Random Forest characterization of upland vegetation and management burning from aerial imagery

Aim The upland moorlands of Great Britain form distinctive landscapes of international conservation importance, comprising mosaics of heathland, acid grassland, blanket bog and bracken. Much of this landscape is managed by rotational burning to create gamebird habitat and there is concern over whether this is driving long‐term changes in upland vegetation communities. However, the inaccessibility and scale of uplands means that monitoring changes in vegetation and burning practices is difficult. We aim to overcome this problem by developing methods to classify aerial imagery into high‐resolution maps of dominant vegetation cover, including the distribution of burns on managed grouse moors. Location  Peak District National Park, England, UK. Methods Colour and infrared aerial photographs were classified into seven dominant land‐cover classes using the Random Forest ensemble machine learning algorithm. In addition, heather (Calluna vulgaris) was further differentiated into growth phases, including sites that were newly burnt. We then analysed the distributions of the vegetation classes and managed burning using detrended correspondence analysis. Results Classification accuracy was c. 95% and produced a 5‐m resolution map for 514 km² of moorland. Cover classes were highly aggregated and strong nonlinear effects of elevation and slope and weaker effects of aspect and bedrock type were evident in structuring moorland vegetation communities. The classification revealed the spatial distribution of managed burning and suggested that relatively steep areas may be disproportionately burnt. Main conclusions Random Forest classification of aerial imagery is an efficient method for producing high‐resolution maps of upland vegetation. These may be used to monitor long‐term changes in vegetation and management burning and infer species–environment relationships and can therefore provide an important tool for effective conservation at the landscape scale.     

The application of multivariate land classification to vegetation survey in the Wicklow Mountains,Ireland

Multivariate land classification and land cover mapping by aerial photographic interpretation were used to model spatial variation of land cover in the Wicklow Mountains, Ireland and to structure a stratified random sampling programme of upland blanket bog vegetation. The total area of blanket bog with gully-erosion features was estimated as 33% of the area studied. Vegetation with hand peat-cutting patterns was estimated at 5%, and there was 35% undissected (intact) vegetation. There were differences between land classes in the estimated area of land cover with gully-erosion features or hand peat-cutting patterns.Sample vegetation quadrats, stratified by land class and aerial photographic land cover type, were grouped by their plant species composition. The groups represented ombrotrophic mire, soligenous mire and shrub heath vegetation. There was significant association between vegetation group and land class, related to variation in regional landscape type, but no significant association between vegetation group and the aerial photographic land cover types, undissected (intact) and dissected (gullied and cut-over) peats. It is proposed that the similarity of vegetation between undissected and dissected blanket bog is related to vegetation regeneration. The need to consider differences in vegetation distribution, composition and dynamics in ecological management strategies is emphasised. The study demonstrated the value of stratified random field sampling for cost-efficient regional ecological assessment in upland blanket bog landscapes typified by the Wicklow mountains, Ireland.     

The Impact of Hydrological Restoration on Benthic Aquatic Invertebrate Communities in a New Zealand Wetland

Drainage is a major disturbance affecting wetlands, as drains lower water tables and convert lentic habitats to lotic ones. Consequently, invertebrate communities in drained wetlands are likely to differ from those in unimpacted wetlands. This study investigated the effect of hydrological restoration on invertebrate communities in small drains in a New Zealand fen. Invertebrates were collected over 4 summers from 10 drains within the wetland, one of which was blocked as part of a restoration program. The sampling protocol thus represented a Before‐After Control‐Impact experiment. Invertebrate community composition varied over the 4 years, but variability was greatest in the manipulated drain before and after it was blocked. Relative abundance of the amphipod Paraleptamphopus decreased after blockage, whereas those of the midges Chironomus zelandicus and Tanypodinae increased. Relative abundances of these taxa in control sites were unchanged. Hydraulic restoration thus had a demonstrable impact on the invertebrate communities. The invertebrate community of the blocked drain was compared to that of natural wetlands in undisturbed catchments. Similarity was very low prior to drain blockage, but increased following drain blockage. Invertebrate communities in the restored drain were more similar to those of low pH wetlands than high pH wetlands. Given the goal of restoring the communities to those similar to natural conditions, this was a beneficial result. These results, coupled with studies that showed a decline in the cover of alien pasture grasses around the blocked drain, suggest that drain blockage represents a cost‐effective way of restoring wetland plant and aquatic invertebrate communities, especially where connectivity allows for the natural recruitment of these organisms into restored areas.     

