A satellite remote-sensing multi-index approach to discriminate pelagic Sargassum in the waters of the Yucatan Peninsula, Mexico

ABSTRACT Recently, the need for quantitative information on the spatiotemporal distribution of floating macroalgae, particularly the two species of genus Sargassum, has grown because of blooms of these species in the Gulf of Mexico and Caribbean Sea. Remote sensing is one of the most frequently used tools to assess pelagic Sargassum distribution. The purpose of this study was to implement a methodological approach to detect floating algae in an efficient and replicable manner at a moderate cost. We analyzed Landsat 8 imagery, from which we calculated four vegetation indices and one floating-algae index to implement a supervised classification, together with the bands 2 and 5, using the Random Forest algorithm. The analysis was performed monthly from 2014 to 2015 for the northeastern Yucatan Peninsula, Mexico, with a total of 91 analyzed images. The quantitative performance metrics of the classifier (overall, Kappa and Tau) were greater than 80%, whereas bands 2 and 5 as well as the atmospherically resistant vegetation index made the greatest contributions to the classifications. During summer 2015, more than 4,000 ha of Sargassum coverage per image were observed, which was substantially greater than that over the rest of the period. This approach constitutes a transferable alternative for the systematic detection of Sargassum, which enables a quantitative semi-automated time series comparison.

[1]  Alistair Grinham,et al.  Random forest algorithm yields accurate quantitative prediction models of benthic light at intertidal sites affected by toxic Lyngbya majuscula blooms , 2012 .

[2]  I. Mariño‐Tapia,et al.  Dispersion in the Yucatan coastal zone: Implications for red tide events , 2010 .

[3]  L. Beal,et al.  Observations of the Florida and Yucatan Currents from a Caribbean Cruise Ship , 2010 .

[4]  J. Ricardo,et al.  Impacto de la acumulación de sargazo del verano del 2015 sobre las tortugas marinas de playa La barca, península de Guanahacabibes/ Impact of sargassum influx during 2015 summer on marine turtles of Playa la Barca, Peninsula de Guanahacabibes , 2016 .

[5]  J. Brock,et al.  Assessment of estuarine water-quality indicators using MODIS medium-resolution bands: initial results from Tampa Bay, FL , 2004 .

[6]  Amanpreet Singh,et al.  A review of supervised machine learning algorithms , 2016, 2016 3rd International Conference on Computing for Sustainable Global Development (INDIACom).

[7]  B. Gavio,et al.  MASSIVE QUANTITIES OF PELAGIC Sargassum ON THE SHORES OF SAN ANDRES ISLAND, SOUTHWESTERN CARIBBEAN , 2014 .

[8]  Chuanmin Hu,et al.  Spectral and spatial requirements of remote measurements of pelagic Sargassum macroalgae , 2015 .

[9]  David S. