Mapping wetlands in Nova Scotia with multi-beam RADARSAT-2 Polarimetric SAR, optical satellite imagery, and Lidar data

Abstract Wetland maps currently in use by the Province of Nova Scotia, namely the Department of Natural Resources (DNR) wetland inventory map and the swamp wetland classes of the DNR forest map, need to be updated. In this study, wetlands were mapped in an area southwest of Halifax, Nova Scotia by classifying a combination of multi-date and multi-beam RADARSAT-2 C-band polarimetric SAR (polSAR) images with spring Lidar, and fall QuickBird optical data using the Random Forests (RF) classifier. The resulting map has five wetland classes (open-water/marsh complex, open bog, open fen, shrub/treed fen/bog, swamp), plus lakes and various upland classes. Its accuracy was assessed using data from 156 GPS wetland sites collected in 2012 and compared to the one obtained with the current wetland map of Nova Scotia. The best overall classification was obtained using a combination of Lidar, RADARSAT-2 HH, HV, VH, VV intensity with polarimetric variables, and QuickBird multispectral (89.2%). The classified image was compared to GPS validation sites to assess the mapping accuracy of the wetlands. It was first done considering a group consisting of all wetland classes including lakes. This showed that only 69.9% of the wetland sites were correctly identified when only the QuickBird classified image was used in the classification. With the addition of variables derived from lidar, the number of correctly identified wetlands increased to 88.5%. The accuracy remained the same with the addition of RADARSAT-2 (88.5%). When we tested the accuracy for identifying wetland classes (e.g. marsh complex vs. open bog) instead of grouped wetlands, the resulting wetland map performed best with either QuickBird and Lidar, or QuickBird, Lidar, and RADARSAT-2 (66%). The Province of Nova Scotia’s current wetland inventory and its associated wetland classes (aerial-photo interpreted) were also assessed against the GPS wetland sites. This provincial inventory correctly identified 62.2% of the grouped wetlands and only 18.6% of the wetland classes. The current inventory’s poor performance demonstrates the value of incorporating a combination of new data sources into the provincial wetland mapping.

[1]  P. Gibbard,et al.  Quaternary glaciations : extent and chronology , 2004 .

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

[3]  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..

[4]  C. Rubec,et al.  Wetlands of Canada. , 1988 .

[5]  John P. Wilson,et al.  Terrain analysis : principles and applications , 2000 .

[6]  Weimin Huang,et al.  Object-Based Classification of Wetlands in Newfoundland and Labrador Using Multi-Temporal PolSAR Data , 2017 .

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

[8]  Eric S. Kasischke,et al.  Using C-Band Synthetic Aperture Radar Data to Monitor Forested Wetland Hydrology in Maryland's Coastal Plain, USA , 2008, IEEE Transactions on Geoscience and Remote Sensing.

[9]  R. King.  Land cover mapping principles: A return to interpretation fundamentals , 2002 .

[10]  Ridha Touzi,et al.  Phase of Target Scattering for Wetland Characterization Using Polarimetric C-Band SAR , 2009, IEEE Transactions on Geoscience and Remote Sensing.

[11]  M. Brinson.  Changes in the functioning of wetlands along environmental gradients , 1993, Wetlands.

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

[13]  Russell G. Congalton,et al.  Assessing the accuracy of remotely sensed data : principles and practices , 1998 .

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

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

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

[17]  Thuy Le Toan,et al.  Polarimetric discriminators for SAR images , 1992, IEEE Trans. Geosci. Remote. Sens..

[18]  N. Poff,et al.  A hydrogeography of unregulated streams in the United States and an examination of scale‐dependence in some hydrological descriptors , 1996 .

[19]  Elijah W. Ramsey,et al.  Using multiple-polarization L-band radar to monitor marsh burn recovery , 1999, IEEE Trans. Geosci. Remote. Sens..

[20]  Robert Leconte,et al.  A review of Canadian remote sensing and hydrology, 1999–2003 , 2005 .

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

[22]  E. C. Pielou,et al.  After the Ice Age : the return of life to glaciated North America , 1991 .

[23]  Koreen Millard,et al.  On the Importance of Training Data Sample Selection in Random Forest Image Classification: A Case Study in Peatland Ecosystem Mapping , 2015, Remote. Sens..

[24]  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.

[25]  Lisa-Maria Rebelo,et al.  Eco-Hydrological Characterization of Inland Wetlands in Africa Using L-Band SAR , 2010, IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing.

[26]  Waldir Renato Paradella,et al.  Use of RADARSAT-1 fine mode and Landsat-5 TM selective principal component analysis for geomorphological mapping in a macrotidal mangrove coast in the Amazon Region , 2005 .

[27]  Laurent Ferro-Famil,et al.  Unsupervised terrain classification preserving polarimetric scattering characteristics , 2004, IEEE Transactions on Geoscience and Remote Sensing.

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

[29]  Ridha Touzi,et al.  Target Scattering Decomposition in Terms of Roll-Invariant Target Parameters , 2007, IEEE Transactions on Geoscience and Remote Sensing.

[30]  Ridha Touzi,et al.  Wetland characterization using polarimetric RADARSAT-2 capability , 2007 .

[31]  Alexis J. Comber,et al.  Random forest classification of salt marsh vegetation habitats using quad-polarimetric airborne SAR, elevation and optical RS data , 2014 .

[32]  A. LaRocque,et al.  Modelling and mapping permafrost at high spatial resolution using Landsat and Radarsat-2 images in Northern Ontario, Canada: Part 2 – regional mapping , 2016 .