Influence of riparian cover on stream temperatures and the growth of the mayfly <Emphasis Type="Italic">Baetis rhodani</Emphasis> in an upland stream

This research aimed to assess whether stream temperature and the size of Baetis rhodani larvae (Insecta: Ephemeroptera) differed between sites with contrasting riparian vegetation cover. Stream temperatures and Baetis sizes were monitored at sites along a 5-km section of a small (30 km²) upland Scottish stream. Some sites had riparian tree cover dominated by mature broadleaved woodland, while others were located in areas of open Calluna moorland. The 13 month data period included two contrasting summers, one typical for the region and the other unusually warm. Daytime water temperatures were generally highest at the open moorland sites. The difference between moorland and broadleaved woodland sites was most apparent in the summer months, with the woodland shading apparently acting to reduce maximum temperatures and mediate temporal fluctuations. Size data suggest that Baetis has a bivoltine life-cycle in the stream. Significant differences in the mean size of Baetis were found between moorland and woodland sites, with differences occurring primarily during the summer months. Mean size of Baetis was significantly greater in the moorland during the time of summer adult emergence. During the months that followed (growth of the next generation), Baetis larvae in the woodland sites were significantly larger. Size distribution data for the population suggest earlier adult emergence in the moorland section and different egg-to-larvae hatch times. Overall, the results indicate that subtle differences in growth and life cycles, linked to differences in riparian cover and streamwater temperature, can occur over short distances in upland streams.     

Warming experiments elucidate the drivers of observed directional changes in tundra vegetation

Few studies have clearly linked long‐term monitoring with in situ experiments to clarify potential drivers of observed change at a given site. This is especially necessary when findings from a site are applied to a much broader geographic area. Here, we document vegetation change at Barrow and Atqasuk, Alaska, occurring naturally and due to experimental warming over nearly two decades. An examination of plant cover, canopy height, and community indices showed more significant differences between years than due to experimental warming. However, changes with warming were more consistent than changes between years and were cumulative in many cases. Most cases of directional change observed in the control plots over time corresponded with a directional change in response to experimental warming. These included increases in canopy height and decreases in lichen cover. Experimental warming resulted in additional increases in evergreen shrub cover and decreases in diversity and bryophyte cover. This study suggests that the directional changes occurring at the sites are primarily due to warming and indicates that further changes are likely in the next two decades if the regional warming trend continues. These findings provide an example of the utility of coupling in situ experiments with long‐term monitoring to accurately document vegetation change in response to global change and to identify the underlying mechanisms driving observed changes.     

Growing season carbon gas exchange from peatlands used as a source of vegetation donor material for restoration

The moss layer transfer technique removes the top layer of vegetation from donor sites as a method to transfer propagules and restore degraded or reclaimed peatlands. As this technique is new, little is known about the impacts of moss layer transfer on vegetation and carbon fluxes following harvest. We monitored growing season carbon dioxide (CO₂) and methane (CH₄) fluxes as well as plant communities at donor sites and neighbouring natural peatland sites in an ombrotrophic bog and minerotrophic fen in Alberta, Canada from which material was harvested between 1 and 6 years prior to the study. Plant recovery at all donor sites was rapid with an average of 72% total plant cover one growing season after harvest at the fen and an average of 87% total plant cover two growing seasons after harvest at the bog. Moss cover also returned, averaging 84% 6 years after harvest at the bog. The majority of natural peatlands in western Canada are treed and tree recruitment at the donor sites was limited. Methane emissions were higher from donor sites compared to natural sites due to the high water table and greater sedge cover. Carbon budgets suggested that the donor fen and bog sites released higher CO₂ and CH₄ over the growing season compared to adjacent natural sites. However, vegetation re-establishment on donor sites was rapid, and it is possible that these sites will return to their original carbon-cycle functioning after disturbance, suggesting that donor sites may recover naturally without implementing management strategies.     

