The Initial Phase of a Longleaf Pine‐Wiregrass Savanna Restoration: Species Establishment and Community Responses

The significant loss of the longleaf pine‐wiregrass ecosystem in the southeastern United States has serious implications for biodiversity and ecosystem functioning. In response to this loss, we have initiated a long‐term and landscape‐scale restoration experiment at the 80,125 ha (310 mi²) Department of Energy Savannah River Site (SRS) located near Aiken, South Carolina. Aristida beyrichiana (wiregrass), an important and dominant grass (i.e., a “matrix” species) of the longleaf pine savanna understory, and 31 other herbaceous “non‐matrix” species were planted at six locations throughout SRS in 2002 and 2003. Of the 36,056 transplanted seedlings, 75% were still alive in June 2004, while mean 1–2 year survival across all planted species was 48%. Lespedeza hirta (hairy lespedeza) exhibited the greatest overall survival per 3 × 3 m cell at 95%, whereas Schizachyrium spp. (little bluestem) exhibited the greatest mean cover among individual species at 5.9%. Wiregrass survival and cover were significantly reduced when planted with non‐matrix species. Aggregate cover of all planted species in restored cells averaged 25.9% in 2006. High rates of survival and growth of the planted species resulted in greater species richness (SR), diversity, and vegetative cover in restored cells. Results suggest that the loss of the longleaf pine‐wiregrass ecosystem may be ameliorated through restoration efforts and illustrate the positive impact of restoration plantings on biodiversity and vegetative cover.     

Resource competition and fire‐regulated nutrient demand in carnivorous plants of wet pine savannas

Question: What are the mechanisms by which fire reduces competition for both a short‐lived and a long‐lived species in old‐growth ground‐cover plant communities of wet pine savannas (originally Pinus palustris, replaced by P. elliottii)? Location: Outer coastal plain of southeastern Mississippi, USA. Methods: I reviewed previous competition experiments and proposed a new hypothesis to explain the relationship between fire, competition, and species co‐existence in wet longleaf pine savannas. The first study is about growth and seedling emergence responses of a short‐lived carnivorous plant, Drosera capillaris, to reduction in below‐ground competition and above‐ plus below‐ground competition. The second study deals with growth and survival responses of a long‐lived perennial carnivorous plant, Sarracenia alata, to neighbour removal and prey‐exclusion to determine if a reduction in nutrient supply increased the intensity of competition in this nutrient‐poor system. Results: Fire increased seedling emergence of the short‐lived species by reducing above‐ground competition through the destruction of above‐ground parts of plants and the combustion of associated litter. Prey exclusion did not increase competitive effects of neighbours on the long‐lived species. However, because the experiment was conducted in a year without fire, shade reduced nutrient demand, which may have obviated competition for soil nutrients between Sarracenia alata and its neighbours. Conclusion: Repeated fires likely interact with interspecific differences in nutrient uptake to simultaneously reduce both above‐ground competition and competition for nutrients in old‐growth ground cover communities in pine savannas. Restoration practitioners should consider the possibility that the composition of the plant community is just as important as fire in ensuring that frequent fires maintain species diversity.     

Seed depredation negates the benefits of midstory hardwood removal on longleaf pine seedling establishment

Midstory hardwoods are traditionally removed to restore longleaf pine on fire‐excluded savannas. However, recent evidence demonstrating midstory hardwood facilitation on longleaf pine seedling survival has brought this practice into question on xeric sites. Also, midstory hardwoods could facilitate longleaf pine seedling establishment, as hardwood litter may conceal seeds from seed predators or improve micro‐environmental conditions for seedling establishment. However, little is known about these potential mechanisms. In this study, we tracked longleaf pine seed depredation and germination in artificially seeded plots (11 seeds/m²) in a factorial design fully crossing hardwood retention or removal with vertebrate seed predator access or exclusion in the Sandhills Ecoregion of North Carolina, U.S.A. Seed depredation averaged 78% across treatments and was greatest in unexcluded plots. Hardwood retention did not affect seed depredation. Longleaf pine averaged 3.6 germinants/4 m² across treatments, and was six times more abundant where vertebrates had been excluded. Hardwood removal had a strong positive effect on seedling germination, likely due to the removal of litter, but only when vertebrates were excluded. Our results indicated midstory hardwoods are not facilitating longleaf pine seedling establishment. Nevertheless, our results indicated that hardwood removal may not increase longleaf pine seedling establishment, as seed depredation diminished the effectiveness of hardwood removal under mast seed availability. Collectively, these results demonstrate the underlying complexity of the longleaf pine ecosystem, and suggest that planting may need to be part of the restoration strategy on sites where seed depredation limits longleaf pine natural regeneration.     

