Journal of Geophysical Research: Biogeosciences

Map-based inventory of wetland biomass and net primary production in western Siberia



[1] We used available field survey and literature data to produce inventory maps of wetland biomass and net primary production (NPP) for western Siberia. Field survey data were obtained for major types of wetland microlandscapes within the boreal (taiga) region. We developed a multiscale approach based on using a regional wetland typology map (1:2,500,000 scale), further refined by satellite image classifications (LANDSAT-7, SPOT, RESURS, 1:200,000 scale). Satellite images on test areas designated in the boreal region of western Siberia were classified by 30 landscape classes. We used aerial photography (1:25,000 scale) to evaluate the fraction of the area occupied by microlandscape elements within patterned wetlands. As a result, we were able to produce a GIS map-based inventory of ecosystem phytomass and NPP in west Siberian wetlands. Using the GIS map, the average and total net primary production and biomass were estimated by ecosystem type, the number of vegetation layers, and climatic gradient. The annual NPP to live biomass ratio increases southward from 0.27 in the tundra to 0.65 in the steppe region. Live biomass of wetlands amounts to only 10–30% of the average biomass of upland forests in the same climatic region, although wetland NPP was found to be equal or higher then that of upland forests. Mosses and the belowground fraction of grasses are the major contributors to NPP. Average live biomass and NPP in wetlands were estimated to be 1600 g/m2 and 790 g/m2/yr, respectively. Total wetland NPP amounts to 530 Tg/yr, and live biomass amounts to 1070 Tg.

1. Introduction

[2] The vast areas of boreal and arctic wetlands play an important role in the global carbon cycle, as reported in a number of studies. Although Siberian wetlands have been accumulating 20–50 g C/m2/yr over the last 5000–10,000 years since deglaciation, they have the potential to become net sources of carbon to the atmosphere under a warmer and drier climate [Tolonen et al., 1992; Botch et al., 1995; Turunen et al., 2001]. Despite covering only 3% of the earth's land surface, the northern peatlands - defined as the area north of 50°N – store approximately 15–30% (200–455 Pg (=1015g)C) of the total terrestrial pool of soil carbon in undecomposed peat [Sjörs, 1981; Post, 1990; Gorham, 1991]. Carbon dating of peat profiles suggests that the area of northern peatlands has been constantly expanding during the last 11,000 years [Neustadt, 1977; Gajewski et al., 2001].

[3] The buildup of CO2 in the atmosphere and the resulting global warming have the potential to affect net primary production (NPP) and carbon storage of terrestrial ecosystems, particularly in northern high latitudes [Denman et al., 2007]. Thus it is important to identify the spatial distribution of NPP and biomass in wetlands, and their responses to climate change under elevated atmospheric CO2 concentration.

[4] Earlier studies [e.g., Smith and Forrest, 1978] (collated in the work of Campbell et al. [2000]) have described the distribution of biomass within peatlands, primarily above ground. However, there have been relatively few measurements of belowground biomass and production. For example, Finér et al. [1993] and Saarinen [1996] have determined root-allocated NPP to vary between 25 and 200 g C/m2/yr. Ignoring the belowground component, which is the usual practice because measurement of root production in peatlands is difficult, excludes about 50% of biomass and a large portion (from 40–90%) of annual production, according to estimations by Coleman [1976], Aerts et al. [1992], Titlyanova et al. [1999], and Moore et al. [2002].

[5] Net primary production in the wetlands has a wide range of variation. Bartsch and Moore [1985] estimated that NPP varies from 50 to 150 g C/m2/yr in the subarctic region and from 100 to 200 g C/m2/yr in the boreal region. Various field studies have shown that the NPP of marshes and swamps in Northern America and Europe ranges between 400 and 1000 g C/m2/yr [Van der Valk and Bliss, 1971; Baradziej, 1974]. In the Canadian wetlands, the aboveground fraction of NPP (including moss layer) was estimated to vary from 400 to 760 g C/m2/yr in the series of oligotrophic-mesotrophic-eutrophic mires [Thormann and Bayley, 1997]. The presence of topography, with such features as hummocks, hollows, ridges, open water pools, and distance from the peatland margin (macrotopographical position) have a profound effect on the plant species distribution and productivity [Andrus et al., 1983; Moore et al., 2002]. Spatial differences in these microsite characteristics (i.e. hydrologic and thermal regimes, nutrient availability) are often as large within an individual peatland as between different peatlands within the same or different climate regions.

[6] Recently, NPP and biomass were estimated for major wetland types within western Siberia (WS) [Bazilevich, 1993; Vasiliev et al., 2001; Peregon et al., 2005]. Additional studies that include both aboveground and belowground NPP and biomass are needed to build a foundation that can be used to determine the overall carbon storage capacity and fluxes of carbon in wetland ecosystems. It is also necessary to develop detailed wetland NPP and biomass inventories as the basis for further research.

[7] Existing global maps of wetland types, CO2 and CH4 emissions, and NPP suffer from crude representation of the wetland typology for western Siberia. Matthews and Fung [1987] grouped the 28 UNESCO wetland vegetation classes into 5 general categories of forested bog, nonforested bog, forested swamp, nonforested swamp, and alluvial wetland formations, and estimated the size of global flooded wetland areas to be 5.3 × 1012 m2. About half of the total area is situated between 50–70°N and is occupied predominantly by forested and nonforested bogs. Using different data sources, Aselmann and Crutzen [1989] also examined the global distribution of wetland CH4 emissions. They grouped 45 different vegetation types into 6 categories: bogs, fens, swamps, marshes, floodplains, and shallow lakes. This information provides global models with a generally realistic estimate of current land cover with coarse spatial resolution. However, these databases suffer from lack of consistency in the vegetation classification used, variable measurement techniques, and mix of spatial sampling resolutions [Running et al., 1995].

