Estimating montane forest above-ground biomass in the upper reaches of the Heihe River Basin using Landsat-TM data

In this work, the results of above-ground biomass (AGB) estimates from Landsat Thematic Mapper 5 (TM) images and field data from the fragmented landscape of the upper reaches of the Heihe River Basin (HRB), located in the Qilian Mountains of Gansu province in northwest China, are presented. Estimates of AGB are relevant for sustainable forest management, monitoring global change, and carbon accounting. This is particularly true for the Qilian Mountains, which are a water resource protection zone. We combined forest inventory data from 133 plots with TM images and Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) global digital elevation model (GDEM) V2 products (GDEM) in order to analyse the influence of the sun-canopy-sensor plus C (SCS+C) topographic correction on estimations of forest AGB using the stepwise multiple linear regression (SMLR) and k-nearest neighbour (k-NN) methods. For both methods, our results indicated that the SCS+C correction was necessary for getting more reliable forest AGB estimates within this complex terrain. Remotely sensed AGB estimates were validated against forest inventory data using the leave-one-out (LOO) method. An optimized k-NN method was designed by varying both mathematical formulation of the algorithm and remote-sensing data input, which resulted in 3000 different model configurations. Following topographic correction, performance of the optimized k-NN method was compared to that of the regression method. The optimized k-NN method (R² = 0.59, root mean square error (RMSE) = 24.92 tonnes ha^–1) was found to perform much better than the regression method (R² = 0.42, RMSE = 29.74 tonnes ha^–1) for forest AGB retrieval over this montane area. Our results indicated that the optimized k-NN method is capable of operational application to forest AGB estimates in regions where few inventory data are available.     

Aboveground forest biomass derived using multiple dates of WorldView-2 stereo-imagery: quantifying the improvement in estimation accuracy

ABSTRACT

The aim of this study was to investigate the capabilities of two date satellite-derived image-based point clouds (IPCs) to estimate forest aboveground biomass (AGB). The data sets used include panchromatic WorldView-2 stereo-imagery with 0.46 m spatial resolution representing 2014 and 2016 and a detailed digital elevation model derived from airborne laser scanning data. Altogether, 332 field sample plots with an area of 256 m² were used for model development and validation. Predictors describing forest height, density, and variation in height were extracted from the IPC 2014 and 2016 and used in k-nearest neighbour imputation models developed with sample plot data for predicting AGB. AGB predictions for 2014 (AGB₂₀₁₄) were projected to 2016 using growth models (AGB_{Projected_2016}) and combined with the AGB estimates derived from the 2016 data (AGB₂₀₁₆). AGB prediction model developed with 2014 data was also applied to 2016 data (AGB_{2016_pred2014}). Based on our results, the change in the 90^th percentile of height derived from the WorldView-2 IPC was able to characterize forest height growth between 2014 and 2016 with an average growth of 0.9 m. Features describing canopy cover and variation in height derived from the IPC were not as consistent. The AGB₂₀₁₆ had a bias of ?7.5% (?10.6 Mg ha^?1) and root mean square error (RMSE) of 26.0% (36.7 Mg ha^?1) as the respective values for AGB_{Projected_2016} were 7.0% (9.9 Mg ha^?1) and 21.5% (30.8 Mg ha^?1). AGB_{2016_pred2014} had a bias of ?19.6% (?27.7 Mg ha^?1) and RMSE of 33.2% (46.9 Mg ha^?1). By combining predictions of AGB₂₀₁₆ and AGB_{Projected_2016} at sample plot level as a weighted average, we were able to decrease the bias notably compared to estimates made on any single date. The lowest bias of ?0.25% (?0.4 Mg ha^?1) was obtained when equal weights of 0.5 were given to the AGB_{Projected_2016} and AGB₂₀₁₆ estimates. Respectively, RMSE of 20.9% (29.5 Mg ha^?1) was obtained using equal weights. Thus, we conclude that combination of two date WorldView-2 stereo-imagery improved the reliability of AGB estimates on sample plots where forest growth was the only change between the two dates.     

