A multi-region assessment of tropical forest biodiversity in a human-modified world

The fate of much of the world’s terrestrial biodiversity is linked to the management of human-modified forest landscapes in the humid tropics. This Special Issue presents the first pan-tropical synthesis of research on the prospects for biodiversity in such systems, with eight individual regional summaries covering Mesoamerica, Amazonia, Atlantic forest of South America, West Africa, Madagascar, Western Ghats, Southeast Asia and Oceania. Two additional papers compare the state of conservation science in tropical forests with both temperate forests and savannah systems. This overview paper provides a comparative analysis of the threats and opportunities facing tropical forest biodiversity, thereby helping to identify the most pressing areas of future research and region-specific factors that contribute towards the effectiveness of individual conservation initiatives. While many of the threats facing tropical forest biodiversity are commonplace they vary markedly in their relative importance across different regions. There is a critical lack of comparable data to understand scale dependent processes, or the relative importance of varying geographic and historical contexts in determining present-day patterns. Conservation science has a key role to play in safeguarding the future of tropical forest biodiversity, but needs to become more effectively embedded in the context of real-world conservation challenges and opportunities. Significant progress can be achieved by improving the cost-effectiveness of research as well as the exchange of ideas and data amongst scientists working in different, often isolated parts of the world. We hope this special issue goes some way top achieving this exchange of knowledge.     

More than just indicators: A review of tropical butterfly ecology and conservation

Roughly 90% of butterfly species live in the tropics. Despite this, we know very little about tropical butterfly ecology particularly when compared to temperate butterfly systems. The relative scarcity of data on tropical butterfly populations hampers our ability to effectively conserve them. In this review we summarize recurring themes from ecological research on tropical butterflies to serve as a framework for understanding their conservation. Key themes include: (1) the tropics represent the evolutionary origins of butterfly diversity, (2) while some tropical butterflies exhibit relatively stable population dynamics, longer-lived adult stages, and more continuous age-specific reproduction compared to temperate zone species, the generality of these patterns is debatable, and (3) complex species interactions (e.g. mimicry, parasitism and predation) can have significantly greater influences on ecological and evolutionary processes in tropical butterflies than in temperate ones. This state of ecological knowledge, combined with scarce resources, has traditionally constrained tropical butterfly conservation efforts to habitat level approaches, unlike the species- and population-specific approaches familiar in North America and Europe. Consequently, much conservation research on butterflies in the tropics has focused on the relationship between habitat quality (e.g. forest fragmentation) and butterfly diversity, though predictive patterns even in this regard remain elusive. We argue that with the increasing threats of habitat destruction, fragmentation and climate change, it is necessary to move beyond this diversity and habitat relationship if we are to improve predictive capabilities when evaluating anthropogenic impacts on tropical butterfly communities. Tropical butterflies are more than just useful indicator species. They represent some of the most spectacular and visually appealing organisms in the world and play many vital roles in tropical ecosystems. We hope that this synthesis will lay the groundwork for future ecological studies of tropical butterfly populations, species, communities and conservation.     

High rates of forest clearance and fragmentation pre- and post-National Park establishment: The case of a Jamaican montane rainforest

The principal defense of tropical forests has been to designate them as protected areas, but few of these protected areas have been studied to determine their effectiveness at preventing forest clearance within their boundaries. We used remote sensing techniques to quantify three fundamental forest cover change processes: forest clearance, re-growth and fragmentation over a 19-year time period, in a National Park of Jamaica. Results show that forest clearance occurred at similar high levels prior and subsequent to the establishment of the area as a National Park in 1993: 1.50% yr⁻¹ (1983–1992) and 1.42% yr⁻¹ (1992–2002), respectively. Fragmentation continued post-establishment, and manifested itself in an increasing number of smaller more vulnerable fragments; the number of fragments increased by 60%, and the mean fragment size decreased by 40%. Core areas decreased with ensuing increases in edge lengths, and fragments became more isolated from one another. After designation as a National Park however, increased forest re-growth occurred, resulting in a 63% decline in the net deforestation rate from 0.80% yr⁻¹ (1983–1992) to 0.26% yr⁻¹ (1992–2002). Image classifications gave accuracies of 90–97% (Kappa 0.80–0.93). Although the decrease in net deforestation may indicate some level of success of this National Park, forest clearance and fragmentation, leading to species shifts and biodiversity loss were still present at pre-National Park levels. We then examined the preponderance of successful protected areas, and found that success may be an artifact of the most widely used method of assessment, which compares deforestation rates inside versus outside protected areas.     

Biodiversity conservation in tropical forest landscapes of Oceania

The fragmented island realm of Oceania includes a relatively small proportion of the world’s tropical forests, but those forests support unusual richness of narrowly endemic species. In common with tropical forests across most of the world, tropical forests in Oceania are declining due to factors associated with increasing human population size, economic drivers and more intensive exploitation. In parts of Oceania, forests are being cleared at unsustainable rates, and replaced with far simpler ecosystems of timber or food crops. To a small degree, the present-day biodiversity of tropical forests in some parts of Oceania may be predisposed to such disturbance, given a history of natural disturbance (particularly through cyclones), and of smaller-scale slash-and-burn agriculture or landscape-scale burning. But, in most places, the current intensity, scale and/or rate of modification far surpass their precedents, and biodiversity is consequently diminishing. Tropical forests in Australia may be an exception to this trend, with now reasonably effective protection. However, more so than for tropical forests in most other continents, the major biodiversity conservation challenges for tropical forests in Oceania are extrinsic. Introduced plants, animals and diseases have collapsed ecological communities through much of Oceania, homogenising the biota from a series of highly distinctive and localised species assemblages to a more impoverished set of ubiquitous disturbance-tolerant exotic species. In many islands, this simplification has occurred regardless of the extent of native forest remaining, such that retention and reservation of primary forest is an insufficient conservation action. The fate of biodiversity in Oceania is also likely to be much affected by climate change, an unbalanced consequence given the region’s relatively small contribution to greenhouse gas emissions. Future hope for biodiversity conservation in tropical forests of Oceania lies in the renewed application of some traditional management constraints, the appropriate delivery of international support (such as may be available through carbon markets), improved quarantine processes, and through some protection naturally offered by the remote scattering of the islands that comprise Oceania.     

Methods for conservation outside of formal reserve systems: The case of ants in the seasonally dry tropics of Veracruz, Mexico

Like most ecosystems of the world, tropical dry forests of the central coast of the Gulf of Mexico are inadequately preserved. Given that reserve expansion is unlikely, it is imperative that the conservation capacity of the matrix surrounding reserves is enhanced. Here, we examine the habitat value of isolated pasture trees and patches of secondary regrowth in terms of their terrestrial and arboreal ant assemblages in both a wet and dry season. These simplified wooded systems increase species densities within the agricultural matrix and provide habitat for some forest ant species. Estimated species richness of arboreal ants was particularly low on isolated trees, highlighting an important limitation. This was not the case for terrestrial ants, which were particularly species rich under isolated trees. We also found that the inter-site variations in species densities and similarity to the forest ant assemblage for terrestrial and arboreal strata were not correlated, suggesting that responses to restoration may not be as uniform as often thought. This has important implications for the use of indicator taxa in suggesting the response of other taxa. In terms of species composition, neither secondary regrowth nor isolated trees were appropriate replacements for forest fragments, even though the studied forest fragments were small (13-32 ha). The ant assemblages did not exhibit a seasonal change in composition. However, season influenced the contrast between habitats, with isolated trees being more distinct from pasture, and regrowth more closely resembling forests, during the wet season. Microclimatic variables indicate that the forests were least affected by the tropical dry season, and this may contribute to their characteristic fauna. We conclude that even small forest patches make a unique contribution to landscape conservation and that, where reserves are limited, conscientious management of the landscape matrix may provide some species with sufficient new habitat to survive outside of reserve systems. These conclusions are influenced by both season and strata studied.     

