Predicting tropical dry forest successional attributes from space: is the key hidden in image texture?

Positive effects of tree diversity on tropical forest restoration in a field-scale experiment

Experiments under controlled conditions have established that ecosystem functioning is generally positively related to levels of biodiversity, but it is unclear how widespread these effects are in real-world settings and whether they can be harnessed for ecosystem restoration. We used remote-sensing data from the first decade of a long-term, field-scale tropical restoration experiment initiated in 2002 to test how the diversity of planted trees affected recovery of a 500-ha area of selectively logged forest measured using multiple sources of satellite data. Replanting using species-rich mixtures of tree seedlings with higher phylogenetic and functional diversity accelerated restoration of remotely sensed estimates of aboveground biomass, canopy cover, and leaf area index. Our results are consistent with a positive relationship between biodiversity and ecosystem functioning in the lowland dipterocarp rainforests of SE Asia and demonstrate that using diverse mixtures of species can enhance their initial recovery after logging.

Remote sensing of savanna woody species diversity: A systematic review of data types and assessment methods

Despite savannas being known for their relatively sparse vegetation coverage compared to other vegetation ecosystems, they harbour functionally diverse vegetation forms. Savannas are affected by climate variability and anthropogenic factors, resulting in changes in woody plant species compositions. Monitoring woody plant species diversity is therefore important to inform sustainable biodiversity management. Remote sensing techniques are used as an alternative approach to labour-intensive field-based inventories, to assess savanna biodiversity. The aim of this paper is to review studies that applied remote sensing to assess woody plant species diversity in savanna environments. The paper first provides a brief account of the spatial distribution of savanna environments around the globe. Thereafter, it briefly defines categorical classification and continuous-scale species diversity assessment approaches for savanna woody plant estimation. The core review section divides previous remote sensing studies into categorical classification and continuous-scale assessment approaches. Within each division, optical, Radio Detection And Ranging (RADAR) and Light Detection and Ranging (LiDAR) remote sensing as applied to savanna woody species diversity is reviewed. This is followed by a discussion on multi-sensor applications to estimate woody plant species diversity in savanna. We recommend that future research efforts should focus strongly on routine application of optical, RADAR and LiDAR remote sensing of physiologically similar woody plant species in savannas, as well as on extending these methodological approaches to other vegetation environments.

Potentials and Limitations of WorldView-3 Data for the Detection of Invasive Lupinus polyphyllus Lindl. in Semi-Natural Grasslands

Semi-natural grasslands contribute highly to biodiversity and other ecosystem services, but they are at risk by the spread of invasive plant species, which alter their habitat structure. Large area grassland monitoring can be a powerful tool to manage invaded ecosystems. Therefore, WorldView-3 multispectral sensor data was utilized to train multiple machine learning algorithms in an automatic machine learning workflow called ‘H2O AutoML’ to detect L. polyphyllus in a nature protection grassland ecosystem. Different degree of L. polyphyllus cover was collected on 3 × 3 m2 reference plots, and multispectral bands, indices, and texture features were used in a feature selection process to identify the most promising classification model and machine learning algorithm based on mean per class error, log loss, and AUC metrics. The best performance was achieved with a binary classification of lupin-free vs. fully invaded 3 × 3 m2 plot classification with a set of 7 features out of 763. The findings reveal that L. polyphyllus detection from WorldView-3 sensor data is limited to large dominant spots and not recommendable for lower plant coverage, especially single plant detection. Further research is needed to clarify if different phenological stages of L. polyphyllus as well as time series increase classification performance.

