Tropical Forest Regions Research Articles

Impacts of climate, phenology, elevation and their interactions on the net primary productivity of vegetation in Yunnan, China under global warming

Both net primary productivity (NPP) and vegetation phenology play essential roles in influencing the carbon sequestration of terrestrial ecosystems. However, the relationship between vegetation phenology and NPP remains unclear under the effects of global warming. This study used Geodetector to analyze the interaction mechanisms between climate, phenology, elevation, and NPP in Yunnan, China. The results are as follows. (1) NPP is positively correlated with NDVI, LOS, EOS, TEMP, PREC and SRAD, and negatively correlated with ELEV, NDBI and SOS. However, the main influencing factors leading to the variation of NPP differ with the region. (2) The spatial distribution of NPP and LOS in Yunnan Province is mainly influenced by the monsoon, showing a pattern of high in the southwest and low in the northeast, and the interaction between PREC and LOS is the main reason for the variation of NPP. (3) In the high-altitude region of northwest Yunnan, the interaction between altitude and PREC is the primary factor affecting NPP variation. (4) In the tropical monsoon forest region of southern Yunnan, the interaction between SRAD and PREC is the main cause of NPP changes. (5) Despite having a small effect on NPP, SOS was the only phenological factor that showed a significant linear relationship with NPP. This study reveals the complex and diverse interactions between climate, phenology, elevation, and NPP in different regions of Yunnan, and provides a new perspective for understanding the carbon cycle and ecological processes of terrestrial ecosystems under global warming.

ECOSTRESS Reveals the Importance of Topography and Forest Structure for Evapotranspiration from a Tropical Forest Region of the Andes

Tropical forests are major sources of global terrestrial evapotranspiration (ET), but these heterogeneous landscapes pose a challenge for continuous estimates of ET, so few studies are conducted, and observation gaps persist. New spaceborne products such as ECOsystem Spaceborne Thermal Radiometer Experiment on Space Station (ECOSTRESS) are promising tools for closing such observation gaps in understudied tropical areas. Using ECOSTRESS ET data across a large, protected tropical forest region (2250 km2) situated on the western slope of the Andes, we predicted ET for different days. ET was modeled using a random forest approach, following best practice workflows for spatial predictions. We used a set of topographic, meteorological, and forest structure variables from open-source products such as GEDI, PROBA-V, and ERA5, thereby avoiding any variables included in the ECOSTRESS L3 algorithm. The models indicated a high level of accuracy in the spatially explicit prediction of ET across different locations, with an r2 of 0.61 to 0.74. Across all models, no single predictor was dominant, and five variables explained 60% of the models’ results, thus highlighting the complex relationships among predictor variables and their influence on ET spatial predictions in tropical mountain forests. The leaf area index, a forest structure variable, was among the three variables with the highest individual contributions to the prediction of ET on all days studied, along with the topographic variables of elevation and aspect. We conclude that ET can be predicted well with a random forest approach, which could potentially contribute to closing the observation gaps in tropical regions, and that a combination of topography and forest structure variables plays a key role in predicting ET in a forest on the western slope of the Andes.

Uncertainty and Emergent Constraints on Enhanced Ecosystem Carbon Stock by Land Greening

AbstractSignificant land greening since the 1980s has been detected through satellite observation, forest inventory, and Earth system modeling. However, whether and to what extent global land greening enhances ecosystem carbon stock remains uncertain. Here, using 40 global models, we first detected a positive correlation between the terrestrial ecosystem carbon stock and leaf area index (LAI) over time. Then, we diagnose the source of uncertainty of simulated the sensitivities of ecosystem carbon stock to LAI based on a traceability analysis. We found that the sensitivity of gross primary productivity (GPP) to LAI is the largest contributor to the model uncertainty in more than 60% of the vegetated grids. Using the ensemble of four long‐term global data sets of GPP and three satellite LAI products from 1982 to 2014, we provided an emergent constraint on the ecosystem carbon stock increase as 0.75 ± 0.46 kg C m−2 per unit LAI over global land areas. Furthermore, the biome‐based results reveal that the tropical forest regions have the highest inter‐model variation and model bias. Overall, this study identifies the uncertainty source and provides constrained estimates of the greening effect on ecosystem carbon stock at the global scale.