Lowland rainforest avifauna and human disturbance: persistence of primary forest birds in selectively logged forests and mixed-rural habitats of southern Peninsular Malaysia

We compared the composition and structure of primary forest avifauna among primary forests, selectively logged forests and mixed-rural areas (e.g. villages and agricultural areas) of Peninsular Malaysia. We found that forests that were selectively logged at least 30 years ago contained only 73-75% of the 159 species of extant primary forest birds, with an increased proportion of dominant species. We estimated that only 28-32% of the primary forest species utilized the mixed-rural habitat, and that the number of species that bred in the agricultural landscapes might be even lower. The microhabitat of different species most affected their vulnerability to disturbance. Most small, arboreal frugivores and omnivores, and insectivores that fed from tree trunks, showed greater persistence in the mixed-rural habitat than ground dwelling bird species, which were affected most by disturbance. Resource abundance and variables that were closely related to forest disturbance such as the density of large trees, density of dead trees, canopy cover density and shrub volume influenced the distribution of the primary forest birds. Large primary forest reserves and a revision of short-cycle logging regimes (ca. 30 years) are needed if we are to conserve the lowland rainforest avifauna of Peninsular Malaysia and other parts of Southeast Asia.     

Modeling species’ geographic distributions for preliminary conservation assessments: an implementation with the spiny pocket mice (Heteromys) of Ecuador

GIS-based modeling of a species’ environmental requirements using known occurrence records can provide estimates of its distribution for conservation assessments when other data are lacking. We used collection records, environmental variables, maps of land cover and protected areas, and the Genetic Algorithm for Rule-Set Prediction (GARP) to estimate the historical, current, and protected ranges of the spiny pocket mice present in Ecuador (Heteromys australis and H. teleus). The results suggest that ca. 52-63% of the distributional areas of H. australis in the country are intact, but suitable habitat in protected areas represents only approximately 11-13% of the species’ historical range there. The distribution of H. teleus has been much more reduced, with only ca. 13-19% of its historical distribution still forested and an estimated 2-3% intact and falling in protected areas. Our work highlights critical areas for future fieldwork and demonstrates an integrated approach to estimating a species’ current distribution for preliminary conservation assessments.     

Litter decomposition in primary forest and adjacent fire-disturbed forests in the Gran Sabana,southern Venezuela

Litter decomposition and nutrient dynamic were studied in tall primary forest (TF) and in adjacent slightly fire-affected (MF) and strongly fire-affected (LF) forests of the Gran Sabana, southern Venezuela. The aim of the study was to compare the mass and nutrient loss of litter in undisturbed forest and adjacent fire-disturbed forests growing under the same soil conditions. The results showed no significant differences in the dry-mass reduction among TF, MF, and LF after 1-year of decomposition. At the end of the decomposition period, the mass loss was 31% in TF, 24% in MF, and 25% in LF. With few exceptions, the initial nutrient content of the litter did not show significant differences among the studied forests. The initial litter was poor in nutrients, especially in P, with C/N ratio extremely high, particularly in LF. Both residual P content and C/N ratio were the most effective predictors of dry-mass loss. The general trend in the three studied forests was net N and to less extend P immobilization and release of K, Mg, and Ca. We concluded that during a 1-year period, the decomposition process was similar in undisturbed and adjacent fire-disturbed forests in the Gran Sabana and that the low litter decomposition seems to be mainly controlled by the low chemical quality of the decomposing litter.     

Seasonal changes in multi-scale spatial structure of soil pH and related parameters along a tropical dry evergreen forest slope

Seasonal changes in multi-scale spatial variation in soil chemical properties, which may be controlled simultaneously by biotic and abiotic factors, have not been studied in tropical dry forests. We evaluated the spatial variation of physico-chemical soil properties, plant litter and terrain attributes at multiple scales in a tropical dry evergreen forest using multivariate geostatistics. Soil samples were collected at different depths using nested interval sampling during 1- and 10-m intervals in both the wet and dry seasons. We measured pH, exchangeable cations (Ex-K⁺ and Ex-Ca²⁺), acidity (Ex-H⁺ and Ex-Al³⁺), particle size (clay and sand contents), and forest floor mass (O_i and O_a). Pronounced spatial variation in pH was observed in surface soil (0-5 cm) but not in deeper soil (5-55 cm). Multi-scale spatial structures with short (20 m) and long (86 m) ranges were observed in the auto- and cross-variograms of soil, litter and slope gradient. Pronounced multi-scale structures were observed simultaneously in pH and Ex-Ca²⁺ both in the wet and dry seasons. Only a short-range structure was observed in Ex-K⁺ and O_a, whereas a long-range structure was pronounced in sand contents and slope gradients. Although the variograms had similar shapes between wet and dry seasons for almost all variables, the short-range structure of the cross-variogram between O_a with pH and base cations was more pronouncedly developed in the wet season than in the dry season. Scale-dependent correlation coefficients suggest that a small-scale spatial variation in pH was connected to heterogeneous litter accumulation via base-cation input, whereas long-range spatial variation was simultaneously linked to particle size and slope gradient. This multivariate geostatistical approach applied within a stand detected biotic and abiotic factors controlling spatial variation in soil properties at both short and long distances.     

