Search

Highlights • Tropical reforestation is a significant component of global environmental change, yet it is far less understood than tropical deforestation. • We delineated regional hotspots of significant net reforestation across Latin America and the Caribbean, and define a typology of these hotspots. • Reforestation hotspots cover 11% of Latin America and the Caribbean, and include 167,667.7 km2 of reforestation occurring between 2001 and 2014. • We identified a typology of five reforestation clusters based on their topography, climate, population trends, and urbanization degree for Latin America and the Caribbean. Abstract Tropical reforestation is a significant component of global environmental change that is far less understood than tropical deforestation, despite having apparently increased widely in scale during recent decades. The regional contexts defining such reforestation have not been well described. They are likely to differ significantly from the geographical profiles outlined by site-specific observations that predominate in the literature. In response, this article determines the distribution, extent, and defining contexts of apparently spontaneous reforestation. It delineates regional 'hotspots' of significant net reforestation across Latin America and the Caribbean and defines a typology of these hotspots with reference to the biophysical and socioeconomic characteristics that unite and distinguish amongst them. Fifteen regional hotspots were identified on the basis of spatial criteria pertaining to the area, distribution, and rate of reforestation 2001–2014, observed using a custom continental MODIS satellite land-cover classification. Collectively, these hotspots cover 11% of Latin America and the Caribbean and they include 167,667.7 km2 of new forests. Comparisons with other remotely sensed estimates of reforestation indicate that these hotspots contain a significant amount of tropical reforestation, continentally and pantropically. The extent of reforestation as a proportion of its hotspot was relatively invariable (3–14%) given large disparities in hotspot areas and contexts. An ordination analysis defined a typology of five clusters, distinguished largely by their topographical roughness and related aspects of agro-ecological marginality, climate, population trends, and degree of urbanization: 'Urban lowlands', 'Mountainous populated areas', 'Rural highlands', 'Rural humid lands' and 'Rural dry lands'. The typology highlights that a range of distinct, even oppositional regional biophysical, demographic, and agricultural contexts have equally given rise to significant, regional net reforestation, urging a concomitant diversification of forest transition science. [ABSTRACT FROM AUTHOR]

Abstract: Biofuels are promoted worldwide as an alternative for the replacement of fossil fuels, especially in the transportation sector, although the extent to which biofuels can meet this need is unclear. Currently, oil palm is the most important crop used for biodiesel production in the world. The growth of oil palm plantations in Asia has often occurred at the expense of forest areas creating environmental concerns. Colombia is the fifth largest oil palm grower worldwide, and policies that provide subsidies have been enacted to ensure that Colombia plays an important role in future biodiesel markets. At that same time, many sectors of society are concerned that the negative effects of biofuels may be worse than the benefits. In this paper we analyze the land use transitions generated by the expansion of oil palm crops between 2002 and 2008; we identify the factors associated with the expansion and project the future expansion of plantations in Colombia by applying spatial regression analysis and econometric models. To model future expansion, we started with a map of oil palm plantations in 2008. An econometric model that incorporates the impact of governmental policies (normative and economic that support the biofuel sector) through a Time Series Intervention Model Analysis was used to estimate the cultivated area in 2020. To spatially project these estimates a spatial logistic regression model that incorporates biophysical and socioeconomic variables was used. Finally, future land use transitions associated with the expansion of oil palm plantations were analyzed. The results show that present and future oil palm expansion is concentrated in areas dominated by pastures, and to a lesser extent areas that are a mix of agricultural land and natural forests. Our results also suggest that it is highly unlikely that the government will reach its goal of over 3 million hectares of oil palm plantations by 2020 and the goal of biodiesel blends supplying 20% of the national fuel needs. [Copyright &y& Elsevier]

Abstract: There is an ongoing debate on the effect different property regimes have on the use of natural resources and land conversion (i.e., deforestation or reforestation). Much of the discussion has been centered on the two main forms of tenure regime: common-pool system and private property. Case studies around the world have provided evidence on whether one is more effective at preventing deforestation than the other, but there is not a clear pattern. Part of the problem is that evidence comes from theoretical models or isolated case studies instead of comparative studies across large areas. This paper helps to fill this gap by analyzing the association between land cover change (2001–2010) and land tenure regimes including private and two types of common-pool systems (communal and ejido) in Mexico at the municipality level. The analyses were conducted for each of the four major biomes (i.e., moist forest, dry forest, coniferous forest, and desert) to control for differences in environmental factors. Municipalities dominated by communal land tenure had the largest increase in woody cover (classified as >80% cover) in the moist forest, dry forest, and coniferous forest biomes, and municipalities classified as private also had an increase in woody cover, particularly in the desert biome. In contrast, municipalities classified as ejidos (common-pool tenure system resulting from the land reform) lost woody cover mostly in moist forest and desert biomes, but gained woody cover in dry forest and coniferous forest biomes. In modeling analyses, environmental variables were the most important variables associated with woody cover change for private and most communal municipalities, while socioeconomic variables were the most important in ejido regimes. These results highlight the importance of land tenure on land cover change, and show that differences in woody cover change between types of common-pool systems can be larger than their differences with private land tenure. During the last 10years, virtually all deforestation has occurred in areas dominated by ejidos; in contrast, communal and private regimes seem to ameliorate the deforestation process. [Copyright &y& Elsevier]