Genetic characterization of a group of commercial African timber species: From genomics to barcoding.

In the last decades, illegal logging has posed a serious threat for the integrity of forest ecosystems and for biodiversity conservation in tropical Africa. Although international treaties and regulatory plans have been implemented to reduce illegal logging, much of the total timber volume is harvested and traded illegally from tropical African forest regions. As a result, the development and the application of analytical tools to enhance the traceability and the identification of wood and related products is critical to enforce international regulations. Among available techniques, DNA barcoding is a promising approach for the molecular identification of plant species. However, although it has been used successfully for the discrimination of animal species, no set of genetic markers is available for the universal identification of plant species. In this work, we firstly characterized the genetic diversity of 17 highly-valuable African timber species from five genera (Afzelia, Guibourtia, Leplea, Milicia, Tieghemella) across their distribution ranges in West and Central Africa using the genome skimming approach in order to reconstruct their chloroplast genomes and nuclear ribosomal DNA. Next, we identified single-nucleotide polymorphisms (SNPs) for the discrimination of closely-related species. In this way, we successfully developed and tested novel species-specific genetic barcodes for species identification.

Distinguishing forest types in restored tropical landscapes with UAV-borne LIDAR

Forest landscape restoration is a global priority to mitigate negative effects of climate change, conserve biodiversity, and ensure future sustainability of forests, with international pledges concentrated in tropical forest regions. To hold restoration efforts accountable and monitor their outcomes, traditional strategies for monitoring tree cover increase by field surveys are falling short, because they are labor-intensive and costly. Meanwhile remote sensing approaches have not been able to distinguish different forest types that result from utilizing different restoration approaches (conservation versus production focus). Unoccupied Aerial Vehicles (UAV) with light detection and ranging (LiDAR) sensors can observe forests` vertical and horizontal structural variation, which has the potential to distinguish forest types. In this study, we explored this potential of UAV-borne LiDAR to distinguish forest types in landscapes under restoration in southeastern Brazil by using a supervised classification method. The study area encompassed 150 forest plots with six forest types divided in two forest groups: conservation (remnant forests, natural regrowth, and active restoration plantings) and production (monoculture, mixed, and abandoned plantations) forests. UAV-borne LiDAR data was used to extract several Canopy Height Model (CHM), voxel, and point cloud statistic based metrics at a high resolution for analysis. Using a random forest classification model we could successfully classify conservation and production forests (90% accuracy). Classification of the entire set of six types was less accurate (62%) and the confusion matrix showed a divide between conservation and production types. Understory Leaf Area Index (LAI) and the variation in vegetation density in the upper half of the canopy were the most important classification metrics. In particular, LAI understory showed the most variation, and may help advance ecological understanding in restoration. The difference in classification success underlines the difficulty of distinguishing individual forest types that are very similar in management, regeneration dynamics, and structure. In a restoration context, we showed the ability of UAV-borne LiDAR to identify complex forest structures at a plot scale and identify groups and types widely distributed across different restored landscapes with medium to high accuracy. Future research may explore a fusion of UAV-borne LiDAR with optical sensors, and include successional stages in the analyses to further characterize and distinguish forest types and their contributions to landscape restoration.

Seed traits and tropical arboreal species conservation: a case study of a highly diverse tropical humid forest region in Southern Costa Rica