Effects of fragmentation on artificial nest predation in a tropical forest in Kenya

Nest predation has been established as a leading source of reproductive failure in songbirds. Numerous studies, conducted primarily in the Northern Hemisphere, have linked high rates of nest predation with fragmentation and habitat edges. To date, virtually no such studies have been conducted in Tropical Africa. Here we investigate the effects of disturbance level and edge on depredation of artificial avian nests in six fragments of Kakamega Forest, Kenya. We found that nest losses were higher at the edge than in the interior as often reported in previous studies. Surprisingly, nest predation was highest in the least disturbed fragments. One possible explanation for these findings is that top predators are missing from Kakamega Forest, allowing nest predator populations to increase in all fragments, while in the highly disturbed fragments, domestic dogs might be surrogates for top predators. Alternatively, human activity may itself lead to a decline in mesopredator populations.     

Simulating anthropogenic impacts to bird communities in tropical rain forests

We used an aggregated modelling approach to simulate the impacts of anthropogenic disturbances on the long-term dynamics of faunal diversity in tropical rain forests. We restricted our study to bird communities even though the approach is more general. We developed a model called BioDiv which simulated the establishment of hypothetical bird species in a forest. Our model was based on the results of a simple matrix model which calculated the spatio-temporal dynamics of a tropical rain forest in Malaysia. We analysed the establishment of bird species in a secondary forest succession and the impacts of 60 different logging scenarios on the diversity of the bird community. Of the three logging parameters (cycle length, method, intensity), logging intensity had the most servere impact on the bird community. In the worst case the number of bird species was reduced to 23% of the species richness found in a primary forest.     

Dung decomposition and pedoturbation in a seasonally dry tropical pasture

Rates of dung decomposition and the associated accumulation of soil transported to the surface were compared for dung deposited during a dry and a wet season in a Costa Rican pasture. Average decomposition rates for the first 140 days after deposition were significantly lower for dung patches deposited at the beginning of the dry season than for patches deposited at the beginning of the wet season (0.73 vs. 1.50 g/day^-1 on a dry weight basis). A strong linear relationship was found between dung removal and soil accumulation at the original soil surface, with an average of 2.0 g soil accumulated for every gram of dung which was removed. This relationship was not affected by deposition season. The lack of a seasonal difference, along with the relatively low decomposition rates during the wet season, were explained by the dominance of termites in the dung patches throughout the year. Evidence of dung beetle activity was never recorded during the dry season and was found in only 18 of the 45 dung patches recovered during the wet season.     

Hotspots in cold climate: Conservation value of woodland key habitats in boreal forests

The concept of Woodland Key Habitats (WKH, small-scaled presumed hotspots of biodiversity) has become an essential component of biodiversity conservation in Fennoscandian and Baltic forests. There have been debates over the importance of WKHs in relation to the conservation of biodiversity in production forests. We applied a systematic review protocol and meta-analysis to summarize knowledge on comparisons of biodiversity qualities, such as dead wood and species richness, between WKHs and production forests in relevant countries. We also summarized the knowledge on the impact of edge effects by comparing WKHs surrounded by production forests to WKHs surrounded by clear cuts. Studies had been conducted in Finland, Norway and Sweden. Based on our meta-analysis, WKHs seem to be relative hotspots for dead wood volume, diversity of dead wood, number of species and number of red-listed species. There were some differences also between countries in these biodiversity qualities. Only two studies compared WKHs surrounded by production forests and clear cuts, respectively. Hence, the capability of WKHs to maintain their original species composition and support species persistence over time remains to be addressed, as well as their role in relation to other conservation tools.     

