Restoring Quercus pyrenaica forests using pioneer shrubs as nurse plants

Question: How to improve reforestation success of Quercus pyrenaica. Location: 1800 m a.s.L, southern Spain. Methods: One‐year‐old Quercus pyrenaica seedlings were planted using two treatments: (1) bare soil, using a 30‐cm diameter augur bit (conventional technique) and (2) under the canopy of a pioneer shrub, Salvia lavandulifolia, using a 12‐cm diameter augur bit. Survival and growth were monitored for six years. Our hypothesis is that the use of shrubs as nurse plants is an alternative technique of reforestation with higher success than traditional techniques, in which pre‐existing vegetation is usually considered a source of competition. The rationale for the study was that for environments with a dry season, pre‐existing vegetation buffers summer drought stress, ameliorates the water status of seedlings and thus usually increases seedling recruitment. Results: Quercus survival was 6.3 × higher when planted under individuals of the pioneer shrub as compared to open areas. Quercus seedlings under shrubs also had shoots 1.8 X longer, while the number of shoots per plant did not differ among treatments. The first summer was the period with the highest mortality (49.1% of seedlings). Summer drought was the main cause of mortality. Conclusions: The use of shrubs as nurse plants for Q.pyrenaica reforestation is a viable technique to increase establishment success. The technique could be similarly useful in other environments with a dry period and for other Quercus species. In addition, this technique offers the advantage of following natural succession, thus minimizing the impact in the community.     

Effects of Planting Depth,Sediment Grain Size,and Nutrients on Ruppia maritima and Potamogeton perfoliatus Seedling Emergence and Growth

Protocols are now available for seed harvest, storage and germination of several mesohaline and polyhaline species; however, low seedling survival rates point to the need for an increased understanding of factors affecting seedling establishment. Depth of seed burial in sediments and initial seedling growth rates are shown to be limiting factors for photosynthetic competency of Ruppia maritima and Potamogeton perfoliatus. Seedling emergence is inversely proportional to planting depth on sediments ranging in grain size from coarse sands (850 μm) to silt (63 μm). Less than 6% of the seeds of either species emerged when buried to a depth of 3 cm in test sediments. Germination was greatest for seeds placed on the surface of sediments; however, these seedlings were subject to displacement because of the weak and fragile roots produced during early growth. Fine sediments may be more favorable for R. maritima seedling establishment, because seedling emergence and height decreased with increasing sediment grain size. Potamogeton perfoliatus seedlings seem to be more tolerant of a wider range of sediment grain sizes than R. maritima as indicated by the lack of an effect of sediment grain size on P. perfoliatus seed emergence, seedling height, and biomass. Increasing nutrients stimulated seedlings of both species; however, even at the highest concentrations tested, growth, as determined by shoot elongation and leaf and root formation, slowed within 7–10 days. This suggests factors other than mineral nutrients and light limit growth or that growth shifts from aboveground biomass production to belowground vegetative spread.     

Factors governing survival of black willow (Salix nigra) cuttings in a streambank restoration project

A field study was conducted at Little Topashaw Creek in northern Mississippi, aimed at expanding the limited database on the survivorship of Salix nigra (black willow) cuttings planted on riparian restoration sites. We tested the hypothesis that sediment moisture availability (deficit, excess) as mediated by sediment texture and depth to the prevailing water table is a major factor governing black willow survival during the initial stage of establishment following transplanting. Replicated plots were established across elevational gradients and a range of soil texture. Each plot contained 16 planted cuttings (2.5 cm diameter × 2.5 m length). Plot depth to water table, soil texture, and soil redox potential were measured. Plant gas exchange, leaf chlorophyll content, growth, and survival were monitored periodically over two growing seasons. Survival was best at low elevation compared to cuttings planted at mid- and high elevations. Poor survival and growth were noted for cuttings that encountered sediment moisture deficits in plots with coarse texture while the best cutting survival was recorded for intermediate sand content plots. Results indicated that plot location on the bank and soil texture are two important factors that influence riverbank restoration success. Therefore, any riparian restoration plan should include careful assessment of these factors prior to undertaking such efforts.     

