Improving the classification of six evergreen subtropical tree species with multi-season data from leaf spectra simulated to WorldView-2 and RapidEye

ABSTRACT Remote sensing offers a feasible means to monitor tree species at a regional level where species distribution and composition is affected by the impacts of global change. Furthermore, the temporal resolution of space-borne multispectral sensors offers the ability to combine phenologically important phases for the optimization of tree species classification. In this study, we determined whether multi-seasonal leaf-level spectral data (winter, spring, summer, and autumn) improved the classification of six evergreen tree species in the subtropical forest region of South Africa when compared to a single season, for hyperspectral data, and reflectance data simulated to the WorldView-2 (WV2) and RapidEye (RE) sensors. Classification accuracies of the test data were assessed using a Partial Least Square Random Forest algorithm. The accuracies were compared between single seasons and multi-season classification and across seasons using analysis of variance and post-hoc Tukey Honest Significant Difference tests. The average overall accuracy (OA) of the leaf-level hyperspectral data ranged from a minimum of 90 ± 3.5% in winter to a maximum of 92 ± 2.7% in summer, outperforming the simulated reflectance data for the WV2 and RE sensors with an average OA of between 8 and 10 percentage points (p < 0.02, Bonferroni corrected). The use of data from multiple seasons increased the average OA and decreased the number of species pair confusions for the simulated multispectral classifications. The producer’s and user’s accuracies of the hyperspectral classification were >82% and showed no significant change using multi-season data. Multiple seasons may therefore be beneficial to multispectral sensors with ≤8 bands, yet remains to be tested at canopy level, for other species and climatic regions.

[1]  I. Valiela,et al.  Mangrove Forests: One of the World's Threatened Major Tropical Environments , 2001 .

[2]  Onisimo Mutanga,et al.  Reducing Leaf-Level Hyperspectral Data to 22 Components of Biochemical and Biophysical Bands Optimizes Tree Species Discrimination , 2015, IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing.

[3]  Andrew K. Skidmore,et al.  Estimation of green grass/herb biomass from airborne hyperspectral imagery using spectral indices and partial least squares regression , 2007, Int. J. Appl. Earth Obs. Geoinformation.

[4]  H. Rodríguez,et al.  Seasonal Trends of Chlorophylls a and b and Carotenoids(x + c) in Native Trees and Shrubs of Northeastern Mexico , 2008 .

[5]  A. Prasad,et al.  Newer Classification and Regression Tree Techniques: Bagging and Random Forests for Ecological Prediction , 2006, Ecosystems.

[6]  Mary E. Martin,et al.  Determining Forest Species Composition Using High Spectral Resolution Remote Sensing Data , 1998 .

[7]  Onisimo Mutanga,et al.  Do seasonal profiles of foliar pigments improve species discrimination of evergreen coastal tree species in KwaZulu- Natal, South Africa? , 2013 .

[8]  C. Vaiphasa.  Remote sensing techniques for mangrove mapping , 2006 .

[9]  R Core Team,et al.  R: A language and environment for statistical computing. , 2014 .

[10]  M. C. Rutherford,et al.  The vegetation of South Africa, Lesotho and Swaziland. , 2006 .

[11]  A. Skidmore,et al.  Narrow band vegetation indices overcome the saturation problem in biomass estimation , 2004 .

[12]  G. Asner,et al.  Cloud cover in Landsat observations of the Brazilian Amazon , 2001 .

[13]  B. Gao.  NDWI—A normalized difference water index for remote sensing of vegetation liquid water from space , 1996 .

[14]  Russell G. Congalton,et al.  A review of assessing the accuracy of classifications of remotely sensed data , 1991 .

[15]  F. Sokolic.  The use of satellite remote sensing to determine the spatial and temporal distribution of surface water on the eastern shores of Lake St. Lucia. , 2006 .