[8] The latest estimates of the area and carbon storage and accumulation rates in mires of the former Soviet Union (FSU) were reported by Botch et al. [1995]. In their study, a major source of uncertainty was attributed to poor accounting of shallow mires that was also not represented in Russian peat inventories. The area of peatlands in the FSU might be twice as large as the area reported by Kivinen and Pakarinen [1981], Tyuremnov [1976], etc., who based their estimates on data from the Peat Fund of the USSR which considers mostly commercial peatlands. Peat and carbon content data for west Siberian peatlands have been summarized recently by Sheng et al. [2004] and Smith et al. [2004].

[9] Basically, two methods are available to estimate the NPP and biomass of a regional terrestrial ecosystem: (1) extrapolating the local field measurements up to a larger region, using a vegetation map and (2) modeling plant productivity at a regional and grid point scale. Modern vegetation models [Woodward, 1987; Prentice et al., 1992] have improved the representations of the global distribution of vegetation, but still have to be corrected by observations. Consequently, Townshend et al. [1991] argued that a use of data from remote sensing of the actual landcover is necessary for more accurate classifications of global vegetation. The theory and rationale for using remote sensing in model-based estimation of photosynthesis and NPP, including those of wetland vegetation, are described by Potter et al. [1993] and Ruimy et al. [1994]. However, since modeling the wetland biomass and NPP with process based models still remains difficult due to the lack of reliable vegetation type maps, thus we decided to use a simple inventory-type approach to combine the GIS and point observation data.

[10] The purposes of the present study are (i) to provide improved an estimation of the extend of the wetland area and distribution of the major mire types in the western Siberia, (ii) to determine the spatial variability of NPP and biomass in relation to macro/micro landscape and position within the bioclimatic division of the Siberian territory, and (iii) to represent this material in the form of a digital vector map at 1:2,500,000 scale.

2. Material and Methods

2.1. Site Description

[11] Western Siberia is usually defined as the area between the Ural Mountains and Yenisey River and is bounded by the Arctic Ocean on the north. The current study covers the area up to the southern edge of the forested steppe zone. See Figure 1 for geographical boundaries and a climatic division map. Western Siberia is an extensive and extremely paludified territory with ∼40% of the global peat deposits and 50–75% of its land area covered with mires [Walter, 1977].

Figure 1.

Geographical division of western Siberia and location of test areas.

[12] To obtain reasonable estimates of the spatial distribution of NPP and biomass, we need to measure them at a representative set of locations along a wide north–south climate (i.e. latitudinal) gradient of WS as shown in Figure 1. Test areas for our field observations were selected in each of three sub-zones of the taiga in the central part of the considered territory (Table 1). Within those test areas, 3, 2 and 3 test sites representing different wetland types and topography were selected within designated test area in the northern, middle and southern taiga, respectively (Table 2).

Table 1. Location of Test Areas
Latitudinal DivisionNumber of Test SitesLocationPrecipitation, mm/yrAverage Annual Temperature, t°CPermafrost
Northern taiga362°50′–63°20′N 75°00′–75°45′E336–360−3,4, −5,440 cm below the surface
Middle taiga260°30′–61°00′N 76°30′–77°00′E400–600−2,8, −3,7-
Southern taiga356°30′–57°00′N 82°30′–83°00′E400–500−0,6, −1,6-
Table 2. General Description of the Test Sites
NoTest Site/Wetland TypeMicrotopographyVegetation CommunityMain Dominants
Northern Taiga
1Ridge-hollow-pool patterned bogRidgePine-dwarf shrub-SphagnumLedum palustre, Betula nana, Sphagnum fuscum
Mesotrophic hollowSedge-SphagnumCarex rostrata, C. limosa, Eriophorum polystachyon, Sphagnum majus, S. jensenii
2“Palsa” bog with permafrostFrozen hillock - “palsa”Shrub-lichen-SphagnumLedum palustre, Betula nana, Rubus chamaemorus, Spagnum fuscum, Cladina rangiferina. C. stellaris, Cetraria spp.
Oligotrophic hollowSedge-SphagnumCarex limosa, Eriophorum russeolum, Sphagnum lindbergii
3Eutrophic alluvial swampTussockSedge-SphagnumCarex rostrata, Eriophorum polystachyon, Sphagnum. fallax, S. riparium
Middle Taiga
4Ridge-hollow-pool patterned bogRyamPine-dwarf shrub-SphagnumPinus sylvestris, Ledum palustre, Andromeda polifolia, Oxycoccus microcarpus, Spagnum fuscum
RidgePine-dwarf shrub-SphagnumLedum palustre, Andromeda polifolia, Oxycoccus microcarpus, Spagnum fuscum
Oligotrophic hollowSedge-SphagnumCarex limosa, Sheuchzeria palustris, Sphagnum lindbergii, S. balticum
5Mesotrophic mire*TussockSedge-SphagnumCarex lasiocarpa, C. rostrata, Betula nana, Sphagnum majus, S. fallax
Southern Taiga
6Ridge-hollow-pool patterned (mesotrophic) bogRidgePine-dwarf shrub-SphagnumPinus sylvestris, Chamaedaphne calyculata, Betula nana, Sphagnum fuscum
RyamPine-dwarf shrub-SphagnumPinus sylvestris, Chamaedaphne calyculata, Betula nana, Rubus chamaemorus, Sphagnum fuscum
Mesotrophic hollowShrub-sedge-SphagnumChamaedaphne calyculata, Betula nana, Equisetum fluviatile, Menyanthes trifoliata, Sphagnum balticum, S. majus, S. fallax, S. jensenii
Mesotrophic hollowSedge-Eriophorum-SphagnumCarex rostrata, Eriophorum vaginatum, Sphagnum majus
7Ridge-hollow-pool patterned (oligotrophic) bogRidgePine-dwarf shrub-SphagnumPinus sylvestris, Chamaedaphne calyculata, Betula nana, Sphagnum fuscum
Oligotrophic hollowSedge-SphagnumCarex limosa, Rhynchospora alba, Sphagnum balticum
8Eutrophic open swampTussockDwarf shrub-grass-mossesDryopteris pinulosa, Equisetum heleocharis, Triglochin palustre, Carex rostrata, Menyanthes trifoliata, Sphagnum centrale, S. wulfianum. S. angustifolium, S. squarrosum