Global tropical dry forest extent and cover: A comparative study of bioclimatic definitions using two climatic data sets

There is a debate concerning the definition and extent of tropical dry forest biome and vegetation type at a global spatial scale. We identify the potential extent of the tropical dry forest biome based on bioclimatic definitions and climatic data sets to improve global estimates of distribution, cover, and change. We compared four bioclimatic definitions of the tropical dry forest biome-Murphy and Lugo, Food and Agriculture Organization (FAO), DryFlor, aridity index-using two climatic data sets: WorldClim and Climatologies at High-resolution for the Earth's Land Surface Areas (CHELSA). We then compared each of the eight unique combinations of bioclimatic definitions and climatic data sets using 540 field plots identified as tropical dry forest from a literature search and evaluated the accuracy of World Wildlife Fund tropical and subtropical dry broadleaf forest ecoregions. We used the definition and climate data that most closely matched field data to calculate forest cover in 2000 and change from 2001 to 2020. Globally, there was low agreement (< 58%) between bioclimatic definitions and WWF ecoregions and only 40% of field plots fell within these ecoregions. FAO using CHELSA had the highest agreement with field plots (81%) and was not correlated with the biome extent. Using the FAO definition with CHELSA climatic data set, we estimate 4,931,414 km2 of closed canopy (≥ 40% forest cover) tropical dry forest in 2000 and 4,369,695 km2 in 2020 with a gross loss of 561,719 km2 (11.4%) from 2001 to 2020. Tropical dry forest biome extent varies significantly based on bioclimatic definition used, with nearly half of all tropical dry forest vegetation missed when using ecoregion boundaries alone, especially in Africa. Using site-specific field validation, we find that the FAO definition using CHELSA provides an accurate, standard, and repeatable way to assess tropical dry forest cover and change at a global scale.

Mapping Tree Species Deciduousness of Tropical Dry Forests Combining Reflectance, Spectral Unmixing, and Texture Data from High-Resolution Imagery

In tropical dry forests, deciduousness (i.e., leaf shedding during the dry season) is an important adaptation of plants to cope with water limitation, which helps trees adjust to seasonal drought. Deciduousness is also a critical factor determining the timing and duration of carbon fixation rates, and affecting energy, water, and carbon balance. Therefore, quantifying deciduousness is vital to understand important ecosystem processes in tropical dry forests. The aim of this study was to map tree species deciduousness in three types of tropical dry forests along a precipitation gradient in the Yucatan Peninsula using Sentinel-2 imagery. We propose an approach that combines reflectance of visible and near-infrared bands, normalized difference vegetation index (NDVI), spectral unmixing deciduous fraction, and several texture metrics to estimate the spatial distribution of tree species deciduousness. Deciduousness in the study area was highly variable and decreased along the precipitation gradient, while the spatial variation in deciduousness among sites followed an inverse pattern, ranging from 91.5 to 43.3% and from 3.4 to 9.4% respectively from the northwest to the southeast of the peninsula. Most of the variation in deciduousness was predicted jointly by spectral variables and texture metrics, but texture metrics had a higher exclusive contribution. Moreover, including texture metrics as independent variables increased the variance of deciduousness explained by the models from R2 = 0.56 to R2 = 0.60 and the root mean square error (RMSE) was reduced from 16.9% to 16.2%. We present the first spatially continuous deciduousness map of the three most important vegetation types in the Yucatan Peninsula using high-resolution imagery.

The potential of UAV-borne spectral and textural information for predicting aboveground biomass and N fixation in legume-grass mixtures

Organic farmers, who rely on legumes as an external nitrogen (N) source, need a fast and easy on-the-go measurement technique to determine harvestable biomass and the amount of fixed N (NFix) for numerous farm management decisions. Especially clover- and lucerne-grass mixtures play an important role in the organic crop rotation under temperate European climate conditions. Multispectral sensors mounted on unmanned aerial vehicles (UAVs) are new promising tools for a non-destructive assessment of crop and grassland traits on large and remote areas. One disadvantage of multispectral information and derived vegetations indices is, that both ignore spatial relationships of pixels to each other in the image. This gap can be filled by texture features from a grey level co-occurrence matrix. The aim of this multi-temporal field study was to provide aboveground biomass and NFix estimation models for two legume-grass mixtures through a whole vegetation period based on UAV multispectral information. The prediction models covered different proportions of legumes (0-100% legumes) to represent the variable conditions in practical farming. Furthermore, the study compared prediction models with and without the inclusion of texture features. As multispectral data usually suffers from multicollinearity, two machine learning algorithms, Partial Least Square and Random Forest (RF) regression, were used. The results showed, that biomass prediction accuracy for the whole dataset as well as for crop-specific models were substantially improved by the inclusion of texture features. The best model was generated for the whole dataset by RF with an rRMSE of 10%. For NFix prediction accuracy of the best model was based on RF including texture (rRMSEP = 18%), which was not consistent with crop specific models.