Tropical areas are vital for conserving endangered tree species. However, barriers to the propagation and storage of plant material still limit preservation outcomes. Seed traits are essential in conservation but have rarely been applied in the tropics. In this work, we compile information about the germination and seed desiccation tolerance behaviour of 953 arboreal species from the Osa Peninsula, Costa Rica. We analysed how those traits are related to the growing strata, seed size, and dispersal mode and how they can be used to improve conservation initiatives based on the species distribution, abundance, extinction risk, and ex situ cultivation. Over a third of the species are probably recalcitrant but this number increase to almost half among the endangered and endemic ones. Most species have seeds dispersed by animals, but wind dispersal is usually orthodox and dominant in secondary forests. All species with tiny seeds and most understorey ones with small seeds are orthodox. Approximately half of the species have dormant seeds, while nearly one-third of nondormant species have cohorts with extended germination. Almost half of the species are conserved ex situ; however, several endangered, endemics could be expected to be considered exceptional. Seed traits could target orthodox species for seed banking in tropical forests, including the potential of secondary forests and a subset of oligarch species. For the recalcitrant ones, approaches based on the risk of extinction, distribution, abundance, and ex situ status allow for prioritizing collection and conservation strategies. Under this approach, we develop a list with priority species.

Reproductive functional organization of woody plant assemblages along regeneration in a Caatinga dry forest

Tropical forest regeneration across old fields has been mainly described as a predictable sequence of functional plant assemblages in response to environmental filtering. However, the way plant reproductive diversity is organized along forest regeneration and how the reproductive profile of woody flora may impact regeneration have been poorly investigated. This paper examines the organization of plant assemblages throughout the regeneration of a Caatinga dry forest in the context of slash-and-burn agriculture with a focus on reproductive plant traits and strategies. We hypothesized a recovery and directional changes in specialist reproductive traits/strategies along forest regeneration but with aridity and chronic anthropogenic disturbances posing contrary forces to this recovery process. This study was carried out in a human-modified landscape of Caatinga dry forest in northeast Brazil. We compiled information on woody plant assemblages inhabiting a forest chronosequence consisting of 28 forest stands of varying age: 14 regenerating (4–50-yr old) and 14 old-growth forests. Woody plant assemblages along forest regeneration had higher occurrence of specialized traits and strategies, such as pollination by vertebrates (16.6%) and dispersion by animals (47.8%). Regenerating forest stands exhibited a higher occurrence of species with specialized floral size (19%) and pollination by vertebrates (26.5%) as compared to old-growth forests. Regenerating forest stands also supported a higher (30%) functional diversity of reproductive traits than old-growth stands. Biomass, aridity, and chronic anthropogenic disturbances affected a myriad of traits/strategies: pollination by birds and animal dispersion increased with increases in biomass while total functional diversity responded negatively; pollination by Sphingids and beetles negatively responded to aridity, while very large fruit size species negatively responded to chronic anthropogenic disturbances. Both forest biomass and aridity correlated negatively with reproductive trait diversity. The Caatinga dry forest regeneration suggests high resilience, no directional/predictable changes (although regenerating vs. old-growth differ in some traits) and aridity imposed to the successional mosaics as a driver. This pattern probably emerges conditioned by the reproductive traits posed by the plant species able to thrive in such stressful environments, including those species that persist through the entire regeneration process, i.e., the regeneration process is marked by high complexity and low predictive power. Although rapid recovery represents good news, the plant-assemblage reproductive profile imposed by disturbance-adapted species may drastically alter resource availability for a myriad of plant attendants (e.g., pollinators and thief) and, consequently, the complex/diverse plant-animal interactions typical of dry forests. This working hypothesis has both theoretical and applied implications in the context of dry forest regeneration in human-modified landscapes.

Genetic Monitoring of the Last Captive Population of Greater Mouse-Deer on the Thai Mainland and Prediction of Habitat Suitability before Reintroduction