Patterns of herbaceous plant diversity in southeastern Louisiana pine savannas

Question: How diverse are Louisiana pine savanna plant communities and how is diversity affected by time since burn and removal of a competitively dominant species? Location: Lake Ramsay, southeastern Louisiana, USA. Methods: Species‐area curves were constructed from nine nested quadrats in open savanna differing in time since burn (6, 18 and 30 months). Species frequency was determined for 100 1 ‐m² quadrats. The dominant gras s, Andropogon virginicus, was removed with herbicide from moist and dry sites to test for possible effects of competition. Results: Slopes of log‐log species‐area relationships were steep (0.195 to 0.379). Time since burn did not affect the richness of herbaceous plants, only woody species. More than half of all species recorded (43/79, 54%) were infrequent (in < 10% of quadrats). After two years, there were no differences in species richness and composition for plots with and without A. virginicus. Conclusions: The high species diversity is typical of other savannas across the coastal plain. The large number of infrequent species indicates that the core‐satellite pattern of species occurrence found in temperate grasslands does not apply to southern pine savannas. The absence of effects due to removal of a dominant may be due to insufficient observation time, or low competition. Most species have traits, such as diminutive life forms, that suggest they are weak competitors for light in the presence of robust matrix grasses and in the absence of fire. Many species in Pinuspalustris savannas are likely either fugitive or peripheral species.     

Modelling responses of pine savannas to climate change and large‐scale disturbance

Global warming can potentially influence ecological communities through altered disturbance regimes in addition to increased temperatures. We investigate the response of pine savannas in the southeastern United States to global warming using a simple Lotka‐Volterra competition model together with predicted changes to fire and hurricane disturbance regimes with global climate change. In the southeastern United States, decreased frequency of both fires and hurricanes with global warming will shift pine savannas toward a forested state. A CO₂ fertilization effect that increases the growth rate of tree populations will also push southeastern landscapes from open savannas towards closed forests. Transient dynamics associated with climate driven changes in vegetation will last on the order of decades to a century. In our model, the sensitivity of savannas to relative changes in the frequency of fire versus hurricanes is linearly dependent on the growth rate and mortality of trees in fire and hurricane disturbances.     

Canopy thinning,not agricultural history,determines early responses of wild bees to longleaf pine savanna restoration

Longleaf pine savannas are highly threatened, fire‐maintained ecosystems unique to the southeastern United States. Fire suppression and conversion to agriculture have strongly affected this ecosystem, altering overstory canopies, understory plant communities, and animal populations. Tree thinning to reinstate open canopies can benefit understory plant diversity, but effects on animal communities are less well understood. Moreover, agricultural land‐use legacies can have long‐lasting impacts on plant communities, but their effects on animal communities either alone or through interactions with restoration are unclear. Resolving these impacts is important due to the conservation potential of fire‐suppressed and post‐agricultural longleaf savannas. We evaluated how historical agricultural land use and canopy thinning affect the diversity and abundance of wild bees in longleaf pine savannas. We employed a replicated, large‐scale factorial block experiment in South Carolina, where canopy thinning was applied to longleaf pine savannas that were either post‐agricultural or remnant (no agricultural history). Bees were sampled using elevated bee bowls. In the second growing season after restoration, thinned plots supported a greater bee abundance and bee community richness. Additionally, restored plots had altered wild bee community composition when compared to unthinned plots, indicating that reduction of canopy cover by the thinning treatment best predicted wild bee diversity and composition. Conversely, we found little evidence for differences between sites with or without historical agricultural land use. Some abundant Lasioglossum species were the most sensitive to habitat changes. Our results highlight how restoration practices that reduce canopy cover in fire‐suppressed savannas can have rapid benefits for wild bee communities.     

