Evaluation of the Landsat-Based Canadian Wetland Inventory Map Using Multiple Sources: Challenges of Large-Scale Wetland Classification Using Remote Sensing

Research output: Contribution to journalArticleResearchpeer review

Authors

  • Meisam Amani
  • Brian Brisco
  • Sahel Mahdavi
  • Arsalan Ghorbanian
  • Armin Moghimi
  • Evan R. Delancey
  • Michael Merchant
  • Raymond Jahncke
  • Lee Fedorchuk
  • Amy Mui
  • Thierry Fisette
  • Mohammad Kakooei
  • Seyed Ali Ahmadi
  • Brigitte Leblon
  • Armand Larocque

External Research Organisations

  • Wood Environment & Infrastructure Solutions
  • Canada Center for Mapping and Earth Observation (CCMEO)
  • K.N. Toosi University of Technology
  • University of Alberta
  • Ducks Unlimited Canada
  • Nova Scotia Department of Lands and Forestry
  • Manitoba Forestry Branch
  • Dalhousie University
  • AgriFood Canada
  • Babol Noshirvani University of Technology
  • University of New Brunswick
View graph of relations

Details

Original languageEnglish
Article number9254004
Pages (from-to)32-52
Number of pages21
JournalIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Volume14
Publication statusPublished - 2021
Externally publishedYes

Abstract

The first Canadian wetland inventory (CWI) map, which was based on Landsat data, was produced in 2019 using the Google Earth Engine (GEE) big data processing platform. The proposed GEE-based method to create the preliminary CWI map proved to be a cost, time, and computationally efficient approach. Although the initial effort to produce the CWI map was valuable with a 71% overall accuracy (OA), there were several inevitable limitations (e.g., low-quality samples for the training and validation of the map). Therefore, it was important to comprehensively investigate those limitations and develop effective solutions to improve the accuracy of the Landsat-based CWI (L-CWI) map. Over the past year, the L-CWI map was shared with several governmental, academic, environmental nonprofit, and industrial organizations. Subsequently, valuable feedback was received on the accuracy of this product by comparing it with various in situ data, photo-interpreted reference samples, land cover/land use maps, and high-resolution aerial images. It was generally observed that the accuracy of the L-CWI map was lower relative to the other available products. For example, the average OA in four Canadian provinces using in situ data was 60%. Moreover, including reliable in situ data, using an object-based classification method, and adding more optical and synthetic aperture radar datasets were identified as the main practical solutions to improve the CWI map in the future. Finally, limitations and solutions discussed in this study are applicable to any large-scale wetland mapping using remote sensing methods, especially to CWI generation using optical satellite data in GEE.

Keywords

    Big data, Canada, Google Earth Engine, Landsat, remote sensing (RS), wetlands

ASJC Scopus subject areas

Sustainable Development Goals

Cite this

Evaluation of the Landsat-Based Canadian Wetland Inventory Map Using Multiple Sources: Challenges of Large-Scale Wetland Classification Using Remote Sensing. / Amani, Meisam; Brisco, Brian; Mahdavi, Sahel et al.
In: IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, Vol. 14, 9254004, 2021, p. 32-52.

Research output: Contribution to journalArticleResearchpeer review

Amani, M, Brisco, B, Mahdavi, S, Ghorbanian, A, Moghimi, A, Delancey, ER, Merchant, M, Jahncke, R, Fedorchuk, L, Mui, A, Fisette, T, Kakooei, M, Ahmadi, SA, Leblon, B & Larocque, A 2021, 'Evaluation of the Landsat-Based Canadian Wetland Inventory Map Using Multiple Sources: Challenges of Large-Scale Wetland Classification Using Remote Sensing', IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, vol. 14, 9254004, pp. 32-52. https://doi.org/10.1109/jstars.2020.3036802
Amani, M., Brisco, B., Mahdavi, S., Ghorbanian, A., Moghimi, A., Delancey, E. R., Merchant, M., Jahncke, R., Fedorchuk, L., Mui, A., Fisette, T., Kakooei, M., Ahmadi, S. A., Leblon, B., & Larocque, A. (2021). Evaluation of the Landsat-Based Canadian Wetland Inventory Map Using Multiple Sources: Challenges of Large-Scale Wetland Classification Using Remote Sensing. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 14, 32-52. Article 9254004. https://doi.org/10.1109/jstars.2020.3036802
Download
@article{122b42350ce84452aa38d3c490594e17,
title = "Evaluation of the Landsat-Based Canadian Wetland Inventory Map Using Multiple Sources: Challenges of Large-Scale Wetland Classification Using Remote Sensing",
abstract = "The first Canadian wetland inventory (CWI) map, which was based on Landsat data, was produced in 2019 using the Google Earth Engine (GEE) big data processing platform. The proposed GEE-based method to create the preliminary CWI map proved to be a cost, time, and computationally efficient approach. Although the initial effort to produce the CWI map was valuable with a 71% overall accuracy (OA), there were several inevitable limitations (e.g., low-quality samples for the training and validation of the map). Therefore, it was important to comprehensively investigate those limitations and develop effective solutions to improve the accuracy of the Landsat-based CWI (L-CWI) map. Over the past year, the L-CWI map was shared with several governmental, academic, environmental nonprofit, and industrial organizations. Subsequently, valuable feedback was received on the accuracy of this product by comparing it with various in situ data, photo-interpreted reference samples, land cover/land use maps, and high-resolution aerial images. It was generally observed that the accuracy of the L-CWI map was lower relative to the other available products. For example, the average OA in four Canadian provinces using in situ data was 60%. Moreover, including reliable in situ data, using an object-based classification method, and adding more optical and synthetic aperture radar datasets were identified as the main practical solutions to improve the CWI map in the future. Finally, limitations and solutions discussed in this study are applicable to any large-scale wetland mapping using remote sensing methods, especially to CWI generation using optical satellite data in GEE.",
keywords = "Big data, Canada, Google Earth Engine, Landsat, remote sensing (RS), wetlands",
author = "Meisam Amani and Brian Brisco and Sahel Mahdavi and Arsalan Ghorbanian and Armin Moghimi and Delancey, {Evan R.} and Michael Merchant and Raymond Jahncke and Lee Fedorchuk and Amy Mui and Thierry Fisette and Mohammad Kakooei and Ahmadi, {Seyed Ali} and Brigitte Leblon and Armand Larocque",
note = "Funding Information: This work was supported in part by the Natural Resources Canada under a Grant to Meisam Amani and in part by the Canada Centre for Mapping and Earth Observation of Natural Resources Canada. Publisher Copyright: {\textcopyright} 2008-2012 IEEE.",
year = "2021",
doi = "10.1109/jstars.2020.3036802",
language = "English",
volume = "14",
pages = "32--52",
journal = "IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing",
issn = "1939-1404",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

