Untangling a Decline in Tropical Forest Resilience: Constraints on the Sustainability of Shifting Cultivation Across the Globe

Shifting cultivators depend on forest biomass inputs to nourish their crops. For them, forest resilience has an immediate impact: it affects crop productivity. A decline in the rate of recovery following shifting cultivation would ultimately affect local, regional and global carbon budgets, with feedbacks to climate. Yet the long-term impacts of shifting cultivation have been quantified in only six locations. In this study, we reanalyze data from these locations to determine whether the rate of biomass recovery is the same from cycle to cycle. Further, using case studies in Southern Yucatan, Mexico and West Kalimantan, Indonesia, we investigate the ecological and socioeconomic factors that affect forest resilience and thus determine whether or not shifting cultivation is sustainable. The reanalysis links aboveground biomass recovery following shifting cultivation to site productivity, forest age, fallow length, history of cultivation, and soil texture. Across locations, biomass accumulation rate declines by 9.3 percent with each cycle of shifting cultivation. Per cycle change in biomass accumulation rate is significantly more negative in younger forests and forests that experience a shorter fallow period. However, more detailed analyses for two case studies suggest that a purely ecological framework is of limited effectiveness in explaining variability in the effect of repeated shifting cultivation. Rather, socioeconomic factors such as migration, subsidies, roads, and settlement history can alter the outcome of shifting cultivation by limiting the accumulation and use of local knowledge.     

Biomass and carbon accumulation in a fire chronosequence of a seasonally dry tropical forest

Seasonally dry tropical forests (SDTF) are a widely distributed vegetation type in the tropics, characterized by seasonal rainfall with several months of drought when they are subject to fire. This study is one of the first attempts to quantify above- and belowground biomass (AGB and BGB) and above- and belowground carbon (AGC and BGC) pools to calculate their recovery after fire, using a chronosequence approach (six forests that ranged form 1 to 29 years after fire and mature forest). We quantified AGB and AGC pools of trees, lianas, palms, and seedlings, and BGB and BGC pools (Oi, Oe, Oa soil horizons, and fine roots). Total AGC ranged from 0.05 to nearly 72 Mg C ha⁻¹, BGC from 21.6 to nearly 85 Mg C ha⁻¹, and total ecosystem carbon from 21.7 to 153.5 Mg C ha⁻¹; all these pools increased with forest age. Nearly 50% of the total ecosystem carbon was stored in the Oa horizon of mature forests, and up to 90% was stored in the Oa-horizon of early successional SDTF stands. The soils were shallow with a depth of <20 cm at the study site. To recover values similar to mature forests, BGC and BGB required <19 years with accumulation rates greater than 20 Mg C ha⁻¹ yr⁻¹, while AGB required 80 years with accumulation rates nearly 2.5 Mg C ha⁻¹ yr⁻¹. Total ecosystem biomass and carbon required 70 and 50 years, respectively, to recover values similar to mature forests. When belowground pools are not included in the calculation of total ecosystem biomass or carbon recovery, we estimated an overestimation of 10 and 30 years, respectively.     

Love of Nurse Plants is Not Enough for Restoring Oak Forests in a Seasonally Dry Tropical Environment

The highest concentration of oak species in the world occurs in Mexico, but human activities have strongly degraded these oak forests. Mexican oaks have high economic, social, and cultural value, and restoring these forests is of paramount importance for the people of Mexico. Here, we propose a method for restoring oak forests using native shrubs that colonize degraded areas as nurse plants for oak seedlings. To test the viability of this proposal, seedling transplant experiments were performed in a degraded area near a protected oak forest relict. Two pioneer shrubs were identified as potential nurse species: Mimosa luisana and Senecio sp. The target oak species was Quercus castanea . Oak seedlings were located beneath the canopies of both shrubs and in the surrounding area without shrub cover. Water is a limiting resource for oak establishment in seasonally dry environments; therefore, we included irrigation systems in our experimental design to determine whether the combination of nurse plants plus watering led to higher rate of survival than the presence of nurse species alone. Seedling survival without watering was less than 20% both beneath nurse species and in the surrounding habitat. When water was supplied, survival rate beneath nurse species increased up to 58% while survival in the surrounding habitat did not differ from that observed in treatments without watering. Our results indicate that survival rate of oak seedlings is increased by the presence of nurse plants only when water is supplied. This suggests that restoration of oak forests in these degraded areas requires both nurse plants and watering.     