Volume estimation in a Eucalyptus plantation using multi-source remote sensing and digital terrain data: a case study in Minas Gerais State,Brazil

In this study, we tested the effectiveness of stand age, multispectral optical imagery obtained from the Landsat 8 Operational Land Imager (OLI), synthetic aperture radar (SAR) data acquired by the Sentinel-1B satellite, and digital terrain attributes extracted from a digital elevation model (DEM), in estimating forest volume in 351 plots in a 1,498 ha Eucalyptus plantation in northern Minas Gerais state, Brazil. A Random Forest (RF) machine learning algorithm was used following the Principal Component Analysis (PCA) of various data combinations, including multispectr al and SAR texture variables and DEM-based geomorphometric derivatives. Using multispectral, SAR or DEM variables alone (i.e. Experiments (ii)–(iv)) did not provide accurate estimates of volume (RMSE (Root Mean Square Error) > 32.00 m³ ha^?1) compared to predictions based on age since planting of Eucalyptus stands (Experiment (i)). However, when these datasets were individually combined with stand age (i.e. Experiments (v)–(vii)), the RF models resulted in better volume estimates than those obtained when using the individual multispectral, SAR and DEM datasets (RMSE < 28.00 m³ ha^?1). Furthermore, a model that integrated the selected variables of these data with stand age (Experiment (viii)) improved volume estimation significantly (RMSE = 22.33 m³ ha^?1). The large and increasing area of Eucalyptus forest plantations in Brazil and elsewhere suggests that this new approach to volume estimation has the potential to support Eucalyptus plantation monitoring and forest management practices.     

Using MODIS time series data to estimate aboveground biomass and its spatio-temporal variation in Inner Mongolia’s grassland between 2001 and 2011

It is critical to understanding grassland biomass and its dynamics to study regional carbon cycles and the sustainable use of grassland resources. In this study, we estimated aboveground biomass (AGB) and its spatio-temporal pattern for Inner Mongolia’s grassland between 2001 and 2011 using field samples, Moderate Resolution Imaging Spectroradiometer normalized difference vegetation index (MODIS-NDVI) time series data, and statistical models based on the relationship between NDVI and AGB. We also explored possible relationships between the spatio-temporal pattern of AGB and climatic factors. The following results were obtained: (1) AGB averaged 19.1 Tg C (1 Tg = 10¹² g) over a total area of 66.01 × 10⁴ km² between 2001 and 2011 and experienced a general fluctuation (coefficient of variation = 9.43%), with no significant trend over time (R² = 0.05, p > 0.05). (2) The mean AGB density was 28.9 g C m^?2 over the whole study area during the 11 year period, and it decreased from the northeastern part of the grassland to the southwestern part, exhibiting large spatial heterogeneity. (3) The AGB variation over the 11 year period was closely coupled with the pattern of precipitation from January to July, but we did not find a significant relationship between AGB and the corresponding temperature changes. Precipitation was also an important factor in the spatial pattern of AGB over the study area (R² = 0.41, p < 0.001), while temperature seemed to be a minor factor (R² = 0.14, p < 0.001). A moisture index that combined the effects of precipitation and temperature explained more variation in AGB than did precipitation alone (R² = 0.45, p < 0.001). Our findings suggest that establishing separate statistical models for different vegetation conditions may reduce the uncertainty of AGB estimation on a large spatial scale. This study provides support for grassland administration for livestock production and the assessment of carbon storage in Inner Mongolia.     

Comparison of different regression models and validation techniques for the assessment of wheat leaf area index from hyperspectral data