2.2. Research Methods

2.2.1. Method 1: Field Observations

[13] At each test site, detailed geobotanical descriptions were recorded and biomass sampling was conducted. When doing so, we made sure that the major topographical features which were typical for the test area were considered. Sampling was repeated two or three times during the growing season at the same test sites for several consecutive years to obtain information on interannual variability and to collect representative data. A total of 5,500 plant samples were collected between 1993 and 2001.

[14] The total amount of biomass was divided into two fractions - live (phytomass) and dead biomass. Dead biomass contains partly-decomposed plant residue, plant litter, dead mosses including the yellow-coloured parts of Sphagnum, and dead roots of grasses and shrubs. Observed live biomass (LB) consists of aboveground (AG), land-surface (LS), and belowground (BG) components: that is LB = AG + LS + BG. The AG component was determined by clipping vegetation (herbs, shrubs and dwarf shrubs) at the top of the moss surface (i.e. the so-called “clipping line”) over a 0.25-m2 square, replicated 7–10 times in a 50-cm interval along the designated transect, which has a length of about 5 m. For further estimations of NPP, the aboveground phytomass was divided into several fractions: the green parts of herbs which are capable of photosynthesis; the shrub leaves of the current and previous years; the shoots and woody parts (both perennial and those that formed in the current year); the upper part of shoots down to the dark line of brown mosses; and the photosynthetically-active upper parts of lichen stems. Living bryophytes and lichens below the clipping line were considered as LS, whereas stems, stem bases, rhizomes, roots of vascular plants, grasses, shrubs and dwarf shrubs, which were located below the clipping line were considered as BG. On the same plot, where the aboveground fraction was clipped, we sampled 1-dm3 cores in 10-cm increments down to a depth of 30 cm below the moss surface to estimate LS and BG. Clipped (raw) material was sorted by species, oven-dried at 108°C and weighed. Hereinafter, all references to mass should be regarded as dry weight.

[15] Like the biomass, NPP also consist of aboveground (ANPP), land-surface (LNP) and belowground (BNP) components. The ANPP of grasses was defined as the maximum seasonal weight of the standing green phytomass. We determined the ANPP of deciduous shrubs as the maximum seasonal weight of leaves and shoots grown in the current year. The weight of green biomass grown in the current year, such as the shoots and foliage, clipped at the end of the growing season corresponds to the ANPP of evergreen shrubs. Production of lichens was estimated as the difference between the maximum and minimum seasonal weights of live biomass. Production of Sphagnum mosses (LNP) was estimated using the measurements of the basal cover as a percentage, linear growth in mm, and shoot density in the number of shoots per dm2. Linear growth of Sphagnum was measured by a method of “individual tags” which allowed us to measure NPP of plants growing obliquely in hollows. In this method, stems are tagged with thin metal rings, just below their crowns, as shown in Figure 2. About 70 to 100 tagged plants per community were sampled one year after tagging. For each sample, the linear shoot growth represents the distance between the level a - just below the plant's crown and level b - where the ring was placed in the previous year. The dry weight of annual growth corresponds to the net primary production of mosses. The total number of Sphagnum stems per unit area was additionally examined to express the LNP in the common unit of g/m2/yr. The sum of these estimated annual increments of different species of bryophytes, corrected for their basal cover, is the LNP.

Figure 2.

Method of “individual tags” used for estimation the linear growth of Sphagnum mosses.

[16] Normally, BNP is conventionally assumed to be 50–80% of NPP based on Wallén [1992]. In this study, however, we used our own method for extraction of belowground components which were formed during the current year. Rhizomes and stem bases were divided into two groups: young (i.e. grown during the sampling year) and old (i.e. the rest). Young stem bases have many fine roots that are lighter in colour and originate directly from the stems. Newer growths of rhizomes are lighter in colour too, and accompanied by vegetative shoots on the top. Roots of grasses and sedges were divided into four groups according to their age and condition (i.e. whether they were live or dead) indicated by a number of visible morphological features, such as length, colour, diameter, turgor, degree of lateral ramification, position on belowground rhizomes, etc. Dwarf shrub roots were divided into three groups based on the same criteria. Accordingly, BNP is defined as the weight of belowground biomass formed in the current year.

[17] In addition to our own NPP and biomass data, for nine wetland types we used 91 biomass and NPP estimations from literature for boreal and northern mires [Kosykh, 2003].

2.2.2. Method 2: Data Synthesis

[18] A “synthesis method” for combining data of several different spatial resolutions was developed to build the regional-scale NPP and live biomass inventory. We applied a multiscale approach to map NPP and live biomass based on an existing wetland typology map at 1:2,500,000 scale [Romanova et al., 1977]. Geometric correction and digitization of this map were done with MapInfo 8.0 software using the topographic map of the Russian Federation (scale at 1:1,000,000) for georeferencing. Obviously, the procedure is prone to error due to the use of hardcopy products and digitization, although the planimetric errors typically appeared to be at most 1.5–2 km. There are 20 wetland types and complexes for the whole western Siberia. Average area fractions for each microlandscape/landscape type, composing the vegetation mosaic of those 20 classes, were derived from remote sensing and ground survey data.