Growth of Slender Banksia from seedling planting,with and without fertiliser

Slender Banksia (Banksia attenuata) is a primary component of declining Banksia woodlands around Perth, south‐western Australia. It is important that its re‐establishment be promoted, but there are little data on its growth rates and response to applied nutrients. To quantify longer‐term growth rate, I periodically measured heights of Slender Banksia planted mid‐2005 over 7½ years. Without fertiliser, these seedlings grew slowly to about 1½ m. In mid‐2009, I planted Slender Banksia on the same site, with and without fertiliser tablets, and evaluated survival, growth and root development over 1½ years. First‐summer Slender Banksia seedling survival of around 30% was not unusual for this species. Low‐phosphorus native plant fertiliser tablets increased growth significantly.     

Improving restoration success through research-driven initiatives: case studies targeting <Emphasis Type="Italic">Pinus pinea</Emphasis> reforestation stock development in Lebanon

Lebanon is located in one of the most ecologically diverse semi-arid regions of the world, characterized by endemism and richness of higher plants, which persist despite centuries of intensive agro-sylvo-pastoral use and more recent urbanization. Extensive deforestation, fires, and potential desertification have advanced concerns over the sustainability of naturally forested ecosystems in Lebanon, with only ca. 13 % forest cover remaining. Despite high numbers of seedlings planted in recent decades, reforestation success in Lebanon has been historically low. Increased success can be achieved through the effective production of high-quality native plant seedlings, capable of survival in increasingly arid climatic and harsh edaphic conditions. Through two trials focused on Pinus pinea as a model species, we aim to use a research-based approach to identify techniques required to produce high-quality seedlings to improve post-transplant performance in Lebanon. The first trial examines the suitability of using controlled release fertilizer for seedling production and evaluates the effect of fertilizer application rates on seedling morphological and tissue nutrient parameters. The second study examines the role of water availability during the three key periods of seedling development in the nursery to identify the optimal irrigation regime needed to achieve functional attributes associated with seedling drought and cold hardiness. Our findings serve as the first data-driven recommendations for improved seedling production in Lebanon and the greater region as well as demonstrate how locally integrated planning can result in effective capacity building.     

Factors Limiting the Survival of Native Tree Seedlings Used in Conservation Efforts at the Edges of Forest Fragments in Upland Madagascar

We tested four reforestation techniques in tropical forest fragments that were damaged by fire in upland Madagascar. We conducted a full‐factorial experiment on the survival of transplanted seedlings of five native tree species in grassland plots adjacent to the forest fragments in the Ambohitantely Forest Reserve. The species studied were Dodonaea madagascariensis, Filicium decipiens, Olea lancea, Podocarpus madagascariensis, and Rhus taratana. A total of 480 seedlings were planted; 207 survived the 15 months of the experiment. The factors examined were distance of the reforestation plots from the forest, mixing of forest soil into the plots, application of chemical fertilizers, experimental shading of plots, and the cover of naturally establishing shrubs. Both increasing the distance of plots from the forest edge and adding chemical fertilizers significantly reduced the survival of all seedlings. The surprising negative effects of fertilization may be partly due to increased competition from naturally establishing shrubs that are adapted to exploit high nitrogen levels. Mixing soil from the forest areas into the plots did not change seedling survival. Shading reduced the survival of D. madagascariensis seedlings and did not increase the survival of any species. These findings suggest that the success of reforestation projects can be increased by planting seedlings close to the existing forest fragments. Reforestation of similar tropical forests is likely to be more successful if efforts are focused on expanding the size of existing fragments of tropical forest, rather than on establishing new fragments in grassland openings.     