Developing successful conservation programs for genetically depleted species is challenging. Survival and adaptive potential are related to genetic and habitat factors; therefore, conservation programs are designed to minimize risks associated with inbreeding and loss of genetic diversity. The greater mouse-deer (Tragulus napu) is a true forest species that contributes to seed distribution dynamics in forests. However, with continuous demographic decline over the last century in the wild, only captive populations of the greater mouse-deer remain on the Thai mainland. A restoration program initiated 20 years ago has increased their population to more than 100 individuals but maintaining high genetic diversity in a small captive population is crucial for successful recovery. Microsatellite genotyping and mitochondrial D-loop and SRY gene sequence analyses were performed to examine the genetic diversity and population structure in 123 greater mouse-deer (64 females and 59 males). Highly reduced effective captive population size with trends of inbreeding were observed. No historical bottleneck was observed. These conditions have reduced their reproductive fitness and ability to adapt to environmental change, increasing the risk of population decline and eventual extinction. Demographic analyses suggested a recent captive population expansion due to effective animal welfare and reproduction. The results also suggested that population size at equilibrium is the main factor of allelic diversity (number of alleles). Large habitat carrying capacity, representing each fixed captive population size can support the genetic diversity of greater mouse-deer. We also identified suitable habitat areas for reintroduction and long-term in situ conservation of greater mouse-deer using maximum entropy modeling. Based on the environmental variables, species distribution modeling for greater mouse-deer indicated lowland tropical forest regions in the Khlong Saeng-Khao Sok forest complexes as most suitable and requiring urgent habitat improvement. These findings highlight the relevance of careful genetic monitoring and habitat suitability for the long-term conservation of greater mouse-deer and enhance the success of future conservation plans.

Biomolecular analyses enable new insights into ancient Egyptian embalming

The ability of the ancient Egyptians to preserve the human body through embalming has not only fascinated people since antiquity, but also has always raised the question of how this outstanding chemical and ritual process was practically achieved. Here we integrate archaeological, philological and organic residue analyses, shedding new light on the practice and economy of embalming in ancient Egypt. We analysed the organic contents of 31 ceramic vessels recovered from a 26th Dynasty embalming workshop at Saqqara1,2. These vessels were labelled according to their content and/or use, enabling us to correlate organic substances with their Egyptian names and specific embalming practices. We identified specific mixtures of fragrant or antiseptic oils, tars and resins that were used to embalm the head and treat the wrappings using gas chromatography–mass spectrometry analyses. Our study of the Saqqara workshop extends interpretations from a micro-level analysis highlighting the socio-economic status of a tomb owner3–7 to macro-level interpretations of the society. The identification of non-local organic substances enables the reconstruction of trade networks that provided ancient Egyptian embalmers with the substances required for mummification. This extensive demand for foreign products promoted trade both within the Mediterranean8–10 (for example, Pistacia and conifer by-products) and with tropical forest regions (for example, dammar and elemi). Additionally, we show that at Saqqara, antiu and sefet—well known from ancient texts and usually translated as ‘myrrh’ or ‘incense’11–13 and ‘a sacred oil’13,14—refer to a coniferous oils-or-tars-based mixture and an unguent with plant additives, respectively.

Genetic diversity and structure in wild Robusta coffee (Coffea canephora A. Froehner) populations in Yangambi (DR Congo) and their relation to forest disturbance.

Degradation and regeneration of tropical forests can strongly affect gene flow in understorey species, resulting in genetic erosion and changes in genetic structure. Yet, these processes remain poorly studied in tropical Africa. Coffea canephora is an economically important species, found in the understorey of tropical rainforests of Central and West Africa, and the genetic diversity harboured in its wild populations is vital for sustainable coffee production worldwide. Here, we aimed to quantify genetic diversity, genetic structure, and pedigree relations in wild C. canephora populations, and we investigated associations between these descriptors and forest disturbance and regeneration. Therefore, we sampled 256 C. canephora individuals within 24 plots across three forest categories in Yangambi (DR Congo), and used genotyping-by-sequencing to identify 18,894 SNPs. Overall, we found high genetic diversity, and no evidence of genetic erosion in C. canephora in disturbed old-growth forest, as compared to undisturbed old-growth forest. In addition, an overall heterozygosity excess was found in all populations, which was expected for a self-incompatible species. Genetic structure was mainly a result of isolation-by-distance, reflecting geographical location, with low to moderate relatedness at finer scales. Populations in regrowth forest had lower allelic richness than populations in old-growth forest and were characterised by a lower inter-individual relatedness and a lack of isolation-by-distance, suggesting that they originated from different neighbouring populations and were subject to founder effects. Wild Robusta coffee populations in the study area still harbour high levels of genetic diversity, yet careful monitoring of their response to ongoing forest degradation remains required.