Leaf area index (LAI) is one of the most important plant parameters when observing agricultural crops and a decisive factor for yield estimates. Remote-sensing data provide spectral information on large areas and allow for a detailed quantitative assessment of LAI and other plant parameters. The present study compared support vector regression (SVR), random forest regression (RFR), and partial least-squares regression (PLSR) and their achieved model qualities for the assessment of LAI from wheat reflectance spectra. In this context, the validation technique used for verifying the accuracy of an empirical–statistical regression model was very important in order to allow the spatial transferability of models to unknown data. Thus, two different validation methods, leave-one-out cross-validation (cv) and independent validation (iv), were performed to determine model accuracy. The LAI and field reflectance spectra of 124 plots were collected from four fields during two stages of plant development in 2011 and 2012. In the case of cross-validation for the separate years, as well as the entire data set, SVR provided the best results (2011: R²_cv = 0.739, 2012: R²_cv = 0.85, 2011 and 2012: R²_cv = 0.944). Independent validation of the data set from both years led to completely different results. The accuracy of PLSR (R²_iv = 0.912) and RFR (R²_iv = 0.770) remained almost at the same level as that of cross-validation, while SVR showed a clear decline in model performance (R²_iv = 0.769). The results indicate that regression model robustness largely depends on the applied validation approach and the data range of the LAI used for model building.     

Predicting Eucalyptus spp. stand volume in Zululand,South Africa: an analysis using a stochastic gradient boosting regression ensemble with multi-source data sets

Accurate, reliable, and up-to-date forest stand volume information is a prerequisite for a detailed evaluation of commercial forest resources and their sustainable management. Commercial forest responses to global climate change remain uncertain, and hence the mapping of stand volume as carbon sinks is fundamentally important in understanding the role of forests in stabilizing climate change effects. The aim of this study was to examine the utility of stochastic gradient boosting (SGB) and multi-source data to predict stand volume of a Eucalyptus plantation in South Africa. The SGB ensemble, random forest (RF), and stepwise multiple-linear regression (SMLR) were used to predict Eucalyptus stand volume and other related tree-structural attributes such as mean tree height and mean diameter at breast height (DBH). Multi-source data consisted of SPOT-5 raw spectral features (four bands), 14 spectral vegetation indices, rainfall data, and stand age. When all variables were used, the SGB algorithm showed that stand volume can be accurately estimated (R² = 0.78 and RMSE = 33.16 m³ ha^?1 (23.01% of the mean)). The competing RF ensemble produced an R² value of 0.76 and a RMSE value of 37.28 m³ ha^?1 (38.28% of the mean). SMLR on the other hand, produced an R² value of 0.65 and an RMSE value of 42.50 m³ ha^?1 (42.50% of the mean). Our study further showed that Eucalyptus mean tree height (R² = 0.83 and RMSE = 1.63 m (9.08% of the mean)) and mean diameter at breast height (R² = 0.74 and RMSE = 1.06 (7.89% of the mean)) can also be reasonably predicted using SGB and multi-source data. Furthermore, when the most important SGB model-selected variables were used for prediction, the predictive accuracies improved significantly for mean DBH (R² = 0.81 and RMSE = 1.21 cm (6.12% of the mean)), mean tree height (R² = 0.86 and RMSE = 1.39 m (7.02% of the mean)), and stand volume (R² = 0.83 and RMSE = 29.58 m³ ha^?1 (17.63% of the mean)). These results underscore the importance of integrating multi-source data with remotely sensed data for predicting Eucalyptus stand volume and related tree-structural attributes.     

Estimating standing biomass in papyrus (Cyperus papyrus L.) swamp: exploratory of in situ hyperspectral indices and random forest regression

Accurate estimates of papyrus (Cyperus papyrus) biomass are critical for an efficient papyrus swamp monitoring and management system. The objective of this study was to test the utility of random forest (RF) regression and two narrow-band vegetation indices in estimating above-ground biomass (AGB) for complex and densely vegetated swamp canopies. The normalized difference vegetation index (NDVI) and enhanced vegetation index (EVI) were calculated from field spectrometry data and fresh AGB was measured in 82 quadrats at three different areas in the iSimangaliso Wetland Park, South Africa. NDVI was calculated from all possible band combinations of the electromagnetic spectrum (350 and 2500 nm), while EVI was calculated from possible band combinations in the blue, red, and near infrared of the spectrum. Backward feature elimination and RF regression were used as variable selection and modelling techniques to predict papyrus AGB. Results showed that the effective portions of electromagnetic spectrum for estimation AGB of papyrus swamp were located within the blue, red, red-edge, and near-infrared regions. The three best selected EVIs were computed from bands located at (i) 445, 682, and 829 nm, (ii) 497, 676, and 1091 nm, and (iii) 495, 678, and 1120 nm. These indices produced better predictive accuracies (R² = 0.90; root mean square error of prediction (RMSEP) = 0.289 kg m^?2; 7.99% of the mean) than the best selected NDVIs (R² = 0.85; RMSEP = 0.343 kgm^?2; 9.49% of the mean) that were calculated from bands located at (i) 739 and 829 nm, (ii) 739 and 814 nm, (iii) 744 and 789 nm, and (iv) 734 and 909 nm. The results of the present study demonstrate the utility of narrow-band vegetation indices (EVI and NDVI) and RF regression in estimating papyrus AGB at high density, a previously challenging task with broadband satellite sensors.     

