Urban Vegetation Mapping from Aerial Imagery Using Explainable AI (XAI).

Urban vegetation mapping is critical in many applications, i.e., preserving biodiversity, maintaining ecological balance, and minimizing the urban heat island effect. It is still challenging to extract accurate vegetation covers from aerial imagery using traditional classification approaches, because urban vegetation categories have complex spatial structures and similar spectral properties. Deep neural networks (DNNs) have shown a significant improvement in remote sensing image classification outcomes during the last few years. These methods are promising in this domain, yet unreliable for various reasons, such as the use of irrelevant descriptor features in the building of the models and lack of quality in the labeled image. Explainable AI (XAI) can help us gain insight into these limits and, as a result, adjust the training dataset and model as needed. Thus, in this work, we explain how an explanation model called Shapley additive explanations (SHAP) can be utilized for interpreting the output of the DNN model that is designed for classifying vegetation covers. We want to not only produce high-quality vegetation maps, but also rank the input parameters and select appropriate features for classification. Therefore, we test our method on vegetation mapping from aerial imagery based on spectral and textural features. Texture features can help overcome the limitations of poor spectral resolution in aerial imagery for vegetation mapping. The model was capable of obtaining an overall accuracy (OA) of 94.44% for vegetation cover mapping. The conclusions derived from SHAP plots demonstrate the high contribution of features, such as Hue, Brightness, GLCM_Dissimilarity, GLCM_Homogeneity, and GLCM_Mean to the output of the proposed model for vegetation mapping. Therefore, the study indicates that existing vegetation mapping strategies based only on spectral characteristics are insufficient to appropriately classify vegetation covers.

New tools for old problems - comparing drone- and field-based assessments of a problematic plant species.

Plant species that negatively affect their environment by encroachment require constant management and monitoring through field surveys. Drones have been suggested to support field surveyors allowing more accurate mapping with just-in-time aerial imagery. Furthermore, object-based image analysis tools could increase the accuracy of species maps. However, only few studies compare species distribution maps resulting from traditional field surveys and object-based image analysis using drone imagery. We acquired drone imagery for a saltmarsh area (18 ha) on the Hallig Nordstrandischmoor (Germany) with patches of Elymus athericus, a tall grass which encroaches higher parts of saltmarshes. A field survey was conducted afterwards using the drone orthoimagery as a baseline. We used object-based image analysis (OBIA) to segment CIR imagery into polygons which were classified into eight land cover classes. Finally, we compared polygons of the field-based and OBIA-based maps visually and for location, area, and overlap before and after post-processing. OBIA-based classification yielded good results (kappa = 0.937) and agreed in general with the field-based maps (field = 6.29 ha, drone = 6.22 ha with E. athericus dominance). Post-processing revealed 0.31 ha of misclassified polygons, which were often related to water runnels or shadows, leaving 5.91 ha of E. athericus cover. Overlap of both polygon maps was only 70% resulting from many small patches identified where E. athericus was absent. In sum, drones can greatly support field surveys in monitoring of plant species by allowing for accurate species maps and just-in-time captured very-high-resolution imagery.

Identifying the vegetation type in Google Earth images using a convolutional neural network: a case study for Japanese bamboo forests.

BACKGROUND: Classifying and mapping vegetation are crucial tasks in environmental science and natural resource management. However, these tasks are difficult because conventional methods such as field surveys are highly labor-intensive. Identification of target objects from visual data using computer techniques is one of the most promising techniques to reduce the costs and labor for vegetation mapping. Although deep learning and convolutional neural networks (CNNs) have become a new solution for image recognition and classification recently, in general, detection of ambiguous objects such as vegetation is still difficult. In this study, we investigated the effectiveness of adopting the chopped picture method, a recently described protocol for CNNs, and evaluated the efficiency of CNN for plant community detection from Google Earth images. RESULTS: We selected bamboo forests as the target and obtained Google Earth images from three regions in Japan. By applying CNN, the best trained model correctly detected over 90% of the targets. Our results showed that the identification accuracy of CNN is higher than that of conventional machine learning methods. CONCLUSIONS: Our results demonstrated that CNN and the chopped picture method are potentially powerful tools for high-accuracy automated detection and mapping of vegetation.

Mapping multi-scale vascular plant richness in a forest landscape with integrated LiDAR and hyperspectral remote-sensing.