Impacts of climate change on the potential distribution of epiphytic cacti in the Caatinga biome, Brazil

The Caatinga biome is the largest dry tropical forest region in South America as well as one of the most vulnerable regions in the world to the climate changes forecast for this century. Climate forecasts for the biome include increased air temperature, reduced rainfall and aridization. This biome does not have a homogeneous landscape; instead it has several rainforest enclaves. This article describes a study to model the potential distribution of four epiphytic cactus species (Epiphyllum phyllanthus (L.) Haw., Rhipsalis floccosa Salm-Dyck ex Pfeiff., Rhipsalis lindbergiana K. Schum and Rhipsalis russellii Britton & Rose.) in the biome under future climate scenarios and traces out a prognosis for the enclaves and the biome. For that purpose, we used the MaxEnt modeling method, considering two future time intervals (2041-2060 and 2061-2080) and the interval 1961-1990 for the current situation, with the RCP4.5 and 8.5 scenarios. The projections for future potential distribution showed a spatial contractions greater than 88% found in the areas of high potential presence for the target species throughout the biome and in all the scenarios. The results strengthen the expectation of aridization in the Caatinga biome, with the loss or shrinkage of rainforest enclaves as time progresses.

Land Use and Landscape Characteristics Are Associated with Core Forest Patches in Ghana

Land uses and terrain characteristics would likely influence the types and spatial arrangements of forest patches, and generally, forest fragmentation. Whereas prior research has focused mainly on direct land use-induced forest fragmentation, this study models the relationship between the spatial distribution of core forest patches, land uses, and terrain variables. Relying on Landsat images from the Atewa Range Forest Reserve (ARFR) in Ghana, we use machine learning geospatial techniques and statistical methods to process satellite images and model the relationship between core forest patches and associated variables. The study finds that a unit reduction in elevation would significantly likely reduce by 0.995 times the possibility of forest patches being core forests, implying that on lower slopes, core forests are less likely to occur. Additionally, we find that a unit increase in slope gradient significantly increases the odds of a forest patch being among the core forest category by 1.35 times. Moreover, our results show that the odds of forest patches being core forests significantly increase by 1.60 and 2.14 times if patches are found beyond 1 km from logging sites and access roads, respectively. This implies that intact forest patches would likely be found on higher slopes, higher elevations, and areas far away from land uses. Based on the results, we suggest that the protection of forest patches should target higher elevations and slopes and most importantly areas far from land uses whereas forest restoration programs should target areas close to land uses and on lower elevations and lower slopes. With this study demonstrating a significant relationship between core forests, land uses and terrain variables, we present important information to land managers for land monitoring and conservation in the ARFR and other tropical forest regions of the world.

Underlying topography and forest height estimation from SAR tomography based on a nonparametric spectrum estimation method with low sidelobes

ABSTRACT The underlying topography and forest height play an indispensable role in many fields, including geomorphology, civil engineering construction, forest investigation, and the modeling of natural disasters. As a new microwave remote sensing technology with three-dimensional imaging capability, synthetic aperture radar (SAR) tomography (TomoSAR) has already been proven to be an important tool for underlying topography and forest height estimation. Many spectrum estimation methods have now been proposed for TomoSAR. However, most of the commonly used methods are susceptible to noise and inevitably produce sidelobes, resulting in a reduced accuracy for the inversion of forest structural parameters. In this paper, to solve this problem, a nonparametric spectrum estimation method with low sidelobes – the G-Pisarenko method – is introduced. This method performs a logarithmic operation on the covariance matrix to obtain the main scattering characteristics of the objects of interest while suppressing the noise as much as possible. The effectiveness of the proposed method is demonstrated by the use of both simulated data and P-band airborne SAR data from a tropical forest region in Gabon, Africa. The results show that the proposed method can reduce the sidelobes and improve the estimation accuracy for the underlying topography and forest height.

Monitoring recovery of tree diversity during tropical forest restoration: lessons from long-term trajectories of natural regeneration.