[16]  Fuan Tsai,et al.  Spectrally segmented principal component analysis of hyperspectral imagery for mapping invasive plant species , 2007 .

[17]  K. Hufkens,et al.  Seasonal patterns of foliar reflectance in relation to photosynthetic capacity and color index in two co-occurring tree species, Quercus rubra and Betula papyrifera , 2012 .

[18]  Åsa Persson,et al.  Species identification of individual trees by combining high resolution LiDAR data with multi‐spectral images , 2008 .

[19]  Galal Omer,et al.  Performance of Support Vector Machines and Artificial Neural Network for Mapping Endangered Tree Species Using WorldView-2 Data in Dukuduku Forest, South Africa , 2015, IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing.

[20]  Lori M. Bruce,et al.  Why principal component analysis is not an appropriate feature extraction method for hyperspectral data , 2003, IGARSS 2003. 2003 IEEE International Geoscience and Remote Sensing Symposium. Proceedings (IEEE Cat. No.03CH37477).

[21]  Jan de Leeuw,et al.  Discriminating species using hyperspectral indices at leaf and canopy scales. The International Arch , 2007 .

[22]  L. Bruzzone,et al.  Tree species classification in the Southern Alps based on the fusion of very high geometrical resolution multispectral/hyperspectral images and LiDAR data , 2012 .

[23]  Henning Buddenbaum,et al.  Comparison of Feature Reduction Algorithms for Classifying Tree Species With Hyperspectral Data on Three Central European Test Sites , 2014, IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing.

[24]  A. Gitelson,et al.  Assessing Carotenoid Content in Plant Leaves with Reflectance Spectroscopy¶ , 2002, Photochemistry and photobiology.

[25]  H. Hotelling.  Analysis of a complex of statistical variables into principal components. , 1933 .

[26]  Tung Fung,et al.  Band Selection Using Hyperspectral Data of Subtropical Tree Species , 2003 .

[27]  S. Oldfield,et al.  The world list of threatened trees , 1998 .

[28]  LarrañagaPedro,et al.  A review of feature selection techniques in bioinformatics , 2007 .

[29]  R. Clark,et al.  Spectroscopic Determination of Leaf Biochemistry Using Band-Depth Analysis of Absorption Features and Stepwise Multiple Linear Regression , 1999 .

[30]  A. Gitelson,et al.  Non-Destructive Assessment of Chlorophyll Carotenoid and Anthocyanin Content in Higher Plant Leaves: Principles and Algorithms , 2004 .

[31]  Onisimo Mutanga,et al.  High density biomass estimation for wetland vegetation using WorldView-2 imagery and random forest regression algorithm , 2012, Int. J. Appl. Earth Obs. Geoinformation.

[32]  M. Kirschbaum.  Forest growth and species distribution in a changing climate. , 2000, Tree physiology.

[33]  I. Sobhan,et al.  Species discrimination from a hyperspectral perspective , 2007 .

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

[35]  Moses Azong Cho,et al.  Assessing the utility WorldView-2 imagery for tree species mapping in South African subtropical humid forest and the conservation implications: Dukuduku forest patch as case study , 2015, Int. J. Appl. Earth Obs. Geoinformation.

[36]  Hiroyuki Oguma,et al.  Seasonal changes in the relationship between photochemical reflectance index and photosynthetic light use efficiency of Japanese larch needles , 2006 .

[37]  W. Barthlott,et al.  Patterns of vascular plant diversity at continental to global scales , 2007 .

[38]  Anne-Laure Boulesteix,et al.  Microarray-based classification and clinical predictors: on combined classifiers and additional predictive value , 2008, Bioinform..

[39]  A. Gitelson,et al.  Three‐band model for noninvasive estimation of chlorophyll, carotenoids, and anthocyanin contents in higher plant leaves , 2006 .

[40]  P. Gong,et al.  Comparison of IKONOS and QuickBird images for mapping mangrove species on the Caribbean coast of Panama , 2004 .