Tree Species (Genera) Identification with GF-1 Time-Series in A Forested Landscape, Northeast China

Forests are the most important component of terrestrial ecosystem; the accurate mapping of tree species is helpful for the management of forestry resources. Moderate- and high-resolution multispectral images have been commonly utilized to identify regional tree species in forest ecosystem, but the accuracy of recognition is still unsatisfactory. To enhance the forest mapping accuracy, this study integrated the land surface phenological metrics and text features of forest canopy on tree species identification based on Gaofen-1 (GF-1) wide field of view (WFV) and time-series images (36 10-day NDVI data), conducted at a forested landscape in Harqin Banner, Northeast China in 2017. The dominant tree species include Pinus tabulaeformis, Larix gmelinii, Populus davidiana, Betula platyphylla, and Quercus mongolica in the study region. The result of forest mapping derived from a 10-day dataset was also compared with the outcome based upon a commonly utilized 30-day dataset in tree species identification. The results indicate that tree species identification accuracy is significantly (p < 0.05) improved with higher temporal resolution (10-day, 79.4%) of images than commonly used monthly data (30-day, 76.14%), and the accuracy can be further increased to 85.13% with a combination of the information derived from principal component analysis (PCA) transformation, phenological metrics (standing for the information of growing season) and texture features. The integration of higher dimensional NDVI data, vegetation growth dynamics and feature of canopy simultaneously will be beneficial to map tree species at the landscape scale.

Combining Radar and Optical Imagery to Map Oil Palm Plantations in Sumatra, Indonesia, Using the Google Earth Engine

Monitoring the expansion of commodity crops in the tropics is crucial to safeguard forests for biodiversity and ecosystem services. Oil palm (Elaeis guineensis) is one such crop that is a major driver of deforestation in Southeast Asia. We evaluated the use of a semi-automated approach with random forest as a classifier and combined optical and radar datasets to classify oil palm land-cover in 2015 in Sumatra, Indonesia, using Google Earth Engine. We compared our map with two existing remotely-sensed oil palm land-cover products that utilized visual and semi-automated approaches for the same year. We evaluated the accuracy of oil palm land-cover classification from optical (Landsat), radar (synthetic aperture radar (SAR)), and combined optical and radar satellite imagery (Combined). Combining Landsat and SAR data resulted in the highest overall classification accuracy (84%) and highest producer’s and user’s accuracy for oil palm classification (84% and 90%, respectively). The amount of oil palm land-cover in our Combined map was closer to official government statistics than the two existing land-cover products that used visual interpretation techniques. Our analysis of the extents of disagreement in oil palm land-cover indicated that our map had comparable accuracy to one of them and higher accuracy than the other. Our results demonstrate that a combination of optical and radar data outperforms the use of optical-only or radar-only datasets for oil palm classification and that our technique of preprocessing and classifying combined optical and radar data in the Google Earth Engine can be applied to accurately monitor oil-palm land-cover in Southeast Asia.

How Can Remote Sensing Help Monitor Tropical Moist Forest Degradation?—A Systematic Review

In the context of the climate and biodiversity crisis facing our planet, tropical forests playing a key role in global carbon flux and containing over half of Earth’s species are important to preserve. They are today threatened by deforestation but also by forest degradation, which is more difficult to study. Here, we performed a systematic review of studies on moist tropical forest degradation using remote sensing and fitting indicators of forest resilience to perturbations. Geographical repartition, spatial extent and temporal evolution were analyzed. Indicators of compositional, structural and regeneration criteria were noted as well as remote sensing indices and metrics used. Tropical moist forest degradation is not extensively studied especially in the Congo basin and in southeast Asia. Forest structure (i.e., canopy gaps, fragmentation and biomass) is the most widely and easily measured criteria with remote sensing, while composition and regeneration are more difficult to characterize. Mixing LiDAR/Radar and optical data shows good potential as well as very high-resolution satellite data. The awaited GEDI and BIOMASS satellites data will fill the actual gap to a large extent and provide accurate structural information. LiDAR and unmanned aerial vehicles (UAVs) form a good bridge between field and satellite data. While the performance of the LiDAR is no longer to be demonstrated, particular attention should be brought to the UAV that shows great potential and could be more easily used by local communities and stakeholders.

