007). In the beginning of June, soil surface started to melt and buds of larch trees had already broken. Larch needles did not flush for a few days after the bud break when air temperature was under 10°C, after which the needle flush began and was almost completed by the middle of June.

Fig. 10.9
figure 9_10

Average diurnal cycles of NEE, PAR, and air temperature. Each line represents the average for 10-day periods: (a) 4-13 June; (b) 24 June-3 July; (c) 14-23 July; (d) 24 July-2 August; (e) 13-22 August; and (f) 2-11 September

Full size image

Daytime minimum of NEE reached −4 µmol m−2 s−1 in late June (see lower graph in Fig.10.9). Furthermore, daytime minimum NEE of −5 µmol m−2 s−1 was observed in mid July when NEE was at its lowest of the year and photosynthesis was likely to be most active. In late July, daytime minimum NEE became −3 µmol m−2 s−1, which is equal to or larger than that in late June. In mid-August, NEE was little larger than in late July, but still negative. The NEE was mostly in the positive range even during the daytime in early September when the leaves began to change color.

During four 10-day periods (10 days beginning on June 4, June 24, July14, and July 24), photosynthetic active radiation (PAR) showed similar values with each other (see middle graph in Fig. 10.9) - air temperature was probably high enough for photosynthesis (see upper graph in Fig. 10.9). However, NEE was clearly different with each other among these periods. Biological factors such as increased photosynthetic ability and CO2 uptake of larch needles must promote Net Ecosystem Production (NEP = −NEE) from early June to early July, but the decrease in radiation and air temperature from mid-August to September should reduce the net uptake.

Later, based on daily integral of Net Ecosystem Production (NEP = −NEE), seasonal variation in NEE from the viewpoint of biological factors (such as photosynthetic ability) and environmental factors (such as air temperature and PAR) will be described again.

The relationship between NEE and absorbed PAR is approximated using a Michaelis-Menten type rectangular hyperbola (Ruimy et al. 1995):

$$ NEE=-P max. \frac{APAR}{APAR+\frac{P max}{\alpha}}+ Reco $$
((10.1))

where Pmax is maximum photosynthetic rate, α is initial light-use efficiency of the ecosystem, and Reco is ecosystem respiration. For light intensity of the ecosystem, absorbed photosynthetic active radiation (APAR) was used here, which was derived as incident PAR above the forest canopy minus reflective PAR. The relationship between daytime NEE and APAR was examined for every 10-day period, then it was approximated by applying Eq. (10.1). As shown in Fig.10.10, CO2 uptake of the ecosystem was significantly lower in early June and early September than in other periods. Saturation of the CO2 uptake against light intensity occurred approximately at APAR >500 µmol m−2 s−1 between mid-June and mid-August.

Fig. 10.10
figure 10_10

Relationships between daytime net ecosystem exchange of CO2 (NEE) and absorbed photosynthetic active radiation (APAR) for 10-day periods: (a) 4-13 June; (b) 24 June-3 July; (c) 14-23 July; (d) 24 July-2 August; (e) 13-22 August; and (f) 2-11 September. Each curve was fitted by Eq. (10.1)

Full size image

To estimate daily and seasonally integrated NEE, missing data of half-hourly NEE were filled with values (the so-called “gap-filling”). Here, gaps mean missing data due to device troubles or rejection by data quality control. The gaps that consisted of one or two points were filled with linear interpolation using data at both sides of each gap. For longer gaps during the daytime, regression models of Eq. (10.1) were used. Also, the longer gaps during nighttime were filled with a single exponential function, as follows: NEE = K0. exp (K1. Ta)(

$$ NEE=K_{0}.exp(K_{1}.Ta) $$
(10.2))

where k 0 and k 1 are coefficients, and Ta is air temperature. As a result of fitting Eq. (10.2) to whole available nighttime data, the most important index of temperature sensitivity of respiration Q10 = exp(k 1 ) was 2.1, and nighttime NEE at air temperature of 10°C was 1.1 µmol m−2 s−1.

Time-series of the daily-integrated NEP (= −NEE), precipitation (Pr), global photosynthetic active radiation (PAR), daily maximum of vapor pressure deficit (VPDmax), and daily average of air temperature and soil temperature are shown in Fig. 10.7. The cumulative NEP for the whole measurement period of 91 days was 76-78 gC m−2. Daily NEP was near zero in early June, sharply increased by the end of June, and reached a seasonal maximum of 2.0 gC m−2 d−1 between late June and mid-July. Thereafter, it decreased by late August, and became negative in early September. This seasonal variation indicates that net uptake by the ecosystem continued from early June to mid-August, reaching a maximum between late June and mid-July, and the ecosystem began to release CO2 after late August.