Given the importance of species diversity as a tool for assessing recovery during forest regeneration and active restoration, robust approaches for assessing changes in tree species diversity over time are urgently needed. We assessed changes in tree species diversity during natural regeneration over 12-20 years in eight 1-ha monitoring plots in NE Costa Rica, six second-growth forests and two old-growth reference forests. We used diversity profiles to show successional trajectories in measures of observed, asymptotic and standardized tree diversity and evenness as well as sample completeness. We randomly subsampled 1-ha plot data to evaluate how well smaller spatial subsamples would have captured temporal trajectories. Annual surveys in eight 1-ha plots were missing substantial numbers of rare or infrequent species. Older second-growth sites showed consistent declines in tree diversity, whereas younger sites showed fluctuating patterns or increases. Subsample areas of 0.5 ha or greater were sufficient to infer the diversity of abundant species, but smaller subsamples failed to capture temporal trajectories of species richness and yielded positively biased estimates of evenness. In tropical forest regions with high levels of diversity, species diversity from small sample plots should be assessed using methods that incorporate abundance information and that standardize for sample coverage. This article is part of the theme issue 'Understanding forest landscape restoration: reinforcing scientific foundations for the UN Decade on Ecosystem Restoration'.

Quantification of Above-Ground Biomass over the Cross-River State, Nigeria, Using Sentinel-2 Data

Higher-resolution wall-to-wall carbon monitoring in tropical Africa across a range of woodland types is necessary in reducing uncertainty in the global carbon budget and improving accounting for Reducing Emissions from Deforestation and forest Degradation Plus (REDD+). This study uses Sentinel-2 multispectral imagery combined with climatic and edaphic variables to estimate the regional distribution of aboveground biomass (AGB) for the year 2020 over the Cross River State, a tropical forest region in Nigeria, using random forest (RF) machine learning. Forest inventory plots were collected over the whole state for training and testing of the RF algorithm, and spread over undisturbed and disturbed tropical forests, and woodlands in croplands and plantations. The maximum AGB plot was estimated to be 588 t/ha with an average of 121.98 t/ha across the entire Cross River State. AGB estimated using random forest yielded an R2 of 0.88, RMSE of 40.9 t/ha, a relRMSE of 30%, bias of +7.5 t/ha and a total woody regional AGB of 0.246 Pg for the Cross River State. These results compare favorably to previous tropical AGB products; with total AGB of 0.290, 0.253, 0.330 and 0.124 Pg, relRMSE of 49.69, 57.09, 24.06 and 56.24% and −41, −48, −17 and −50 t/ha bias over the Cross River State for the Saatchi, Baccini, Avitabile and ESA CCI maps, respectively. These are all compared to the current REDD+ estimate of total AGB over the Cross River State of 0.268 Pg. This study shows that obtaining independent reference plot datasets, from a variety of woodland cover types, can reduce uncertainties in local to regional AGB estimation compared with those products which have limited tropical African and Nigerian woodland reference plots. Though REDD+ biomass in the region is relatively larger than the estimates of this study, REDD+ provided only regional biomass rather than pixel-based biomass and used estimated tree height rather than the actual tree height measurement in the field. These may cast doubt on the accuracy of the estimated biomass by REDD+. These give the biomass map of this current study a comparative advantage over others. The 20 m wall-to-wall biomass map of this study could be used as a baseline for REDD+ monitoring, evaluation, and reporting for equitable distribution of payment for carbon protection benefits and its management.

Analysis of the Influence of Deforestation on the Microphysical Parameters of Clouds in the Amazon