Prediction of Biomass and N Fixation of Legume-Grass Mixtures Using Sensor Fusion

European farmers and especially organic farmers rely on legume-grass mixtures in their crop rotation as an organic nitrogen (N) source, as legumes can fix atmospheric N, which is the most important element for plant growth. Furthermore, legume-grass serves as valuable fodder for livestock and biogas plants. Therefore, information about aboveground biomass and N fixation (NFix) is crucial for efficient farm management decisions on the field level. Remote sensing, as a non-destructive and fast technique, provides different methods to quantify plant trait parameters. In our study, high-density point clouds, derived from terrestrial laser scanning (TLS), in combination with unmanned aerial vehicle-based multispectral (MS) data, were collected to receive information about three plant trait parameters (fresh and dry matter, nitrogen fixation) in two legume-grass mixtures. Several crop surface height metrics based on TLS and vegetation indices based on the four MS bands (green, red, red edge, and near-infrared) were calculated. Furthermore, eight texture features based on mean crop surface height and the four MS bands were generated to measure horizontal spatial heterogeneity. The aim of this multi-temporal study over two vegetation periods was to create estimation models based on biomass and N fixation for two legume-grass mixtures by sensor fusion, a combination of both sensors. To represent conditions in practical farming, e.g., the varying proportion of legumes, the experiment included pure stands of legume and grass of the mixtures. Sensor fusion of TLS and MS data was found to provide better estimates of biomass and N _Fix than separate data analysis. The study shows the important role of texture based on MS and point cloud data, which contributed greatly to the estimation model generation. The applied approach offers an interesting method for improvements in precision agriculture.

Determining the K coefficient to leaf area index estimations in a tropical dry forest

Vegetation indices are useful tools to remotely estimate several important parameters related to ecosystem functioning. However, improving and validating estimations for a wide range of vegetation types are necessary. In this study, we provide a methodology for the estimation of the leaf area index (LAI) in a tropical dry forest (TDF) using the light diffusion through the canopy as a function of the successional stage. For this purpose, we estimated the K coefficient, a parameter that relates the normalized difference vegetation index (NDVI) to LAI, based on photosynthetically active radiation (PAR) and solar radiation. The study was conducted in the Mata Seca State Park, in southeastern Brazil, from 2012 to 2013. We defined four successional stages (very early, early, intermediate, and late) and established one optical phenology tower at one plot of 20 × 20 m per stage. Towers measured the incoming and reflected solar radiation and PAR for NDVI calculation. For each plot, we established 24 points for LAI sampling through hemispherical photographs. Because leaf cover is highly seasonal in TDFs, we determined ΔK (leaf growth phase) and K_max (leaf maturity phase). We detected a strong correlation between NDVI and LAI, which is necessary for a reliable determination of the K coefficient. Both NDVI and LAI varied significantly between successional stages, indicating sensitivity to structural changes in forest regeneration. Furthermore, the K values differed between successional stages and correlated significantly with other environmental variables such as air temperature and humidity, fraction of absorbed PAR, and soil moisture. Thus, we established a model based on spectral properties of the vegetation coupled with biophysical characteristics in a TDF that makes possible to estimate LAI from NDVI values. The application of the K coefficient can improve remote estimations of forest primary productivity and gases and energy exchanges between vegetation and atmosphere. This model can be applied to distinguish different successional stages of TDFs, supporting environmental monitoring and conservation policies towards this biome.