Natural disturbances and the physiognomy of pine savannas: A phenomenological model

Abstract. Question: The decline of the Pinus palustris ecosystems has resulted from anthropogenic influences, such as conversion to pine plantation forestry, agriculture and land development, all of which are closely related to increases in human populations. Other effects, however, have arisen from alterations in disturbance regimes that maintain the structure and function of these ecosystems. How have alterations of the disturbance regime altered the physiognomy of ‘old‐growth’ stands, and what are the implications for ecosystem conservation and restoration? Methods: In contrast to models that emphasize close interactions among the vertically complex strata, we develop a conceptual phenomenological model for the physiognomic structure of Pinus palustris stands. We relate two natural disturbances (tropical storms and fire) that affect different stages of the life cycle to different aspects of the physiognomic structure. We then compare overstorey stand structure and ground cover composition of two old‐growth longleaf stands near the extremes of different composite disturbance regimes: the Wade Tract (frequent hurricanes and fire) and the Boyd Tract (infrequent hurricanes and long‐term fire exclusion). Results: We predict that tropical storms and fires have different effects on stand physiognomy. Tropical storms are periodic, and sometimes intense, whereas fires are more frequent and less intense. Hurricanes directly influence the overstorey via wind‐caused damage and mortality, and indirectly influence the herb layer by altering the spatial distribution of shading and litter accumulation. Fire exerts direct effects on juvenile stages and indirect effects on the herb layer via fine fuel consumption and selective mortality of potential competitors of P. palustris juveniles. These differences in effects of disturbances can result in widely different physiognomies for P. palustris stands. Finally, some global climate change scenarios have suggested that changes may occur in tropical storm and fire regimes, altering frequency and severity. Such changes may greatly affect pine stands, and ultimately entire pine savanna ecosystems. Conclusions: Our phenomenological model of disturbance regimes in Pinus palustris old‐growth produces very different physiognomies for different disturbances regimes that reflect natural process and human management actions. This model can be used to derive restoration strategies for pine savannas that are linked to reinstitution of important ecological processes rather than specific physiognomic states.     

Distribution of native legumes (Leguminoseae) in frequently burned longleaf pine (Pinaceae)-wiregrass (Poaceae) ecosystems

Legume species distribution and abundance and selected environmental variables were quantified across a complex gradient (varying in both water-holding capacity and fertility) for frequently burned longleaf pine (Pinus palustris)–wiregrass (Aristida stricta) ecosystems. Legumes were present in all months; however, abundance peaked in June and was minimal after killing frosts in October. Legume species were prominent in the flora (43 species encountered) ubiquitous (94% of 2-m² subplots had at least one legume species), and abundant (nearly 120 000 stems/ha). Although most species were widely distributed throughout the gradient, Lespedeza angustifolia was distinctly associated with the more hydric end of the gradient, while both Petalostemon pinnatum and Galactia microphylla were located in the more xeric extreme. The percentage variation in species that could be accounted for by environmental variation was low (27%). Of the variation that could be accounted for, a number of environmental variables were important, including soil moisture, pine basal area (i.e., light), and bivalent base cations (e.g., Ca²⁺). Although gradients in resource availability among sites did not affect the distribution of species or abundance of legumes strongly, variation in resources are likely to regulate N₂-fixation rates of the various native legume species, and thereby affect ecological functions such as maintenance of N capital and productivity.     