Sprouting success of shrubs after fire: height-dependent relationships for different strategies

The sprouting success of co-occurring populations of shrub species in a temperate woodland of semi-arid Australia was investigated and related to population survival strategies. Straw was added to 21 × 15 m plots in the woodland, burnt and the pre-fire characteristics of shrubs were used to determine the basis for sprouting success. Species differed widely (4–94%) in sprouting success; a high percentage of established seedlings of all species were killed by fire but survival increased with height reaching a maximum at 25–60 cm (depending on the species). Thickness of bark at stem bases increased with height growth but sprouting success was not related to bark thickness; sprouting success of shrubs at similar thickness varied greatly between species. All species were able to initiate sprouts after cutting through their basal stems, so lack of active meristems was not a limitation. Species differed in the height at which shrubs began flowering but this was always after maximum sprouting success was reached. It is proposed that differences between individual shrubs in supply of nutrients, carbohydrates, and/or water to activated meristems would account for patterns of in ter- and intra-specific sprouting success. The data are consistent with recognised fire survival strategies. `Sprouters', the species relying more on sprouting than recruitment for population persistence, maintained maximum sprouting success with height growth and gained sprouting ability along stems once they reached 1 m in height. In contrast, `non-sprouters', the species largely relying on recruitment from seed to maintain populations, were either not able to sprout after seedling establishment or steadily lost the ability to maintain sprouts with growth beyond 60 cm and did not develop axillary buds along stems at any height. Received: 19 July 1997 / Accepted: 8 February 1998     

Planting methods matter for cost‐effective rainforest restoration

Forest restoration efforts in Australia's Wet Tropics establish <100 ha/year, compared with 20‐year average clearing rates of 1661 ha/year. Establishment costs are serious impediments to restoration efforts. Costs range from $25,000 to $67,000/ha, compared with less than $5,000/ha in other areas using other methods. Some of this difference stems from methods used to plant trees. Augered holes are used commonly in environmental plantings, whereas planting spades are used mostly in forestry plantings. To determine the most cost‐effective planting method between auger and spade planting methods, we compared the planting costs and the survival and growth of seedlings of local rainforest species, The speed of planting with a spade is on average four times quicker than with an auger, whereas the survival (range = 89–94%, spade vs auger respectively), and growth (slightly greater height growth for auger planted individuals) rates were only marginally different. Given these results, using planting spades is a cost‐effective alternative to augering holes.     

Soil conditioning effects of Phragmites australis on native wetland plant seedling survival

Interactions between introduced plants and soils they colonize are central to invasive species success in many systems. Belowground biotic and abiotic changes can influence the success of introduced species as well as their native competitors. All plants alter soil properties after colonization but, in the case of many invasive plant species, it is unclear whether the strength and direction of these soil conditioning effects are due to plant traits, plant origin, or local population characteristics and site conditions in the invaded range. Phragmites australis in North America exists as a mix of populations of different evolutionary origin. Populations of endemic native Phragmites australis americanus are declining, while introduced European populations are important wetland invaders. We assessed soil conditioning effects of native and non‐native P. australis populations on early and late seedling survival of native and introduced wetland plants. We further used a soil biocide treatment to assess the role of soil fungi on seedling survival. Survival of seedlings in soils colonized by P. australis was either unaffected or negatively affected; no species showed improved survival in P. australis‐conditioned soils. Population of P. australis was a significant factor explaining the response of seedlings, but origin (native or non‐native) was not a significant factor. Synthesis: Our results highlight the importance of phylogenetic control when assessing impacts of invasive species to avoid conflating general plant traits with mechanisms of invasive success. Both native (noninvasive) and non‐native (invasive) P. australis populations reduced seedling survival of competing plant species. Because soil legacy effects of native and non‐native P. australis are similar, this study suggests that the close phylogenetic relationship between the two populations, and not the invasive status of introduced P. australis, is more relevant to their soil‐mediated impact on other plant species.     

