Community structure of ground beetles (Coleoptera: Carabidae) along an altitudinal gradient on Mt. Sobaeksan,Korea

We conducted this study to determine the diversity patterns, community structures, and seasonality of ground beetle assemblages along an altitudinal gradient (437–1420 m) on Mt. Sobaeksan, Korea. Ground beetles were collected by pitfall traps installed along an altitudinal gradient (437, 757, 1100, and 1420 m). A total of 32 species belonging to seven subfamilies were identified from 3259 collected ground beetles. The diversity pattern of the ground beetle assemblage according to altitude was neither monotonic decreased nor hump-shaped. However, subfamily assemblages and wing form diversity patterns differed according to altitude and may be correlated with altitude or some other environmental variables. The dominant species were Synuchus cycloderus (29.4%) and Eucarabus sternbergi sobaeksanensis (15.4%) and their seasonal activities according to altitude were similar. According to non-metric multidimensional scaling, ground beetles and altitudes could be divided into two distinct groups: a low altitude group (437–757 m) and a high altitude group (1100–1420 m). Some species were particularly abundant at high altitudes, such as Aulonocarabus koreanus kwonileeique and Poecilus nitidicollis, while others were highly abundant at low altitudes, such as Pterostichus audax, Pterostichus ishikawai, and Synuchus species.     

Impacts of forest gaps on butterfly diversity in a Bornean peat-swamp forest

Forest degradation is leading to widespread negative impacts on biodiversity in South-east Asia. Tropical peat-swamp forests are one South-east Asian habitat in which insect communities, and the impacts of forest degradation on them, are poorly understood. To address this information deficit, we investigated the impacts of forest gaps on fruit-feeding butterflies in the Sabangau peat-swamp forest, Central Kalimantan, Indonesia. Fruit-baited traps were used to monitor butterflies for 3 months during the 2009 dry season. A network of 34 traps (n_gap = 17, n_shade = 17) was assembled in a grid covering a 35 ha area. A total of 445 capture events were recorded, comprising 384 individuals from 8 species and 2 additional species complexes classified to genera. On an inter-site scale, canopy traps captured higher species richness than understory traps; however, understory traps captured higher diversity within each site. Species richness was positively correlated with percent canopy cover and comparisons of diversity indices support these findings. Coupled with results demonstrating morphological differences in thorax volume and forewing length between species caught in closed-canopy traps vs. those in gaps, this indicates that forest degradation has a profound effect on butterfly communities in this habitat, with more generalist species being favored in disturbed conditions. Further studies are necessary to better understand the influences of macro-habitat quality and seasonal variations on butterfly diversity and community composition in South-east Asian peat-swamp forests.     

How do evergreen and deciduous species respond to shade?—Tolerance and plasticity of subtropical tree and shrub species of South-East China

Different plant species make use of resource gradients such as light in different ways. First, plant species specialize in using different parts of the gradient, resulting in niche partitioning, Second, within the section of the gradient used by a species, plants respond to a different resource supply by plasticity. Our study addressed both of these strategies, with the main objective to relate mean responses and plasticity indices of seedlings of woody species to species characteristics such as leaf habit and to variation in branch lengths and local frequency of adult trees in forest communities.A greenhouse experiment with 36 deciduous and 35 evergreen subtropical tree and shrub species was carried out to test the influence of light and nutrient availability on trait expression and plasticity of the species. The greenhouse responses of seedlings were compared with adult individuals in the field, based on a set of 46 species that occurred also in the 27 permanent plots in a secondary subtropical broadleaved forest in Zhejiang Province (SE-China).In the greenhouse experiment, most variables showed significant differences between unshaded (250 μE m⁻² s⁻¹) and shaded (10 μE m⁻² s⁻¹) treatments as well as between high and low nutrient supply. Deciduous species were more plastic than evergreen species in their response to light. No significant correlations were detected in mean responses and plasticity between juvenile plants in the greenhouse and adult trees in the field. However, some trait values such as biomass and node density as well as plasticity of several traits were positively related to the species’ abundance in the field, suggesting that locally abundant species tend to be less “plastic” than locally rare species.     