Estimating plot volume using lidar height and intensity distributional parameters

This study explored the feasibility of height distributional metrics and intensity values extracted from low-density airborne light detection and ranging (lidar) data to estimate plot volumes in dense Korean pine (Pinus koraiensis) plots. Multiple linear regression analyses were performed using lidar height and intensity distributional metrics. The candidate variables for predicting plot volume were evaluated using three data sets: total, canopy, and integrated lidar height and intensity metrics. All intensities of lidar returns used were corrected by the reference distance. Regression models were developed using each data set, and the first criterion used to select the best models was the corrected Akaike Information Criterion (AIC_c). The use of three data sets was statistically significant at R² = 0.75 (RMSE = 52.17 m³ ha^?1), R² = 0.84 (RMSE = 45.24 m³ ha^?1), and R² = 0.91 (RMSE = 31.48 m³ ha^?1) for total, canopy, and integrated lidar distributional metrics, respectively. Among the three data sets, the integrated lidar metrics-derived model showed the best performance for estimating plot volumes, improving errors up to 42% when compared to the other two data sets. This is attributed to supplementing variables weighted and biased to upper limits in dense plots with more statistical variables that explain the lower limits. In all data sets, intensity metrics such as skewness, kurtosis, standard deviation, minimum, and standard error were employed as explanatory variables. The use of intensity variables improved the accuracy of volume estimation in dense forests compared to prior research. Correction of the intensity values contributed up to a maximum of 58% improvement in volume estimation when compared to the use of uncorrected intensity values (R² = 0.78, R² = 0.53, and R² = 0.63 for total, canopy, and integrated lidar distributional metrics, respectively). It is clear that the correction of intensity values is an essential step for the estimation of forest volume.     

Height and biomass of mangroves in Africa from ICESat/GLAS and SRTM

The accurate quantification of the three-dimensional (3-D) structure of mangrove forests is of great importance, particularly in Africa where deforestation rates are high and the lack of background data is a major problem. The objectives of this study are to estimate (1) the total area, (2) canopy height distributions, and (3) above-ground biomass (AGB) of mangrove forests in Africa. To derive the 3-D structure and biomass maps of mangroves, we used a combination of mangrove maps derived from Landsat Enhanced Thematic Mapper Plus (ETM+), lidar canopy height estimates from ICESat/GLAS (Ice, Cloud, and land Elevation Satellite/Geoscience Laser Altimeter System), and elevation data from SRTM (Shuttle Radar Topography Mission) for the African continent. The lidar measurements from the large footprint GLAS sensor were used to derive local estimates of canopy height and calibrate the interferometric synthetic aperture radar (InSAR) data from SRTM. We then applied allometric equations relating canopy height to biomass in order to estimate AGB from the canopy height product. The total mangrove area of Africa was estimated to be 25,960 km² with 83% accuracy. The largest mangrove areas and the greatest total biomass were found in Nigeria covering 8573 km² with 132 × 10⁶ Mg AGB. Canopy height across Africa was estimated with an overall root mean square error of 3.55 m. This error includes the impact of using sensors with different resolutions and geolocation error. This study provides the first systematic estimates of mangrove area, height, and biomass in Africa.     