Improving restoration success through microsite selection: an example with planting sagebrush seedlings after wildfire

Post‐fire restoration of foundation plant species, particularly non‐sprouting shrubs, is critically needed in arid and semi‐arid rangeland, but is hampered by low success. Expensive and labor‐intensive methods, including planting seedlings, can improve restoration success. Prioritizing where these more intensive methods are applied may improve restoration efficiency. Shrubs in arid and semi‐arid environments can create resource islands under their canopies that may remain after fire. Seedlings planted post‐fire in former canopy and between canopies (interspace) may have different survival and growth. We compared planting Wyoming big sagebrush (Artemisia tridentata Nutt. ssp. wyomingensis Beetle & Young) seedlings post‐fire in former sagebrush canopy and interspace microsites at five locations. Four growing seasons after planting, seedling survival was 46 and 7% in canopy and interspace microsites, respectively. Sagebrush cover was 5.8 times greater in canopy compared to interspace microsites. Sagebrush survival and cover were likely greater because of less competition from herbaceous vegetation as well as benefiting from resource island effects in canopy microsites. Initially, post‐fire abundance of exotic annual grasses was less in canopy microsites, but by the third year post‐fire it was substantially greater in canopy microsites, indicating that resource availability to seedlings was greater, at least initially, in canopy microsites. These results suggest microsites with greater likelihood of success should be identified and then utilized to improve restoration success and efficiency. This is important as the need for restoration greatly exceeds resources available for restoration.     

Benefits of Using Shrubs as Nurse Plants for Reforestation in Mediterranean Mountains: A 4-Year Study

Shrubs are commonly considered competitors of planted seedlings in reforestation programs. However, shrubs can facilitate the establishment of understory seedlings in environments that, like Mediterranean‐type ecosystems, are characterized by harsh environmental conditions. In 1997, an experiment was set up in the Sierra Nevada mountains (southeast Spain) to test the use of shrubs as nurse plants for an alternative reforestation technique. Two‐year‐old seedlings of Pinus sylvestris and Pinus nigra were planted in four microhabitats: (1) open interspaces without vegetation (which is the usual method employed in reforestation programs), (2) under individuals of the shrub Salvia lavandulifolia, (3) under the north side of spiny shrubs, and (4) under the south side of spiny shrubs. Seedlings were also distributed in plots with and without ungulates to test the effect of herbivore damage. We report here the results of survival and growth after four growing seasons, a time span long enough to draw robust conclusions concerning the suitability of this technique. Pine survival was remarkably higher when planted under individuals of S. lavandulifolia as compared with open areas (2.6 times for P. sylvestris and 1.8 for P. nigra). The survival of both pine species was also higher when planted on the north side of spiny shrubs, while mortality on the south side was similar to that found in open areas. The reduction of solar radiation by the canopy of shrubs was likely the main factor determining shrub facilitation. The growth of the pines differed among years. However, growth was not inhibited when planted with shrubs as compared with open areas in any of the years. Herbivore damage was low but was mostly concentrated in the leader shoot, exacerbating the deleterious effect of ungulate herbivores on pine growth. We conclude that the use of shrubs as nurse plants for reforestation is a viable technique to increase establishment success of reforestation in Mediterranean‐type ecosystems and that it might be similarly useful in other water‐stressed environments. In addition, this technique offers the advantage of following natural succession, thus minimizing the impact in the community.     

Legacy of thinning on woody species composition and structure in southern Appalachian Mountain hardwood forests: restoration implications