Species richness,abundance, distributional pattern and trait composition of butterfly assemblage change along an altitudinal gradient in the Gulmarg region of Jammu & Kashmir,India

Despite enormous diversity, abundance, and role in ecosystem processes, little is known about how butterflies differ across altitudinal gradients. For this, butterfly communities were investigated along an altitudinal gradient of 2700–3200 m a.s.l, along the Gulmarg region of Jammu & Kashmir, India. We aimed to determine how the altitudinal gradient and environmental factors affect the butterfly diversity and abundance. Our findings indicate that species richness and diversity are mainly affected by the synergism between climate and vegetation. Alpha diversity indices showed that butterfly communities were more diverse at lower elevations and declined significantly with increase in elevation. Overall, butterfly abundance and diversity is stronger at lower elevations and gradually keep dropping towards higher elevations because floristic diversity decreased on which butterflies rely for survival and propagation. A total of 2023 individuals of butterflies were recorded belonging to 40 species, represented by 27 genera and 05 families. Six survey sites (S I- S VI) were assessed for butterfly diversity from 2018 to 2020 in the Gulmarg region of Jammu & Kashmir. Across the survey, Nymphalidae was the most dominant family represented by 16 genera and 23 species, while Papilionidae and Hesperiidae were least dominant represented by 01 genera and 01 species each. Among the six collection sites selected, Site I was most dominant, represented by 16 genera and 21 species, while Site VI was least dominant, represented by 04 genera and 04 species.     

The impact of forest encroachment after agricultural land abandonment on passerine bird communities: The case of Greece

Agricultural land abandonment is one of the main drivers of land use change, leading to various responses of farmland ecological communities. In an effort to better understand the effect of agricultural land abandonment on passerine bird communities, we sampled 20 randomly selected sites [1 km × 1 km] in remote Greek mountains, reflecting an abandonment gradient, in terms of forest encroachment. We sampled 169 plots using the point count method of fixed distance (47 passerine species), and we investigated bird diversity and community structure turnover along the gradient. We found that grazing intensity has a beneficial effect hampering forest encroachment that follows progressively land abandonment. Habitat composition changes gradually with forests developing at the expense of open meadows and heterogeneous grasslands. Forest encroachment has a significant negative effect on bird diversity and species richness, affecting in particular typical farmland and Mediterranean shrubland species. Birds form five distinct ecological clusters after land abandonment: species mostly found in pinewoods and cavity-dwelling species; species that prefer open forests forest edges or ecotones; species that prefer shrubland or open habitats with scattered woody vegetation; Mediterranean farmland birds that prefer semi-open habitats with hedges and/or woodlots; and, generalist forest-dwelling or shrubland species. We extracted a set of 22 species to represent the above ecological communities, as a new monitoring tool for agricultural land use change and conservation. We suggest that the maintenance of rural mosaics should be included in the priorities of agricultural policy for farmland bird diversity conservation.     

Taxonomic and functional diversity of stream invertebrates along an environmental stress gradient

Anthropogenic stress has been identified as main driver of freshwater biodiversity loss. Adverse effects on the biodiversity of freshwater organisms, such as macroinvertebrates, may propagate to associated ecosystem functions, such as organic matter breakdown (OMB). In this context, the functional diversity (FD) of communities has been suggested to be a more suitable predictor of changes in ecosystem functions than taxonomic diversity (TD). We investigated the response of TD and FD of invertebrate communities to an environmental stress gradient and the relation of both metrics to the rate of organic matter breakdown. For this, we sampled macroinvertebrates and determined OMB using leaf bags along an environmental stress gradient (i.e. changes in physicochemical and hydromorphological conditions) in 29 low-order streams. Taxonomic richness decreased with increasing environmental stress (r = −0.55) but was not related to OMB. Conversely, the Simpson diversity of communities was not associated with the gradient but correlated moderately (r = 0.41) with OMB. Of three functional diversity indices (functional richness, evenness and divergence), only functional richness correlated moderately with the stress gradient (r = −0.41) and any of the indices correlated with OMB. Nevertheless, functional metrics such as specific trait modalities and the total abundance of the dominant shredders correlated higher (r = 0.46 and 0.48) with OMB than the TD indices. Given a relatively small species pool in our study and methodical constraints such as the limited resolution of autecological information, the FD might perform better in other contexts and if focusing on response and effect traits for the stressor and ecosystem function under scrutiny, respectively.     