An efficient model for the prediction of SMAP sea surface salinity using machine learning approaches in the Persian Gulf

ABSTRACT

Sea Surface Salinity (SSS) is a pre-eminent parameter in oceanology causing extreme climate and weather events such as floods and droughts. Therefore, knowledge discovery of SSS is increasingly becoming a fundamental problem in recent years. However, not only the inadequacy of in-situ SSS data in large ocean basins are hampering conduction of detailed analyses of patterning SSS variations but also conventional data-gathering techniques for SSS estimation are often too expensive and time-consuming to meet the amount of data required in SSS estimation studies. Conversely, the brand-new Soil Moisture Active-Passive (SMAP) mission could provide validated SSS data along with its main objective soil moisture retrieval. As a result, collecting a candidate data set of surface’s parameters as inputs to SSS with the aid of Pearson correlation and Boruta feature selection techniques, this paper aims to study the predictive skills of machine learning approaches to estimate SMAP radiometer SSS in the Persian Gulf region from April 2015 to April 2017. Thus, four machine learning methods including Support Vector Regression (SVR), artificial neural network (ANN), random forest (RF) and gradient boosting machine (GBM) were adopted to model the SSS. Two approaches of GBM and RF provided scarcely equivalent predictions for both the calibration and validation data sets that were distinguishably substantiated by experimental results and simulations, nonetheless, slightly superior results were attained with the GBM model by correlation coefficient (r) = 0.734, root mean squared error (RMSE) = 0.906 and mean absolute error (MAE) = 0.627. The findings demonstrate promising SSS estimation from SMAP, which could provide a baseline to perceive the large-scale changes in SSS.     

Impact of spatial variability of tropical forest structure on radar estimation of aboveground biomass

Understanding the spatial variability of tropical forest structure and its impact on the radar estimation of aboveground biomass (AGB) is important to assess the scale and accuracy of mapping AGB with future low frequency radar missions. We used forest inventory plots in old growth, secondary succession, and forest plantations at the La Selva Biological Station in Costa Rica to examine the spatial variability of AGB and its impact on the L-band and P-band polarimetric radar estimation of AGB at multiple spatial scales. Field estimation of AGB was determined from tree size measurements and an allometric equation developed for tropical wet forests. The field data showed very high spatial variability of forest structure with no spatial dependence at a scale above 11 m in old-growth forest. Plot sizes of greater than 0.25 ha reduced the coefficients of variation in AGB to below 20% and yielded a stationary and normal distribution of AGB over the landscape. Radar backscatter measurements at all polarization channels were strongly positively correlated with AGB at three scales of 0.25 ha, 0.5 ha, and 1.0 ha. Among these measurements, PHV and LHV showed strong sensitivity to AGB < 300 Mg ha^− 1 and AGB < 150 Mg ha^− 1 respectively at the 1.0 ha scale. The sensitivity varied across forest types because of differences in the effects of forest canopy and gap structure on radar attenuation and scattering. Spatial variability of structure and speckle noise in radar measurements contributed equally to degrading the sensitivity of the radar measurements to AGB at spatial scales less than 1.0 ha. By using algorithms based on polarized radar backscatter, we estimated AGB with RMSE = 22.6 Mg ha^− 1 for AGB < 300 Mg ha^− 1 at P-band and RMSE = 23.8 Mg ha^− 1 for AGB < 150 Mg ha^− 1 at L-band and with the accuracy optimized at 1-ha scale within 95% confidence interval. By adding the forest height, estimated from the C-band Interferometry data as an independent variable to the algorithm, the AGB estimation improved beyond the backscatter sensitivity by 20% at P-band and 40% at L-band. The results suggested the estimation of AGB can be improved substantially from the fusion of lidar or InSAR derived forest height with the polarimetric backscatter.     