In the Appalachian Mountains, Liriodendron tulipifera monocultures are widespread, with these forests lacking both species and structural diversity. In this study, we developed models that described the effects of thinning treatments, conducted almost 60 years ago, on the density, composition, and functional identity of the woody understory in L. tulipifera forests. The woody understory of these thinned L. tulipifera forests was diverse, with the small seedling (<1.4 m), large seedling (≥1.4 m and <2.54 cm dbh), and sapling (≥2.54 and <12.7 cm) layers possessing 38, 32, and 23 species, respectively. Although model performance was low to moderate (r² = 0.05–0.40), we found that legacy effects, alone or in combination with environmental variables, explained, in part, the variability associated with the density, composition, and functional identity of the small seedling, large seedling, and sapling size classes, with the relative influence of legacy versus environmental effects varying by metric and size class. Post-thinning basal area and/or percent of basal area removed were not the primary legacy effects influencing the woody understory. Instead, legacy effects associated with species composition of the overstory before and/or after thinning along with average stem diameter post-thinning, variation in stem diameter post-thinning, and age at the time of thinning were more influential than density or thinning intensity. This study provides evidence that conserving species diversity during forest management activities can have positive long-term effects on composition and function of the woody understory and increase restoration potential.     

Use of shelter tubes,grass‐specific herbicide,and herbivore exclosures to reduce stressors and improve restoration of semiarid thornscrub forests

In semiarid south Texas, land conversion has reduced thornscrub forests by greater than 95%, and stressors, including competition with invasive grasses, mammalian herbivory, and drought, threaten the success of restoration efforts. This study assessed the effectiveness of multiple restoration treatments aimed at improving survival and growth of thornscrub forest seedlings planted in old agricultural fields. In January 2013, we treated greater than 1,100 seedlings with grass‐specific herbicide, herbivore exclosures, and shelter tubes, used separately or combined. We further evaluated the effects of shelter tube duration (0, 6, 12, and 18 months). For each seedling, we quantified surrounding invasive grass cover, browse intensity, height, and basal diameter every 4 months until September 2014. Herbicide application decreased invasive grass cover approximately 5‐fold and increased seedling survival (23%) and basal diameter (26%). Shelter tube application for 12 and 18 months increased seedling survival (10%) and height (43 and 74%, respectively), whereas seedlings treated with tubes for only 6 months performed similar to those left untreated. Exclosures had no impact on seedling survival but increased seedling height (23%) and basal diameter (26%). We found no significant interactive effects of treatments. Overall, herbicide most effectively increased seedling survival and basal diameter growth, whereas shelter tubes proved most useful for promoting height growth. Combined, these treatments increased implementation and maintenance costs 2‐fold, but minimized seedling mortality and maximized restoration potential. These findings highlight the necessity of post‐planting seedling management to reduce stress from invasive grasses, mammalian herbivory, and drought and improve restoration potential in semiarid thornscrub forests.     

Planting Seedlings in Tree Islands Versus Plantations as a Large‐Scale Tropical Forest Restoration Strategy

Planting tree seedlings in small patches (islands) has been proposed as a method to facilitate forest recovery that is less expensive than planting large areas and better simulates the nucleation process of recovery. We planted seedlings of four tree species at 12 formerly agricultural sites in southern Costa Rica in two designs: plantation (entire 50 × 50 m area planted) and island (six patches of three sizes). We monitored seedling survival, height, and canopy area over 3 years. To elucidate mechanisms influencing survival and growth, we measured soil and foliar nutrients, soil compaction, and photosynthesis. Survival of all species was similar in the two planting designs. Seedling height and canopy area were greater in plantations than islands at most sites, and more seedlings in islands decreased in height due to damage incurred during plot maintenance. Survival, height, and canopy area were both site‐ and species‐specific with the two N‐fixing species (Inga edulis and Erythrina poeppigiana) greater than the other species (Terminalia amazonia and Vochysia guatemalensis). Foliar N was higher in Terminalia and Vochysia in sites where Inga growth was greater. Soil nutrients, however, explained a small amount of the large differences in growth across sites. Leaf mass per area was higher in islands, and P use efficiency was higher in plantations. Our results show advantages (good seedling survival, cheaper) and disadvantages (more seedling damage, slightly lower growth) to the island planting design. Our study highlights the importance of replicating restoration strategies at several sites to make widespread management recommendations.     