Holocene vegetation and climate change from near Lake Pedder, south-west Tasmania, Australia

Aim To use surface pollen and vegetation relationships to aid the interpretation of a Holocene pollen record. Location South‐west Tasmania, Australia. Methods A survey was undertaken of surface‐pollen samples from the major regional vegetation types: alpine, rain forest and moorland. Relationships between vegetation type and surface‐pollen representation were analysed using twinspan classification and ordination. A core was retrieved from moorland vegetation, and interpretation of the fossil pollen sequence was aided using relationships detected in our surface‐pollen analysis. Results Regional vegetation types are reflected in the pollen rain of south‐west Tasmania, despite the over‐representation of important rain forest tree species in samples from non‐forest sites. twinspan classification of the surface‐pollen samples identified the following indicator pollen taxa for each vegetation type: Astelia alpina (alpine); Lagarostrobos franklinii (rain forest); Leptospermum and Melaleuca (moorland). Detrended correspondence analysis of the surface‐pollen samples clearly separates samples from each vegetation type. Correlation of the ordination axes with environmental data identified a dominant temperature/altitudinal gradient in the surface‐pollen data (R = 0.852/0.844). Application of the results of the surface‐pollen analysis to the fossil sequence revealed that fire‐promoted moorland has dominated the local environment around the core site for the entire Holocene. Changes in fossil pollen composition also suggest that temperatures increased through the Late Glacial to peak in the mid‐Holocene and declined thereafter, a trend consistent with other sites in the region. Main conclusions Pollen spectra can successfully be used to predict local vegetation in south‐west Tasmania. At least this part of inland south‐west Tasmania has remained forest‐free throughout the Holocene, conflicting with the dominant palaeoecological paradigm of a mid‐Holocene dominated by rain forest. A comparison with pollen records from moorland vegetation across the region suggests that fire‐promoted moorland has dominated the landscape since the Late Glacial. We suggest that burning by people through the Late Glacial (if not earlier) facilitated the spread of moorland throughout the region, greatly restricting the expansion of rain forest. The continued influence of fire throughout the Holocene in this perennially wet landscape argues for a revision of the dominant human‐occupation model that depicts an abandonment of the interior of south‐west Tasmania in the Late Glacial in response to the expansion of rain forest.     

Assessing long‐term hydrological and ecological responses to drainage in a raised bog using paleoecology and a hydrosequence

Question: We studied vegetation succession after drainage in a bog, as an analogue for potential persistent water table drawdown due to climate change. We asked: (1) how does bog vegetation change following a long‐term water table lowering and (2) how are effects of drainage on hydrology and vegetation distributed temporally and spatially? Location: Mer Bleue peatland, Ontario, Canada (45.41°N, 75.48°W). Methods: Analyses of changes in vegetation and hydrology associated with drainage were examined spatially along a hydrosequence and temporally using paleoecological reconstructions from peat cores (testate amoebae, pollen) in a drained portion of a peatland untouched for 85 years following drainage. Relationships between modern vegetation and water table were assessed through clustering and ordination analyses of vegetation relevés. Results: Post‐drainage increases in tree cover, especially Betula and Larix, decreases in Sphagnum cover and shifts in species composition of dominant shrubs were observed. Present‐day vegetation patterns along the hydrosequence were primarily related to seasonal variability of water table depth. Paleoecological records reveal that where the present‐day vegetation has been impacted by drainage, persistent water table lowering occurred in response to drainage. However, in an area with relatively natural vegetation, a transient drop in water table depth occurred at the time of drainage. Conclusions: Temporal and spatial patterns revealed that the bog response to drainage was spatially and temporally heterogeneous, and probably mediated by feedbacks among vegetation, peat structure and hydrology. Spatial patterns along the hydrosequence were similar to those observed in paleoecological reconstructions, but the use of the two complementary techniques provides additional insights.     

A mesocosm study of the effect of restoration on methane (CH4) emissions from blanket peat

In a mesocosm study, we investigated the effect of different restoration methods on methane (CH₄) emissions from, and the global warming potential (GWP) of, blanket peat. The controlled laboratory study involved two distinct components: Experiment 1 focused on greenhouse gas exchanges from blocked drains (grips) and evaluated the effects of restoration method, water-level dynamics and climate on CH₄ emissions and GWP. Experiment 2 assessed the role of plant functional type (PFT) on CH₄ emissions from restored peat outside of the grip. A nine month meteorological simulation (April–December) was completed, testing five hypotheses across the two experiments. We found that the method of grip blocking/damming does make a difference with respect to CH₄ emissions and GWP. Of the methods considered, damming with no infill between the dams is preferred to either of the methods involving infilling (heather bale and re-profiling). GWP of all within-grip restoration outcomes was positive (i.e., indicating a net warming effect), and was not influenced by climate or water-level regime. PFT influences CH₄ emissions but not GWP in restored blanket bog. When considering radiative forcing, this finding suggests that it does not matter which PFT dominates a restored area. It is noted that the laboratory findings are, in some senses, preliminary because the experiments consider only a relatively short period immediately after restoration.     