Tree species diversity and community composition in a human-dominated tropical forest of Western Ghats biodiversity hotspot,India

The structure, function, and ecosystem services of tropical forest depend on its species richness, diversity, dominance, and the patterns of changes in the assemblages of tree populations over time. Long-term data from permanent vegetation plots have yielded a wealth of data on the species diversity and dynamics of tree populations, but such studies have only rarely been undertaken in tropical forest landscapes that support large human populations. Thus, anthropogenic drivers and their impacts on species diversity and community structure of tropical forests are not well understood. Here we present data on species diversity, community composition, and regeneration status of tropical forests in a human-dominated landscape in the Western Ghats of southern India. Enumeration of 40 plots (50 m × 20 m) results a total of 106 species of trees, 76 species of saplings and 79 species of seedlings. Detrended Correspondence Analysis ordination of the tree populations yielded five dominant groups, along disturbance and altitudinal gradients on the first and second axes respectively. Abundant species of the area such as Albizia amara, Nothopegia racemosa and Pleiospermum alatum had relatively few individuals in recruiting size classes. Our data indicate probable replacement of rare, localized, and old-growth ‘specialists’ by disturbance-adapted generalists, if the degradation is continuing at the present scale.     

Species diversity of butterflies in Changbai Mountain in China

Using a modified belt transect method, we investigated the butterfly communities in five different vertical vegetation belts of Changbai Mountain in China from 1992 to 2009; these belts were broadleaf deciduous forest, coniferous–deciduous mixed forest, coniferous forest, erman’s birch forest and alp tundra. We determined the number of species and abundance of butterflies in each belt and in the coniferous–deciduous mixed forest belt, we also compared these parameters among different months. Preston’s lognormal distribution was used to model the species abundance distributions and five indicators (Shannon–Wiener diversity index (H′), Pielou uniformity index (J), Simpson predominance centralization index (C), Margalef abundance index (E) and Jaccard similarity coefficients) were used to analyze the butterfly community diversity. We found four main results. (1) Across all five vertical vegetation belts, 9641 butterflies were collected, belonging to 7 families, 98 genera and 196 species. As altitude increased, the number of butterfly genera and species gradually reduced. There was a relationship between the distribution of dominant species and the total species between each belt and the distribution of vascular plants. (2) The species abundance distribution was successfully modeled as a Preston’s lognormal distribution; the best fit was obtained when α = 0.326, the determinant coefficient of the equation was 0.74798. The species abundance distribution indicates that Changbai Mountain provides a suitable environment for butterflies; there was high species richness and an even distribution of butterfly species. There were few very common and very rare species, with most species having an intermediate abundance. (3) As altitude increased, H′ and E gradually became smaller, while C showed the opposite pattern, and J did not significantly change. The similarity coefficients analysis demonstrated a clear difference among belts; the farther apart any two belts, the smaller the similarity coefficient, indicating less similarity in the butterfly communities. The similarity coefficient between the deciduous forest and the coniferous–deciduous mixed forest belt was the largest (0.651) while that between the deciduous forest and the alp tundra was the smallest (0.141). (4) Comparison of the butterfly species communities among different months in the coniferous–deciduous mixed forest found that H′ and E showed similar directional changes, while the opposite pattern was found with C; the changes in J did not necessarily reflect the actual change in diversity.     

Recovery of ciliate communities from an oligotrophic mountain stream after a catastrophic wind damage

We studied the recovery of ciliate taxoceonoses in a mountain stream after a catastrophic windstorm that strongly affected mountainous regions of Slovakia in 2004. To this end, we analyzed changes in the community structure of ciliate assemblages from the Zubrovica stream in the Low Tatra Mts during the time frame between 2003 and 2008 by various statistical methods, including diversity and similarity indices, hierarchical clustering, multidimensional scaling, and PCA. The climax ciliate communities were characteristic for oligotrophic running waters, i.e., they were equitable (E_H = 0.75–1.00) and harbored a comparatively low number of species (2–9) with typically low abundances (10–125 ind./ml). However, the community structure changed dramatically after the devastating windstorm. There was a significant increase of species number (10–30) and abundance (260–1480 ind./ml), concomitant with a decrease of the equitability (E_H = 0.43–0.83). These changed quantitative and qualitative community parameters wore off comparatively quickly, i.e., about six months after the catastrophic windstorm, the ciliate taxocoenoses had reached a community structure similar to that before the wind damage. The present observations and those from terrestrial habitats indicate that ciliate communities have a good capability to comparatively quickly reach a climax even after a strong disturbance.     

