WETLAND MAPPING IN NEW BRUNSWICK, CANADA WITH LANDSAT5-TM, ALOS-PALSAR, AND RADARSAT-2 IMAGERY

Abstract. Several maps of wetland areas in central New Brunswick, Canada, were produced by applying the Random Forests classifier to different combinations of optical Landsat-5 TM images, dual-polarized (HH, HV) Radarsat-2 C-band and Alos-1 PalSAR L-band Synthetic Aperture Radar (SAR) images and digital elevation data. The resulting maps were compared to 199 GPS wetland sites that were visited between 2012 and 2018 as well as to a combination of two wetland maps currently used by the Province of New Brunswick. The number of correctly identified GPS wetland sites was the highest when both the Alos-PalSAR and Radarsat-2 images are used (97.9%). This percentage of correctly identified sites were well above the accuracy of the official New Brunswick wetland maps (44.7 %). With the best-classified image, the misidentifications were due to wetlands not being classified in the right wetland class, and just one case was a wetland site being classified in a non-wetland class. For the NB wetland map, about a quarter of the wetland validation sites were classified in a non-wetland class, and about the same number of sites were classified in the wrong wetland class.

[1]  Jacob Cohen,et al.  Weighted kappa: Nominal scale agreement provision for scaled disagreement or partial credit. , 1968 .

[2]  Danny Lo Seen,et al.  Mapping wetlands in Nova Scotia with multi-beam RADARSAT-2 Polarimetric SAR, optical satellite imagery, and Lidar data , 2018, Int. J. Appl. Earth Obs. Geoinformation.

[3]  B. Brisco,et al.  Evaluation of C-band polarization diversity and polarimetry for wetland mapping , 2011 .

[4]  Mahesh Pal,et al.  Random forest classifier for remote sensing classification , 2005 .

[5]  R. Tiner,et al.  Wetland Indicators: A Guide to Wetland Identification, Delineation, Classification, and Mapping , 1999 .

[6]  Brian Brisco,et al.  Wetland Classification Using Multi-Source and Multi-Temporal Optical Remote Sensing Data in Newfoundland and Labrador, Canada , 2017 .

[7]  Robert G. Bryant,et al.  Mapping the effects of water stress on Sphagnum: Preliminary observations using airborne remote sensing , 2006 .

[8]  Eric S. Kasischke,et al.  Mapping boreal peatland ecosystem types from multitemporal radar and optical satellite imagery , 2017 .

[9]  Ridha Touzi,et al.  Wetland Characterization using Polarimetric RADARSAT-2 Capability , 2006, 2006 IEEE International Symposium on Geoscience and Remote Sensing.

[10]  Philip A. Townsend,et al.  Relationships between forest structure and the detection of flood inundation in forested wetlands using C-band SAR , 2002 .

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

[12]  T. J. Pultz,et al.  Case studies demonstrating the hydrological applications of C-band multipolarized and polarimetric SAR , 2004 .

[13]  Achim Zeileis,et al.  BMC Bioinformatics BioMed Central Methodology article Conditional variable importance for random forests , 2008 .

[14]  B. Markham,et al.  Summary of Current Radiometric Calibration Coefficients for Landsat MSS, TM, ETM+, and EO-1 ALI Sensors , 2009 .

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

[16]  Michael Battaglia,et al.  Development of a Bi-National Great Lakes Coastal Wetland and Land Use Map Using Three-Season PALSAR and Landsat Imagery , 2015, Remote. Sens..

[17]  Mahta Moghaddam,et al.  Mapping vegetated wetlands of Alaska using L-band radar satellite imagery , 2009 .

[18]  C. Rubec,et al.  The Canadian Wetland Classification System , 2016 .

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

[20]  Mohamed Shehata,et al.  A Multiple Classifier System to improve mapping complex land covers: a case study of wetland classification using SAR data in Newfoundland, Canada , 2018 .

[21]  Floyd M. Henderson,et al.  Radar detection of wetland ecosystems: a review , 2008 .

[22]  Joseph F. Knight,et al.  Influence of Multi-Source and Multi-Temporal Remotely Sensed and Ancillary Data on the Accuracy of Random Forest Classification of Wetlands in Northern Minnesota , 2013, Remote. Sens..

[23]  Björn Waske,et al.  Classifier ensembles for land cover mapping using multitemporal SAR imagery , 2009 .

[24]  L. Hess,et al.  Radar detection of flooding beneath the forest canopy - A review , 1990 .

[25]  Junhua Li,et al.  A rule-based method for mapping Canada's wetlands using optical, radar and DEM data , 2005 .

[26]  Y. Yamagata,et al.  Classification of wetland vegetation by texture analysis methods using ERS-1 and JERS-1 images , 1993, Proceedings of IGARSS '93 - IEEE International Geoscience and Remote Sensing Symposium.

[27]  Andreas Schmitt,et al.  SAR polarimetric change detection for flooded vegetation , 2013, Int. J. Digit. Earth.

[28]  Brigitte Leblon,et al.  Surficial materials mapping in Nunavut, Canada with multibeam RADARSAT-2 dual-polarization C-HH and C-HV, LANDSAT-7 ETM+, and DEM data , 2012 .

[29]  Gilles Louppe,et al.  Understanding variable importances in forests of randomized trees , 2013, NIPS.

[30]  Mary Ellen Miller,et al.  Spectral detection of near-surface moisture content and water-table position in northern peatland ecosystems , 2014 .

[31]  Brian Brisco,et al.  The integration of optical, topographic, and radar data for wetland mapping in northern Minnesota , 2011 .

[32]  Brian Brisco,et al.  An Assessment of Simulated Compact Polarimetric SAR Data for Wetland Classification Using Random Forest Algorithm , 2017 .

[33]  Johannes R. Sveinsson,et al.  Random Forests for land cover classification , 2006, Pattern Recognit. Lett..

[34]  Laura L. Bourgeau-Chavez,et al.  Use of Radarsat-2 and ALOS-PALSAR SAR images for wetland mapping in New Brunswick , 2014, 2014 IEEE Geoscience and Remote Sensing Symposium.

[35]  Laura L. Bourgeau-Chavez,et al.  Improving Wetland Characterization with Multi-Sensor, Multi-Temporal SAR and Optical/Infrared Data Fusion , 2009 .