Estimating aboveground woody biomass change in Kalahari woodland: combining field,radar, and optical data sets

Maps that accurately quantify aboveground vegetation biomass (AGB) are essential for ecosystem monitoring and conservation. Throughout Namibia, four vegetation change processes are widespread, namely, deforestation, woodland degradation, the encroachment of the herbaceous and grassy layers by woody strata (woody thickening), and woodland regrowth. All of these vegetation change processes affect a range of key ecosystem services, yet their spatial and temporal dynamics and contributions to AGB change remain poorly understood. This study quantifies AGB associated with the different vegetation change processes over an 8-year period, for a region of Kalahari woodland savannah in northern Namibia. Using data from 101 forest inventory plots collected during two field campaigns (2014–2015), we model AGB as a function of the Advanced Land Observing Satellite Phased Array L-band synthetic aperture radar (PALSAR and PALSAR-2) and dry season Landsat vegetation index composites, for two periods (2007 and 2015). Differences in AGB between 2007 and 2015 were assessed and validated using independent data, and changes in AGB for the main vegetation processes are quantified for the whole study area (75,501 km²). We find that woodland degradation and woody thickening contributed a change in AGB of ?14.3 and 2.5 Tg over 14% and 3.5% of the study area, respectively. Deforestation and regrowth contributed a smaller portion of AGB change, i.e. ?1.9 and 0.2 Tg over 1.3% and 0.2% of the study area, respectively.     

Modelling above-ground live trees biomass and carbon stock estimation of tropical lowland Dipterocarp forest: integration of field-based and remotely sensed estimates

Carbon exists as carbon dioxide (CO₂) which is one of the greenhouse gases (GHG) in the atmosphere that has an enormous influence on the impact of climate change. Therefore, the forest plays an undeniably pivotal role as a carbon sink, which absorbs carbon dioxide from the atmosphere. This research aims to develop allometric equation for above-ground live tree biomass (AGB) by combining field-based, combination of field data observation and technology (WV-3 and light detection and ranging (lidar)) and by using only technology derivation. The independent predictor was induced based on the literature review and theories, and an ordinary least square (OLS) estimator will be used to develop multiple linear regression models. During model selection, the best model fit was selected by calculating statistical parameters such as residual of the coefficient of determination (R²) selection methods, adjusted coefficient of determination (R²_adj), root mean square error, graphical analysis of the residuals, standard error (S_yx), and Akaike information criterion. An allometric equation of this research was developed using carbon stocks as dependent variables, and four of the predictor’s variables: diameter at breast height (DBH); total height observed at field (h_F); total height derived from airborne lidar (h_L); and morphometric variables of the crown projection area (CPA). Based on the statistic indicators, the most suitable model is Model 1, ln (S_c) = – β₀ + β₁ ln (h_L) + β₂ ln (DBH) + β₃ ln (CPA) for the combination of remote sensing and field observation; ln (S_c) = – β₀ + β₁ ln (h_F) + β₂ ln (DBH) for field inventory only; and ln (S_c) = – β₀ + β₁ ln (h_L) + β₂ ln (CPA) for remote sensing only. This model is reliable in forest management to estimate the AGB and carbon stock estimation using a selection of variable sources.     

Moso bamboo forest extraction and aboveground carbon storage estimation based on multi-source remotely sensed images

Using a combination of moso bamboo forest thematic maps derived from Landsat Thematic Mapper (TM) images, field inventory data, and Moderate Resolution Imaging Spectroradiometer (MODIS) normalized difference vegetation index (NDVI) images, moso bamboo forest was extracted using the matched filtering (MF) technique and its aboveground carbon storage (AGC) was then estimated. This study presents a feasible method for extracting large-scale moso bamboo forests and for estimating moso bamboo forest AGC based on low-spatial resolution MODIS images. The results showed that moso bamboo forests in the majority of counties can be accurately estimated between actual area and estimates, with an R ² of 0.8453. The fitted accuracy of the AGC model was high (R ² = 0.491). The prediction accuracy of the AGC model was also evaluated using validation samples collected from Lin'an City, with an R ² and root mean square error prediction of 0.4778 and 3.06 Mg C ha^?1, respectively. The AGC in the majority of counties or cities in Zhejiang Province was between 0 and 15 Mg C ha^?1, and to a certain extent the predicted AGC estimates were close to observed ground truth data and representative of the study area.     

