Mapping forest successional stages in the Brazilian Amazon using forest heights derived from TanDEM-X SAR interferometry

Bispo, Polyanna da Conceição; Pardini, Matteo; Papathanassiou, Konstantinos; Kugler, Florian; Balzter, Heiko; Rains, Dominik; Santos, João Roberto dos; Rizaev, Igor G.; Tansey, Kevin; Santos, M. N.; Araujo, L. S. de

doi:10.1016/j.rse.2019.05.013

Cited by 26 publications

(16 citation statements)

References 62 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Among the 12 HV-TDX models, R 2 increase was the largest in two LiDAR3 models of (1) HV + CHM12-LiDAR3 and (3) HV + CHM5-LiDAR3. Previous studies have also shown that a CHM obtained from InSAR data and an ALS-derived DTM retrieved forest properties well [51,52]. Our three LiDAR3 models, (1), (2) and 3, retrieved higher AGB values > 300 Mg/ha effectively.…”

Section: High Agb Estimation Accuracies Of Lidar3 Modelssupporting

confidence: 65%

“…Our test site is located in a hemi-boreal zone with medium forest cover. Although most studies have treated the SRTM DEM as a DSM [50][51][52], it is necessary to examine the possibility for the SRTM DEM to serve as a DTM in the hemi-boreal region. If the SRTM DEM was found to serve as a DTM in our test site, it enhances the opportunity of producing a large-scale AGB map with a global TDX DSM.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Synthesis of L-Band SAR and Forest Heights Derived from TanDEM-X DEM and 3 Digital Terrain Models for Biomass Mapping

et al. 2020

View full text Add to dashboard Cite

In this study, we compared the accuracies of above-ground biomass (AGB) estimated by integrating ALOS (Advanced Land Observing Satellite) PALSAR (Phased-Array-Type L-Band Synthetic Aperture Radar) data and TanDEM-X-derived forest heights (TDX heights) at four scales from 1/4 to 25 ha in a hemi-boreal forest in Japan. The TDX heights developed in this study included nine canopy height models (CHMs) and three model-based forest heights (ModelHs); the nine CHMs were derived from the three digital surface models (DSMs) of (I) TDX 12 m DEM (digital elevation model) product, (II) TDX 90 m DEM product and (III) TDX 5 m DSM, which we developed from two TDX–TSX (TerraSAR-X) image pairs for reference, and the three digital terrain models (DTMs) of (i) an airborne Light Detection and Ranging (LiDAR)-based DTM (LiDAR DTM), (ii) a topography-based DTM and (iii) the Shuttle Radar Topography Mission (SRTM) DEM; the three ModelHs were developed from the two TDX-TSX image pairs used in (III) and the three DTMs (i to iii) with the Sinc inversion model. In total, 12 AGB estimation models were developed for comparison. In this study, we included the C-band SRTM DEM as one of the DTMs. According to Walker et al. (2007), the SRTM DEM serves as a DTM for most of the Earth’s surface, except for the areas with extensive tree and/or shrub coverage, e.g., the boreal and Amazon regions. As our test site is located in a hemi-boreal zone with medium forest cover, we tested the ability of the SRTM DEM to serve as a DTM in our test site. This study especially aimed to analyze the capability of the two TDX DEM products (I and II) to estimate AGB in practice in the hemi-boreal region, and to examine how the different forest height creation methods (the simple DSM and DTM subtraction for the nine CHMs and the Sinc inversion model-based approach for the three ModelHs) and the different spatial resolutions of the three DSMs and three DTMs affected the AGB estimation results. We also conducted the slope-class analysis to see how the varying slopes influenced the AGB estimation accuracies. The results show that the combined use of the PALSAR data and the CHM derived from (I) TDX 12 m DEM and (i) LiDAR DTM achieved the highest AGB estimation accuracies across the scales (R2 ranged from 0.82 to 0.97), but the CHMs derived from (I) TDX 12 m DEM and another two DTMs, (ii) and (iii), showed low R2 values at any scales. In contrast, the two CHMs derived from (II) TDX 90 m DEM and both (i) LiDAR DTM and (iii) SRTM DEM showed high R2 values > 0.87 and 0.78, respectively, at the scales > 9.0 ha, but they yielded much lower R2 values at smaller scales. The three ModelHs gave the lowest R2 values across the scales (R2 ranged from 0.39 to 0.60). Analyzed by slope class at the 1.0 ha scale, however, all the 12 AGB estimation models yielded high R2 values > 0.66 at the lowest slope class (0° to 9.9°), including the three ModelHs (R2 ranged between 0.68 to 0.69). The two CHMs derived from (II) TDX 90 m DEM and both (i) LiDAR DTM and (iii) SRTM DEM showed R2 values of 0.80 and 0.71, respectively, at the lowest slope class, while the CHM derived from (I) TDX 12 m DEM and (i) LiDAR DTM showed high R2 values across the slope classes (R2 > 0.82). The results show that (I) TDX 12 m DEM had a high capability to estimate AGB, with a high accuracy across the scales and the slope classes in the form of CHM, but the use of (i) LiDAR DTM was required. On the other hand, (II) TDX 90 m DEM was able to achieve high AGB estimation accuracies not only with (i) LiDAR DTM, but also with (iii) SRTM DEM in the form of CHM, but it was limited to large scales > 9.0 ha; however, all the models developed in this study have the possibility to achieve higher AGB estimation accuracies at the 1.0 ha scale in flat terrains with slope < 10°. The analysis showed the strengths and limitations of each model, and it also indicates that the data creation methods, the spatial resolutions of datasets and topographic features affects the effective spatial scales for AGB mapping, and the optimal combinations of these features should be chosen to obtain high AGB estimation accuracies.