Assessment of plant species diversity based on hyperspectral indices at a fine scale

Fast and nondestructive approaches of measuring plant species diversity have been a subject of excessive scientific curiosity and disquiet to environmentalists and field ecologists worldwide. In this study, we measured the hyperspectral reflectances and plant species diversity indices at a fine scale (0.8 meter) in central Hunshandak Sandland of Inner Mongolia, China. The first-order derivative value (FD) at each waveband and 37 hyperspectral indices were used to assess plant species diversity. Results demonstrated that the stepwise linear regression of FD can accurately estimate the Simpson (R² = 0.83), Pielou (R² = 0.87) and Shannon-Wiener index (R² = 0.88). Stepwise linear regression of FD (R² = 0.81, R² = 0.82) and spectral vegetation indices (R² = 0.51, R² = 0.58) significantly predicted the Margalef and Gleason index. It was proposed that the Simpson, Pielou and Shannon-Wiener indices, which are widely used as plant species diversity indicators, can be precisely estimated through hyperspectral indices at a fine scale. This research promotes the development of methods for assessment of plant diversity using hyperspectral data.

Modelling patterns of pollinator species richness and diversity using satellite image texture

Assessing species richness and diversity on the basis of standardised field sampling effort represents a cost- and time-consuming method. Satellite remote sensing (RS) can help overcome these limitations because it facilitates the collection of larger amounts of spatial data using cost-effective techniques. RS information is hence increasingly analysed to model biodiversity across space and time. Here, we focus on image texture measures as a proxy for spatial habitat heterogeneity, which has been recognized as an important determinant of species distributions and diversity. Using bee monitoring data of four years (2010–2013) from six 4 × 4 km field sites across Central Germany and a multimodel inference approach we test the ability of texture features derived from Landsat-TM imagery to model local pollinator biodiversity. Textures were shown to reflect patterns of bee diversity and species richness to some extent, with the first-order entropy texture and terrain roughness being the most relevant indicators. However, the texture measurements accounted for only 3–5% of up to 60% of the variability that was explained by our final models, although the results are largely consistent across different species groups (bumble bees, solitary bees). While our findings provide indications in support of the applicability of satellite imagery textures for modeling patterns of bee biodiversity, they are inconsistent with the high predictive power of texture metrics reported in previous studies for avian biodiversity. We assume that our texture data captured mainly heterogeneity resulting from landscape configuration, which might be functionally less important for wild bees than compositional diversity of plant communities. Our study also highlights the substantial variability among taxa in the applicability of texture metrics for modelling biodiversity.

The Effects of GLCM parameters on LAI estimation using texture values from Quickbird Satellite Imagery

When the leaf area index (LAI) of a forest reaches 3, the problem of spectrum saturation becomes the main limitation to improving the accuracy of the LAI estimate. A sensitivity analysis of the Grey Level Co-occurrence Matrix (GLCM) parameters which can be applied to satellite image processing and analysis showed that the most important parameters included orientation, displacement and moving window size. We calculated the values of Angular Second Moment (ASM), Entropy (ENT), Correlation (COR), Contrast (CON), Dissimilarity (DIS) and Homogeneity (HOM) from Quickbird panchromatic imagery using a GLCM method. Four orientations, seven displacements and seven window sizes were considered. An orientation of 90° was best for estimating the LAI of black locust forest, regardless of moving window size, displacement and texture parameters. Displacements of 3 pixels appeared to be best. The orientation and window size had only a little influence on these settings. The highest adjusted r² values were obtained using a 3 × 3 moving window size for ASM and ENT. The tendency of CON, COR, DIS and HOM to vary with window size was significantly affected by orientation. This study can help with parameter selection when texture features from high resolution imagery are used to estimate broad-leaved forest structure information.