The central role of floristic diversity in maintaining habitat integrity and ecosystem function has propelled efforts to map and monitor its distribution across forest landscapes. While biodiversity studies have traditionally relied largely on ground-based observations, the immensity of the task of generating accurate, repeatable, and spatially-continuous data on biodiversity patterns at large scales has stimulated the development of remote-sensing methods for scaling up from field plot measurements. One such approach is through integrated LiDAR and hyperspectral remote-sensing. However, despite their efficiencies in cost and effort, LiDAR-hyperspectral sensors are still highly constrained in structurally- and taxonomically-heterogeneous forests - especially when species' cover is smaller than the image resolution, intertwined with neighboring taxa, or otherwise obscured by overlapping canopy strata. In light of these challenges, this study goes beyond the remote characterization of upper canopy diversity to instead model total vascular plant species richness in a continuous-cover North Carolina Piedmont forest landscape. We focus on two related, but parallel, tasks. First, we demonstrate an application of predictive biodiversity mapping, using nonparametric models trained with spatially-nested field plots and aerial LiDAR-hyperspectral data, to predict spatially-explicit landscape patterns in floristic diversity across seven spatial scales between 0.01-900 m2 . Second, we employ bivariate parametric models to test the significance of individual, remotely-sensed predictors of plant richness to determine how parameter estimates vary with scale. Cross-validated results indicate that predictive models were able to account for 15-70% of variance in plant richness, with LiDAR-derived estimates of topography and forest structural complexity, as well as spectral variance in hyperspectral imagery explaining the largest portion of variance in diversity levels. Importantly, bivariate tests provide evidence of scale-dependence among predictors, such that remotely-sensed variables significantly predict plant richness only at spatial scales that sufficiently subsume geolocational imprecision between remotely-sensed and field data, and best align with stand components including plant size and density, as well as canopy gaps and understory growth patterns. Beyond their insights into the scale-dependent patterns and drivers of plant diversity in Piedmont forests, these results highlight the potential of remotely-sensible essential biodiversity variables for mapping and monitoring landscape floristic diversity from air- and space-borne platforms.

Modeling plant composition as community continua in a forest landscape with LiDAR and hyperspectral remote sensing.

In light of the need to operationalize the mapping of forest composition at landscape scales, this study uses multi-scale nested vegetation sampling in conjunction with LiDAR-hyperspectral remotely sensed data from the G-LiHT airborne sensor to map vascular plant compositional turnover in a compositionally and structurally complex North Carolina Piedmont forest. Reflecting a shift in emphasis from remotely sensing individual crowns to detecting aggregate optical-structural properties of forest stands, predictive maps reflect the composition of entire vascular plant communities, inclusive of those species smaller than the resolution of the remotely sensed imagery, intertwined with proximate taxa, or otherwise obscured from optical sensors by dense upper canopies. Stand-scale vascular plant composition is modeled as community continua: where discrete community-unit classes at different compositional resolutions provide interpretable context for continuous gradient maps that depict n-dimensional compositional complexity as a single, consistent RGB color combination. In total, derived remotely sensed predictors explain 71%, 54%, and 48% of the variation in the first three components of vascular plant composition, respectively. Among all remotely sensed environmental gradients, topography derived from LiDAR ground returns, forest structure estimated from LiDAR all returns, and morphological-biochemical traits determined from hyperspectral imagery each significantly correspond to the three primary axes of floristic composition in the study site. Results confirm the complementarity of LiDAR and hyperspectral sensors for modeling the environmental gradients constraining landscape turnover in vascular plant composition and hold promise for predictive mapping applications spanning local land management to global ecosystem modeling.

Comparison modeling for alpine vegetation distribution in an arid area.