Abstract: Although over 40years of satellite imagery has greatly helped in documenting the location and extent of human impact, especially deforestation, our ability to confidently detect current patterns of land change at broad spatial scales needs improvement. To address this challenge, we have developed a cost-effective mapping procedure based on 250-m MODIS imagery that produces annual land-use/land-cover (LULC) maps for Latin America and the Caribbean (LAC). This procedure uses annual spectral statistics, collection of references samples with a Web-based tool, and tree-based Random Forests classifiers stratified by biome map regions to produce wall-to-wall, annual LULC maps for 2001 to 2010 that cover all of LAC. Across 26 map regions, overall pixel-level accuracy averaged 80.2±8.1% for eight basic LULC classes, and 84.6±6.5% for a five-class scheme. Municipality-scale area change between 2001 and 2010 in the three dominant classes (woody vegetation, mixed woody/plantation, and agriculture/herbaceous vegetation) was then estimated using regression models fit to 10years of data, thus minimizing the impact of inter-annual class variation on change statistics. Closed-canopy forest area and change between 2001 and 2009 were well correlated with high-resolution maps of the Brazilian Legal Amazon (PRODES project). Our LAC-wide analysis of significant change revealed the recent extent and magnitude of deforestation hotspots, such as in the Amazon moist forests and the dry forests of Argentina, Paraguay and Bolivia. Our data also revealed biome-specific clusters of municipalities with increasing woody vegetation due to forest recovery, reforestation, or woody encroachment. Taken as a whole, our MODIS-based mapping and trend modeling methodology can provide reliable land change data, not just for the tropics or for forest cover, but for all biomes and municipalities in LAC and including multiple LULC classes. Because this information can be produced quickly on an annual time scale, with internally-consistent data sources, it is a very useful tool for resource managers, policy makers, scientists and conservationists interested in tracking recent land change across broad-scale, political and environmental gradients. [Copyright &y& Elsevier]

Abstract: Land use and land cover (LULC) maps from remote sensing are vital for monitoring, understanding and predicting the effects of complex human–nature interactions that span local, regional and global scales. We present a method to map annual LULC at a regional spatial scale with source data and processing techniques that permit scaling to broader spatial and temporal scales, while maintaining a consistent classification scheme and accuracy. Using the Dry Chaco ecoregion in Argentina, Bolivia and Paraguay as a test site, we derived a suite of predictor variables from 2001 to 2007 from the MODIS 250m vegetation index product (MOD13Q1). These variables included: annual statistics of red, near infrared, and enhanced vegetation index (EVI), phenological metrics derived from EVI time series data, and slope and elevation. For reference data, we visually interpreted percent cover of eight classes at locations with high-resolution QuickBird imagery in Google Earth. An adjustable majority cover threshold was used to assign samples to a dominant class. When compared to field data, we found this imagery to have georeferencing error <5% the length of a MODIS pixel, while most class interpretation error was related to confusion between agriculture and herbaceous vegetation. We used the Random Forests classifier to identify the best sets of predictor variables and percent cover thresholds for discriminating our LULC classes. The best variable set included all predictor variables and a cover threshold of 80%. This optimal Random Forests was used to map LULC for each year between 2001 and 2007, followed by a per-pixel, 3-year temporal filter to remove disallowed LULC transitions. Our sequence of maps had an overall accuracy of 79.3%, producer accuracy from 51.4% (plantation) to 95.8% (woody vegetation), and user accuracy from 58.9% (herbaceous vegetation) to 100.0% (water). We attributed map class confusion to limited spectral information, sub-pixel spectral mixing, georeferencing error and human error in interpreting reference samples. We used our maps to assess woody vegetation change in the Dry Chaco from 2002 to 2006, which was characterized by rapid deforestation related to soybean and planted pasture expansion. This method can be easily applied to other regions or continents to produce spatially and temporally consistent information on annual LULC. [Copyright &y& Elsevier]