Long-term effects of grazing on subalpine and alpine grasslands in the Central Alps,Austria

The effects of grazing exclusion on species diversity and functional diversity were analyzed along an elevation gradient from the subalpine (1960 m a.s.l.) to the lower and upper alpine zone (2275 m–2650 m a.s.l.) in the Austrian Central Alps for 15 years. Nine sites were chosen, including grasslands at different elevations, a bog and a glacier foreland site at the lower alpine zone and a snowbed at the upper alpine zone. Data were acquired by frequency counts in 1 m² permanent plots inside each fenced area (three plots per site) and outside in grazed areas (three plots). Diversity indices and functional diversity were analyzed by means of generalized linear mixed models (GLMMs). Exclosure, duration of exclusion and exclosure*years (interaction effect) were defined as predictor variables. Multivariate ordination techniques were used to (i) determine species responses to grazing exclusion (pRDA, partial redundancy analysis) and (ii) to create a distance matrix representing the changes between exclosures and control plots per year (NMDS, non-metric multidimensional scaling). At the subalpine grassland, first differences between exclosure and control plots occurred already only after three years, at the upper alpine zone after four and five years. Contrary to our expectation, dwarf shrubs did not increase within the exclosures of the subalpine grassland. Instead, mainly the tall forb Geranium sylvaticum increased. Species richness significantly decreased at the exclosures of the subalpine zone, the snowbed and at one upper alpine grassland sites. The communities of the glacier foreland and the bog were hardly affected by grazing exclusion.We conclude that plant species and communities react individually depending on elevation and grazing animals. Grazing exclusion studies at high elevations should definitely be carried out in the long-term.     

Plants as indicators of snow layer duration in southern Norwegian mountains

The main aim of this paper was to study the responses of mountain plants in relation to the time of snowmelt. Three mountain areas situated along an oceanic–continental gradient were selected as study sites, and the sample plots ranged from 182 m below to 473 m above the climatic forest limit. In total, 185 quadrats (2 m × 2 m), stratified to include only oligotrophic and mesotrophic mountain vegetation types, were selected to represent a topographic range along altitudinal gradients. In each quadrat, the percentage groundcover of the species was recorded. From the beginning of April until July 2004, snow thickness was monitored, and the Julian day when the snow had completely melted was determined for all plots. The relationship between species abundances and Julian day of snowmelt were analysed by two different numerical methods: (1) relative values for species optimum and tolerance were given by Detrended Canonical Correspondence Analysis (DCCA) with Julian day of snowmelt as the constraining variable. (2) Species responses were modelled by Generalized Linear Models (GLM). For species with significant unimodal responses, optimum and tolerance were calculated. For species with significant linear models, different species response models were identified by the regression intercepts. One hundred and twenty six species (taxa) were tested, and 103 evidenced statistically significant (p < 0.05) distribution responses. Several common alpine plants had a distribution that appeared to be independent of snow. On the basis of the results of the numerical methods, the species were separated into nine Snow Indicator (SI) classes, as a parallel to the Ellenberg indicator values. The species’ SI values were used to calculate weighted average SI values to examine the relationships between previously described plant communities and vegetation transects which experience different snow conditions.     

Vegetation mapping and multivariate approach to indicator species of a forest ecosystem: A case study from the Thandiani sub Forests Division (TsFD) in the Western Himalayas

QuestionsDoes the plant species composition of Thandiani sub Forests Division (TsFD) correlate with edaphic, topographic and climatic variables? Is it possible to identify different plant communities in relation to environmental gradients with special emphasis on indicator species? Can this approach to vegetation classification support conservation planning?LocationThandiani sub Forests Division, Western Himalayas.MethodsQuantitative and qualitative characteristics of species along with environmental variables were measured using a randomly stratified design to identify the major plant communities and indicator species of the Thandiani sub Forests Division. Species composition was recorded in 10 × 2.5 × 2 and 0.5 × 0.5 m square plots for trees, shrubs and herbs, respectively. GPS, edaphic and topographic data were also recorded for each sample plot. A total of 1500 quadrats were established in 50 sampling stations along eight altitudinal transects encompassing eastern, western, northern and southern aspects (slopes). The altitudinal range of the study area was 1290 m to 2626 m above sea level using. The relationships between species composition and environmental variables were analyzed using Two Way Cluster Analysis (TWCA) and Indicator Species Analysis (ISA) via PCORD version 5.ResultsA total of 252 plant species belonging to 97 families were identified. TWCA and ISA recognized five plant communities. ISA additionally revealed that mountain slope aspect, soil pH and soil electrical conductivity were the strongest environmental factors (p ≤ 0.05) determining plant community composition and indicator species in each habitat. The results also show the strength of the environment-species relationship using Monte Carlo procedures.ConclusionsAn analysis of vegetation along an environmental gradient in the Thandiani sub Forests Division using the Braun-Blanquet approach confirmed by robust tools of multivariate statistics identified indicators of each sort of microclimatic zones/vegetation communities which could further be used in conservation planning and management not only in the area studied but in the adjacent regions exhibit similar sort of environmental conditions.     