Vicariance and endemism in a Neotropical savanna hotspot: distribution patterns of Cerrado squamate reptiles

Aim To test predictions of the vicariance model, to define basic biogeographical units for Cerrado squamates, and to discuss previous biogeographical hypotheses. Location Cerrado; South American savannas south of the Amazon, extending across central Brazil, with marginal areas in Bolivia and Paraguay and isolated relictual enclaves in adjacent regions. Methods We compiled species occurrence records via field sampling and revision of museum specimens and taxonomic literature. All species were mapped according to georeferenced locality records, and classified as (1) endemic or non‐endemic, (2) typical of plateaus or depressions, and (3) typical of open or forested habitats. We tested predictions of the vicariance model using biotic element analysis, searching for non‐random clusters of species ranges. Spatial congruence of biotic elements was compared with putative areas of endemism revealed by sympatric restricted‐range species. Effects of topographical and vegetational mosaics on distribution patterns were studied according to species composition in biotic elements and areas of endemism. Results We recorded 267 Cerrado squamates, of which 103 (39%) are endemics, including 20 amphisbaenians (61% endemism), 32 lizards (42%) and 51 snakes (32%). Distribution patterns corroborated predictions of the vicariance model, revealing groups of species with significantly clustered ranges. An analysis of endemic species recovered seven biotic elements, corroborating results including non‐endemics. Sympatric restricted‐range taxa delimited 10 putative areas of endemism, largely coincident with core areas of biotic elements detected with endemic taxa. Distribution patterns were associated with major topographical and vegetational divisions of the Cerrado. Endemism prevailed in open, elevated plateaus, whereas faunal interchange, mostly associated with forest habitats, was more common in peripheral depressions. Main conclusions Our results indicate that vicariant speciation has strongly shaped Cerrado squamate diversity, in contrast to earlier studies emphasizing faunal interchange and low endemism in the Cerrado vertebrate fauna. Levels of squamate endemism are higher than in any other Cerrado vertebrate group. The high number of recovered endemics revealed previously undetected areas of evolutionary relevance, indicating that biogeographical patterns in the Cerrado were poorly represented in previous analyses. Although still largely undocumented, effects of vicariant speciation may be prevalent in a large fraction of Cerrado and Neotropical biodiversity.     

Influence of CO2 enrichment and nitrogen fertilization on tissue chemistry and carbon allocation in longleaf pine seedlings

One-year old, nursery-grown longleaf pine (Pinus palustris Mill.) seedlings were grown in 45-L pots containing a coarse sandy medium and were exposed to two concentrations of atmospheric CO₂ (365 or 720 mol^-1) and two levels of nitrogen (N) fertility (40 or 400 kg N ha^-1 yr^-1) within open top chambers for 20 months. At harvest, needles, stems, coarse roots, and fine roots were separated and weighed. Subsamples of each tissue were frozen in liquid N, lyophilized at -50°C, and ground to pass a 0.2 mm sieve. Tissue samples were analyzed for carbon (C), N, nonpolar extractives (fats, waxes, and oils = FWO), nonstructural carbohydrates (total sugars and starch), and structural carbohydrates (cellulose, lignin, and tannins). Increased dry weights of each tissue were observed under elevated CO₂ and with high N; however, main effects of CO₂ were significant only on belowground tissues. The high N fertility tended to result in increased partitioning of biomass aboveground, resulting in significantly lower root to shoot ratios. Elevated CO₂ did not affect biomass allocation among tissues. Both atmospheric CO₂ and N fertility tended to affect concentration of C compounds in belowground, more than aboveground, tissues. Elevated CO₂ resulted in lower concentrations of starch, cellulose, and lignin, but increased concentrations of FWO in root tissues. High N fertility increased the concentration of starch, cellulose, and tannins, but resulted in lower concentrations of lignin and FWO in roots. Differences between CO₂ concentrations tended to occur only with high N fertility. Atmospheric CO₂ did not affect allocation patterns for any compound; however the high N treatment tended to result in a lower percentage of sugars, cellulose, and lignin belowground.Joseph W. Jones     

Effects of Hardwood Reduction Techniques on Longleaf Pine Sandhill Vegetation in Northwest Florida