[19] As a first step, a map of land cover classification for each selected test area was constructed via visual interpretation of satellite images with high spatial resolution. SPOT HRV-images for the test areas in southern and middle taiga regions (both were acquired in 1995) as well as LANDSAT-7 ETM+ (2000) and Russian satellite - RESURS (1999) images covering the northern part of boreal region were used for this purpose. Spatial resolution is around 30 m for LANDSAT-7 ETM+ and RESURS images, and around 20 m for SPOT HRV images. All images were taken during the mid- or late growing season and provide almost cloud-free coverage over considered areas. Pre-processing and interpretation were carried out using GIS-software (MapInfo Professional 6.5, MapInfo Corp.). Only false-color satellite imagery in a digital format composed of near infrared (NIR), and visible red and green spectral bands (equivalent to LANDSAT bands 4, 3 and 2) were used in our study. Images were interpreted from a computer monitor for colour, tone and texture. The mapping scale is 1:100,000–1:200,000 and classification accuracy is approximately 100 m because of blurred boundaries between different vegetation classes. About 30 classes of forest, paludified forest and wetland ecosystems were derived for the boreal region. At this scale, the satellite images were classified into 10 wetland classes, compared with a large-scale wetland typology map [Romanova et al., 1977], which used only 3 classes within the boreal region. We adopted the classification system developed by Lapshina and Vasiliev [2001], in which all the landscapes in the taiga zone are subdivided into 8 general categories according to the humidity index, source of nutrients and nutrient availability. Several types of upland forests, flood-plain and paludified forests, as well as the wide range of wetland landscapes are recognized in this classification. For example, homogeneous (non-patterned) wetlands are subdivided into pine-dwarf shrub-Sphagnum bog, forested Sphagnum swamp, sedge-Sphagnum fen, brown moss and sedge-brown moss fen within the considered area. But there are also elements of a composite nature which are present in this classification, such as open patterned wetlands, that differ in the area fraction occupied by ridges, hollows and small lakes. This wetland type is widely distributed in WS.

[20] Finally, we evaluated the fractional area coverage of microlandscape elements (ridges, hollows, and lakes) for each type of patterned wetland by interpreting the 1:25,000 scale aerial photographs, where available. We applied this information to elucidate the structures of patterned wetlands on satellite image-based maps for the same climatic region as shown in Figure 1. We had no aerial photographs for the test areas in the northern taiga; therefore, we used ground survey data for the estimation of the relative extents of ridges, hollows, hillocks and lakes. Then, the ground survey NPP and biomass data were scaled up corresponding to wetland type on the regional scale map.

3. Results

[21] Inventory estimates are presented in Tables 3 and 4 in the form of average NPP and biomass for each ecosystem, vegetation layer, and component; then scaled up for all wetland types on the wetland typology map of West Siberia (Table 5).

Table 3. Biomass Pools and Standard Deviations (Boreal Region, Western Siberia)a
EcosystemsLive BiomassLive Biomass (Total)Dead BiomassTotal BiomassNd
Green PhytomassBelowground
M + LbGrasscShrubsc
  • a

    Units are in g/m2.

  • b

    M + L = mosses and lichens correspond to land-surface biomass (LS).

  • c

    Grasses and shrubs correspond to aboveground biomass (AG).

  • d

    Here n is number of test plots.

  • e


  • f


  • g

    Pine-dwarf shrub-Sphagnum, ryam.

  • h


Northern Taiga
Eutr. swampe371 ± 304352873 ± 8313399971 ± 96211,310 ± 14309
Palsaf555 ± 2411106788 ± 164146013,606 ± 168715,066 ± 18629
Ridgeg410 ± 3892961349 ± 17620646658 ± 2038722 ± 23912
Mesotr. hollowe200 ± 207414986 ± 7712748865 ± 87010,139 ± 152412
Oligotr. hollowe351 ± 50192397 ± 8376910,973 ± 195011,742 ± 188918
Middle Taiga
Ryamg303 ± 335285931 ± 7215248419 ± 3019943 ± 3205
Ridgeg345 ± 367184850 ± 20138611,203 ± 58412,596 ± 5795
Mesotr. hollowe539 ± 4579163967 ± 18317484112 ± 2065860 ± 2013
Oligotr. hollowe436 ± 504026468 ± 579704566 ± 2415536 ± 2205
Southern Taiga
Ryamg572 ± 51533831284 ± 13022925506 ± 10507798 ± 92023
Ridgeg317 ± 7066139952 ± 12814744450 ± 10245924 ± 14296
Eutr. swamph415 ± 531171822590 ± 15433045079 ± 2558383 ± 1009
Mesotr. hollowe332 ± 49461211594 ± 4820933468 ± 3985561 ± 39425
Oligotr. hollowe420 ± 252761375 ± 418832983 ± 873866 ± 8821
Table 4. Net Primary Production by Layer and Site Type, Long-Term Average in Boreal Region, Western Siberiaa
EcosystemsBiomassAnnual ShootsMosses and Lichens (Land-Surface)Belowground Parts of VegetationTotal
  • a

    Units are in g/m2/yr.

  • b


  • c


  • d

    Pine-dwarf shrub-Sphagnum, Ryam.