Changes of butterfly communities after forest fire

Fires change the diversity and composition of insects in forest ecosystems. In the present study, we examined the change of butterfly communities after a fire including the increase of butterfly richness, grassland species, and generalist species, and more changed communities. Butterflies were surveyed for 5 years after the big Uljin fire in 2007. During each year, butterflies were counted monthly by the line transect method from April to October at two sites (burned vs. unburned, ~ 1.5 km routes). Specialist grassland species decreased in the year of the fire but generalist species did not increase significantly. Butterfly richness did not change but butterfly diversity decreased due to a sudden increase of a species, Polygonia c-aureum. The butterfly community in the year of the fire was different from those in later years, showing temporary change of community in the year of the fire. Species composition was significantly different between burned and unburned sites, but this phenomenon cannot be interpreted as an influence of fire due to highly variable species composition of local butterfly assemblages and the non-repetitive sampling site of the present study.     

Pattern of ant diversity in Korea: An empirical test of Rapoport's altitudinal rule

Two diversity patterns (hump-shaped and monotonic decrease) frequently occur along altitude or latitude gradients. We examined whether patterns of ant species richness along altitudes in South Korea can be described by these patterns and whether ranges of ant species follow Rapoport's altitudinal rule. Ants on 12 high mountains (> 1100 m) throughout South Korea (from 33° N to 38° N) were surveyed using pitfall traps at intervals of 200–300 m altitude. The temperatures at the sampling sites were determined from digital climate maps. Ant species richness decreased monotonically along the altitudinal gradient and increased along the temperature gradient. However, species richness of cold-adapted species (highland species) showed a hump-shaped pattern along altitude and temperature gradients. The altitude and temperature ranges of ant species followed Rapoport's rule. Sampling site temperature ranges were significantly correlated with coldness. Therefore, Rapoport's rule can be explained by high cold-tolerance of species inhabiting high altitudes or latitudes.     

Effects of plant diversity on microbial biomass and community metabolic profiles in a full-scale constructed wetland

There has been less understanding of relations of microbial community patterns with plant diversity in constructed wetlands. We conducted a single full-scale subsurface vertical flow constructed wetland (SVFCW, 1000 m²) study focusing on domestic wastewater processing. This study measured the size and structure of microbial community using fumigation extraction and BIOLOG Ecoplate? techniques, to examine the effects of macrophyte diversity on microbial communities that are critical in treatment efficiency of constructed wetlands. We also determined the relationship of plant diversity (species richness) with its biomass production under disturbance of the same wastewater supply. Linear regression analysis showed that plant biomass production strongly correlated with plant species richness (R = 0.407, P < 0.001). Increase in plant species richness increased microbial biomass carbon and nitrogen (R = 0.494, P < 0.001; R = 0.465, P < 0.001) and utilization of amino acids on Ecoplates (R = 0.235, P = 0.03), but limited the utilization of amine/amides (R = ?0.338, P = 0.013). Principal components analysis (PCA) showed that the diversity and community-level physiological profiles (CLPP) of microbial community at 168 h of incubation strongly depended on the presence or absence of plant species in the SVFCW system, but not on the species richness. This is the first step toward understanding relations of plant diversity with soil microbial community patterns in constructed wetlands, but the effect of species diversity on microbial community should be further studied.     

Tree species composition rather than diversity triggers associational resistance to the pine processionary moth

The reduction of insect herbivory is one of the services provided by tree diversity in forest ecosystems. While it is increasingly acknowledged that the compositional characteristics of tree species assemblages play a major role in triggering associational resistance to herbivores, underlying mechanisms are less well known. We addressed this question in the ORPHEE experiment by assessing pine processionary moth infestations (Thaumetopoea pityocampa) across a tree diversity gradient from pine monocultures to five species mixtures. We showed that tree species richness per se had no effect on the probability of attack by this pest. By contrast, the infestation rate was strongly dependent on plot composition. Mixtures of pines (Pinus pinaster) and birches (Betula pendula) were less prone to T. pityocampa infestations, whereas mixtures of pines and oaks (Quercus spp.) were more often attacked than pine monocultures. By taking into account the relative height of pines and associated broadleaved species, this effect could be explained by pine apparency. Pines were on average 343 ± 5 cm height. Birches, as fast growing trees, were slightly taller than pines (363 ± 6 cm), while oak trees were significantly smaller (74 ± 1 cm). Host trees of T. pityocampa were then partly hidden in mixtures of pines and birches but more apparent in mixtures with oaks. We suggest that reduced pine apparency disrupted visual cues used by female moths to select host trees prior to oviposition. This study highlights the need to take into account tree traits such as growth rate when selecting the tree species that have to be associated in order to improve forest resistance to pest insects.     