Mapping and modeling vegetation distribution are fundamental topics in vegetation ecology. With the rise of powerful new statistical techniques and GIS tools, the development of predictive vegetation distribution models has increased rapidly. However, modeling alpine vegetation with high accuracy in arid areas is still a challenge because of the complexity and heterogeneity of the environment. Here, we used a set of 70 variables from ASTER GDEM, WorldClim, and Landsat-8 OLI (land surface albedo and spectral vegetation indices) data with decision tree (DT), maximum likelihood classification (MLC), and random forest (RF) models to discriminate the eight vegetation groups and 19 vegetation formations in the upper reaches of the Heihe River Basin in the Qilian Mountains, northwest China. The combination of variables clearly discriminated vegetation groups but failed to discriminate vegetation formations. Different variable combinations performed differently in each type of model, but the most consistently important parameter in alpine vegetation modeling was elevation. The best RF model was more accurate for vegetation modeling compared with the DT and MLC models for this alpine region, with an overall accuracy of 75 % and a kappa coefficient of 0.64 verified against field point data and an overall accuracy of 65 % and a kappa of 0.52 verified against vegetation map data. The accuracy of regional vegetation modeling differed depending on the variable combinations and models, resulting in different classifications for specific vegetation groups.

Invasion by Cedrela odorata threatens long distance migration of Galapagos tortoises.

Invasive alien species are among the most pervasive threats to biodiversity. Invasive species can cause catastrophic reductions in populations of native and endemic species and the collapse of ecosystem function. A second major global conservation concern is the extirpation of large-bodied mobile animals, including long-distance migrants, which often have keystone ecological roles over extensive spatial extents. Here, we report on a potentially catastrophic synergy between these phenomena that threatens the endemic biota of the Galapagos Archipelago. We used GPS telemetry to track 140 migratory journeys by 25 Western Santa Cruz Island Galapagos tortoises. We plotted the spatial interaction between tortoise migrations and recently established non-native forest dominated by the invasive tree Cedrela odorata (Cedrela forest). We qualified (a) the proportion of migratory journeys that traversed Cedrela forest, and (b) the probability that this observed pattern occurred by chance. Tortoise migrations were overwhelmingly restricted to small corridors between Cedrela forest blocks, indicating clear avoidance of those blocks. Just eight of 140 migrations traversed extensive Cedrela stands. Tortoises avoid Cedrela forest during their migrations. Further expansion of Cedrela forest threatens long-distance migration and population viability of critically endangered Galapagos tortoises. Applied research to determine effective management solutions to mitigate Cedrela invasion is a high priority.

Enhancing vegetation formation classification: Integrating coarse-scale traditional mapping knowledge and advanced machine learning.

Mapping vegetation formation types in large areas is crucial for ecological and environmental studies. However, this is still challenging to distinguish similar vegetation formation types using existing predictive vegetation mapping methods, based on commonly used environmental variables and remote sensing spectral data, especially when there are not enough training samples. To solve this issue, we proposed a predictive vegetation mapping method by integrating an advanced machine learning algorithm and knowledge in an early coarse-scale vegetation map (VMK). First, we implemented classification using the random forest algorithm by integrating the early vegetation map as an auxiliary feature (VMF). Then, we determined the rationality of classified vegetation types and distinguished the confusing types, respectively, based on the knowledge of the spatial distributions and hierarchies of vegetation. Finally, we replaced each recognized unreasonable vegetation type with its corresponding reasonable vegetation type. We implemented the new method in upstream of the Yellow River based on GaoFen-1 satellite images and other environmental variables (i.e., topographical and climate variables). Results showed that the overall accuracy using the VMK method ranged from 67.7 % to 76.8 %, which was 10.9 % to 13.4 % and 3.2 % to 6.6 %, respectively, higher than that of the method without the early vegetation map (NVM) and the VMF method, based on cross-validation with 20 % to 60 % random training samples. The spatial details of the vegetation map using the VMK method were also more reasonable compared to the NVM and VMF methods. These results indicated that the VMK method can distinctly improve the mapping accuracy at the vegetation formation level by integrating knowledge of existing vegetation maps. The proposed method can largely reduce the requirements on the number of field samples, which is especially important for alpine mountains and arctic region, where collecting training samples is more difficult due to the harsh natural environment.

U-Net Convolutional Neural Network for Mapping Natural Vegetation and Forest Types from Landsat Imagery in Southeastern Australia.