Forest Regeneration in a Chronosequence of Tropical Abandoned Pastures: Implications for Restoration Ecology

During the mid‐1900s, most of the island of Puerto Rico was deforested, but a shift in the economy from agriculture to small industry beginning in the 1950s resulted in the abandonment of agricultural lands and recovery of secondary forest. This unique history provides an excellent opportunity to study secondary forest succession and suggest strategies for tropical forest restoration. To determine the pattern of secondary succession, we describe the woody vegetation in 71 abandoned pastures and forest sites in four regions of Puerto Rico. The density, basal area, aboveground biomass, and species richness of the secondary forest sites were similar to those of the old growth forest sites (>80 yr) after approximately 40 years. The dominant species that colonized recently abandoned pastures occurred over a broad elevational range and are widespread in the neotropics. The species richness of Puerto Rican secondary forests recovered rapidly, but the species composition was quite different in comparison with old growth forest sites, suggesting that enrichment planting will be necessary to restore the original composition. Exotic species were some of the most abundant species in the secondary forest, but their long‐term impact depended on life history characteristics of each species. These data demonstrate that one restoration strategy for tropical forest in abandoned pastures is simply to protect the areas from fire, and allow natural regeneration to produce secondary forest. This strategy will be most effective if remnant forest (i.e., seed sources) still exist in the landscape and soils have not been highly degraded. Patterns of forest recovery also suggest strategies for accelerating natural recovery by planting a suite of generalist species that are common in recently abandoned pastures in Puerto Rico and throughout much of the neotropics.     

Distribution Patterns of Tropical Dry Forest Trees Along a Mesoscale Water Availability Gradient

Tropical dry forests (TDFs) host a large diversity of tree species but little is known of potential mechanisms that contribute to its maintenance. Given the paramount importance of water availability in such forests, tree species would be expected to show nonrandom patterns along water availability gradients, as well as differential individual species responses. In this work we explored whether that was true for 50 dominant tree species. Within a total area of 5.2 ha, divided into 26 transects each with ten 20 × 10 m plots, we registered presence–absence of these tree species with diameter at breast height ≥5 cm. We assessed the response of trees to four environmental variables differentially related to water availability in three steps: (1) identifying the shape of the response to individual environmental variables, (2) testing for artifacts in previous patterns due to spatial autocorrelation of presence–absence, and (3) identifying the environmental variable or combination of variables that best explained the pattern. We then classified the species with respect to their probability of occurrence along the gradient, and explored which components of the water cycle were likely to be driving the observed patterns. We found that 14 species were generalists, 16 drought tolerant, 9 intermediate, 3 water demanding and 8 showed mixed responses. Lateral flow and access to ground water most likely underlie such patterns. Our results confirm the key role played by water availability in tree species distribution. Water‐related niche differentiation seems to be crucial for maintaining the high diversity of this TDF. Abstract in Spanish is available at http://www.blackwell‐synergy.com/loi/btp .     

Effects of Season and Vegetation Type on Community Organization of Dung Beetles in a Tropical Dry Forest1

Dung beetles are important components of most terrestrial ecosystems. In tropical rain forests, dung beetle communities can be very rich in number of species and individuals, and they are known to be useful bioindicators of habitat disturbance. In contrast, very little is known about the organization of dung beetle communities in tropical dry forests. The aim of this study was to describe in detail the dung beetle community of a Mexican tropical dry forest and to assess the relative importance of rainfall seasonality and forest structure in affecting the temporal and spatial dynamics of this community. Dung beetles were captured with pitfall traps at the beginning of the rainy season, the middle of the rainy season, and the middle of the dry season, in two distinct forest types: deciduous forest (DF) and semideciduous forest (SDF) at the Estación de Biología Chamela. Both rainfall seasonality and forest structure affected the community organization of dung beetles. During both rainy periods, 14 species were captured, but only three during the dry season. Dung beetles captured during the dry season were only found in the SDF. When comparing the beginning and the middle of the rainy season, differences in abundance and guild structure were also observed between both periods and between forest types, but these differences were much less pronounced.     

Vegetation Structure,Composition, and Species Richness Across a 56‐year Chronosequence of Dry Tropical Forest on Providencia Island,Colombia1

We compared vegetation structure and species richness across a 56‐yr chronosequence of six replicated age classes of dry tropical forest on the island of Providencia, Colombia, in the Southwest Caribbean. Stand age classes were determined using sequential, orthorectified panchromatic aerial photos acquired between 1944 and 1996 and Landsat 7 ETM + satellite imagery from 2000. Along the chronosequence we established 59 plots of 2 × 50 m (0.01 ha) to document changes in species richness, basal area, tree height, stem density, and sprouting ability. All woody trees and shrubs >2.5 cm diameter at breast height (DBH) were censused and measured. Although woody species density reached a peak in stands from 32 to 56 yr old, rarefaction analysis showed that species richness increased linearly with stand age and was highest in stands 56 yr old or greater. Nonparametric, abundance‐based estimators of species richness also showed positive and linear associations with age. Basal area and mean tree height were positively associated with age since abandonment, while sprouting ability showed a negative relationship. Our results indicate rapid recovery of woody species richness and structural characteristics along this tropical dry forest chronosequence.     

Recovery of Amphibian and Reptile Assemblages During Old‐Field Succession of Tropical Rain Forests

Conversion of tropical forests to agriculture affects vertebrate assemblages, but we do not know how fast or to what extent these assemblages recover after field abandonment. We addressed this question by examining amphibians and reptiles in secondary forests in southeastern Mexico. We used chronosequence data (12 secondary forests fallow for 1–23 yr and 3 old‐growth forest sites) to analyze successional trajectories and estimate recovery times of assemblage attributes for amphibians and reptiles. We conducted 6 surveys at each site over 14 mo (1200 person‐hours) and recorded 1552 individuals, including 25 species of amphibians and 36 of reptiles, representing 96 and 74 percent of the expected regional number of species, respectively. Abundance, species richness, and species diversity of amphibians increased rapidly with successional age, approaching old‐growth forest values in < 30 yr. Species richness and species diversity of reptiles reached old‐growth forest values in < 20 yr. By contrast, the abundance of reptiles and the assemblage composition of amphibians and reptiles recovered more slowly. Along the chronosequence, we observed more species replacement in reptile assemblages than in amphibian assemblages. Several species in the old‐growth forest were absent from secondary forests. Dispersal limitation and harsh conditions prevailing in open sites and early successional environments appear to preclude colonization by old‐growth forest species. Furthermore, short fallow periods and isolation of forest remnants lead to the formation of new assemblages dominated by species favored by human disturbances.     

Seedling and Sprout Response to Slash-and-Burn Agriculture in a Tropical Deciduous Forest1

Survival and growth of seedlings and sprouts were assessed in three plots for 16 mo following the slashing and burning of a tropical deciduous forest in Jalisco, Mexico. We encountered a total of 47 species: 21 seedling species and 35 sprout species. Calliandra formosa and Piptadenia flava were the most common seedling species; Bursera arborea, Cordia alliodora, and Piptadenia constricta were the most common sprouts. Colubrina triflora, Diphysa occidentalis, and Cnidoscolus spinosus had limited sprouting ability. Twenty-six species were represented by one seedling or one sprout. Thirty-eight percent of the seedlings were tree species, 59 percent were shrub species, and 2 percent were vines species. In contrast, 86 percent of the sprouts were from trees and 13 percent from shrubs. One year after the initial measurements, 29 percent of the seedlings and 13 percent of the sprouts were dead. Each of the seedling means (number of stems/individual, height and diameter of the tallest stem, and elliptical crown area) was significantly smaller (P < 0.05) than that of sprouts at all three measurement periods, but relative growth rates were similar. Total canopy area of seedlings had a larger relative increase than did the canopy area of sprouts. The presence of seedlings increased species diversity compared to calculated diversity excluding seedlings. Timing of fruit dispersal in relation to the date of burning and the high number of Leguminosae species in the forest appeared to favor seedling establishment for some species.     

Testing Chronosequences through Dynamic Approaches: Time and Site Effects on Tropical Dry Forest Succession

Chronosequences, commonly used to assess succession, have been questioned because of their failure to project successional trajectories. Here, we develop a simple analytical approach combining both chronosequence and dynamic data to test the power of age of abandonment and site factors to explain and predict succession. The approach proceeds by first fitting statistical models relating age to attribute values (the chronosequence model) and their observed changes (the dynamic model) to test explanatory power. Predictive power is then tested by bootstrapping the chronosequence model to derive confidence intervals for expected changes and comparing them with the dynamic model. Finally, residuals from both models are tested against site factors. The procedure was applied to six attributes (basal area, plant density, mean plant height, species richness, evenness, and composition) of the woody community (plants >1 cm dbh within 0.1‐ha plots) in nine abandoned cattle pastures (0–12 yr) and three old growth tropical dry forests monitored over 6 yr. Age explained 60–97 percent of the variance in community attributes and only 32–57 percent in observed changes. It significantly overestimated basal area and mean height, while species richness and composition were highly predicted. Besides age, management history also explained successional dynamics. Our results suggest age is not necessarily a reliable predictor of short‐term successional dynamics, and explanatory power is not indicative of predictive power. Because of this low reliability, caution is needed when applying chronosequences to evaluate ecosystem services' recovery. The analytical approach developed here contributes to a better exploration of those possible limitations.     

Forest Succession in Tropical Hardwood Hammocks of the Florida Keys: Effects of Direct Mortality from Hurricane Andrew1

A tree species replacement sequence for dry broadleaved forests (tropical hardwood hammocks) in the upper Florida Keys was inferred from species abundances in stands abandoned from agriculture or other anthropogenic acitivities at different times in the past. Stands were sampled soon after Hurricane Andrew, with live and hurricane‐killed trees recorded separately; thus it was also possible to assess the immediate effect of Hurricane Andrew on stand successional status. We used weighted averaging regression to calculate successional age optima and tolerances for all species, based on the species composition of the pre‐hurricane stands. Then we used weighted averaging calibration to calculate and compare inferred successional ages for stands based on (1) the species composition of the pre‐hurricane stands and (2) the hurricane‐killed species assemblages. Species characteristic of the earliest stages of post‐agricultural stand development remains a significant component of the forest for many years, but are gradually replaced by taxa not present, even as seedlings, during the first few decades. This compositional sequence of a century or more is characterized by the replacement of deciduous by evergreen species, which is hypothesized to be driven by increasing moisture storage capacity in the young organic soils. Mortality from Hurricane Andrew was concentrated among early‐successional species, thus tending to amplify the long‐term trend in species composition.     

Insect Herbivores and Leaf Damage along Successional and Vertical Gradients in a Tropical Dry Forest

The availability and quality of resources for herbivores in tropical dry forests (TDFs) vary in time and space, affecting herbivore guilds differently across spatial scales (both horizontally and vertically), with consequences to the distribution of leaf damage in these forests. We attempted to elucidate the distribution patterns of herbivorous insect guilds and leaf damage throughout the secondary succession and vertical stratification along the rainy season in a Brazilian TDF. With the advance of the succession, a greater richness and abundance of herbivorous insects were found, resulting in higher leaf damage in intermediate and late stages. This pattern, however, was not observed for the frequency of leaf miners. At a smaller spatial scale, the host tree height positively affected the richness and abundance of insects. The higher leaf damage was found in canopy, which also harbored a greater richness and abundance of chewing herbivores compared to the understory at both the beginning and the end of the rainy season. Although for sap‐sucking insects, this was only true at the beginning of the season. We detected a decrease in insect richness and abundance at the end of the rainy season, probably due to a synchronization of insect activity with the availability of young, highly nutritious plant tissues. These results are consistent with other studies that found a general trend of increasing richness and abundance of herbivorous insects and leaf damage throughout the secondary succession (early to late stages) and between vertical strata (understory to canopy), suggesting that forest complexity positively affects herbivores.     

Effects of Selective Vegetation Thinning on Seed Removal in Secondary Forest Succession1

Seed removal was assessed for two tree species in three forest types: (1) secondary forest with and (2) without selective vegetation thinning, and (3) mature forest. Selective vegetation thinning meant the removal of all stems ≤3 cm in diameter of secondary‐forest species and was intended as a management technique to accelerate succession toward mature forest. Thinning did not have an effect on seed removal. One of the species showed lower seed removal in mature forest compared to secondary forest.     

