Mapping of Shorea robusta Forest Using Time Series MODIS Data

Bhoj Ghimire, Masahiko Nagai, Nitin Tripathi, Apichon Witayangkurn, Bhogendra Mishra, Nophea Sasaki

Research output: Contribution to journalArticleAcademicpeer-review

3 Citations (Scopus)
30 Downloads (Pure)

Abstract

Mapping forest types in a natural heterogeneous forest environment using remote sensing data is a long-standing challenge due to similar spectral reflectance from different tree species and significant time and resources are required for acquiring and processing the remote sensing data. The purpose of this research was to determine the optimum number of remote sensing images and map the Sal forest through the analysis of Vegetation Index (VI) signatures. We analyzed the eight days’ composite moderate resolution imaging spectroradiometer (MODIS) time series normalized differential vegetation index (NDVI), and enhanced vegetation index (EVI) for the whole year of 2015. Jeffries-Matusita (J-M) distance was used for the separability index. Performance of EVI and NDVI was tested using random forest (RF) and support vector machine (SVM) classifiers. Boruta algorithm and statistical analysis were performed to identify the optimum set of imageries. We also performed data level five-fold cross validation of the model and field level accuracy assessment of the classification map. The finding confirmed that EVI with SVM (F-score of Sal 0.88) performed better than NDVI with either SVM or RF. The optimum 12 images during growing and post monsoon season significantly decreased processing time (to one-fourth) without much deteriorating accuracy. Accordingly, we were able to map the Sal forest whose area is accounted for about 36% of the 82% forest cover in the study area. The proposed methodology can be extended to produce a temporal forest type classification map in any other location.
Original languageEnglish
Article number384
JournalForests
Volume8
Issue number10
DOIs
Publication statusPublished - 2017

Fingerprint

Shorea robusta
moderate resolution imaging spectroradiometer
vegetation index
MODIS
time series analysis
time series
remote sensing
forest types
taxonomy
monsoon season
accuracy assessment
spectral reflectance
model validation
forest cover
reflectance
statistical analysis
monsoon
imagery
fold
methodology

Keywords

  • ITC-ISI-JOURNAL-ARTICLE
  • ITC-GOLD

Cite this

Ghimire, B., Nagai, M., Tripathi, N., Witayangkurn, A., Mishra, B., & Sasaki, N. (2017). Mapping of Shorea robusta Forest Using Time Series MODIS Data. Forests, 8(10), [384]. https://doi.org/10.3390/f8100384
Ghimire, Bhoj ; Nagai, Masahiko ; Tripathi, Nitin ; Witayangkurn, Apichon ; Mishra, Bhogendra ; Sasaki, Nophea. / Mapping of Shorea robusta Forest Using Time Series MODIS Data. In: Forests. 2017 ; Vol. 8, No. 10.
@article{be389d552bc244c1abdaf65096633c80,
title = "Mapping of Shorea robusta Forest Using Time Series MODIS Data",
abstract = "Mapping forest types in a natural heterogeneous forest environment using remote sensing data is a long-standing challenge due to similar spectral reflectance from different tree species and significant time and resources are required for acquiring and processing the remote sensing data. The purpose of this research was to determine the optimum number of remote sensing images and map the Sal forest through the analysis of Vegetation Index (VI) signatures. We analyzed the eight days’ composite moderate resolution imaging spectroradiometer (MODIS) time series normalized differential vegetation index (NDVI), and enhanced vegetation index (EVI) for the whole year of 2015. Jeffries-Matusita (J-M) distance was used for the separability index. Performance of EVI and NDVI was tested using random forest (RF) and support vector machine (SVM) classifiers. Boruta algorithm and statistical analysis were performed to identify the optimum set of imageries. We also performed data level five-fold cross validation of the model and field level accuracy assessment of the classification map. The finding confirmed that EVI with SVM (F-score of Sal 0.88) performed better than NDVI with either SVM or RF. The optimum 12 images during growing and post monsoon season significantly decreased processing time (to one-fourth) without much deteriorating accuracy. Accordingly, we were able to map the Sal forest whose area is accounted for about 36{\%} of the 82{\%} forest cover in the study area. The proposed methodology can be extended to produce a temporal forest type classification map in any other location.",
keywords = "ITC-ISI-JOURNAL-ARTICLE, ITC-GOLD",
author = "Bhoj Ghimire and Masahiko Nagai and Nitin Tripathi and Apichon Witayangkurn and Bhogendra Mishra and Nophea Sasaki",
year = "2017",
doi = "10.3390/f8100384",
language = "English",
volume = "8",
journal = "Forests",
issn = "1999-4907",
publisher = "Multidisciplinary Digital Publishing Institute",
number = "10",