Abstract: Land use/land cover change is an important driver of global change and changes in carbon stocks. Estimating the changes in carbon stocks due to tropical deforestation has been difficult, mainly because of uncertainties in estimating deforestation rates and the biomass in the forest that have been cut. In this study, we combined detailed land-use change over a 27-year period based on satellite images and forest inventory data to estimate changes in biomass carbon stocks in the Xishuangbanna prefecture (1.9millionha) of China. Xishuangbanna is located in southwestern China in the upper watershed of the Mekong River, and the major forest types are tropical seasonal rain forest, mountain rain forest, and subtropical evergreen broadleaf forest. In the past when the region was completely forested the total biomass carbon would have been approximately 212.65±8.75TgC. By 1976 forest cover had been reduced to 70%, and in addition many forests had been degraded resulting in a large decrease in the total biomass carbon stocks (86.97±3.70TgC). From 1976 to 2003, the mean deforestation rate was 13 722hayear−1 (1.12%), and this resulted in the loss of 370,494ha of forest, and by 2003 total biomass carbon stocks had been reduced to 80.85±2.64TgC. The annual carbon emissions due to land-use change, mainly forest conversion to agriculture and rubber plantations, were 0.37±0.03TgCyear−1 between 1976 and 1988 and 0.13±0.04TgCyear−1 between 1988 and 2003. During the next 20 years, if rubber plantations expand into forests outside of reserves, shrublands, grasslands, and shifting cultivation below 1500m the total biomass carbon stocks of Xishuangbanna will decrease to 76.45±1.49TgC in 2023. This would reflect a loss of 4.13±1.14TgC between 2003 and 2023, or an annual loss of 0.21±0.06TgCyear−1. Alternatively, if rubber plantations only expand into areas of shifting cultivation below 1500m, and all areas presently in shrublands and grasslands are allowed to recover into secondary forests, total biomass carbon stock of the region would increase to 92.65±3.80TgC in 2023. Under this scenario, the growth of existing forests and the expansion of new forests would result in a net sequestration of 0.60±0.06TgCyear−1. This study demonstrates that the uncertainty of biomass estimates can be greatly reduced if detailed land-use analyses are combined with forest inventory data, and that slight changes in future land-use practices can have large implications for carbon fluxes. [Copyright &y& Elsevier]

Abstract: Rapid urban growth has increased the importance of restoring degraded vegetation patches within these areas. In this study, we reforested a site that was previously dominated by exotic grasses within an urban area. The goal of this study was to evaluate restoration success in a reforested site using four variables of vegetation structure, five groups of organisms, and eight variables of ecosystem processes, and compare these values with a pre-reforested site and a forested reference site using the Subjective Bray Curtis Ordination. The change in vegetation structure provided arboreal habitats that increased species diversity and ecosystem processes in the reforested site. Specifically, the development of a vertical vegetation structure was associated with: (1) a decrease in herbaceous cover, which allowed the colonization of woody seedlings; (2) a change in microclimatic conditions, which enhanced the colonization of ants and amphibians; (3) colonization of arboreal reptiles and birds; and (4) an increase in litter production, which enhanced nutrient inputs. Moreover, the Subjective Bray Curtis Ordination demonstrated an overall recovery of approximately 70%. Planting woody species was sufficient to stimulate rapid recovery of many ecosystem attributes. Future restoration projects should include multiple variables that reflect important ecosystem attributes to determine the success of a project and to direct future management efforts. [Copyright &y& Elsevier]

Abstract: Most restoration projects have focused on recovery of vegetation to assess restoration success. Nevertheless if the goal of a restoration project is to create an ecosystem that is self-supporting and resilient to perturbation, we also need information on the recovery of other trophic levels and ecosystem processes. To provide an example on how to assess restoration success, we compared four measures of vegetation structure, four measures of species diversity, and six measures of ecosystem processes among pre-reforested, reforested, and reference sites. In addition, we described how Bray Curtis Ordination could be used to evaluate restoration success. Vegetation structure recovered rapidly due to the increase in vegetation height and the decrease in herbaceous cover. Other measures such as litter cover, number of litter layers, and DBH size class values are recovering at slower rates, but they also have increased vegetation heterogeneity in the reforested site. Species diversity recovered rapidly. The increase in vegetation structure changed the local conditions in the reforested site facilitating the colonization of woody seedlings, ants, reptiles, and amphibians. Ecosystem processes, particularly litter production and turnover, have enhanced the incorporation of nutrients and organic matter in the soil. By including vegetation structure, species diversity, and ecosystem processes measures we have better information to determine the success of a restoration project. Moreover, the Subjective Bray Curtis Ordination is a useful approach for evaluating different restoration techniques or identifying measures that are recovering slowly and would benefit from additional management. [Copyright &y& Elsevier]

Focuses on the regeneration of forest following the abandonment of pastures and coffee plantations located in the karst region of Puerto Rico. Density of woody stems; Use of canonical correspondence analysis; Seedling composition in coffee sites; Influence of land history on the pattern of regeneration.