Effects of plant diversity on biomass production and substrate nitrogen in a subsurface vertical flow constructed wetland

Most biodiversity experiments have been conducted in grassland ecosystems with nitrogen limitation, while little research has been conducted on relationships between plant biomass production, substrate nitrogen retention and plant diversity in wetlands with continuous nitrogen supply. We conducted a plant diversity experiment in a subsurface vertical flow constructed wetland for treating domestic wastewater in southeastern China. Plant aboveground biomass production ranged from 20 to 3121 g m^?2 yr^?1 across all plant communities. In general, plant biomass production was positively correlated with species richness (P = 0.001) and functional group richness (P = 0.001). Substrate nitrate concentration increased significantly with increasing plant species richness (P = 0.046), but not with functional group richness (P = 0.550). Furthermore, legumes did not affect biomass production (P = 0.255), retention of substrate nitrate (P = 0.280) and ammonium (P = 0.269). Compared to the most productive of the corresponding monocultures, transgressive overyielding of mixed plant communities did not occur in most polycultures. Because greater diversity of plant community led to higher biomass production and substrate nitrogen retention, thus we recommend that plant biodiversity should be incorporated in constructed wetlands to improve wastewater treatment efficiency.     

Effect of fire on soil nutrients and under storey vegetation in Chir pine forest in Garhwal Himalaya,India

The study was carried out in the Pinus roxburghii Sargent (Chir pine) forest in the sub-tropical region of Garhwal Himalaya to assess the effect of fire on soil nutrient status at different altitudes (700 m, 800 m and 1000 m), soil depths (0–20 cm, 20–40 cm and 40–60 cm) and on under storey vegetation. The soil nutrients and under storey vegetation were assessed before fire (pre-fire) and after fire (post-fire). The results of the study indicate that fire plays an important role in soil nutrient status and under storey vegetation. The nutrients (soil organic carbon, nitrogen, phosphorus and potassium), decreased in post-fire assessment and with increasing altitudes, and soil depths, compared to pre-fire assessment. The under storey vegetation diminished after fire in all forest sites. The study concludes that in Chir pine forest, fire plays a role in reducing soil nutrients along the altitudinal gradient, soil depths and under storey vegetation. Thus, these nutrients can be saved through some management practices e.g. by early controlled burning and by educating local villagers about the negative impacts of severe wild fires on soil and vegetation.     

Post-fire dynamics of woody vegetation in seasonally flooded forests (impucas) in the Cerrado-Amazonian Forest transition zone

Fire disturbance alters the structural complexity of forests, above-ground biomass stocks and patterns of growth, recruitment and mortality that determine temporal dynamics of communities. These changes may also alter forest species composition, richness, and diversity. We compared changes in plant recruitment, mortality, and turnover time over three years between burned and unburned sites of two seasonally flooded natural forest patches in a predominantly savanna landscape (regionally called ‘impucas’) in order to determine how fire alters forest dynamics and species composition. Within each impuca, 50 permanent plots (20 m × 10 m) were established and all individuals ≥5 cm diameter at breast height (DBH) identified and measured in two censuses, the first in 2007 and the second in 2010. Unplanned fires burned 30 plots in impuca 1 and 35 in impuca 2 after the first census, which enabled thereafter the comparison between burned and unburned sites. The highest mortality (8.0 and 24.3% year⁻¹ for impuca 1 and 2) and turnover time (69 and 121.5 years) were observed in the burned sites, compared to 3.7 and 5.2% year⁻¹ (mortality), and 28.4 and 40.9 years (turnover), respectively, for the unburned sites. Although these seasonally flooded impuca forests are embedded in a fire-adapted savanna landscape, the impucas vegetation appears to be sensitive to fire, with burned areas having higher mortality and turnover than unburned areas. This indicates that these forest islands are potentially at risk if regional fire frequency increases.