An evaluation of the use of translocated blanket bog vegetation for heathland restoration

Abstract. The decision was taken by an opencast coal mining company to translocate on-site blanket bog vegetation, on completion of mining, at a site in Co. Durham, UK, both to preserve it and to use it to enhance recolonization. The vegetation of the treatments was monitored for seven years after site completion and this paper reports on the progress of the translocated material and its effect on recolonization. Translocation of large turves of blanket bog into carefully prepared receptor cells preserved most of the vegetation intact, but resulted in severe decline in the frequency of Sphagnum, while the design of the receptor site as strips of translocated vegetation enclosing strips of spread, stored peat accelerated recolonization of the intervening bare peat by Calluna vulgaris, but not of other target species. This attempt to translocate blanket bog vegetation and at the same time use it to accelerate recolonization was only partly successful. It was concluded that the ecological requirements of species known to be significant for ecosystem function, such as Sphagnum, must be fulfilled if translocation of blanket bog is to be attempted in future.     

Vegetation and soils of the Meikle Kilrannoch ultramafic sites

Summary

The vegetation of two ultramafic sites (MK1 and MK2) at Meikle Kilrannoch are described. MK1 is dome shaped and has much weathering bedrock whilst MK2 is flatter, lacks weathering bedrock and has probably been entirely peat covered. Six vegetation Groups were recognized: I, high-level blanket bog; II and III, grass heath (with II more sedge-rich); IV, dwarf shrub heath; V and VI, debris (an open vegetation on stony skeletal soils). Vegetation maps were produced which had five mapping units: Group I, Groups II and IH combined, Group IV, Groups V and VI combined, and non-vegetated eroding peat. Three main soil types occur: peat, which underlies Group I; a complex of freely draining magnesian brown soils and imperfectly drained magnesian gleys which bear Groups II–IV; and skeletal soils, derived either from weathering bedrock (MK1 only) or ultramafic drift, which bear vegetation in Groups V or VI. The rare plants are commonest in, but not restricted to, debris vegetation on skeletal soils. The likely plant toxicity of soil magnesium at the sites is reaffirmed but it is suggested that the low plant cover in the debris is more likely to result from low nutrients or intensive frost action or both. Floristic differences between MK1 and MK2 are discussed and quantified for Lychnis alpina (which had 68000individuals on the former site and 46 on the latter) but the causes of the differences remain unexplained.     

Animal and vegetation patterns in natural and man-made bog pools: implications for restoration

1. Peatlands have suffered great losses following drainage for agriculture, forestry, urbanisation, or peat mining, near inhabited areas. We evaluated the faunal and vegetation patterns after restoration of a peatland formerly mined for peat. We assessed whether bog pools created during restoration are similar to natural bog pools in terms of water chemistry, vegetation structure and composition, as well as amphibian and arthropod occurrence patterns. 2. Both avian species richness and peatland vegetation cover at the site increased following restoration. Within bog pools, however, the vegetation composition differed between natural and man‐made pools. The cover of low shrubs, Sphagnum moss, submerged, emergent and floating vegetation in man‐made pools was lower than in natural pools, whereas pH was higher than in typical bog pools. Dominant plant species also differed between man‐made and natural pools. 3. Amphibian tadpoles, juveniles and adults occurred more often in man‐made pools than natural bog pools. Although some arthropods, including Coleoptera bog specialists, readily colonised the pools, their abundance was two to 26 times lower than in natural bog pools. Plant introduction in bog pools, at the stocking densities we applied, had no effect on the occurrence of most groups. 4. We conclude that our restoration efforts were partially successful. Peatland‐wide vegetation patterns following restoration mimicked those of natural peatlands, but 4 years were not sufficient for man‐made pools to fully emulate the characteristics of natural bog pools.     