We tested whether the intensity of hardwood midstory reduction causes commensurate improvements of herbaceous groundcover in fire‐suppressed Pinus palustris (longleaf pine) sandhills. Using a complete randomized block design, we compared the effects of three hardwood reduction techniques (spring burning, application of the ULW^® form of the herbicide hexazinone, chainsaw felling/girdling) and a no‐treatment control on plant species richness, and on life form and common species densities at Eglin Air Force Base, Florida, U.S.A., from 1995 to 1998. ULW^® and felling/girdling plots were burned for fuel reduction two years after initial treatment application. We also sampled the same variables in frequently‐burned reference sandhills to establish targets for restoration. Spring burns achieved partial topkill of oaks (17.6–41.1% from 1995 to 1998) compared to reductions of 69.1–94% accomplished by ULW^® and of 93.2–67.8% by felling/girdling treatments. We predicted that plant species richness and densities of herbaceous groundcover life forms would increase according to the percent hardwood reductions. Predictions were not supported by treatment effects for species richness because positive responses to fire best explained increases in plant richness, whereas ULW^® effects accounted for the largest initial decreases. Legumes, non‐legume forbs, and graminoids did not respond to treatments as predicted by the hypothesis. Again, positive responses to fire dominated the results, which was supported by greater herbaceous densities observed in reference plots. Overall, we found that the least effective and least expensive hardwood midstory reduction method, fire, resulted in the greatest groundcover improvements as measured by species richness and herbaceous groundcover plant densities.     

Carbon exchange of a mature,naturally regenerated pine forest in north Florida

We used eddy covariance and biomass measurements to quantify the carbon (C) dynamics of a naturally regenerated longleaf pine/slash pine flatwoods ecosystem in north Florida for 4 years, July 2000 to June 2002 and 2004 to 2005, to quantify how forest type, silvicultural intensity and environment influence stand‐level C balance. Precipitation over the study periods ranged from extreme drought (July 2000–June 2002) to above‐average precipitation (2004 and 2005). After photosynthetic photon flux density (PPFD), vapor pressure deficit (VPD) >1.5 kPa and air temperature <10 °C were important constraints on daytime half‐hourly net CO₂ exchange (NEE_day) and reduced the magnitude of midday CO₂ exchange by >5 μmol CO₂ m^?2 s^?1. Analysis of water use efficiency indicated that stomatal closure at VPD>1.5 kPa moderated transpiration similarly in both drought and wet years. Night‐time exchange (NEE_night) was an exponential function of air temperature, with rates further modulated by soil moisture. Estimated annual net ecosystem production (NEP) was remarkably consistent among the four measurement years (range: 158–192 g C m^?2 yr^?1). In comparison, annual ecosystem C assimilation estimates from biomass measurements between 2000 and 2002 ranged from 77 to 136 g C m^?2 yr^?1. Understory fluxes accounted for approximately 25–35% of above‐canopy NEE over 24‐h periods, and 85% and 27% of whole‐ecosystem fluxes during night and midday (11:00–15:00 hours) periods, respectively. Concurrent measurements of a nearby intensively managed slash pine plantation showed that annual NEP was three to four times greater than that of the Austin Cary Memorial Forest, highlighting the importance of silviculture and management in regulating stand‐level C budgets.     

New evidence for hybrid zones of forest and savanna elephants in Central and West Africa

The African elephant consists of forest and savanna subspecies. Both subspecies are highly endangered due to severe poaching and habitat loss, and knowledge of their population structure is vital to their conservation. Previous studies have demonstrated marked genetic and morphological differences between forest and savanna elephants, and despite extensive sampling, genetic evidence of hybridization between them has been restricted largely to a few hybrids in the Garamba region of northeastern Democratic Republic of Congo (DRC). Here, we present new genetic data on hybridization from previously unsampled areas of Africa. Novel statistical methods applied to these data identify 46 hybrid samples – many more than have been previously identified – only two of which are from the Garamba region. The remaining 44 are from three other geographically distinct locations: a major hybrid zone along the border of the DRC and Uganda, a second potential hybrid zone in Central African Republic and a smaller fraction of hybrids in the Pendjari–Arli complex of West Africa. Most of the hybrids show evidence of interbreeding over more than one generation, demonstrating that hybrids are fertile. Mitochondrial and Y chromosome data demonstrate that the hybridization is bidirectional, involving males and females from both subspecies. We hypothesize that the hybrid zones may have been facilitated by poaching and habitat modification. The localized geography and rarity of hybrid zones, their possible facilitation from human pressures, and the high divergence and genetic distinctness of forest and savanna elephants throughout their ranges, are consistent with calls for separate species classification.     