Seasonal change in NEP during the early growing season could be primarily related to needle phenology of the larch trees. Namely, during a few-week-period of needle growth after bud-break in early and mid-June, daily net uptake sharply increased to the seasonal maximum (Fig. 10.7c), though no visible signs of seasonal growth were observed for the most species of ground vegetation (i.e., woody shrubs). This coincidence between the patterns of NEP and plant phenology suggests that net CO2 uptake in the early growing season largely depends on photosynthetic activity of the larch trees. On the other hand, daily net CO2 uptake decreased gradually from late July to the end of August. In this period, canopy photosynthetic activity may still have continued at least until late August, because senescence of larch needles occurred in early September or later. However, both air temperature and incident radiation have already begun to decline in late July (Fig. 10.7b, d), suggesting that NEE of the larch ecosystem might be mostly governed by seasonal changes in air temperature and incoming radiation rather than photosynthetic activities of larch needles during latter half of the growing season (late July to August).

The bud-break and flushing of L. gmelinii needles may be triggered by soil surface thawing in the spring (Kujansuu et al. 2007; also see Chap. 17, this Vol.). Air temperature and soil temperature data from 2004 to 2005 at Tura Hydro-meteorological Station (Fig. 10.6) show that thawing of the soil surface continued for several days (indicated as “melting period” in Fig. 10.6 when soil temperature remains near zero) and was completed in early May 2005, after which bud break occurred in mid-May near the station. Consequently, close relationships between the energy budgets over the soil surface and phenological characteristics could be a key issue in the CO2 budget of larch ecosystems.

Daily maximum of half-hourly or hourly CO2 uptake rates during midsummer reached 3-6 µmol m−2 s−1. Table 10.3 lists the maximum values of half-hourly or hourly NEP (=−NEE) mainly for old larch forests and other boreal forests. The NEP value of L. gmelinii forest at Tura is considerably lower than that of other boreal forests. This may be primarily due to smaller leaf area index (LAI = ca. 0.6 ha ha−1) of a typical old stand on permafrost compared to that of other forests (LAI = 1-4 ha ha−1). Thus, cumulative NEP (70-80 gC m−2 for 91 days) during a growing season was also much lower than that (290 gC m−2 for 100 days) of the larch forest near Yakutsk with LAI of ca. 3.7 ha ha−1 (Dolman et al. 2004).

Table 10.3 Maximum NEP (half-hourly or hourly) at midday of midsummer, annual precipitation, and stand characteristics of boreal forests
Full size table

Lower cumulative value of CO2 uptake could be associated not only with low LAI, but also with other environmental factors, such as short growing periods, typically beginning in late May or early June and ending in early September (i.e., only three months or less). Moreover, low soil temperature and shallow depth of soil active layer may make productivity of the larch forest lower. In fact, the maximum depth of soil active layer is shallower (ca. 0.7 m) in the present study site (Kajimoto et al. 2007) than in the larch forest near Yakutsk (1.2 m) (Dolman et al. 2004). In winter, NEE may be slightly positive due to gradual release of CO2 from accumulated snow or frozen soils to the atmosphere as suggested in many forest ecosystems. Sum of such CO2 release in winter is probably negligibly small in the stand of the present study (T. Morishita et al. personal communication). For further discussion of seasonal and annual NEE of typical old larch forests in the region, multi-year measurements are needed to ascertain how representative is the seasonal NEE during a single growing season. For evaluation of annual NEE, it is also necessary to estimate CO2 exchange during winter by applying technically more difficult chamber measurements and continuous meteorological observations.

Additionally, water balance in the region of continuous permafrost is also globally an important theme (see Chap. 13, this Vol.). According to preliminary analysis, cumulative precipitation and evapotranspiration (probably without interception evaporation) were mostly comparable at the present study site during the measurement period. Daily evapotranspiration averaged 1.3 mm d−1 and reached the maximum of 2.4 mm d−1 (Y. Nakai, unpubl. data). Further analysis of water vapor fluxes, especially gap-filling, would still be needed for discussion of long-term evapotranspiration.

5 Conclusions

NEE of CO2 in permafrost Larix gmelinii forests in Central Siberia can be summarized as follows.

  • Information on carbon budget of larch forests is still limited compared to boreal forests in other regions. However, stable operation of tower flux measurements at remote sites is possible during growing seasons using the eddy covariance technique and solar power.

  • Typical rates of CO2 uptake and release by this 105-year-old are as follows: Maximum of half-hourly net CO2 uptake rates during midsummer ranged from 3 to 5 µmol m−2 s−1; Maximum of daily net uptake of CO2 during the growing season was about 2 gC m−2 d−1, occurring between late June and mid-July.

  • In comparison to other boreal forests, magnitude of net CO2 uptake and cumulative net CO2 uptake were low. Lower net CO2 uptake of this permafrost larch forest may be primarily associated with its small leaf area index, and is further influenced by the environmental factors characteristics to the permafrost region (e.g., low soil temperature, short growing season). Initiation and following increase in CO2 uptake were closely related to phenological development of larch needles.

Directions that should be needed for further understanding of NEE of CO2 in typical open-canopy old larch forests in the permafrost region of Siberia would be:

  • To continue longer-term and multiyear measurements. This could clarify functions of permafrost larch forests.

  • To partition entire ecosystem exchange into contributions of ground vegetation and of larch trees.

  • To clarify and model phenology of larch trees in the permafrost Siberia.