Studies have shown that deforestation can cause changes in energy, moisture, and precipitation flows, with implications for local and regional climate. These studies generally focus on understanding how the hydrological cycle is impacted by deforestation, but few studies have investigated these impacts on cloud microphysics in tropical forest regions. The objective of this study was to quantitatively evaluate the impacts of deforestation on the microphysical parameters of clouds, based on data extracted from active and passive orbital sensors from the TRMM satellite. The study area comprised the state of Rondônia, Brazil. The analyses of the microphysical parameters extracted from the Microwave Imager (TMI) and Precipitation Radar (PR) sensors of the 2A-CLIM and 2A25 products were performed considering a period of 14 years. The parameters analyzed were Rain Water Path (RWP), Ice Water Path (IWP), Surface Precipitation (SP), Freezing Level Height (FH), and Rainfall Type (RT). Land cover type data were extracted from the Project to Monitor Deforestation in the Legal Amazon (PMDA). Our results showed that local deforestation significantly altered the microphysical parameters of the study region. In general, the values of the microphysical parameters of the clouds in the transition areas (locations where forest pixels are neighbors to deforested pixels) were about 5–25% higher compared to forested and deforested areas associated with a higher frequency of episodes of convective rainfall possibly driven by mesoscale circulations. Correspondingly, forested areas had higher rainfall rates compared to deforested areas. Meanwhile, deforested areas had higher amounts for IWP, of around 1–16%, and FH, of around 2–8%, in relation to forested areas. Conversely, the RWP showed a decrease of around 2–20%. These results suggest that the microphysical structure of clouds has different characteristics when related to forested and deforested areas in the Amazon. This is useful for evaluation of simulations of cloud microphysical parameters in numerical models of weather and climate.

Saccharides in atmospheric PM2.5 in tropical forest region of southwest China: Insights into impacts of biomass burning on organic carbon aerosols

Biomass burning (BB) in South and Southeast Asia has a strong impact on regional air quality, yet its impact on atmospheric PM2.5 of tropical rainforest regions, a background region occupying a large area in South Asia, has rarely been investigated. In this work, we performed one-year PM2.5 sampling during December 2018 to October 2019 at a tropical rainforest site in southwest China. PM2.5 mass concentration, major chemical components, and ten saccharides were determined to study seasonal variations of PM2.5 chemical composition, and further to understand possible impacts of BB to organic carbon (OC) aerosols at this background region. The concentration levels of PM2.5, major PM2.5 components, and total saccharides were significantly higher in dry season than in wet season. Besides, PM2.5, OC, and total saccharides were highly correlated (R2 > 0.64, p < 0.001) during the sampling period, suggesting they might share common sources. Source apportionment of saccharides revealed that BB was the main source in both seasons. Furthermore, the contributions of BB to OC (BB/OC) were estimated using levoglucosan as a molecular tracer while levoglucosan's chemical degradation was considered. It was found that over 80% of LG was degraded in both seasons, suggesting BB sources were largely from the transport of external air mass. The estimated BB/OC were over 50%, indicating BB was an important source of OC and likely of PM2.5 in both seasons. The air-mass backward trajectory analysis and active fire spots data indicate intense BB emission sources were from South and Southeast Asia in dry season and the BB emissions in southern region of China could impact on the studied area in wet season.

Using deep convolutional neural networks to forecast spatial patterns of Amazonian deforestation

Abstract Tropical forests are subject to diverse deforestation pressures while their conservation is essential to achieve global climate goals. Predicting the location of deforestation is challenging due to the complexity of the natural and human systems involved but accurate and timely forecasts could enable effective planning and on‐the‐ground enforcement practices to curb deforestation rates. New computer vision technologies based on deep learning can be applied to the increasing volume of Earth observation data to generate novel insights and make predictions with unprecedented accuracy. Here, we demonstrate the ability of deep convolutional neural networks (CNNs) to learn spatiotemporal patterns of deforestation from a limited set of freely available global data layers, including multispectral satellite imagery, the Hansen maps of annual forest change (2001–2020) and the ALOS PALSAR digital surface model, to forecast deforestation (2021). We designed four model architectures, based on 2D CNNs, 3D CNNs, and Convolutional Long Short‐Term Memory (ConvLSTM) Recurrent Neural Networks (RNNs), to produce spatial maps that indicate the risk to each forested pixel (~30 m) in the landscape of becoming deforested within the next year. They were trained and tested on data from two ~80,000 km2 tropical forest regions in the Southern Peruvian Amazon. The networks could predict the location of future forest loss to a high degree of accuracy (F1 = 0.58–0.71). Our best performing model (3D CNN) had the highest pixel‐wise accuracy (F1 = 0.71) when validated on 2020 forest loss (2014–2019 training). Visual interpretation of the mapped forecasts indicated that the network could automatically discern the drivers of forest loss from the input data. For example, pixels around new access routes (e.g. roads) were assigned high risk, whereas this was not the case for recent, concentrated natural loss events (e.g. remote landslides). Convolutional neural networks can harness limited time‐series data to predict near‐future deforestation patterns, an important step in harnessing the growing volume of satellite remote sensing data to curb global deforestation. The modelling framework can be readily applied to any tropical forest location and used by governments and conservation organisations to prevent deforestation and plan protected areas.