Vegetation Dynamics on an Abandoned Vacuum-Mined Peatland: 5 Years of Monitoring

We studied from 1998 to 2002 the fine-scale vegetation dynamics of a poorly regenerated vacuum-mined bog located in southern Quebec. We selected mined sites that have been abandoned for 14 years and monitored the vascular and non-vascular plants, and some hydrological characteristics. We focussed our study on the monitoring of cotton-grass (Eriophorum vaginatum L.) tussocks. Major changes in the plant cover were observed during the five-year period, such as a decrease (26–31%) in the number of cotton-grass tussocks and an increase in the ericaceous shrub cover. The water table level (lower than 40 cm below the soil surface) and frost heaving appear to be the main factors explaining the failure of cotton-grass and of other typical bog plant species to colonize abandoned mined surfaces. The ericaceous shrub cover, although increasing, was still sparse even after two decades of abandonment, and it may take several additional decades before a complete shrub cover establishes itself. While the natural revegetation process of this vacuum-mined bog is still dynamic after two decades of abandonment, there is no evidence to suggest that vegetation assemblages will eventually resemble those of undisturbed peatlands.     

Moorland pools as refugia for endangered species characteristic of raised bog gradients

In intact raised bog landscapes transitions from ombrotrophic into minerotrophic conditions occur. These gradients are lost from many bogs due to peat harvesting and drainage, and are difficult to restore. To determine which endangered species are characteristic of pristine raised bog gradients and their current status in degraded bogs, plants and macroinvertebrates were surveyed in Estonian intact raised bogs and Dutch degraded bog remnants. Dutch national distribution data were used to determine whether communities with these species occurred outside bog habitats. Water chemistry data were used to describe associated environmental conditions. Intact bog gradients were the preferred habitat for six plant species and fifteen macroinvertebrate species, all of which are endangered. In degraded bogs these species were scarce or not recorded. In intact bogs these species lived at sites where runoff from the bog massif came into contact with regional ground water resulting in a gradient in pH, alkalinity, Ca, Fe and ionic ratio. Analysis of Dutch national distribution data revealed aggregations of these endangered species in moorland pools. These pools contained water chemistry gradients similar to those found in pristine bogs, which occurred at sites were groundwater seepage and stream water came in contact. In the past, stream water has been used to increase pH and trophic status of moorland pools facilitating fisheries. Today, this practice offers a conservation strategy for the protection of endangered species for which no short-term alternatives are available.     

The Regeneration of a Highly Disturbed Ecosystem: A Mined Peatland in Southern Québec

We studied the natural regeneration of an ombrotrophic peatland (Cacouna bog) located in southern Québec that was disturbed by peat mining and other anthropogenic activities over a 200-year period. Using an extensive collection of historical documents, as well as dendrochronological data, we reconstructed the history of the peatland. We also sampled vegetation and environmental variables, and integrated the data in a geographic information system. More than 60% of the total area of the bog was mined between 1942 and 1975, and 98 km of ditches were dug to drain the site. The peatland lost 34% of its initial peat volume between 1946 and 1998. Although the bog was severely disturbed, the spontaneous revegetation of the site by vascular plants was successful (90%–100% cover). However, only 10% of the total mined area has been recolonized by Sphagnum species, mainly because drainage ditches are still operational and contribute to drying out the bog. Water table level, peat deposit thickness, and pH are abiotic factors strongly influencing the vegetation composition in the bog. Spatial and historical factors are also important components in this study since they explain, either alone or in interaction with abiotic factors, 44% of the variation of the species data. The intensity of mining activities and the pattern of abandonment of mined sectors strongly influenced abiotic factors, which in turn affected the revegetation process. Even if the Sphagnum cover of the bog is low, the rapid “recovery” of the vegetation cover in the peatland indicates that after the reestablishment of an appropriate hydrological regime, a highly disturbed peatland has a considerable potential for regeneration. Received 24 April 2001; accepted 30 October 2001.     

Comparative landscape genetics of pond‐breeding amphibians in Mediterranean temporal wetlands: The positive role of structural heterogeneity in promoting gene flow

Comparative landscape genetics studies can provide key information to implement cost‐effective conservation measures favouring a broad set of taxa. These studies are scarce, particularly in Mediterranean areas, which include diverse but threatened biological communities. Here, we focus on Mediterranean wetlands in central Iberia and perform a multi‐level, comparative study of two endemic pond‐breeding amphibians, a salamander (Pleurodeles waltl) and a toad (Pelobates cultripes). We genotyped 411 salamanders from 20 populations and 306 toads from 16 populations at 18 and 16 microsatellite loci, respectively, and identified major factors associated with population connectivity through the analysis of three sets of variables potentially affecting gene flow at increasingly finer levels of spatial resolution. Topographic, land use/cover, and remotely sensed vegetation/moisture indices were used to derive optimized resistance surfaces for the two species. We found contrasting patterns of genetic structure, with stronger, finer scale genetic differentiation in Pleurodeles waltl, and notable differences in the role of fine‐scale patterns of heterogeneity in vegetation cover and water content in shaping patterns of regional genetic structure in the two species. Overall, our results suggest a positive role of structural heterogeneity in population connectivity in pond‐breeding amphibians, with habitat patches of Mediterranean scrubland and open oak woodlands (“dehesas”) facilitating gene flow. Our study highlights the usefulness of remotely sensed continuous variables of land cover, vegetation and water content (e.g., NDVI, NDMI) in conservation‐oriented studies aimed at identifying major drivers of population connectivity.     