Download

TY - JOUR

T1 - Evaluation of the Landsat-Based Canadian Wetland Inventory Map Using Multiple Sources

T2 - Challenges of Large-Scale Wetland Classification Using Remote Sensing

AU - Amani, Meisam

AU - Brisco, Brian

AU - Mahdavi, Sahel

AU - Ghorbanian, Arsalan

AU - Moghimi, Armin

AU - Delancey, Evan R.

AU - Merchant, Michael

AU - Jahncke, Raymond

AU - Fedorchuk, Lee

AU - Mui, Amy

AU - Fisette, Thierry

AU - Kakooei, Mohammad

AU - Ahmadi, Seyed Ali

AU - Leblon, Brigitte

AU - Larocque, Armand

N1 - Funding Information: This work was supported in part by the Natural Resources Canada under a Grant to Meisam Amani and in part by the Canada Centre for Mapping and Earth Observation of Natural Resources Canada. Publisher Copyright: © 2008-2012 IEEE.

PY - 2021

Y1 - 2021

N2 - The first Canadian wetland inventory (CWI) map, which was based on Landsat data, was produced in 2019 using the Google Earth Engine (GEE) big data processing platform. The proposed GEE-based method to create the preliminary CWI map proved to be a cost, time, and computationally efficient approach. Although the initial effort to produce the CWI map was valuable with a 71% overall accuracy (OA), there were several inevitable limitations (e.g., low-quality samples for the training and validation of the map). Therefore, it was important to comprehensively investigate those limitations and develop effective solutions to improve the accuracy of the Landsat-based CWI (L-CWI) map. Over the past year, the L-CWI map was shared with several governmental, academic, environmental nonprofit, and industrial organizations. Subsequently, valuable feedback was received on the accuracy of this product by comparing it with various in situ data, photo-interpreted reference samples, land cover/land use maps, and high-resolution aerial images. It was generally observed that the accuracy of the L-CWI map was lower relative to the other available products. For example, the average OA in four Canadian provinces using in situ data was 60%. Moreover, including reliable in situ data, using an object-based classification method, and adding more optical and synthetic aperture radar datasets were identified as the main practical solutions to improve the CWI map in the future. Finally, limitations and solutions discussed in this study are applicable to any large-scale wetland mapping using remote sensing methods, especially to CWI generation using optical satellite data in GEE.

AB - The first Canadian wetland inventory (CWI) map, which was based on Landsat data, was produced in 2019 using the Google Earth Engine (GEE) big data processing platform. The proposed GEE-based method to create the preliminary CWI map proved to be a cost, time, and computationally efficient approach. Although the initial effort to produce the CWI map was valuable with a 71% overall accuracy (OA), there were several inevitable limitations (e.g., low-quality samples for the training and validation of the map). Therefore, it was important to comprehensively investigate those limitations and develop effective solutions to improve the accuracy of the Landsat-based CWI (L-CWI) map. Over the past year, the L-CWI map was shared with several governmental, academic, environmental nonprofit, and industrial organizations. Subsequently, valuable feedback was received on the accuracy of this product by comparing it with various in situ data, photo-interpreted reference samples, land cover/land use maps, and high-resolution aerial images. It was generally observed that the accuracy of the L-CWI map was lower relative to the other available products. For example, the average OA in four Canadian provinces using in situ data was 60%. Moreover, including reliable in situ data, using an object-based classification method, and adding more optical and synthetic aperture radar datasets were identified as the main practical solutions to improve the CWI map in the future. Finally, limitations and solutions discussed in this study are applicable to any large-scale wetland mapping using remote sensing methods, especially to CWI generation using optical satellite data in GEE.

KW - Big data

KW - Canada

KW - Google Earth Engine

KW - Landsat

KW - remote sensing (RS)

KW - wetlands

UR - http://www.scopus.com/inward/record.url?scp=85098761839&partnerID=8YFLogxK

U2 - 10.1109/jstars.2020.3036802

DO - 10.1109/jstars.2020.3036802

M3 - Article

AN - SCOPUS:85098761839

VL - 14

SP - 32

EP - 52

JO - IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing

JF - IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing

SN - 1939-1404

M1 - 9254004

ER -