Effect of biostimulation on the microbial community in PCB-contaminated sediments through periodic amendment of sediment with iron

Reductive dehalogenation of polychlorinated biphenyls (PCBs) by indigenous dehalorespiring microorganisms in contaminated sediments may be enhanced via biostimulation by supplying hydrogen generated through the anaerobic corrosion of elemental iron added to the sediment. In this study, the effect of periodic amendment of sediment with various dosages of iron on the microbial community present in sediment was investigated using phospholipid fatty acid analysis (PLFA) over a period of 18 months. Three PCB-contaminated sediments (two freshwater lake sediments and one marine sediment) were used. Signature biomarker analysis of the microbial community present in all three sediments revealed the enrichment of Dehalococcoides species, the population of which was sustained for a longer period of time when the sediment microcosms were amended with the lower dosage of iron (0.01 g iron per g dry sediment) every 6 months as compared to the blank system (without iron). Lower microbial stress levels were reported for the system periodically amended with 0.01 g of iron per g dry sediment every 6 months, thus reducing the competition from other hydrogen-utilizing microorganisms like methanogens, iron reducers, and sulfate reducers. The concentration of hydrogen in the system was found to be an important factor influencing the shift in microbial communities in all sediments with time. Periodic amendment of sediment with larger dosages of iron every 3 months resulted in the early prevalence of Geobacteraceae and sulfate-reducing bacteria followed by methanogens. An average pH of 8.4 (range of 8.2–8.6) and an average hydrogen concentration of 0.75% (range of 0.3–1.2%) observed between 6 and 15 months of the study were found to be conducive to sustaining the population of Dehalococcoides species in the three sediments amended with 0.01 g iron per g dry sediment. Biostimulation of indigenous PCB dechlorinators by the periodic amendment of contaminated sediments with low dosages of iron metal may therefore be an effective technology for remediation of PCB-contaminated sediments.     

Disentangling the effects of facilitation on restoration of the Atlantic Forest

Facilitation is an important ecological mechanism with potential applications to forest restoration. We hypothesized that different facilitation treatments, distance from the forest edge and time since initiation of the experiment would affect forest restoration on abandoned pastures. Seed and seedling abundance, species richness and composition were recorded monthly during two years under isolated trees, bird perches and in open pasture. Seed arrival and seedling establishment were measured at 10 m and 300 m from the forest edge. We sampled a total of 131,826 seeds from 115 species and 487 seedlings from 46 species. Isolated trees and bird perches increased re-establishment of forest species; however, species richness was higher under isolated trees. Overall, abundance and richness of seeds and seedlings differed between sampling years, but was unaffected by distance from the forest edge. On the other hand, species composition of seeds and seedlings differed among facilitation treatments, distance from the forest edge and between years. Seedling establishment success rate was larger in large-seeded species than medium- and small-seeded species. Our results suggest that isolated trees enhance forest re-establishment, while bird perches provide a complementary effort to restore tree abundance in abandoned pastures. However, the importance of seed arrival facilitation shifts toward establishment facilitation over time. Arriving species may vary depending on the distance from the forest edge and disperser attractors. Efforts to restore tropical forests on abandoned pastures should take into account a combination of both restoration strategies, effects of time and proximity to forest edge to maximize regeneration.     

Natural regeneration of selected timber species in the Republic of Congo

Natural regeneration of timber species is critical to the sustainable management of tropical forests. To understand what determines regeneration success of timber species in the Congo Basin, we evaluated whether seedling recruitment rates differed between forest logged 30 years previously and unlogged forest and determined the environmental factors that influence seedling density, growth and survival. We monitored the fate of 2186 seedlings of seven timber species within 462, 25‐m² plots located along 21 transects. We characterized seedling plots by light availability, soil nutrient availability and pH, and abundance of mammalian herbivores and then used linear and generalized linear mixed models to evaluate the variables that influenced seedling density, growth and survival. Light availability and canopy openness were 18% and 81% higher in logged than unlogged forest, and concentration of soil nutrients varied between sites. Seedling density was 32% higher in unlogged than logged forest. Taking all species together, seedling survival was positively correlated with calcium and negatively with magnesium and available phosphorus. Rates of seedling growth increased with available light. Taken separately, seedlings of the selected timber species responded differently to abiotic and biotic factors, demonstrating species‐specific regeneration requirements.     