• At the country scale, the extent of pastures declined and the area of croplands and woody areas increased. • Pasture expansion was the main driver of deforestation in the tropical moist forest biome. • A large proportion of cropland expansion has occurred in desert ecoregions where there is access to irrigation. • The large increase in woody vegetation was concentrated in arid and dry forest regions due to pasture and cropland abandonment. Conversion of land cover is one of the main causes of global environmental change and identifying the regions where sustained trends of land change are occurring provides useful information for land and resources management. For all ecoregions in Mexico, we analyzed land use changes over 14 years (2001–2014) using MODIS images (250 m) and identified regions that had significant gains or loss of woody vegetation, pasturelands, or croplands. The land use patterns varied greatly among the 40 major ecoregions, but in general, woody vegetation and cropland cover increased while pastures decreased. In contrast to previous studies, much of the increase in croplands did not correspond with hotspots of decline in pastures but occurred in the Sonora and Chihuahuan deserts ecoregions in northern Mexico. Industrial cotton, sorghum, and pecans production for export where the major crops responsible for the increase in these ecoregions. Similar to patterns in the rest of Latin America, pasture expansion mainly occurred in ecoregions in the tropical moist forest biome (e.g. Veracruz and Peten-Veracruz ecoregions). The ecoregions that experienced the greatest increase in woody vegetation were the Balsas dry forest ecoregion along the Pacific coast and Trans Mexican volcanic oak-pine ecoregion. In both regions, rural-urban migration appears to be an important driver of the transition from grasslands to woody vegetation, while the conversion of croplands to woody vegetation in the Tamaulipan mezquital ecoregion was clearly associated with drug related violence. This study documents how the complex interactions among national and international demand for agricultural products, national policies, demography, violence and climate change are affecting land change across Mexico. [ABSTRACT FROM AUTHOR]

Natural resource extraction is increasing rapidly in tropical forests, but we lag behind in understanding the impacts of these disturbances on biodiversity. In high diversity tropical habitats, acoustic monitoring is an efficient tool for sampling a large proportion of the fauna across varied spatial and temporal scales. We used passive acoustic monitoring in a pre-montane forest in Peru to investigate how soundscape composition and richness of acoustic frequencies varied with distance from a natural gas exploratory well and with operational phase (construction and drilling). We also evaluated how anuran and avian species richness and vocal activity varied with distance and between phases. Soundscape analyses showed that acoustic frequency similarity was greatest among sites closer to (≤250 m) and farther from (≥500 m) the platform. Soundscapes revealed more frequencies were used during construction and showed a weak trend of increasing frequency richness with increasing distance from the disturbance. Avian species richness and detections increased with distance from the platform, but anuran richness and detections declined with distance. Operational phase did not play a significant role in overall richness or activity patterns of either group. Among birds, insectivore detections increased with distance from the platform, and nectarivores were detected more frequently during the drilling phase. Results demonstrate that acoustic monitoring and soundscape analyses are useful tools for evaluating the impact of development activity on the vocalizing community, and should be implemented as a best practice in monitoring biodiversity and for guiding specific mitigation strategies. [ABSTRACT FROM AUTHOR]

In this study, we evaluated deep convolutional neural networks for classifying the calls of 24 birds and amphibian species detected in ambient field recordings from the tropical mountains of Puerto Rico. Training data were collected using a template-based detection algorithm followed by a manual validation process. As preparing sufficient training data is a major challenge for many deep learning applications, we propose a novel approach that combines transfer learning of a pre-trained deep convolutional neural network (CNN) model and a semi-supervised pseudo-labeling method with a custom loss function to meet this challenge. Our proposed methodology enables the network to be trained in a supervised fashion with labeled and unlabeled data simultaneously, which effectively increases the size of training set and thus boosts the model performance. In classifying a test set of manually validated positive and negative template-based detections, our proposed model achieves 97.7% sensitivity (true positive rate), 96.4% specificity (true negative rate) and 99.5% Area Under a Curve (AUC). This multi-label multi-species classification methodology and its framework can be easily adopted by other acoustic classification problems. [ABSTRACT FROM AUTHOR]