show abstract

Section: High Agb Estimation Accuracies Of Lidar3 Modelssupporting

confidence: 65%

Section: Introductionmentioning

confidence: 99%

Synthesis of L-Band SAR and Forest Heights Derived from TanDEM-X DEM and 3 Digital Terrain Models for Biomass Mapping

et al. 2020

View full text Add to dashboard Cite

show abstract

“…The microwave pulses transmitted by a SAR system, especially at longer wavelengths such as the L-or P-band, interact with the branches and trunks, providing information about the forest structure, which is highly correlated to AGB [37][38][39]. Airborne LiDAR provides information on the canopy height and canopy cover [40][41][42][43], which are good proxies for woody AGB estimations [44]. The main limitations of LiDAR are that it is still an expensive technology and it is typically not available for large areas [45].…”

Section: Introductionmentioning

confidence: 99%

Woody Aboveground Biomass Mapping of the Brazilian Savanna with a Multi-Sensor and Machine Learning Approach

et al. 2020

Self Cite

View full text Add to dashboard Cite

The tropical savanna in Brazil known as the Cerrado covers circa 23% of the Brazilian territory, but only 3% of this area is protected. High rates of deforestation and degradation in the woodland and forest areas have made the Cerrado the second-largest source of carbon emissions in Brazil. However, data on these emissions are highly uncertain because of the spatial and temporal variability of the aboveground biomass (AGB) in this biome. Remote-sensing data combined with local vegetation inventories provide the means to quantify the AGB at large scales. Here, we quantify the spatial distribution of woody AGB in the Rio Vermelho watershed, located in the centre of the Cerrado, at a high spatial resolution of 30 metres, with a random forest (RF) machine-learning approach. We produced the first high-resolution map of the AGB for a region in the Brazilian Cerrado using a combination of vegetation inventory plots, airborne light detection and ranging (LiDAR) data, and multispectral and radar satellite images (Landsat 8 and ALOS-2/PALSAR-2). A combination of random forest (RF) models and jackknife analyses enabled us to select the best remote-sensing variables to quantify the AGB on a large scale. Overall, the relationship between the ground data from vegetation inventories and remote-sensing variables was strong (R2 = 0.89), with a root-mean-square error (RMSE) of 7.58 Mg ha−1 and a bias of 0.43 Mg ha−1.