  • e


Northern Taiga
Eutrophic swampb65 ± 516 ± 98 ± 1236 ± 20634 ± 65959 ± 91
Palsac12 ± 446 ± 812 ± 2198 ± 47293 ± 40561 ± 111
Ridged7 ± 294 ± 935 ± 3278 ± 10239 ± 120653 ± 165
Mesotrophic hollowb73 ± 170.30.3217 ± 37405 ± 63696 ± 94
Oligotrophic hollowb15 ± 2--196 ± 24143 ± 41354 ± 51
Middle Taiga
Ridged8 ± 156 ± 318 ± 1268 ± 10207 ± 93557 ± 90
Ryamd7 ± 280 ± 323 ± 2310 ± 28217 ± 80637 ± 146
Mesotrophic hollowb59 ± 387 ± 338 ± 2346 ± 47357 ± 143887 ± 190
Oligotrophic hollowb25 ± 45 ± 14 ± 1268 ± 17198 ± 45500 ± 170
Southern Taiga
Ridgesd12 ± 144 ± 312 ± 2397 ± 40303 ± 42768 ± 42
Ryamd5 ± 186 ± 838 ± 2273 ± 33532 ± 152934 ± 205
Eutrophic swampb160 ± 7208 ± 40101 ± 5190 ± 441310 ± 111969 ± 131
Mesotrophic hollowe78 ± 1250 ± 1214 ± 2285 ± 11879 ± 551306 ± 121
Oligotrophic hollowb56 ± 512 ± 210 ± 1250 ± 9199 ± 55527 ± 57
Table 5. Biomass (Tg)/NPP (Tg/yr) and Area (106 ha) by Microlandscape Type
Wetland TypeaGrass-Shrubs-Mossy LayerWoody Layer
Total (Without Woody Layer)Area Fraction of Microlandscapes
ElevatedDepressedLakes, 106 ha,
Hillocks, RollersRidges and RyamsOligotrophic HollowsMesotrophic HollowsEutrophic Mires
  • a

    Wetland types according to Romanova et al. [1977]. A. Humid zone. I. Oligotrophic and meso-eutrophic lowland polygon mires: 1 - Polygonal-roller and polygonal-fissure mires, 2 - Polygonal mires combined with grass- and moss-dominated mires. II. Oligotrophic and mesotrophic flat-palsa mires: 3 - Patterned (hollow and hollow-pool) flat-palsa bogs, 4 - Flat-palsa and high-palsa bogs, 5 - Shrub-dominated tussock mires. III. High-palsa oligo-mesotrophic and oligo-eutrophic mires: 6 - High palsa-hollow and pool-hollow patterned mires, 7 - High-palsa and flat-palsa mires. IV. Oligotrophic (Sphagnum dominated) domed bogs: 8 - Sphagnum-dominated bogs with pools and open stand of trees, 9-10-Ridge-hollow, ridge-hollow-pool and ridge-pool patterned bogs, 11 - Forested (treed) shrubs- and moss-dominated mires, 12 - Moss-dominated treed (Pinus) mires. B. Semi-arid zone. V. Eutrophic and mesotrophic (sedge-brown mosses treed) flat mires: 13 - Ridge-hollow patterned bogs, 14 - Grass-mossy mires and oligotrophic pine-dwarf shrub-Sphagnum domed bogs - ryams, 15 - Grass- and grass-moss-dominated mires, 16 - Grass-sedge and Sphagnum-sedge hardwood swamps. C. Arid zone. VI. Eutrophic reed and grass-dominated salt marches: 17 - Reed and sedge-reed mires concerning oligotrophic domed bogs (ryam), 18 - Reed-sedge and grassy mires, 19 - Grass-dominated mires on salted soils, 20 - Unidentified mire type.

  • b

    For wetland types 1–2, and 13–19, estimates are based on literature survey.

Tundra and Forested Tundrab
1109.2/29.1 (4.0)26.0/5.8 (1.0)-2.4/0.7 (0.2)-80.8/22.6 (2.8)--
258.3/15.9 (2.0)8.0/1.8 (0.3)---50.3/14.1 (1.7)--
Northern Taiga
366.7/24.2 (6.2)48.7/15.9 (2.8)-18.0/8.3 (2.3)--(1.1)-
416.7/7.9 (1.5)11.4/5.6 (0.7)-4.3/2.0 (0.6)1.0/0.3 (0.01)-(0.2)-
528.6/8.5 (1.9)28.6/8.5 (1.9)------
630.0/11.9 (2.2)16.9/6.2 (0.8)4.9/1.3 (0.2)5.5/2.8 (0.6)2.7/1.6 (0.2)-(0.4)-
77.6/3.7 (0.7)5.2/2.5 (0.3)-2.0/1.1 (0.3)0.4/0.1 (0.01)-(0.1)-
Middle and Southern Taiga
8104.3/50.4 (10.8)-44.5/19.5 (2.9)59.8/30.9 (6.2)--(1.7)0.6/0.1 (0.3)
965.9/32.7 (7.0)-26.6/12.7 (1.9)39.3/20.0 (4.0)--(1.1)3.8/0.3 (1.9)
1077.1/42.7 (3.8) 12.5/6.0 (0.9)7.3/3.8 (0.8)13.8/7.0 (0.8)43.5/25.9 (1.3)--
11282.6/123.9 (18.5)-282.6/123.9 (18.5)----2.0/87.3 (18.5)
1234.0/14.9 (2.2)-34.0/14.9 (2.2)-----
Forested Steppe and Steppeb
1321.9/16.1 (0.8)-12.3/8.6 (0.4)--9.6/7.5 (0.4)--
143.1/1.5 (0.1)-0.2/0.1 (0.01)--2.9/1.4 (0.1)-0.2/0.01 (0.008)
1559.6/28.9 (1.4)----59.6/28.9 (1.4)--
1642.6/20.7 (1.0)----42.6/20.7 (1.0)-118.7/6.1 (1.0)
174.2/5.1 (0.3)-0.6/0.1 (0.01)--3.6/5.0 (0.3)-0.75/0.1 (0.03)
1857.2/89.8 (2.1)----57.2/89.8 (2.1)--
191.9/3.2 (0.1)----1.9/3.2 (0.1)--
Total1071.5/531.0 (68.5)144.8/46.3 (7.8)418.2/187.0 (27.0)138.6/69.6 (15.0)17.9/9.0 (1.0)352.0/219.1 (11.2)(4.6)126.0/93.9 (21.7)

3.1. Biomass Pools

[22] Phytomass or live biomass (LB) - is the most important fraction of biomass, which controls of the ecosystem function, whereas dead biomass (DB) is the parent material for peat formation. Field survey data for both the live and dead biomass at boreal wetlands in western Siberia are presented in Table 3. The live biomass appears quite variable without apparent latitudinal gradient, so we found an easy way to consider biomass variability within each plant community.