Growth responses of African savanna trees implicate atmospheric [CO2] as a driver of past and current changes in savanna tree cover

Atmospheric CO₂ has more than doubled since the last glacial maximum (LGM) and could double again within this century, largely due to anthropogenic activity. It has been suggested that low [CO₂] contributed to reduced tree cover in savanna and grassland biomes at LGM, and that increasing [CO₂] over the last century promoted increases in woody plants in these ecosystems over the past few decades. Despite the implications of this idea for understanding global carbon cycle dynamics and key global role of the savanna biome, there are still very few experimental studies quantifying the effects of CO₂ on tree growth and demography in savannas and grasslands. In this paper we present photosynthetic, growth and carbon allocation responses of African savanna trees (Acacia karroo and Acacia nilotica) and a C₄ grass, Themeda triandra, exposed to a gradient of CO₂ concentrations from 180 (typical of LGM) to 1000 µmol mol^?1 in open‐top chambers in a glasshouse as a first empirical test of this idea. Photosynthesis, total stem length, total stem diameter, shoot dry weight and root dry weight of the acacias increased significantly across the CO₂ gradient, saturating at higher CO₂ concentrations. After clipping to simulate fire, plants showed an even greater response in total stem length, total stem diameter and shoot dry weight, signalling the importance of re‐sprouting following disturbances such as fire or herbivory in savanna systems. Root starch (per unit root mass and total root starch per plant) increased steeply along the CO₂ gradient, explaining the re‐sprouting response. In contrast to the strong response of tree seedlings to the CO₂ gradient, grass productivity showed little variation, even at low CO₂ concentrations. These results suggest that CO₂ has significant direct effects on tree recruitment in grassy ecosystems, influencing the ability of trees to recover from fire damage and herbivory. Fire and herbivore regimes that were effective in controlling tree increases in grassy ecosystems could thus be much less effective in a CO₂‐rich world, but field‐based tests are needed to confirm this suggestion.     

Climate and land-use interactively shape butterfly diversity in tropical rainforest and savanna ecosystems of southwestern China

Human-induced habitat conversion and degradation, along with accelerating climatic change, have resulted in considerable global biodiversity loss. Nevertheless, how local ecological assemblages respond to the interplay between climate and land-use change remains poorly understood. Here, we examined the effects of climate and land-use interactions on butterfly diversity in different ecosystems of southwestern China. Specifically, we investigated variation in the alpha and beta diversities of butterflies in different landscapes along human-modified and climate gradients. We found that increasing land-use intensity not only caused a dramatic decrease in butterfly alpha diversity but also significantly simplified butterfly species composition in tropical rainforest and savanna ecosystems. These findings suggest that habitat modification by agricultural activities increases the importance of deterministic processes and leads to biotic homogenization. The land-use intensity model best explained species richness variation in the tropical rainforest, whereas the climate and land-use intensity interaction model best explained species richness variation in the savanna. These results indicate that climate modulates the effects of land-use intensity on butterfly alpha diversity in the savanna ecosystem. We also found that the response of species composition to climate varied between sites: specifically, species composition was strongly correlated with climatic distance in the tropical rainforest but not in the savanna. Taken together, our long-term butterfly monitoring data reveal that interactions between human-modified habitat change and climate change have shaped butterfly diversity in tropical rainforest and savanna. These findings also have important implications for biodiversity conservation under the current era of rapid human-induced habitat loss and climate change.