Coupling potential of ICESat/GLAS and SRTM for the discrimination of forest landscape types in French Guiana

The Shuttle Radar Topography Mission (SRTM) has produced the most accurate nearly global elevation dataset to date. Over vegetated areas, the measured SRTM elevations are the result of a complex interaction between radar waves and tree crowns. In this study, waveforms acquired by the Geoscience Laser Altimeter System (GLAS) were combined with SRTM elevations to discriminate the five forest landscape types (LTs) in French Guiana. Two differences were calculated: (1) penetration depth, defined as the GLAS highest elevations minus the SRTM elevations, and (2) the GLAS centroid elevations minus the SRTM elevations. The results show that these differences were similar for the five LTs, and they increased as a function of the GLAS canopy height and of the SRTM roughness index. Next, a Random Forest (RF) classifier was used to analyze the coupling potential of GLAS and SRTM in the discrimination of forest landscape types in French Guiana. The parameters used in the RF classification were the GLAS canopy height, the SRTM roughness index, the difference between the GLAS highest elevations and the SRTM elevations and the difference between the GLAS centroid elevations and the SRTM elevations. Discrimination of the five forest landscape types in French Guiana was possible, with an overall classification accuracy of 81.3% and a kappa coefficient of 0.75. All forest LTs were well classified with an accuracy varying from 78.4% to 97.5%. Finally, differences of near coincident GLAS waveforms, one from the wet season and one from the dry season, were analyzed. The results showed that the open forest LT (LT12), in some locations, contains trees that lose leaves during the dry season. These trees allow LT12 to be easily discriminated from the other LTs that retain their leaves using the following three criteria: (1) difference between the GLAS centroid elevations and the SRTM elevations, (2) ratio of top energy in the wet season to top energy in the dry season, or (3) ratio of ground energy in the wet season to ground energy in the dry season.

[1]  J. Chambers,et al.  Tree allometry and improved estimation of carbon stocks and balance in tropical forests , 2005, Oecologia.

[2]  Qi Chen.  Retrieving vegetation height of forests and woodlands over mountainous areas in the Pacific Coast region using satellite laser altimetry , 2010 .

[3]  Zhi-Hua Zhou,et al.  Exploratory Undersampling for Class-Imbalance Learning , 2009, IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics).

[4]  Valéry Gond,et al.  Broad-scale spatial pattern of forest landscape types in the Guiana Shield , 2011, Int. J. Appl. Earth Obs. Geoinformation.

[5]  J. Bryan Blair,et al.  Validation of SRTM Elevations Over Vegetated and Non-vegetated Terrain Using Medium-Footprint Lidar , 2006 .

[6]  A. Hudak,et al.  Nearest neighbor imputation of species-level, plot-scale forest structure attributes from LiDAR data , 2008 .

[7]  P. Addo‐Fordjour,et al.  Mixed Species Allometric Models for Estimating above-Ground Liana Biomass in Tropical Primary and Secondary Forests, Ghana , 2013 .

[8]  S. Fritz,et al.  A new land‐cover map of Africa for the year 2000 , 2004 .

[9]  Daniel Sabatier,et al.  Tropical forest phenology in French Guiana from MODIS time series , 2011 .

[10]  S. S. Alia,et al.  FUSION OF REMOTELY SENSED MULTISPECTRAL IMAGERY AND LIDAR DATA FOR FOREST STRUCTURE ASSESSMENT AT THE TREE LEVEL , 2008 .

[11]  Jacob T. Mundt,et al.  Mapping Sagebrush Distribution Using Fusion of Hyperspectral and Lidar Classifications , 2006 .

[12]  Wenze Yang,et al.  Physically based vertical vegetation structure retrieval from ICESat data: Validation using LVIS in White Mountain National Forest, New Hampshire, USA , 2011 .

[13]  W. Cohen,et al.  Estimates of forest canopy height and aboveground biomass using ICESat , 2005 .

[14]  R. Dubayah,et al.  Lidar Remote Sensing for Forestry , 2000, Journal of Forestry.

[15]  Andy Liaw,et al.  Classification and Regression by randomForest , 2007 .

[16]  Nicolas Baghdadi,et al.  Testing Different Methods of Forest Height and Aboveground Biomass Estimations From ICESat/GLAS Data in Eucalyptus Plantations in Brazil , 2014, IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing.