[23] In the pine-dwarf shrub-Sphagnum community, which is the main vegetation formation in “ryam” (oligotrophic raised bog) and builds up the ridges in patterned ridge-hollow-pool wetlands, live biomass is mainly created by belowground organs of shrubs - accounting up to 65% of its value. Most of the remaining 35% of live biomass is created by mosses and lichens; grasses account for minor contributions to both aboveground and belowground LB components. Dead biomass varies over a wide range depending on the geographical location of the test area.

[24] The lowest amount of live biomass is contained in oligotrophic sedge-Sphagnum hollows, located in soggy and mineral-limited conditions. Regardless of where in the taiga zone they are found, the biomass content of this type of ecosystem is about 3 times lower than in other plant communities. Mosses and belowground components make up the major part of live biomass. The distinguishing feature of these ecosystems is the nearly equal values of aboveground and belowground components of biomass. There is no considerable difference in LB values between oligotrophic hollows in the northern and southern taiga, although the total amount of dead biomass is 4 times higher in northern hollows.

[25] Mesotrophic sedge-Sphagnum hollows and eutrophic sedge-Eriophorum-Sphagnum swamps produce a large amount of live biomass, especially in the middle and southern taiga. In these ecosystems, belowground organs (roots) of grasses and dwarf shrubs contribute to over 70% of LB content. About 15–25% of live biomass is made up of land-surface vegetation components (bryophytes and lichens), while 5–10% of LB was created by aboveground shrubs and 2–5% by grasses.

[26] The live biomass pool in shrub-lichen-Sphagnum communities (“palsa” bog in northern taiga) is 1460 g/m2 of which 54% is due to the weight of belowground organs. The land-surface vegetation fraction contributes 38% of LB. Grasses and shrubs account for only 8% of LB due to inclement climate and the presence of permafrost. The value of the perennial LB of shrubs growing on “palsa” is half that of pine-dwarf shrubs-Sphagnum communities, but their annual parts (leaves) contents are similar. Palsa stores a large amount of dead biomass.

[27] Estimates presented in Table 3 also show that live biomass varies more dramatically among the different ecosystems of one climatic region than among the same types of ecosystem of different climatic regions. The ecosystem types, listed in the order of increasing LB contents, are: oligotrophic hollow, “palsa”, ridge and “ryam”, mesotrophic hollow and eutrophic swamp. The major contribution to the live biomass is the belowground part of shrubs and grasses in all climatic regions (60% of LB). Mosses (land surface parts) produce the majority of live biomass only in the oligotrophic hollows as well as “palsa” bogs with permafrost (30% of LB). Aboveground vegetation parts account to only 10% of live biomass.

[28] Total (live and dead) wetland biomass of the northern taiga region is twice as higher as in southern taiga, while the live biomass increases accordingly. Live biomass amounts to about 12% of the total biomass in northern taiga, while its contribution is 2.5 times higher in the southern taiga (Figure 3). However, the dead biomass dominates over the live biomass in all regions.

Figure 3.

Total biomass (mean values, g/m2) and its fractions (%). Regions: A, southern taiga; B, middle taiga; C, northern taiga.

3.2. Net Primary Production

[29] NPP of wetlands in northern taiga varies in the range of 350–960 g/m2/yr (Table 4). The eutrophic alluvial (valley) swamps are characterized by rich biodiversity and the highest value of NPP with a large contribution from roots. Aboveground parts of vascular plants, mainly dwarf shrubs, make up about 20% of the annual increment of biomass on the ridges, while the contribution of land-surface component amounts to 43%. Belowground organs make up about 37% of the total NPP. The ratio between aboveground and belowground components is 1:2. The NPP composition in the “palsa” ecosystems is similar to that of mesotrophic hollows, but the total NPP of palsa is lower than the NPP of mesotrophic hollows. Belowground NPP of palsa accounts for 52% of its total NPP and is four times larger than the aboveground NPP. Although both are topographically similar (i.e. both are topographically depressed), mesotrophic hollows can be distinguished from oligotrophic hollows by their more abundant mineral supply and higher biodiversity. The NPP of mesotrophic hollows is two times higher due to the weight of belowground organs (more than 400 g/m2/yr of dry matter or 58% of total NPP) and land-surface component contributing 31% of its total NPP. The belowground NPP exceeds the aboveground NPP six-fold in mesotrophic and ten-fold in oligotrophic hollows.

[30] Mesotrophic hollows produce a large portion of NPP in the middle taiga. Specifically, ecosystem types can be ordered as follows in the direction of increasing NPP: oligotrophic hollow, ridge, “ryam”, and mesotrophic hollow. The contribution from aboveground components in the total NPP increases, following the same order of ecosystem types (i.e. from aboveground NPP accounting for 7% of the total NPP in oligotrophic hollow to 20% in mesotrophic hollow). The land-surface fraction of the NPP remains uniform across all the ecosystems and amounts to 40–53% of the total. Roots make up about 33–40% of the total NPP.

[31] The largest NPP values can be found in the ecosystems of southern taiga. The total and belowground NPPs are found to increase from oligotrophic hollows to eutrophic swamps, following the trend similar to that mentioned above for the middle taiga. Generally speaking, the average NPP of an ecosystem in southern taiga is 1.5 times larger than the NPP of the same ecotype in middle taiga. The highest NPP is observed in eutrophic wetlands. Sedge and grass roots make the largest contribution to NPP in hollows (see Table 4), while dwarf shrub roots are the dominants in “ryam” ecosystems. A considerable part (from 10 to 52% of the total NPP) is contributed by mosses and lichens. The aboveground fraction is the lowest.