Wildlife Recovery During Tropical Forest Succession: Assessing Ecological Drivers of Community Change

Despite the high proportion of secondary forests in the tropics, their conservation value remains poorly understood, particularly with regard to animals. Most theoretical studies of succession have focused on plants, linking life history trade‐offs to well‐known patterns of community change. However, the same trade‐offs proposed for plants should apply to animals, and indeed, animal studies show a change in community dominance from habitat generalist to forest specialist species during succession. Focusing on the diverse terrestrial small mammals of the endangered Atlantic Forest, we assessed which ecological drivers (habitat structure and food availability) affect community changes during succession. If the change in community dominance is driven by trade‐offs between productivity and efficiency, it should be mainly associated with a decrease in food availability. As expected, from younger to older forest, habitat generalists decreased in richness and total abundance, concurrent with a decrease in arthropod biomass. By contrast, the increase in richness and total abundance of forest specialists was not clearly supported by the data; however, this group was not affected by food availability. These results are congruent with a trade‐off between competitive ability and ability to use abundant resources, and indicate that the major community change during succession involves habitat generalists. Secondary forests may thus be valuable for conservation, at least where habitat loss and fragmentation are not high, and old growth forest is available.     

Seasonal Patterns of Flowering and Fruiting in a Dry Tropical Forest in Jamaica1

Tropical dry forests occupy more area and are more endangered than rainforests, yet their regeneration ecology has received less study and is consequently poorly understood. We recorded the flowering and fruiting phenology of a tropical dry forest in Jamaica over a period of 26 mo within ten 15 × 15‐m plots. Community‐wide recruitment reached a maximum in the wet season, whereas no recruitment occurred during the dry season. We observed a unimodal peak in rainfall and fruit production, and the periodicity and intensity of seed production were significantly correlated with rainfall seasonality (the optimal time for germination). Flowering at the community and system levels lagged behind a significant increase and subsequent decrease in rainfall by 7 and 3 mo, respectively, indicating that the dominant factor controlling flowering periodicity is the passage of the major (4‐mo long) rainy season and changes in soil moisture conditions. Fruiting lagged behind flowering by 2 mo and a significant increase in fruiting occurred 2 mo prior to a significant increase in rainfall. At the population level, a correspondence analysis identified a major dichotomy in the patterns of flowering and fruiting between species and indicated two broad species groups based on their time of peak fruiting and the number of times they were in fruit. These were either individuals which were usually in peak fruit 1–2 mo prior to the start of the major rainy season or those that were in fruit more or less continuously throughout the year with no peak fruiting time. This study supports the view that seasonal variation in rainfall and hence soil water availability constitutes both the proximate and the ultimate cause of flowering periodicity in tropical dry forests.     

Recovery of a Subtropical Dry Forest After Abandonment of Different Land Uses1

We studied the ecological characteristics of 45–50‐yr‐old subtropical dry forest stands in Puerto Rico that were growing on sites that had been deforested and used intensively for up to 128 yr. The study took place in the Guánica Commonwealth Forest. Our objective was to assess the long‐term effects of previous land use on this forest—i.e., its species composition, structure, and functioning. Previous land‐use types included houses, farmlands, and charcoal pits. Stands with these land uses were compared with a nearby mature forest stand. The speed and path of forest recovery after deforestation and land‐use abandonment depended on the conditions of the land. Study areas where land uses had removed the forest canopy and altered soil conditions (houses and farmlands) required a longer time to recover and had a different species composition than study areas where land uses retained a forest canopy (charcoal pits). Different forest attributes recovered at different rates. Crown area index, stem density, and litterfall rate recovered faster than stemwood and root, biomass, tree height, and basal area. Where previous land uses removed the canopy, Leucaena leucocephala, a naturalized alien pioneer species, dominated the regrowth. Native species dominated abandoned charcoal pits and mature forest. The change in species composition, including the invasion of alien species, appears to be the most significant long‐term effect of human use and modification of the landscape.     