Comparison of cavity modification and community involvement as strategies for hornbill conservation in Thailand

We monitored breeding success of hornbills in Thailand for four sympatric species at Khao Yai National Park during 1981-2002 and six species at Budo-Sungai Padi National Park during 1994-2002. Within a 60 km² study area at Khao Yai, the well protected area, use of available cavities ranged from 94% in 1984 to 50% in 1993. Competition for nesting cavities was 40% of cavities available indicating the shortage of suitable cavity. We monitored the cavity condition and modified 48 cavities and, between 1996 and 2000, these contributed annually 23-45% of nestings that were successfully fledged (n = 352). Within the 90 km² at Budo Mountain, with heavy human disturbance, use of available cavities for nesting decreased from 81% in 1997 to 30% by 2001 and of successful fledging from 96% in 1997 to 55% in 1999, but both increased once our involvement with local communities had eradicated poaching and reduced disturbance. Competition for nesting cavities here was 26% of cavities available indicating the effect of poaching. Both strategies were successful in the short term but long-term management of cavities in large forest trees is required at Khao Yai and of forest disturbance at Budo.     

The value of primary, secondary and plantation forests for Amazonian birds

Secondary forests and exotic tree plantations are rapidly expanding across tropical landscapes, yet we currently have a very poor understanding of the value of these human-dominated forest landscapes for biodiversity conservation. Mist netting, point counts and transect walks were used to compare the bird communities of these habitats and neighboring primary forest in north-east Brazilian Amazonia. The extensive spatial scale of plantations and second-growth in our study area enabled us to implement a robust replicated design, with survey plots approximately two to three orders of magnitude larger than most previous studies of land-use change in the tropics, thus minimising the influence of the surrounding landscape. Species richness was highest in primary forest and lowest in Eucalyptus plantations, and community turnover between habitats was very high whether based upon matrices of relative abundance or species presence-absence data, and for both point count and mist net data. Monthly line-transect censuses conducted over an annual cycle showed an increase in the detection of canopy frugivores and seed predators during the peak of flower and fruit availability in primary forest, but failed to suggest that second-growth or Eucalyptus stands provide suitable foraging habitat at any time of the year. The conservation value of both secondary forest and plantations was low compared to conclusions from previous studies. Our results indicate that while large-scale reforestation of degraded land can increase regional levels of diversity, it is unlikely to conserve most primary forest species, such as understorey insectivores and canopy frugivores.     

Effects of land use on soil erosion in a tropical dry forest ecosystem,Chamela watershed,Mexico

During the past few decades, Mexico has been converting tropical dry forest (TDF) into cropland and pasture, with land degradation expressed as soil erosion being the main environmental consequence. The factors and processes influencing soil erosion are related to scale. At a microscale, the stability of soil aggregates has a significant impact on soil erodibility and strongly influences other soil properties. However, at plot and watershed scales, these relationships are less well known. The relationships between the distribution of soil aggregate size, soil properties and soil erosion were examined for two soil geomorphological units (hillslopes over granite and hillslopes over tuffs) and three land uses (TDF, unburned pasture and burned pasture) within the Chamela watershed of west–central Mexico. To evaluate soil aggregation as a parameter for upscaling soil erosion, the researchers measured microtopographic features at plot scales and interpreted 1:35,000 panchromatic aerial photographs at a watershed scale. Analysis of variance indicated significant differences in soil organic carbon (P < 0.05) and soil moisture (P < 0.01) contents between the two soil geomorphological units, and field tests showed differences in soil texture and structure.     

Importance of reservoirs for the conservation of freshwater molluscs in a tropical urban landscape

Freshwater species and ecosystems are gravely imperiled, particularly within urban landscapes of tropical Asia. In one of the region’s most urbanized landscapes (i.e., Singapore), we determined: (1) the importance of six different habitats (i.e., catchment reservoirs, estuarine reservoirs, forest streams, rural streams, ponds and monsoon canals) for conserving the diversity of freshwater molluscs; (2) key environmental factors (e.g., pH) affecting molluscan distribution; (3) important biogeographical determinants (e.g., area) of molluscan richness within each habitat; and (4) the habitat affinities of introduced species. High sampling saturation was achieved at most study habitats with minimal sampling effort, suggesting that the utilization of molluscs as bioindicators can expedite freshwater conservation initiatives. Estuarine reservoirs (6.0 ± 2.0) had the highest molluscan richness, vis-à-vis catchment reservoirs, forest streams, rural streams, ponds and monsoon canals (3.0 ± 1.5; 0; 3.3 ± 2.0; 1.8 ± 0.5 and 3.5 ± 0.5 respectively). Both reservoir types possessed species compositions distinct from other habitats and contained majority (76%) of the sampled species. Reservoirs therefore serve to conserve the bulk of local freshwater malacofauna, especially if they are maintained at near-neutral pH levels (i.e., ∼7.3) and contain large substrates (i.e., rocks). Area was the best predictor of molluscan richness across all habitats, implying that larger freshwater habitats require higher conservation priorities than smaller ones. Introduced (non-native) species (e.g., Pomacea canaliculata) had high affinities for reservoirs, which are in need of monitoring to curb population expansions. The interminable growth of human settlements urgently requires a reconciliatory approach, which includes the ecologically-sound design and management of modified habitats to complement reserves in sustaining native freshwater species.     