Predicting the spectral information of future land cover using machine learning

Application of machine learning models to study land-cover change is typically restricted to the change detection of categorical, i.e. classified, land-cover data. In this study, our aim is to extend the utility of such models to predict the spectral band information of satellite images. A Random Forests (RF)-based machine learning model is trained using topographic and historical climatic variables as inputs to predict the spectral band values of high-resolution satellite imagery across two large sites in the western United States, New Mexico (10,570 km²), and Washington (9400 km²). The model output is used to obtain a true colour photorealistic image and an image showing the normalized difference vegetation index values. We then use the trained model to explore what the land cover might look like for a climate change scenario during the 2061–2080 period. The RF model achieves high validation accuracy for both sites during the training phase, with the coefficient of determination (R²) = 0.79 for New Mexico site and R² = 0.73 for Washington site. For the climate change scenario, prominent land-cover changes are characterized by an increase in the vegetation cover at the New Mexico site and a decrease in the perennial snow cover at the Washington site. Our results suggest that direct prediction of spectral band information is highly beneficial due to the ability it provides for deriving ecologically relevant products, which can be used to analyse land-cover change scenarios from multiple perspectives.     

TanDEM-X elevation model data for canopy height and aboveground biomass retrieval in a tropical peat swamp forest

It was demonstrated in the past that radar data is useful to estimate aboveground biomass due to their interferometric capability. Therefore, the potential of a globally available TanDEM-X digital elevation model (DEM) was investigated for aboveground biomass estimation via canopy height models (CHMs) in a tropical peat swamp forest. However, CHMs based on X-band interferometers usually require external terrain models. High accurate terrain models are not available on global scale. Therefore, an approach exclusively based on TanDEM-X and the decrease of accuracy compared to an approach utilizing a high accurate terrain model is assessed. In addition, the potential of X-band interferometric heights in tropical forests needs to be evaluated. Therefore, two CHMs were derived from an intermediate TanDEM-X DEM (iDEM; as a precursor for WorldDEM^TM) alone and in combination with lidar measurements used as terrain model. The analysis showed high accuracies (root mean square error [RMSE] = 5 m) for CHMs based on iDEM and reliable estimation of aboveground biomass. The iDEM CHM, exclusively based on TanDEM-X, achieved a poor R² of 0.2, nonetheless resulted in a cross-validated RMSE of 54 t ha^?1 (16%). The low R² suggested that the X-band height alone was not sufficient to estimate an accurate CHM, and thus the need for external terrain models was confirmed. A CHM retrieved from the difference of iDEM and an accurate lidar terrain model achieved a considerably higher correlation with aboveground biomass (R² = 0.68) and low cross-validated RMSE of 24.5 t ha^?1 (7.5%). This was higher or comparable to other aboveground biomass estimations in tropical peat swamp forests. The potential of X-band interferometric heights for CHM and biomass estimation was thus confirmed in tropical forest in addition to existing knowledge in boreal forests.     

Mapping of flooded vegetation by means of polarimetric Sentinel-1 and ALOS-2/PALSAR-2 imagery

This article presents for the first time the combination of dual-polarimetric C-band Sentinel-1 synthetic aperture radar (SAR) data and quad-polarimetric L-band ALOS-2/PALSAR-2 imagery for mapping of flooded areas with a special focus on flooded vegetation. L-band SAR data is well suited for mapping of flooded vegetation, while C-band enables an accurate extraction open water areas. Polarimetric decomposition-based unsupervised Wishart classification is combined with object-based post-classification refinement and the integration of spatial contextual information and global auxiliary data. In eight different scenarios, focusing on single datasets or fusion of classification results of several ones, respectively, different polarimetric decomposition and classification principles, including the entropy/anisotropy/alpha and the Freeman–Durden–Wishart classification, were investigated. The helix scattering component of the Yamaguchi decomposition, derived from ALOS-2 imagery, showed high suitability to refine the Sentinel-1-based detection of flooded vegetation. A test site at the Evros River (Greek/Turkish border region) was chosen, which was affected by a flooding event that occurred in spring 2015. The validation was based on high spatial resolution optical WorldView-2 imagery acquired with short temporal delay to the SAR data.     