Effect of Grazing on Restoration of Endemic Dwarf Pine (Pinus culminicola Andresen et Beaman) Populations in Northeastern Mexico

A pilot experiment designed to test the effect of cattle, small mammals, and elevation on the success of reforestation of an endemic dwarf pine species in northeastern Mexico was implemented. Pinus culminicola (Andresen et Beaman) grows only in four high peaks in the Sierra Madre Oriental and is under pressure from grazing, wildfires, and human activities such as mining, road development for timber extraction, and telecommunication and aerial navigation devices. We planted and monitored 2‐year‐old seedlings at three elevations within the natural distribution range of this species at Cerro El Potosí in Nuevo León, Mexico. At each elevation three treatments were established: (1) seedlings protected from cattle plus small mammals, (2) seedlings protected from cattle, and (3) seedlings with free access to cattle and small mammals. Seedling survival was approximately 50% in (1) after 4 years, but there were no surviving seedlings with free access to cattle. Elevation in general did not account for variation in survival. Seedling growth was poor during the 4 years, which implies that seedlings remain susceptible to grazing and trampling by cattle and small mammals. The implications for a large‐scale restoration program are discussed.     

Effect of the grass Leersia hexandra on the dispersal,seed germination,and establishment of Pachira aquatica seedlings

1. In wetlands, hydrochory is one of the main mechanisms of seed dispersal and there is often synchrony between propagule production and the flood season. Different sources of disturbance can prevent seed dispersal to suitable sites, and if environmental conditions are not adequate for germination and seedling establishment, recruitment will be limited, affecting succession.
2. We worked in a disturbed tropical freshwater swamp where the native grass Leersia hexandra has dominated open areas, creating a grass matrix that surrounds patches of swamp forest. Leersia grows vigorously, forming cushions of dry matter that cover the soil, forming a potential obstacle to seed dispersal. We asked whether the vegetative growth of this grass prevents the entry and dispersal of seeds of the tropical swamp tree Pachira aquatica, thwarting seed germination and seedling establishment, and arresting succession. We set up transects in the grass matrix in two zones: close to the river (R) and bordering the tree patches or fragments (F). We quantified tree seed and seedling presence, survival and growth in situ and experimentally introduced seeds and seedlings in the field and monitored seed germination and the survival and growth of their seedlings, as well as that of transplanted seedlings.
3. There was a negative relationship between the number of seeds and established seedlings, and the distance to river or fragment (r = −0.86, p < 0.001 for zone R; and r = −0.77, p < 0.001 for zone F) and with the grass cushion (r = −0.68, p = 0.005 for zone R; and r = −0.66, p = 0.007 for zone F); the grass creating a barrier to dispersal. When seeds were sown after clearing the grass cushion, germination success was high, so this stage is not limited. The transplanted seedlings had better survival and a greater final height than the seedlings of the sown seeds. Grass cover had a negative effect on both types of seedlings. Seedling survival rates were inversely related to grass cover, showing that seedlings overgrown by grass had low survival rates. Flooding is a stress factor for seedlings and produced mortality, in addition to the effects of the grass.
4. Together, the field survey and the experiment show that succession is being arrested in two ways: (1) by limiting seed dispersal because the grass cushion slows the dispersal and penetration of seeds into the vegetation; and (2) by limiting seedling establishment because the grass competes for space and light. Our results show that even where the grass is native, slower growing, seed-dependent species may struggle to compete and establish. If grass cover is increasing, these swamps are very vulnerable to a decrease in area because it is very difficult for them to regenerate naturally.