}

Ghimire, B, Nagai, M, Tripathi, N, Witayangkurn, A, Mishra, B & Sasaki, N 2017, 'Mapping of Shorea robusta Forest Using Time Series MODIS Data' Forests, vol. 8, no. 10, 384. https://doi.org/10.3390/f8100384

Mapping of Shorea robusta Forest Using Time Series MODIS Data. / Ghimire, Bhoj; Nagai, Masahiko; Tripathi, Nitin; Witayangkurn, Apichon; Mishra, Bhogendra ; Sasaki, Nophea.

In: Forests, Vol. 8, No. 10, 384, 2017.

Research output: Contribution to journalArticleAcademicpeer-review

TY - JOUR

T1 - Mapping of Shorea robusta Forest Using Time Series MODIS Data

AU - Ghimire, Bhoj

AU - Nagai, Masahiko

AU - Tripathi, Nitin

AU - Witayangkurn, Apichon

AU - Mishra, Bhogendra

AU - Sasaki, Nophea

PY - 2017

Y1 - 2017

N2 - Mapping forest types in a natural heterogeneous forest environment using remote sensing data is a long-standing challenge due to similar spectral reflectance from different tree species and significant time and resources are required for acquiring and processing the remote sensing data. The purpose of this research was to determine the optimum number of remote sensing images and map the Sal forest through the analysis of Vegetation Index (VI) signatures. We analyzed the eight days’ composite moderate resolution imaging spectroradiometer (MODIS) time series normalized differential vegetation index (NDVI), and enhanced vegetation index (EVI) for the whole year of 2015. Jeffries-Matusita (J-M) distance was used for the separability index. Performance of EVI and NDVI was tested using random forest (RF) and support vector machine (SVM) classifiers. Boruta algorithm and statistical analysis were performed to identify the optimum set of imageries. We also performed data level five-fold cross validation of the model and field level accuracy assessment of the classification map. The finding confirmed that EVI with SVM (F-score of Sal 0.88) performed better than NDVI with either SVM or RF. The optimum 12 images during growing and post monsoon season significantly decreased processing time (to one-fourth) without much deteriorating accuracy. Accordingly, we were able to map the Sal forest whose area is accounted for about 36% of the 82% forest cover in the study area. The proposed methodology can be extended to produce a temporal forest type classification map in any other location.

AB - Mapping forest types in a natural heterogeneous forest environment using remote sensing data is a long-standing challenge due to similar spectral reflectance from different tree species and significant time and resources are required for acquiring and processing the remote sensing data. The purpose of this research was to determine the optimum number of remote sensing images and map the Sal forest through the analysis of Vegetation Index (VI) signatures. We analyzed the eight days’ composite moderate resolution imaging spectroradiometer (MODIS) time series normalized differential vegetation index (NDVI), and enhanced vegetation index (EVI) for the whole year of 2015. Jeffries-Matusita (J-M) distance was used for the separability index. Performance of EVI and NDVI was tested using random forest (RF) and support vector machine (SVM) classifiers. Boruta algorithm and statistical analysis were performed to identify the optimum set of imageries. We also performed data level five-fold cross validation of the model and field level accuracy assessment of the classification map. The finding confirmed that EVI with SVM (F-score of Sal 0.88) performed better than NDVI with either SVM or RF. The optimum 12 images during growing and post monsoon season significantly decreased processing time (to one-fourth) without much deteriorating accuracy. Accordingly, we were able to map the Sal forest whose area is accounted for about 36% of the 82% forest cover in the study area. The proposed methodology can be extended to produce a temporal forest type classification map in any other location.

KW - ITC-ISI-JOURNAL-ARTICLE

KW - ITC-GOLD

UR - https://ezproxy2.utwente.nl/login?url=https://webapps.itc.utwente.nl/library/2017/isi/mishra_map.pdf

U2 - 10.3390/f8100384

DO - 10.3390/f8100384

M3 - Article

VL - 8

JO - Forests

JF - Forests

SN - 1999-4907

IS - 10

M1 - 384

ER -

Ghimire B, Nagai M, Tripathi N, Witayangkurn A, Mishra B, Sasaki N. Mapping of Shorea robusta Forest Using Time Series MODIS Data. Forests. 2017;8(10). 384. https://doi.org/10.3390/f8100384