[17]  Mario Chica-Olmo,et al.  An assessment of the effectiveness of a random forest classifier for land-cover classification , 2012 .

[18]  Juha Hyyppä,et al.  Retrieval of Forest Aboveground Biomass and Stem Volume with Airborne Scanning LiDAR , 2013, Remote. Sens..

[19]  G. Vosselman,et al.  Single and two epoch analysis of ICESat full waveform data over forested areas , 2008 .

[20]  David J. Harding,et al.  SRTM C-band and ICESat Laser Altimetry Elevation Comparisons as a Function of Tree Cover and Relief , 2006 .

[21]  K. Ranson,et al.  Forest vertical structure from GLAS : An evaluation using LVIS and SRTM data , 2008 .

[22]  Alan H. Strahler,et al.  Assessing general relationships between aboveground biomass and vegetation structure parameters for improved carbon estimate from lidar remote sensing , 2009 .

[23]  Yi-Min Huang,et al.  Weighted support vector machine for classification with uneven training class sizes , 2005, 2005 International Conference on Machine Learning and Cybernetics.

[24]  Thomas R. Crow,et al.  Estimating aboveground biomass using Landsat 7 ETM+ data across a managed landscape in northern Wisconsin, USA , 2004 .

[25]  K. Becek.  INVESTIGATION OF ELEVATION BIAS OF THE SRTM C- AND X-BAND DIGITAL ELEVATION MODELS , 2008 .

[26]  Shu-Xin Du,et al.  Weighted support vector machine for classification , 2005, 2005 IEEE International Conference on Systems, Man and Cybernetics.

[27]  Clement Atzberger,et al.  Tree Species Classification with Random Forest Using Very High Spatial Resolution 8-Band WorldView-2 Satellite Data , 2012, Remote. Sens..

[28]  M. Lefsky,et al.  Mapping tropical forest biomass with radar and spaceborne LiDAR in Lopé National Park, Gabon: Overcoming problems of high biomass and persistent cloud , 2012 .

[29]  Alexander Braun,et al.  Verification of the Vertical Error in C-Band SRTM DEM Using ICESat and Landsat-7, Otter Tail County, MN , 2007, IEEE Transactions on Geoscience and Remote Sensing.

[30]  R. Colwell.  Remote sensing of the environment , 1980, Nature.

[31]  Lorenzo Bruzzone,et al.  Fusion of Hyperspectral and LIDAR Remote Sensing Data for Classification of Complex Forest Areas , 2008, IEEE Transactions on Geoscience and Remote Sensing.

[32]  Michelle Hofton,et al.  Laser altimeter return pulse correlation: a method for detecting surface topographic change , 2002 .

[33]  M. Lefsky,et al.  Laser altimeter canopy height profiles: methods and validation for closed-canopy, broadleaf forests , 2001 .

[34]  Nicolas Baghdadi,et al.  External Geophysics, Climate and Environment Assessment of C-band SRTM DEM in a dense equatorial forest zone , 2005 .

[35]  Hongjie Xie,et al.  Estimating vertical error of SRTM and map-based DEMs using ICESat altimetry data in the eastern Tibetan Plateau , 2011 .

[36]  Lindsay B. Hutley,et al.  Allometry for estimating aboveground tree biomass in tropical and subtropical eucalypt woodlands: towards general predictive equations , 2005 .

[37]  E. Rodríguez,et al.  A Global Assessment of the SRTM Performance , 2006 .

[38]  Leo Breiman,et al.  Random Forests , 2001, Machine Learning.

[39]  W. Cohen,et al.  Lidar Remote Sensing for Ecosystem Studies , 2002 .

[40]  Panayiotis E. Pintelas,et al.  Mixture of Expert Agents for Handling Imbalanced Data Sets , 2003 .

[41]  J. Rosette.  SATELLITE LIDAR ESTIMATION OF STEMWOOD VOLUME; A METHOD USING WAVEFORM DECOMPOSITION , 2008 .

[42]  Junjie Zhang,et al.  An improved method for estimating forest canopy height using ICESat-GLAS full waveform data over sloping terrain: A case study in Changbai mountains, China , 2010, Int. J. Appl. Earth Obs. Geoinformation.