3.3. Mapping of NPP and Live Biomass

[32] According to the wetland typology map of Romanova et al. [1977], wetlands cover a total of 68.5 Mha or 27% of western Siberia, with a maximum area in taiga regions (30%) and a minimum (about 7%) over northern tundra (Table 5). The main twenty wetland types were classified into three categories according to their hydrological conditions: areas with excessive, intermediate, or insufficient water supply. Tundra and forested tundra, which contain 6 Mha of wetland area, are classified in the humid (excessive water supply) climatic zone. On the other hand, 55 Mha of wetlands are situated in the boreal region (also excessive water supply). Regions with intermediate water supply (forest steppe and sub-taiga) include 3.3 Mha of wetlands, while about 2.5 Mha of wetlands are located in the region with unsufficient water supply (steppes) with eutrophic reed and grass-dominated ecosystems.

[33] The annual NPP and biomass pool of WS wetlands vary depending on their geographical locations. The greatest amounts of NPP and biomass are made by the ecosystems in middle and southern taiga, where wetlands occupy more then 30% of the area (Figure 4). Wetlands of northern taiga and tundra produce the lowest NPP. The lowest amount of biomass is observed in the steppe region.

Figure 4.

Spatial distribution of NPP and live biomass created by wetlands in different climatic regions.

[34] The major part of live biomass (713.0 Tg) is located in boreal wetlands. Ridges and pine-dwarf shrub-Sphagnum communities (ryam) contain about 57% of that biomass in middle and southern taiga. About 15% of LB is contained in hillocks and rollers in northern taiga. The contribution made by other microlandscapes is minor. A similar pool of biomass was observed at wetlands in tundra and forested tundra (167.5 Tg, or 16% of the total), and forested steppe and steppe (190.2 Tg, or 18% of the total). The major LB is created by grass-dominated eutrophic mires in both the tundra and steppe regions.

[35] Most of the NPP is created by wetlands in the boreal region (320.8 Tg/yr, or 60% of the total value), with the maximum contribution by the southern subzone. Other regions produce much less: only 10% of the total annual NPP is produced by wetlands in northern taiga. Forested wetlands and ridges are particularly productive with their contributions adding up to about 56% of the NPP for the entire boreal region. Mesotrophic hollows and eutrophic swamps are also relatively productive, producing 15% (35 Tg/yr total) of the total NPP of taiga in spite of their minor area coverage. The NPP of wetlands is 45 Tg/yr in tundra, while it is about 165 Tg/yr (30% of western Siberian wetland NPP) in the forested steppe and steppe area. Even though the steppe region has insufficient water supply and minor wetland coverage, they produce even more NPP (97.7 Tg/yr) than the forested steppe region which is known to possess more favourable hydrothermal conditions for the development of wetlands (67.2 Tg/yr). A unified live biomass and NPP map (excluding the data for the woody layer) is presented for 20 wetland classes within western Siberia in Figure 5.

Figure 5.

Live biomass (LB, t/ha) and net primary production (NPP, t/ha/yr) distributions in west Siberian wetlands (excluding woody layer). 1 ton DOM = 0.5 ton C (DOM - dry organic matter).

4. Discussion

[36] Our biomass estimations appear to be comparable to those for treeless peatlands in the boreal regions of North America and Europe, although the amount of aboveground biomass is at the low end of the range reported in the literature [Wallén and Malmer, 1992; Campbell et al., 2000; Moore et al., 2002]. The amount of grass biomass tends to be in the good agreement with the published values. Shrub biomass is extremely variable, but is consistently at the low end of the previous estimates. The West Siberian wetlands have a high belowground biomass of 0.9–2.6 kg m−2 that is also similar to those reported for Canadian wetlands: 1.2–2.4 kg m−2 [Moore et al., 2002]. In our case, the only exception is oligotrophic hollows, where the belowground biomass was found to be relatively low despite of the climatic conditions (0.4–0.5 kg m−2), since oligotrophic hollows have low amount of the total biomass, in general.

[37] On the basis of our results, we assume that the variability of the wetland NPP in the West Siberian taiga region is mainly controlled by the nutrient supply and availability which confirms the statement made by Magnani et al. [2007] for temperate and boreal forests. A weak latitudinal gradient also exists in its distribution, i.e. the NPP of wetlands increases from north to south in the taiga zone (Table 6). The increase of belowground NPP and the reduction of the contribution of moss species have been identified in the same direction. Mosses (land-surface component) make a large contribution to NPP in all west Siberian wetlands, accounting for the major part of the total NPP in the poor conditions found in ridges and oligotrophic hollows. On the other hand, the belowground component dominates in meso- and eutrophic conditions.

Table 6. Range of NPP and Live Biomass by Ecosystem, and Average Distribution of NPP by Vegetation Layer
Latitudinal DivisionNPP, g/m2/yrDistribution of NPP by Vegetation LayerLive Biomass, g/m2
Aboveground, %Land-Surface (Mosses), %Belowground, %
Northern taiga354–958113851769–2064
Middle taiga500–887134047970–1748
Southern taiga527–1970102862883–3304

[38] Overall, our study revealed the high NPP values in west Siberian wetlands. How do our estimates compare with other direct field measurements? In fact, there are very few direct measurements of the total (aboveground and belowground) NPP, largely because of the difficulty in measuring contributions from belowground parts of vascular plants, although the latter is considered very important [Wallén, 1992]. For example, the NPP of 1424 g C/m2/yr was found in boreal sedge fen [Saarinen, 1999], with 90% of the total found belowground. The total NPP was also estimated at several sites in southern Manitoba, northern Minnesota, and northwestern Ontario by Belyea and Warner [1996]. They found wide variations in NPP depending on species, microhabitat, and peatland type. The lowest NPP ranging from 90 to 310 g C/m2/yr was found in the Lichen depressions, whereas it could be 370–500 and 630–800 g C/m2/yr in Sphagnum hollows and ridges, respectively. NPP reaches its maximum at 1943 g C/m2/yr in the bog ecosystems.