show abstract

“…The average annual rainfall is 1909 mm, with a dry season between July and November [46]. The site is situated in a relatively flat area with an elevation between 80 m and 180 m [47,48]. Vegetation in the TNF is formed mainly by dense and open ombrophilous rainforests.…”

Section: Study Areamentioning

confidence: 99%

“…Vegetation in the TNF is formed mainly by dense and open ombrophilous rainforests. These forests are composed of a continuous canopy of perennial trees with heights between 25 m and 30 m and occasional emergent trees that reach up to 50-60 m height [48]. This type of forest has several layers and sublayers under the canopy, including small trees, shrubby, and herbaceous plants.…”

Section: Study Areamentioning

confidence: 99%

Discriminating Forest Successional Stages, Forest Degradation, and Land Use in Central Amazon Using ALOS/PALSAR-2 Full-Polarimetric Data

et al. 2020

Self Cite

View full text Add to dashboard Cite

We discriminated different successional forest stages, forest degradation, and land use classes in the Tapajós National Forest (TNF), located in the Central Brazilian Amazon. We used full polarimetric images from ALOS/PALSAR-2 that have not yet been tested for land use and land cover (LULC) classification, neither for forest degradation classification in the TNF. Our specific objectives were: (1) to test the potential of ALOS/PALSAR-2 full polarimetric images to discriminate LULC classes and forest degradation; (2) to determine the optimum subset of attributes to be used in LULC classification and forest degradation studies; and (3) to evaluate the performance of Random Forest (RF) and Support Vector Machine (SVM) supervised classifications to discriminate LULC classes and forest degradation. PALSAR-2 images from 2015 and 2016 were processed to generate Radar Vegetation Index, Canopy Structure Index, Volume Scattering Index, Biomass Index, and Cloude–Pottier, van Zyl, Freeman–Durden, and Yamaguchi polarimetric decompositions. To determine the optimum subset, we used principal component analysis in order to select the best attributes to discriminate the LULC classes and forest degradation, which were classified by RF. Based on the variable importance score, we selected the four first attributes for 2015, alpha, anisotropy, volumetric scattering, and double-bounce, and for 2016, entropy, anisotropy, surface scattering, and biomass index, subsequently classified by SVM. Individual backscattering indexes and polarimetric decompositions were also considered in both RF and SVM classifiers. Yamaguchi decomposition performed by RF presented the best results, with an overall accuracy (OA) of 76.9% and 83.3%, and Kappa index of 0.70 and 0.80 for 2015 and 2016, respectively. The optimum subset classified by RF showed an OA of 75.4% and 79.9%, and Kappa index of 0.68 and 0.76 for 2015 and 2016, respectively. RF exhibited superior performance in relation to SVM in both years. Polarimetric attributes exhibited an adequate capability to discriminate forest degradation and classes of different ecological succession from the ones with less vegetation cover.

show abstract

Mapping forest successional stages in the Brazilian Amazon using forest heights derived from TanDEM-X SAR interferometry

Cited by 26 publications

References 62 publications

Synthesis of L-Band SAR and Forest Heights Derived from TanDEM-X DEM and 3 Digital Terrain Models for Biomass Mapping

Synthesis of L-Band SAR and Forest Heights Derived from TanDEM-X DEM and 3 Digital Terrain Models for Biomass Mapping

Woody Aboveground Biomass Mapping of the Brazilian Savanna with a Multi-Sensor and Machine Learning Approach

Discriminating Forest Successional Stages, Forest Degradation, and Land Use in Central Amazon Using ALOS/PALSAR-2 Full-Polarimetric Data

Contact Info

Product

Resources

About