Accurate and comparable annual mapping is critical to understanding changing vegetation distribution and informing land use planning and management. A U-Net convolutional neural network (CNN) model was used to map natural vegetation and forest types based on annual Landsat geomedian reflectance composite images for a 500 km × 500 km study area in southeastern Australia. The CNN was developed using 2018 imagery. Label data were a ten-class natural vegetation and forest classification (i.e., Acacia, Callitris, Casuarina, Eucalyptus, Grassland, Mangrove, Melaleuca, Plantation, Rainforest and Non-Forest) derived by combining current best-available regional-scale maps of Australian forest types, natural vegetation and land use. The best CNN generated using six Landsat geomedian bands as input produced better results than a pixel-based random forest algorithm, with higher overall accuracy (OA) and weighted mean F1 score for all vegetation classes (93 vs. 87% in both cases) and a higher Kappa score (86 vs. 74%). The trained CNN was used to generate annual vegetation maps for 2000-2019 and evaluated for an independent test area of 100 km × 100 km using statistics describing accuracy regarding the label data and temporal stability. Seventy-six percent of pixels did not change over the 20 years (2000-2019), and year-on-year results were highly correlated (94-97% OA). The accuracy of the CNN model was further verified for the study area using 3456 independent vegetation survey plots where the species of interest had ≥ 50% crown cover. The CNN showed an 81% OA compared with the plot data. The model accuracy was also higher than the label data (76%), which suggests that imperfect training data may not be a major obstacle to CNN-based mapping. Applying the CNN to other regions would help to test the spatial transferability of these techniques and whether they can support the automated production of accurate and comparable annual maps of natural vegetation and forest types required for national reporting.

The vegetation of Holocene coastal dunes of the Cape south coast, South Africa.

The vegetation of calcareous coastal dunes of Holocene age along the south coast of South Africa's Cape Floristic Region is poorly described. This vegetation comprises a mosaic of communities associated with two biomes, Fynbos and Subtropical Thicket. Previously, expert knowledge rather than quantitative floristic analysis has been used to identify and delimit vegetation units. In many areas, mapped units conflate vegetation on Holocene sand with that on unconsolidated sediments of late Pleistocene age, despite pronounced species turnover across this edaphic boundary. Despite dominance by Cape lineages and fynbos vegetation, dune vegetation in the eastern part of the region has been included in the Subtropical Thicket Biome rather than the Fynbos Biome. The high levels of local plant endemism associated with this dune vegetation and the small and fragmented configuration of these habitats, makes it an urgent conservation priority especially when placed in the context of rising sea levels, increasing development pressures and numerous other threats. Here we provide a quantitative analysis of 253 plots of the 620 km2 of Holocene dune vegetation of the study area using phytosociological and multivariate methods. We identified six fynbos and two thicket communities based on the occurrences of 500 species. Following a long tradition in Cape vegetation typology, we used the Strandveld (beach vegetation) concept as our first-order vegetation entity and identified six units based on the fynbos floras. These were, from east to west, Southeastern Strandveld, St Francis Strandveld, Goukamma Strandveld, Southwestern Strandveld and Grootbos Strandveld. Each unit was differentiated by a suite of differential species, most being Holocene dune endemics. The two thicket communities-Mesic and Xeric Dune Thicket-showed limited variation across the study area and were subsumed into the Strandveld units. We discussed our findings in terms of vegetation-sediment relationships, emphasizing the need for a greater geographical coverage of sediment ages to facilitate a better understanding of deposition history on vegetation composition. We also discussed the role of soil moisture and fire regime on structuring the relative abundance of fynbos and thicket across the Holocene dune landscape. Finally, we address the conservation implications of our study, arguing that all remaining Holocene dune habitat should be afforded the highest conservation priority in regional land-use planning processes.

Requirements for a global lidar system: spaceborne lidar with wall-to-wall coverage.

Lidar is the optimum technology for measuring bare-Earth elevation beneath, and the structure of, vegetation. Consequently, airborne laser scanning (ALS) is widely employed for use in a range of applications. However, ALS is not available globally nor frequently updated due to its high cost per unit area. Spaceborne lidar can map globally but energy requirements limit existing spaceborne lidars to sparse sampling missions, unsuitable for many common ALS applications. This paper derives the equations to calculate the coverage a lidar satellite could achieve for a given set of characteristics (released open-source), then uses a cloud map to determine the number of satellites needed to achieve continuous, global coverage within a certain time-frame. Using the characteristics of existing in-orbit technology, a single lidar satellite could have a continuous swath width of 300 m when producing a 30 m resolution map. Consequently, 12 satellites would be needed to produce a continuous map every 5 years, increasing to 418 satellites for 5 m resolution. Building 12 of the currently in-orbit lidar systems is likely to be prohibitively expensive and so the potential of technological developments to lower the cost of a global lidar system (GLS) are discussed. Once these technologies achieve a sufficient readiness level, a GLS could be cost-effectively realized.