Solar radiation and soil moisture drive tropical forest understory responses to experimental and natural hurricanes

AbstractTropical forest understory regeneration occurs rapidly after disturbance with compositional trajectories that depend on species availability and environmental conditions. To predict future tropical forest regeneration dynamics, we need a deeper understanding of how pulse disturbance events, like hurricanes, interact with environmental variability to affect understory demography and composition. We examined fern and sapling mortality, recruitment, and community composition in relation to solar radiation and soil moisture using 17 years of forest dynamics data (2003–2019) from the Canopy Trimming Experiment in the Luquillo Experimental Forest, Puerto Rico. Solar radiation increased 150% and soil moisture increased 40% following canopy trimming of experimental plots relative to control plots. All plots were disturbed in 2017 by Hurricanes Irma and Maria, so experimentally trimmed plots presented the opportunity to study the effects of multiple hurricanes, while control plots isolated the effects of a single natural hurricane. Recruitment rates maximized at 0.14 individuals/plot/month for ferns and 0.20 stems/plot/month for saplings. Recruitment and mortality were distributed more evenly over the 17 years of monitoring in experimentally trimmed plots than in control plots; however, following Hurricane Maria demographic rates substantially increased in control plots only. In experimentally trimmed plots, the largest community compositional shifts occurred as a result of the trimming events, and compositional changes were greatest for control plots after Hurricane Maria in 2017. Pioneer tree and fern species increased in abundance in response to both simulated and natural hurricanes. Following Hurricane Maria, two dominant pioneer species, Cyathea arborea and Cecropia schreberiana, recruited abundantly, but only in control plots. In trimmed plots, increased solar radiation and soil moisture shifted understory species composition steadily toward pioneer and secondary‐successional species, with soil moisture interacting strongly with canopy trimming. Thus, both solar radiation and soil moisture are environmental drivers affecting pioneer species recruitment following disturbance, which interact with canopy opening following hurricanes. Our results suggest that if hurricane disturbances increase in frequency and severity, as suggested by climate change predictions, the understory regeneration of late‐successional species, such as Manilkara bidentata and Sloanea berteroana, which prefer deeper shade and slightly drier soil microsites, may become imperiled.

Small mammals reduce distance dependence and increase seed predation risk in tropical rainforest fragments

AbstractSeed predation and reduced predation risk with distance from conspecific trees are important influences on tree regeneration in tropical forests. Shifts in animal communities, such as an increase in rodents and other small mammals due to forest fragmentation, could alter patterns of seed predation and affect tree regeneration and community dynamics in forest fragments. We performed a field experiment on four native rainforest tree species in the Western Ghats, India, to test whether fragmentation increases seed predation by mammals and alters the distance dependence of seed predation. We monitored seed predation within open and mammal exclosure plots, near and far from the canopies of conspecific trees, in contiguous and fragmented forests. Seed predation of Cullenia exarillata, Ormosia travancorica, and Syzygium rubicundum was markedly higher in forest fragments, and more so within open plots than exclosures, while the predominantly insect‐predated Acronychia pedunculata experienced similar predation in contiguous forests and fragments. Seed predation of C. exarillata and S. rubicundum was unrelated to distance from conspecific trees in open plots in both contiguous forests and fragments, in contrast to exclosures that showed marked near versus far differences in seed predation. Our findings suggest that by increasing overall seed predation risk and imposing similar seed predation risk near and far from adults variably across the tree species, small mammals could alter processes that shape tree diversity and species composition in fragmented tropical rainforests.