Recovery of Moorland Vegetation after Aerial Spraying of Bracken (Pteridium aquilinum (L.) Kuhn) with Asulam

Bracken (Pteridium aquilinum) is a major weed of seminatural vegetation in Great Britain, as in many other countries. As a consequence, large areas have been subject to control with the intention of restoring the former vegetation. The use of aerial spraying with asulam, a narrow spectrum, carbamate herbicide, has become a common method of control. However, its long‐term efficacy has not been assessed in terms of either controlling the bracken or in producing suitable restored habitats. This study undertook such an assessment by comparing the results of a 2002 survey of the sites subject to bracken control by aerial spraying of asulam with previous (1990/1991 and 1994) surveys of the same sites. This showed that a single application of asulam was effective in eradicating bracken (<1% cover remaining) on a third of sites. However, on 10% of the sites, the bracken had regenerated completely (cover >80%) and on the remainder it was still present in patches (>20% of quadrats), often at high density. More than half the sprayed sites had seen good recovery of moorland vegetation, the target of the restoration, because they were now classified as having upland heathland vegetation within the National Vegetation Classification. Considerable amounts of bracken control are grant aided as part of agri‐environment schemes. These schemes should be adapted to encourage good practice, namely, intensive follow‐up treatment by spraying any emerging fronds, and to encourage treatment of previously sprayed areas rather than spraying of new areas in order to protect previous investment of grant aid.     

Vegetation development on moorland after control of Pteridium aquilinum with asulam

Abstract. A range of moorland sites in the North York Moors National Park were surveyed where Pteridium aquilinum has been sprayed once with the herbicide asulam between 1 and 7 yr before. The regeneration of both P. aquilinum and the moorland vegetation were studied. The latter was variable and much slower than reported for lowland grassland or heathland, with a large proportion of the developing plant cover consisting of mosses, especially Campylopus introflexus. This slow rate of revegetation was primarily due to the paucity of micro-sites offered for germination in dense P. aquilinum litter, the poor establishment of seedlings in this litter and its slow breakdown. The consequences of this slow and variable development of vegetation to moorland restoration are discussed.     

Effects of shallow flooding on vegetation and carbon pools in boreal peatlands

Question: What are the effects of shallow flooding on boreal peatlands on vegetation composition and size of carbon pools in the living and dead vegetation? Location: Lake 979, Experimental Lakes Area, northwestern Ontario, Canada. Methods: A boreal basin peatland complex with treed bog, open bog, and open water was experimentally flooded by raising water level ca. 1.3 m. Vegetation and above‐ground biomass were compared between pre‐flood conditions and those nine years after flooding. Peat accumulation since flooding was also quantified. Results: Flooding caused almost all trees to die, leading to a net loss of 86% of the above‐ground living plant biomass after nine years of the flooding. Floating up of peat was rapid in the central part of the basin, and the floating peat mats were characterized by newly established open bog community. Wetland types were diversified from bog into open bog, fen, and marsh, accompanied with great species turnover. Floating open bog community accumulated the greatest amount of peat since flooding. Conclusions: This study shows that shallow flooding of bog vegetation can lead to quick re‐establishment of open bog vegetation upon the floating up of peat mats as well as changes to more diverse vegetation over decadal time spans. We estimate that the carbon pools in 2002 in living and dead plant biomass since 1992 are comparable to what they were in the above‐ground biomass in 1992. Flooding caused an initial net decrease in carbon stores, but carbon in the pre‐flood living plant biomass was replaced by both carbon in dead biomass of the pre‐flood vegetation and newly sequestered carbon in new peat growth and post‐flood living plant biomass. Possible vegetation change toward bog‐dominated system could lead to increasing rate of new peat growth, which could affect future carbon sink/source strength of the system.