[33]  Laura L. Bourgeau-Chavez,et al.  Remotely Monitoring Great Lakes Coastal Wetlands with Multi-Sensor, Multi-Temporal SAR and Multi-Spectral Data , 2008, IGARSS 2008 - 2008 IEEE International Geoscience and Remote Sensing Symposium.

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

[35]  Haydee Karszenbaum,et al.  Influence of Flood Conditions and Vegetation Status on the Radar Backscatter of Wetland Ecosystems , 2001 .

[36]  Oguz Gungor,et al.  Evaluation of random forest method for agricultural crop classification , 2012 .

[37]  Richard A. Fournier,et al.  An object-based method to map wetland using RADARSAT-1 and Landsat ETM images: test case on two sites in Quebec, Canada , 2007 .

[38]  Brian Brisco,et al.  Wetland classification in Newfoundland and Labrador using multi-source SAR and optical data integration , 2017 .

[39]  Achim Zeileis,et al.  Conditional variable importance for random forests , 2008, BMC Bioinformatics.

[40]  T. M. Lillesand,et al.  Remote Sensing and Image Interpretation , 1980 .

[41]  Ron Kwok,et al.  Classification of multi-look polarimetric SAR imagery based on complex Wishart distribution , 1994 .

[42]  John R. Schott,et al.  Remote Sensing: The Image Chain Approach , 1996 .

[43]  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..

[44]  Eric Pottier,et al.  An entropy based classification scheme for land applications of polarimetric SAR , 1997, IEEE Trans. Geosci. Remote. Sens..

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

[46]  Eric S. Kasischke,et al.  Influence of incidence angle on detecting flooded forests using C-HH synthetic aperture radar data , 2008 .

[47]  H. Zebker,et al.  Imaging radar polarization signatures: Theory and observation , 1987 .

[48]  Robert G. Bryant,et al.  Detecting near-surface moisture stress in Sphagnum spp. , 2005 .

[49]  Stacy L. Ozesmi,et al.  Satellite remote sensing of wetlands , 2002, Wetlands Ecology and Management.

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

[51]  S. Sader,et al.  Accuracy of landsat-TM and GIS rule-based methods for forest wetland classification in Maine , 1995 .

[52]  Brian Brisco,et al.  Mapping and Monitoring Surface Water and Wetlands with Synthetic Aperture Radar , 2015 .

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

[54]  Stephen L. Durden,et al.  A three-component scattering model for polarimetric SAR data , 1998, IEEE Trans. Geosci. Remote. Sens..

[55]  H. Zebker,et al.  Imaging radar polarimetry from wave synthesis , 1986 .

[56]  R. J. Brown,et al.  Evaluation of Multidate ERS-1 and Multispectral Landsat Imagery for Wetland Detection in Southern Ontario , 1998 .

[57]  Brian Brisco,et al.  Water resource applications with RADARSAT-2 – a preview , 2008, Int. J. Digit. Earth.

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

[59]  J. Henry,et al.  Envisat multi‐polarized ASAR data for flood mapping , 2006 .

[60]  K. Todd,et al.  Automated discrimination of upland and wetland using terrain derivatives , 2007 .

[61]  Nova Scotia,et al.  ECOLOGICAL LAND CLASSIFICATION for NOVA SCOTIA , 2005 .

[62]  Accuracy assessment of digital elevation model using stochastic simulation , 2006 .

[63]  J. Goodman.  Some fundamental properties of speckle , 1976 .

[64]  Richard A. Fournier,et al.  Towards a strategy to implement the Canadian Wetland Inventory using satellite remote sensing , 2007 .

[65]  R. Kettig,et al.  Classification of Multispectral Image Data by Extraction and Classification of Homogeneous Objects , 1976, IEEE Transactions on Geoscience Electronics.

[66]  E. Rodríguez,et al.  Theory and design of interferometric synthetic aperture radars , 1992 .

[67]  Eric S. Kasischke,et al.  Monitoring South Florida Wetlands Using ERS-1 SAR Imagery , 1997 .

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

[69]  Robert Woodruff,et al.  Detecting seasonal flooding cycles in marshes of the Yucatan Peninsula with SIR-C polarimetric radar imagery , 1997 .

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

[71]  Carlos López-Martínez,et al.  Statistical Assessment of Eigenvector-Based Target Decomposition Theorems in Radar Polarimetry , 2005, IEEE Trans. Geosci. Remote. Sens..

[72]  Brian Brisco,et al.  RADARSAT-2 Beam Mode Selection for Surface Water and Flooded Vegetation Mapping , 2014 .

[73]  I. Moore,et al.  Digital terrain modelling: A review of hydrological, geomorphological, and biological applications , 1991 .

[74]  Andrew Baird,et al.  Detecting water stress in Sphagnum sp. , 2005 .

[75]  B. R. M. Rao,et al.  Monitoring the spatial extent of coastal wetlands using ERS-1 SAR data , 1999 .

[76]  Thomas L. Ainsworth,et al.  Unsupervised classification using polarimetric decomposition and the complex Wishart classifier , 1999, IEEE Trans. Geosci. Remote. Sens..

[77]  K. Millard,et al.  Wetland mapping with LiDAR derivatives, SAR polarimetric decompositions, and LiDAR–SAR fusion using a random forest classifier , 2013 .

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

[79]  Björn Waske,et al.  RANDOM FORESTS FOR CLASSIFYING MULTI-TEMPORAL SAR DATA , 2007 .