[41]  P. Curran.  Remote sensing of foliar chemistry , 1989 .

[42]  Sushma Panigrahy,et al.  Hyperspectral leaf signature as an added dimension for species discrimination: case study of four tropical mangroves , 2011, Wetlands Ecology and Management.

[43]  R. Congalton,et al.  Accuracy assessment: a user's perspective , 1986 .

[44]  Dar A. Roberts,et al.  Species-Level Differences in Hyperspectral Metrics among Tropical Rainforest Trees as Determined by a Tree-Based Classifier , 2012, Remote. Sens..

[45]  J. Peñuelas,et al.  The Role of Plants in the Effects of Global Change on Nutrient Availability and Stoichiometry in the Plant-Soil System1[W] , 2012, Plant Physiology.

[46]  D. Roy,et al.  The availability of cloud-free Landsat ETM+ data over the conterminous United States and globally , 2008 .

[47]  C. Elvidge.  Visible and near infrared reflectance characteristics of dry plant materials , 1990 .

[48]  Mary Ann Fajvan,et al.  A Comparison of Multispectral and Multitemporal Information in High Spatial Resolution Imagery for Classification of Individual Tree Species in a Temperate Hardwood Forest , 2001 .

[49]  Karl Pearson F.R.S..  LIII. On lines and planes of closest fit to systems of points in space , 1901 .

[50]  R. Pu,et al.  A comparative analysis of high spatial resolution IKONOS and WorldView-2 imagery for mapping urban tree species , 2012 .

[51]  Frank Veroustraete,et al.  Seasonal variations in leaf area index, leaf chlorophyll, and water content; scaling-up to estimate fAPAR and carbon balance in a multilayer, multispecies temperate forest. , 1999, Tree physiology.

[52]  Pedro Larrañaga,et al.  A review of feature selection techniques in bioinformatics , 2007, Bioinform..

[53]  S. Wold,et al.  PLS-regression: a basic tool of chemometrics , 2001 .

[54]  Onisimo Mutanga,et al.  Capability of models to predict leaf N and P across four seasons for six sub-tropical forest evergreen trees , 2015 .

[55]  Alexandre Carleer,et al.  Exploitation of Very High Resolution Satellite Data for Tree Species Identification , 2004 .

[56]  Kathy Steppe,et al.  Seasonal leaf dynamics for tropical evergreen forests in a process-based global ecosystem model , 2012 .

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

[58]  G. D. Fuller.  Vegetation of South Africa , 1917, Botanical Gazette.

[59]  M. Cho,et al.  Classification of savanna tree species, in the Greater Kruger National Park region, by integrating hyperspectral and LiDAR data in a Random Forest data mining environment , 2012 .

[60]  W. Oechel,et al.  Seasonal patterns of reflectance indices, carotenoid pigments and photosynthesis of evergreen chaparral species , 2002, Oecologia.

[61]  Emilie B. Grossmann,et al.  Mapping ecological systems with a random forest model: tradeoffs between errors and bias , 2010 .

[62]  J. Knight.  Water Resources in South Africa , 2018, World Regional Geography Book Series.

[63]  Onisimo Mutanga,et al.  Exploiting machine learning algorithms for tree species classification in a semiarid woodland using RapidEye image , 2013 .

[64]  Ş. Procheş.  Pooley's Trees of Eastern South Africa — A Complete Guide, Richard Boon, 2010, Flora & Fauna Publications Trust. Price: R295.00, PVC Cover, ISBN: 978-0-620-46019-4, E-mail: woodpamela@telkomsa.net, Website: www.briza.co.za , 2010 .

[65]  S. Adelabu,et al.  Employing ground and satellite-based QuickBird data and random forest to discriminate five tree species in a Southern African Woodland , 2015 .

[66]  Piermaria Corona,et al.  VALUTAZIONE DELLE RISORSE FORESTALI A LIVELLO GLOBALE , 2013 .