Forest attribute imputation using machine-learning methods and ASTER data: comparison of k-NN,SVR and random forest regression algorithms

In a small case study of mixed hardwood Hyrcanian forests of Iran, three non-parametric methods, namely k-nearest neighbour (k-NN), support vector machine regression (SVR) and tree regression based on random forest (RF), were used in plot-level estimation of volume/ha, basal area/ha and stems/ha using field inventory and Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) data. Relevant pre-processing and processing steps were applied to the ASTER data for geometric and atmospheric correction and for enhancing quantitative forest parameters. After collecting terrestrial information on trees in the 101 samples, the volume, basal area and tree number per hectare were calculated in each plot. In the k-NN implementation using different distance measures and k, the cross-validation method was used to find the best distance measure and optimal k. In SVR, the best regularized parameters of four kernel types were obtained using leave-one-out cross-validation. RF was implemented using a bootstrap learning method with regularized parameters for decision tree model and stopping. The validity of performances was examined using unused test samples by absolute and relative root mean square error (RMSE) and bias metrics. In volume/ha estimation, the results showed that all the three algorithms had similar performances. However, SVR and RF produced better results than k-NN with relative RMSE values of 28.54, 25.86 and 26.86 (m³ ha^–1), respectively, using k-NN, SVR and RF algorithms, but RF could generate unbiased estimation. In basal area/ha and stems/ha estimation, the implementation results of RF showed that RF was slightly superior in relative RMSE (18.39, 20.64) to SVR (19.35, 22.09) and k-NN (20.20, 21.53), but k-NN could generate unbiased estimation compared with the other two algorithms used.     

Estimating NO2 dry deposition using satellite data in eastern China

The spatial and temporal characters of ground-level NO₂ concentration over eastern China were retrieved from the monthly averaged tropospheric NO₂ column densities from Global Ozone Monitoring Experiment (GOME, data used in this study are from April 1996 to December 2002) and Scanning Imaging Absorption Spectrometer for Atmospheric Chartography (SCIAMACHY, data used in this study are from January 2003 to December 2011) measurements. Together with the NO₂ concentration and the dry deposition velocity maps of eastern China, the fluxes of NO₂ dry deposition were estimated for three specific regions. The results indicated that the surface NO₂ concentration in eastern China increased dramatically from 1996 to 2011, and it showed distinct regional and seasonal variational characteristics. The highest concentration occurred in winter while the lowest occurred in summer. There was also variation in the spatial distribution with the peak value of NO₂ concentration appearing in the plains of north China (R1), the Yangtze River delta (R2), and the Pearl River delta (R3). A sharp increase of NO₂ concentration appeared in R1 and R2, while it was invariant or showed an obvious decrease in R3 during the period of 1996–2011. Furthermore, we compared the NO₂ dry deposition fluxes estimated from the ground-level NO₂ concentration and the dry deposition velocity of NO₂ with the mass concentration of NO₂ dry deposition that, measured from the control experiments and by consulting the published literature, showed a significant correlation (P < 0.001) and had a high R value (R = 0.73). The results also indicated that the NO₂ dry deposition fluxes increased over eastern China, with a maximum value of 8.25 kg N ha^?1 yr^?1 from 1996 to 2011 in R3, while the value was characterized by fluxes of less than 2.27 kg N ha^?1 yr^?1 in R2. When comparing the NO₂ dry deposition over different land covers, the values distinctly peaked over artificial surfaces and evergreen forests, with maximum values of 10.07 and 9.49 kg N ha^?1 yr^?1 in R1, 5.05 and 4.94 kg N ha^?1 yr^?1 in R2, and 20.95 and 23.15 kg N ha^?1 yr^?1 in R3. However, the lowest value of NO₂ dry deposition flux appeared over needleleaf forests, with 0.53, 0.24, and 1.29 kg N ha^?1 yr^?1 for R1, R2, and R3, respectively.