Effects of ground fires on understorey bird assemblages in Amazonian forests

Avifaunal responses to understorey fire disturbance and subsequent changes in habitat structure were examined within 20 burnt and unburnt forest plots of 0.25 ha (10×250 m), 10-15 months after an unprecedented understorey fire swept through the Tapajós-Arapiuns Extractive Reserve of central Brazilian Amazonia following the severe 1997-1998 El-Niño dry season. Although these surface fires in the previously undisturbed primary forest were relatively mild, they resulted in dramatic changes in forest structure consistent with those found elsewhere in Amazonia. Bird species negatively affected by these changes tended to be the least common, the most disturbance-sensitive, and habitat specialists. Considering different guilds, ant followers, dead-leaf gleaners, terrestrial gleaners, and arboreal sallying insectivores were the most negatively affected, whereas nectivores and arboreal granivores became more abundant in the burnt forest. The results highlight the severe consequences of even relatively mild surface fires in neotropical forests, and the importance of controlling haphazard frontier expansion for the conservation of susceptible species that are endemic to fire-prone regions.     

The spatial patterns of indigenous wildlife use in western Panama: Implications for conservation management

Measuring the impacts of indigenous hunting on neotropical wildlife populations remains a difficult task. Significant insights can be gained by analyzing the spatial patterns of hunting yields. This paper presents a geographic analysis of game captured by hunters in 59 households belonging to five neighbouring villages in western Panama over a period of 232 days. The locations of 1269 game kill sites within a hunting zone of 131 km² were documented with trained local investigators and then entered into a geographic information system with associated attribute data for analysis. Results show a heavy concentration of kill sites around the study communities. Nearly 90% of the total harvest by weight was captured within just 2 km of hunters’ homes. Yet, while several species are captured in large numbers close to home, others are caught closer to the outer peripheries of the shared hunting zone, suggesting that some degree of localized depletion may have occurred. The findings demonstrate the value of mapping the boundaries of hunting zones and game kill sites to assess the impacts of hunting on game species in tropical and other ecosystems and to provide an empirical basis for delimiting conservation zones that balance the use and protection of wildlife in tropical forest regions.     

Sustaining biodiversity conservation in human-modified landscapes in the Western Ghats: Remnant forests matter

Human-modified tropical landscapes under semi-natural or agro-ecosystems often harbor biodiversity of significant conservation value. In the Western Ghats of India, these ecosystems also provide connectivity between protected areas and other remnant forests. We investigated the conservation value of these landscapes and agro-ecosystems using results from 35 studies covering 14 taxonomic groups. Large, conspicuous taxonomic groups and tree-covered land-use types have received much focus in this area of research in the Western Ghats. We computed a response ratio defined as the log ratio of species richness in human land use to species richness in forest control site from 17 studies. In a meta-analysis, we investigated variation of this ratio across studies with respect to three variables: taxonomic group, the land-use type sampled and the extent of forest cover within the study landscape. Higher forest cover within the landscape emerged as a major positive influence on biodiversity in human-modified landscapes for vertebrates and vegetation while no patterns emerged for invertebrates. Our results suggest that loss of remnant forest patches from these landscapes is likely to reduce biodiversity within agro-ecosystems and exacerbate overall biodiversity loss across the Western Ghats. Conservation of these remnant forest patches through protection and restoration of habitat and connectivity to larger forest patches needs to be prioritized. In the densely populated Western Ghats, this can only be achieved by building partnerships with local land owners and stakeholders through innovative land-use policy and incentive schemes for conservation.