Regional‐Scale Variation in Litter Production and Seasonality in Tropical Dry Forests of Southern Mexico1

Highly seasonal rainfall creates a pulse of litterfall in the southern Yucatan peninsula region, with cascading effects on the timing of essential nutrient fluxes, microbial dynamics, and vegetation growth. I investigated whether forest age or a regional environmental gradient related to rainfall has a greater effect on patterns of litterfall in this increasingly human‐dominated landscape. Litterfall was sampled in 10–13 stands in each of three locations spanning a rainfall gradient of ca 900–1400 mm/yr. Litter was collected monthly from November 1998 through January 2000 in mature forests and in secondary forests aged 2–25 yr. Despite a substantial precipitation gradient, age was the only significant predictor of annual litter mass. Two‐ to five‐yr‐old forests produced significantly less litter than 12–25‐yr‐old secondary forests (4.6 vs. 6.2 Mg/ha/yr), but the difference between older secondary forests and mature forests (9 percent) was not significant. Litter production increased with rainfall, but not significantly so. The pattern of litterfall was similar across locations and age classes, with a peak during late March or early April. However, litterfall seasonality was most pronounced in the old secondary and mature forests. Litterfall was more evenly distributed throughout the year in forests under 10 yr old. Seasonality of litterfall was also less pronounced at the wettest site, with less disparity between peak litterfall and off‐peak months. Seasonality was not related to soil texture. Forest age and rainfall are important drivers of litterfall dynamics; however, both litter mass and degree of seasonality depended more strongly on forest age. Thus, the impact of land‐use change on litter nutrient cycling is as great, if not greater, than the constraint imposed by the major natural environmental factor affecting tropical dry forests.     

Microclimate Change and Effect on Fire Following Forest-Grass Conversion in Seasonally Dry Tropical Woodland1

We tested the hypothesis that, where fire has historically been infrequent, wooded areas that have been invaded by grasses and converted to grassland by fire are predisposed to future fire compared to adjacent areas that remain wooded; thus, an initial forest fire may promote future fires. We compared microclimate between a grass-dominated burned area and a nearby grass-invaded woodland that has not burned in recent history, both located in the submontane dry forest of Hawaii Volcanoes National Park. The results were used to parameterize BEHAVE, a fire behavior prediction model developed by the USDA Forest Service. The model's predictions include probability of ignition, intensity, rate of spread, and tree mortality. Contrary to expectations, daytime hourly mean temperatures were higher and relative humidity was lower in the woodland site. However, the differences in temperature and humidity were not great enough to affect spread rate or probability of ignition. Wind speeds were substantially greater in the grassland, and this was most important in driving differences in modeled fire spread. Given similar synoptic conditions, a fire started in the grassland can be expected to spread an order of magnitude faster than one started in the woodland.     

Patterns and Correlates of Tropical Dry Forest Structure and Composition in a Highly Replicated Chronosequence in Yucatan,Mexico

Research on tropical dry forest (TDF) succession i0s needed for effective conservation and management of this threatened and understudied ecosystem. We used a highly replicated chronosequence within a 37,242‐ha TDF landscape to investigate successional patterns by plant size class and to evaluate the influence of stand age, topographic position, soil properties and spatial autocorrelation on forest structure and composition. We used a SPOT5 satellite image to obtain a land‐cover thematic map, and sampled woody vegetation (adults: >5 cm diam; saplings: 1–5 cm) and soil properties in 168 plots distributed among four vegetation classes: VC1 (3–8‐yr‐old forest), VC2 (9–15‐yr‐old forest), VC3 (>15‐yr‐old forest on flat areas), VC4 (>15‐yr‐old forest on hills). Stem density decreased with stand age and was lowest in VC3, while height, basal area and species density increased with age and were higher in older than in younger forests. Topographic position also influenced forest structure and composition. Basal area and height were largely determined by stand age, whereas stem and species density, and composition were influenced mostly by soil variables associated with fertility, and by spatial autocorrelation. Adults and saplings showed contrasting patterns and correlates of community structure, but similar patterns and correlates of composition, possibly due to the prevalence of coppicing. Our results show that our sampling approach can overcome several limitations of chronosequence studies, and provide insights in the patterns and drivers of succession, as well as guidelines for forest management and conservation. Abstract in Spanish is available at http://www.blackwell‐synergy.com/loi/btp .