Carbon sequestration potential of second-growth forest regeneration in the Latin American tropics

Regrowth of tropical secondary forests following complete or nearly complete removal of forest vegetation actively stores carbon in aboveground biomass, partially counterbalancing carbon emissions from deforestation, forest degradation, burning of fossil fuels, and other anthropogenic sources. We estimate the age and spatial extent of lowland second-growth forests in the Latin American tropics and model their potential aboveground carbon accumulation over four decades. Our model shows that, in 2008, second-growth forests (1 to 60 years old) covered 2.4 million km(2) of land (28.1% of the total study area). Over 40 years, these lands can potentially accumulate a total aboveground carbon stock of 8.48 Pg C (petagrams of carbon) in aboveground biomass via low-cost natural regeneration or assisted regeneration, corresponding to a total CO2 sequestration of 31.09 Pg CO2. This total is equivalent to carbon emissions from fossil fuel use and industrial processes in all of Latin America and the Caribbean from 1993 to 2014. Ten countries account for 95% of this carbon storage potential, led by Brazil, Colombia, Mexico, and Venezuela. We model future land-use scenarios to guide national carbon mitigation policies. Permitting natural regeneration on 40% of lowland pastures potentially stores an additional 2.0 Pg C over 40 years. Our study provides information and maps to guide national-level forest-based carbon mitigation plans on the basis of estimated rates of natural regeneration and pasture abandonment. Coupled with avoided deforestation and sustainable forest management, natural regeneration of second-growth forests provides a low-cost mechanism that yields a high carbon sequestration potential with multiple benefits for biodiversity and ecosystem services.

Inferring species richness and turnover by statistical multiresolution texture analysis of satellite imagery

Background

The quantification of species-richness and species-turnover is essential to effective monitoring of ecosystems. Wetland ecosystems are particularly in need of such monitoring due to their sensitivity to rainfall, water management and other external factors that affect hydrology, soil, and species patterns. A key challenge for environmental scientists is determining the linkage between natural and human stressors, and the effect of that linkage at the species level in space and time. We propose pixel intensity based Shannon entropy for estimating species-richness, and introduce a method based on statistical wavelet multiresolution texture analysis to quantitatively assess interseasonal and interannual species turnover.

Methodology/Principal Findings

We model satellite images of regions of interest as textures. We define a texture in an image as a spatial domain where the variations in pixel intensity across the image are both stochastic and multiscale. To compare two textures quantitatively, we first obtain a multiresolution wavelet decomposition of each. Either an appropriate probability density function (pdf) model for the coefficients at each subband is selected, and its parameters estimated, or, a non-parametric approach using histograms is adopted. We choose the former, where the wavelet coefficients of the multiresolution decomposition at each subband are modeled as samples from the generalized Gaussian pdf. We then obtain the joint pdf for the coefficients for all subbands, assuming independence across subbands; an approximation that simplifies the computational burden significantly without sacrificing the ability to statistically distinguish textures. We measure the difference between two textures' representative pdf's via the Kullback-Leibler divergence (KL). Species turnover, or diversity, is estimated using both this KL divergence and the difference in Shannon entropy. Additionally, we predict species richness, or diversity, based on the Shannon entropy of pixel intensity.To test our approach, we specifically use the green band of Landsat images for a water conservation area in the Florida Everglades. We validate our predictions against data of species occurrences for a twenty-eight years long period for both wet and dry seasons. Our method correctly predicts 73% of species richness. For species turnover, the newly proposed KL divergence prediction performance is near 100% accurate. This represents a significant improvement over the more conventional Shannon entropy difference, which provides 85% accuracy. Furthermore, we find that changes in soil and water patterns, as measured by fluctuations of the Shannon entropy for the red and blue bands respectively, are positively correlated with changes in vegetation. The fluctuations are smaller in the wet season when compared to the dry season.

Conclusions/Significance

Texture-based statistical multiresolution image analysis is a promising method for quantifying interseasonal differences and, consequently, the degree to which vegetation, soil, and water patterns vary. The proposed automated method for quantifying species richness and turnover can also provide analysis at higher spatial and temporal resolution than is currently obtainable from expensive monitoring campaigns, thus enabling more prompt, more cost effective inference and decision making support regarding anomalous variations in biodiversity. Additionally, a matrix-based visualization of the statistical multiresolution analysis is presented to facilitate both insight and quick recognition of anomalous data.