Robinia pseudoacacia-dominated vegetation types of Southern Europe: Species composition, history, distribution and management.

Knowledge of the species composition of invaded vegetation helps to evaluate an ecological impact of aliens and design an optimal management strategy. We link a new vegetation analysis of a large dataset to the invasion history, ecology and management of Robinia pseudoacacia stands across Southern Europe and provide a map illustrating Robinia distribution. Finally, we compare detected relationships with Central Europe. We show that regional differences in Robinia invasion, distribution, habitats and management are driven both by local natural conditions (climate and soil properties, low competitive ability with native trees) and socioeconomic factors (traditional land-use). Based on the classification of 467 phytosociological relevés we distinguished five broad vegetation types reflecting an oceanity-continentality gradient. The stands were heterogeneous and included 824 taxa, with only 5.8% occurring in more than 10% of samples, representing mainly hemerobic generalists of mesophilous, nutrient-rich and semi-shady habitats. The most common were dry ruderal stands invading human-made habitats. Among native communities, disturbed mesic and alluvial forests were often invaded throughout the area, while dry forests and scrub dominated in Balkan countries. Continuous, long-term and large-scale cultivation represent a crucial factor driving Robinia invasions in natural habitats. Its invasion should be mitigated by suitable management taking into account adjacent habitats and changing cultivation practices to select for native species. Robinia invasion has a comparable pattern in Central and Southern Europe, but there is a substantial difference in management and utilization causing heterogeneity of many South-European stands.

Monitoring recent changes of vegetation in Fildes Peninsula (King George Island, Antarctica) through satellite imagery guided by UAV surveys.

Mapping accurately vegetation surfaces in space and time in the ice-free areas of Antarctica can provide important information to quantitatively describe the evolution of their ecosystems. Spaceborne remote sensing is the adequate way to map and evaluate multitemporal changes on the Antarctic vegetation at large but its nature of occurrence, in relatively small and sparse patches, makes the identification very challenging. The inclusion of an intermediate scale of observation between ground and satellite scales, provided by Unmanned Aerial Vehicles (UAV) imagery, is of great help not only for their effective classification, but also for discriminating their main communities (lichens and mosses). Thus, this paper quantifies accurately recent changes of the vegetated areas in Fildes Peninsula (King George Island, Antarctica) through a novel methodology based on the integration of multiplatform data (satellite and UAV). It consists of multiscale imagery (spatial resolution of 2 m and 2 cm) from the same period to create a robust classifier that, after intensive calibration, is adequately used in other dates, where field reference data is scarce or not available at all. The methodology is developed and tested with UAV and satellite data from 2017 showing overall accuracies of 96% and kappa equal to 0.94 with a SVM classifier. These high performances allow the extrapolation to a pair of previous dates, 2006 and 2013, when atmospherically clear very high-resolution satellite imagery are available. The classification allows verifying a loss of the total area of vegetation of 4.5% during the 11-year time period under analysis, which corresponds to a 10.3% reduction for Usnea sp. and 9.8% for moss formations. Nevertheless, the breakdown analysis by time period shows a distinct behaviour for each vegetation type which are evaluated and discussed, namely for Usnea sp. whose decline is likely to be related to changing snow conditions.

Impact of ecological redundancy on the performance of machine learning classifiers in vegetation mapping.

Vegetation maps are models of the real vegetation patterns and are considered important tools in conservation and management planning. Maps created through traditional methods can be expensive and time-consuming, thus, new more efficient approaches are needed. The prediction of vegetation patterns using machine learning shows promise, but many factors may impact on its performance. One important factor is the nature of the vegetation-environment relationship assessed and ecological redundancy. We used two datasets with known ecological redundancy levels (strength of the vegetation-environment relationship) to evaluate the performance of four machine learning (ML) classifiers (classification trees, random forests, support vector machines, and nearest neighbor). These models used climatic and soil variables as environmental predictors with pretreatment of the datasets (principal component analysis and feature selection) and involved three spatial scales. We show that the ML classifiers produced more reliable results in regions where the vegetation-environment relationship is stronger as opposed to regions characterized by redundant vegetation patterns. The pretreatment of datasets and reduction in prediction scale had a substantial influence on the predictive performance of the classifiers. The use of ML classifiers to create potential vegetation maps shows promise as a more efficient way of vegetation modeling. The difference in performance between areas with poorly versus well-structured vegetation-environment relationships shows that some level of understanding of the ecology of the target region is required prior to their application. Even in areas with poorly structured vegetation-environment relationships, it is possible to improve classifier performance by either pretreating the dataset or reducing the spatial scale of the predictions.

