Spectral Discrimination of Insect Defoliation Levels in Mopane Woodland Using Hyperspectral Data

Mopane woodland are a source of valuable resources that contribute substantially to rural economies and nutrition across Southern Africa. However, a number of factors have, of late, brought the sustainability of the mopane woodland resources into question. One of such factors is the difficulty in monitoring of defoliation process within the woodland. In this study we set out to discriminate the levels of change in forest canopy cover detectable after insect defoliation using ground based hyperspectral measurements in mopane woodland. Canopy spectral measurements were taken from three levels of defoliation: Undefoliated (UD), Partly defoliated (PD) and Refoliating plants (R) using ASD FieldSpec HandHeld 2. A pre-filtering approach (ANOVA) was compared with random forest independent variable selector in selecting the significant wavelengths for classification. Furthermore, a backward feature elimination method was used to select optimal wavelengths for discriminating the different levels of defoliation in mopane woodland. Results show that optimal wavelengths located at 707 nm, 710 nm, 711 nm, 712 nm, 713 nm, 714 nm, 727 nm, and 1066 nm were able to discriminate between the three levels of defoliation. The results further show that there was no significant difference in the overall accuracy of classification when random forest variable selector was used 82.42% (Kappa = 0.64) and the pre-filtering approach (ANOVA) 81.21% (Kappa = 0.68) used before building the classification. Overall, the study clearly demonstrated that the dynamic process of defoliation in mopane woodland can be assessed and detected using hyperspectral dataset and effective algorithm for discrimination.

[1]  Ned Horning,et al.  Random Forests : An algorithm for image classification and generation of continuous fields data sets , 2010 .

[2]  W. Cohen,et al.  Comparison of Methods for Detecting Conifer Forest Change with Thematic Mapper Imagery , 2007 .

[3]  Clement Atzberger,et al.  Vegetation Structure Retrieval in Beech and Spruce Forests Using Spectrodirectional Satellite Data , 2012, IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing.

[4]  Nigel P. Fox,et al.  Progress in Field Spectroscopy , 2006, 2006 IEEE International Symposium on Geoscience and Remote Sensing.

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

[6]  Philip A. Townsend,et al.  Estimating the effect of gypsy moth defoliation using MODIS , 2008 .

[7]  Akin Ozçift,et al.  Random forests ensemble classifier trained with data resampling strategy to improve cardiac arrhythmia diagnosis. , 2011, Computers in biology and medicine.

[8]  M. Cho,et al.  A new technique for extracting the red edge position from hyperspectral data: The linear extrapolation method , 2006 .

[9]  G. Carter.  Ratios of leaf reflectances in narrow wavebands as indicators of plant stress , 1994 .

[10]  O. Mutanga,et al.  Spectral discrimination of papyrus vegetation (Cyperus papyrus L.) in swamp wetlands using field spectrometry , 2009 .

[11]  J. Skinner,et al.  The influence of large mammalian herbivores on growth form and utilization of mopane trees, Colophospermum mopane, in Botswana's Northern Tuli Game Reserve. , 2000 .

[12]  A. Skidmore,et al.  Predicting in situ pasture quality in the Kruger National Park, South Africa, using continuum-removed absorption features , 2004 .

[13]  K. Ranson,et al.  Evidence of Evergreen Conifer Invasion into Larch Dominated Forests During Recent Decades in Central Siberia , 2007 .

[14]  Onisimo Mutanga,et al.  A Review of Remote Sensing of Insect Defoliation and its Implications for the Detection and Mapping of Imbrasia belina Defoliation of Mopane Woodland , 2012 .

[15]  O. Mutanga,et al.  Discriminating the papyrus vegetation (Cyperus papyrus L.) and its co-existent species using random forest and hyperspectral data resampled to HYMAP , 2012 .

[16]  Rick L. Lawrence,et al.  Mapping invasive plants using hyperspectral imagery and Breiman Cutler classifications (RandomForest) , 2006 .

[17]  K. Olaf Niemann,et al.  Addressing the Effects of Canopy Structure on the Remote Sensing of Foliar Chemistry of a 3-Dimensional, Radiometrically Porous Surface , 2012, IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing.

[18]  G. Carter,et al.  Leaf optical properties in higher plants: linking spectral characteristics to stress and chlorophyll concentration. , 2001, American journal of botany.

[19]  A. Skidmore,et al.  A hyperspectral band selector for plant species discrimination , 2007 .

[20]  G. Carter,et al.  Early detection of plant stress by digital imaging within narrow stress-sensitive wavebands , 1994 .

[21]  W. Mojeremane,et al.  Seed Treatments for Enhancing Germination of Colophospermum mopane Seeds: A Multipurpose Tree in Botswana , 2005 .

[22]  William M. Ciesla,et al.  European woodwasp: a potential threat to North America's conifer forests. , 2003 .

[23]  Onisimo Mutanga,et al.  Discriminating the early stages of Sirex noctilio infestation using classification tree ensembles and shortwave infrared bands , 2011 .

[24]  D. Wessels,et al.  Use of Mopane Woodland Resources and Associated Woodland Management Challenges in Rural Areas of South Africa , 2012 .

[25]  N. Coops,et al.  Assessment and monitoring of damage from insects in Australian eucalypt forests and commercial plantations , 2004 .

[26]  Jonathan Cheung-Wai Chan,et al.  Multiple Criteria for Evaluating Machine Learning Algorithms for Land Cover Classification from Satellite Data , 2000 .

[27]  C. Stone,et al.  Crown‐scale evaluation of spectral indices for defoliated and discoloured eucalypts , 2008 .

[28]  Johannes R. Sveinsson,et al.  Random Forest classification of multisource remote sensing and geographic data , 2004, IGARSS 2004. 2004 IEEE International Geoscience and Remote Sensing Symposium.

[29]  O. Mutanga,et al.  Imaging spectroscopy (hyperspectral remote sensing) in southern Africa: an overview , 2010 .

[30]  Christine Stone,et al.  Assessment of crown condition in eucalypt vegetation by remotely sensed optical indices. , 2004, Journal of environmental quality.

[31]  Robert H. Fraser,et al.  Mapping insect‐induced tree defoliation and mortality using coarse spatial resolution satellite imagery , 2005 .

[32]  Onisimo Mutanga,et al.  DETERMINING THE OPTIMAL SPATIAL RESOLUTION OF REMOTELY SENSED DATA FOR THE DETECTION OF SIREX NOCTILIO INFESTATIONS IN PINE PLANTATIONS IN KWAZULU-NATAL, SOUTH AFRICA , 2008 .

[33]  J. Eitel,et al.  Quantifying aboveground forest carbon pools and fluxes from repeat LiDAR surveys , 2012 .

[34]  C van der Waal,et al.  Induced chemical defences in Colophospermum mopane trees , 2007 .

[35]  Bjoern H. Menze,et al.  A comparison of random forest and its Gini importance with standard chemometric methods for the feature selection and classification of spectral data , 2009, BMC Bioinformatics.

[36]  B. Lundén,et al.  MAPPING OF COLOPHOSPERMUM MOPANE USING LANDSAT TM IN EASTERN BOTSWANA , 2008 .