[39] As an alternative to direct field measurements, the aboveground NPP was used to estimate the total aboveground and belowground NPP based on the means of pooled aboveground NPP [Campbell et al., 2000; Vitt et al., 2001]. Following a literature review, values of 50% and 30% of the aboveground net primary production (including moss layer) were considered as reasonable estimates of belowground NPP for wetlands with and without, respectively, accumulation of peat. In this manner, the total NPP was estimated in a range from 264 g C/m2/yr in permafrost bogs to 1602 C/m2/yr in shrubby swamps (excluding trees). Our results, however, show the different ratio between the aboveground and belowground NPP components, from 1: 0.6 at oligotrophic wetland sites (ridges, hollows and ryams) to 1:2 at eutrophic and mesotrophic grass- and shrub-dominated wetlands. Overall, our NPP estimates are close to those obtained for boreal wetlands in continental North America and Canada (summarized by Moore et al. [2002]), and show the similar range and variability.

[40] NPP correlates well with live biomass. These parameters in grassy wetlands are nearly proportional to each other. Ecosystems dominated by dwarf-shrub vegetation feature a different type of relationship between NPP and phytomass. Shrub ecosystems accumulate a large amount of aboveground phytomass, but their NPP is often lower than in grass-dominated wetlands. On the other hand, larger belowground biomass always corresponds to a larger NPP. The NPP to live biomass ratio increases southward from 0.27 in tundra dominated by polygonal wetlands, to 0.38 in northern taiga, 0.46 in southern and middle taiga, 0.53 in forest-steppe areas (eutrophic and mesotrophic wetlands) and 0.65 in steppe areas (grass-dominated wetlands).

[41] Our results (Table 3) suggest that the live biomass of wetlands amounts to only 10–30% of the total biomass content of upland forest ecosystems in the same climatic region, as estimated by Bazilevich [1993] (not shown). Amounts of the wetland biomass are close to those of upland grasslands, and they exhibit similar latitudinal variability.

[42] On the other hand, NPP values on the wetlands were found to be similar or even sometimes higher than upland forests in all the climatic conditions. The highest NPP was observed in southern grassland-type (meso- and eutrophic) wetlands, where NPP is 1.5–2 times higher than in coniferous forests in the same region.

[43] There is appreciable variability observed for both biomass and NPP data obtained in our study, as indicated by standard deviations, SD, calculated by long term field observations (presented in Tables 3 and 4). Our results show that the mean values of NPP and biomass substantially implicate the interannual variability ranging from 5 to 20%. The maximum variability was observed in the aboveground fraction of NPP and live biomass for vascular plants (with an SD of up to 40% of the mean value) as well as the belowground fraction of NPP (up to 50%). The biomass and NPP of mosses and lichens showed low interannual variability - SD does not exceed 15% in most of the considered ecosystems. It is very likely that our results incorporate even fewer uncertainties than the data presented by Campbell et al. [2000], who estimated the standard deviations of measured NPP and standing biomass between years for individual sites in a Canadian wetland to be often greater than 50%.

[44] We should note the somewhat higher live biomass and NPP values to the north of the boreal region. There are two possible causes for these conflicting observations. The first one is the difference in observation methods, because we used data published by several different research groups. Our estimations of live biomass and NPP in tundra and forested tundra are based on the data of Bazilevich [1993] which indicate considerably high values of NPP and biomass in arctic wetlands. Another reason for obtaining such an unexpected result could be the dominance of mesotrophic and eutrophic fens in arctic tundra regions, with higher total NPP values and a large contribution from the belowground NPP fraction as compared to oligotrophic mires in taiga zones.

5. Conclusions

[45] As a result of our study, we have produced a regional scale NPP/biomass inventory in western Siberia in the form of a digital vector map, where field observations were scaled up to represent the average value for the whole area. Inevitably, generalization and extrapolation of a relatively small scale observation to a much wider area can result in some errors, which we can not reliably estimate given the limited number of test areas within the large region of the same broad wetland class.

[46] The average live biomass in west Siberian wetlands was estimated to be 1600 g/m2, and the average NPP amounts to 790 g/m2/yr. The annual NPP to live biomass ratio increases southward from 0.27 in the tundra to 0.65 in the steppe region. The live biomass of wetlands is several times lower than the total biomass of upland forests in the same climatic region. However, wetland NPP was found to be close to or sometimes even higher than that of upland forests. Mosses and the belowground fraction of grasses form the major contribution to NPP.

[47] Adding the woody layer increases the total NPP and biomass. Of course, net primary production of the woody layer varies with its density and canopy height, but their overall contributions to the total biomass and NPP values are relatively minor compared to the wetland contributions. The total NPP of west Siberian wetlands was estimated to be 530 Tg (teragram dry matter)/yr when woody layers are ignored, and 625 Tg/yr when woody parts are included. Live biomass changes from 1070 to about 1200 Tg when woody layers are included in the calculation.


[48] We wish to acknowledge the many members of the carbon cycle research community who have published data, which we compiled for this report. This work was supported by a cooperative agreement between the National Institute for Environmental Studies (Tsukuba, Japan) and the Institute of Soil Science and Agrochemistry SB RAS (Novosibirsk, Russia). It was partly supported by a grant (S-1) from the Global Environment Research Fund, Japan, and Grants-in-Aid for Creative Scientific Research (2005/17GS0203) from the Ministry of Education, Science, Sports and Culture, Japan. We would like to thank all collaborators from ISSA for their help with the field work and laboratory analysis and also our colleagues at Frontier Research Center for Global Change, JAMSTEC (Yokohama, Japan) for providing us with useful discussions and support. We also greatly appreciate the valuable suggestions of the two reviewers.