Upland vegetation mapping using Random Forests with optical and radar satellite data.

Uplands represent unique landscapes that provide a range of vital benefits to society, but are under increasing pressure from the management needs of a diverse number of stakeholders (e.g. farmers, conservationists, foresters, government agencies and recreational users). Mapping the spatial distribution of upland vegetation could benefit management and conservation programmes and allow for the impacts of environmental change (natural and anthropogenic) in these areas to be reliably estimated. The aim of this study was to evaluate the use of medium spatial resolution optical and radar satellite data, together with ancillary soil and topographic data, for identifying and mapping upland vegetation using the Random Forests (RF) algorithm. Intensive field survey data collected at three study sites in Ireland as part of the National Parks and Wildlife Service (NPWS) funded survey of upland habitats was used in the calibration and validation of different RF models. Eight different datasets were analysed for each site to compare the change in classification accuracy depending on the input variables. The overall accuracy values varied from 59.8% to 94.3% across the three study locations and the inclusion of ancillary datasets containing information on the soil and elevation further improved the classification accuracies (between 5 and 27%, depending on the input classification dataset). The classification results were consistent across the three different study areas, confirming the applicability of the approach under different environmental contexts.

Small unmanned aerial vehicles (micro-UAVs, drones) in plant ecology.

PREMISE OF THE STUDY: Low-elevation surveys with small aerial drones (micro-unmanned aerial vehicles [UAVs]) may be used for a wide variety of applications in plant ecology, including mapping vegetation over small- to medium-sized regions. We provide an overview of methods and procedures for conducting surveys and illustrate some of these applications. METHODS: Aerial images were obtained by flying a small drone along transects over the area of interest. Images were used to create a composite image (orthomosaic) and a digital surface model (DSM). Vegetation classification was conducted manually and using an automated routine. Coverage of an individual species was estimated from aerial images. RESULTS: We created a vegetation map for the entire region from the orthomosaic and DSM, and mapped the density of one species. Comparison of our manual and automated habitat classification confirmed that our mapping methods were accurate. A species with high contrast to the background matrix allowed adequate estimate of its coverage. DISCUSSION: The example surveys demonstrate that small aerial drones are capable of gathering large amounts of information on the distribution of vegetation and individual species with minimal impact to sensitive habitats. Low-elevation aerial surveys have potential for a wide range of applications in plant ecology.

Transformaciones espacio-temporales de la cobertura vegetal en el Parque Nacional Corcovado, 1960-2014 / Space-time transformations of vegetation coverage in Corcovado National Park, 1960-2014

Resumen Los estudios a nivel regional que evalúan las dinámicas espacio-temporales de la vegetación en Costa Rica, especialmente, dentro de los Parques Nacionales son escasos. Así, considerando aportar en este vacío de conocimiento, este artículo analiza la distribución espacio-temporal de la vegetación dentro del periodo 1960-1976, 1992, 1997 y 2012 en Parque Nacional Corcovado, localizado en la Península de Osa y catalogado como el bosque tropical lluvioso más septentrional en la costa pacífica de América. Además, esta área contiene una riqueza de biodiversidad, fundamentada en su antigüedad geológica, el aislamiento que presentó durante largos periodos; así como las condiciones climáticas particulares que generan ecosistemas únicos como bosques nubosos relacionados con brisa marina a alturas de más de 500 msnm. Este estudio evalúa la distribución espacial de la vegetación a partir de mapas resultantes del proceso de fotointerpretación de imágenes del 1960, 1976, 1997 y 2012, así como del análisis del índice de paisaje. Se concluye que las transformaciones espacio-temporales de la vegetación durante el periodo de estudio han sido mínimas, y el hecho de que hayan sido escasas las áreas impactadas por la actividad antrópica, generó una restauración ecológica importante durante las últimas décadas. Se encontró una relación de expansión y contracción entre el bosque nuboso y bosque, así como este último y el bosque inundado, en función de la recuperación de la cobertura boscosa dentro del Parque Nacional y de la Península de Osa, y el volumen y distribución de la precipitación. Asimismo, este estudio propone la necesidad de establecer el monitoreo permanente de la vegetación para esclarecer las relaciones que se establecen entre estos tipos de vegetación.

Abstract Regional studies evaluating spatial-temporal transformations of vegetation in Costa Rica, especially within National Parks, are scarce. Therefore, this paper analyses the vegetation distribution during 1960, 1976, 1997 and 2014 in Corcovado National Park. This protected area is located in the Osa Peninsula, Costa Rica, and represents the Northern most tropical rain forest on the Pacific coast of America. This area offers a great wealth of biodiversity due to its geological formation, isolation for long time periods, and its particular climatic conditions that generate unique ecosystems such as cloud forests associated with ocean situated close to hill breezes located over 500 masl, as well as dense tropical forest. This study evaluates the spatial distribution of vegetation based on maps that resulted from the process of photo-interpretation of 1960, 1976, 1997 and 2012, as well as from the landscape index analysis. It concludes that during the study period, the vegetation changes have been minimal. Instead, in the few areas impacted by human activity (small-scale agriculture and pasture lands) an ecological restoration has occurred during recent decades. In addition, this research suggests that the recovering forest cover within the park and even within the Osa Peninsula has been expanding the cloud forest. An increase and contraction relationship between the different categories (Cloud forest and forests as well of flooded forest and forest in flat zones) was found. Furthermore, this study suggests the need of permanent plots in order to monitor vegetation and identify the factors that explain this previous process. Rev. Biol. Trop. 66(1): 352-367. Epub 2018 March 01.

Mapping tropical dry forest habitats integrating Landsat NDVI, Ikonos imagery, and topographic information in the Caribbean Island of Mona

Assessing the status of tropical dry forest habitats using remote sensing technologies is one of the research priorities for Neotropical forests. We developed a simple method for mapping vegetation and habitats in a tropical dry forest reserve, Mona Island, Puerto Rico, by integrating the Normalized Difference vegetation Index (NDvI) from Landsat, topographic information, and high-resolution Ikonos imagery. The method was practical for identifying vegetation types in areas with a great variety of plant communities and complex relief, and can be adapted to other dry forest habitats of the Caribbean Islands. NDvI was useful for identifying the distribution of forests, woodlands, and shrubland, providing a natural representation of the vegetation patterns on the island. The use of Ikonos imagery allowed increasing the number of land cover classes. As a result, sixteen land-cover types were mapped over the 5 500 ha area, with a kappa coefficient of accuracy equal to 79 %. This map is a central piece for modeling vertebrate species distribution and biodiversity patterns by the Puerto Rico Gap Analysis Project, and it is of great value for assisting research and management actions in the island. Rev. Biol. Trop. 56 (2): 625-639. Epub 2008 June 30.

El estudio y evaluación de los bosques tropicales secos mediante herramientas de teledetección es una de las prioridades de investigación en los ambientes neotropicales. Desarrollamos una metodología simple para mapear la vegetación de la isla de Mona, Puerto Rico, mediante el uso del índice de vegetación normalizado (NDVI por sus siglas en inglés) de Landsat, información topográfica, e imágenes auxiliares de alta resolución Ikonos. La metodología fue útil para identificar las clases de vegetación en un área de gran variedad de comunidades vegetales y relieve complejo, y puede ser adaptada a otras regiones de bosque seco de las islas del Caribe. El NDVI permitió identificar la distribución de los bosques cerrados, abiertos, y arbustos, proveyendo una representación natural de los patrones de vegetación en la isla. Las imágenes de Ikonos permitieron incrementar el número de clases detectadas. Como resultado, mapeamos 16 clases de cobertura del terreno en las 5 500 hectáreas de la isla de Mona, con un coeficiente de concordancia kappa de un 79%. La información obtenida en este estudio será utilizada para modelar la distribución de los vertebrados terrestres y patrones de biodiversidad en la isla, como parte del proyecto Gap Análisis de Puerto Rico, y es de gran valor para asistir en las actividades de investigación y manejo en la isla.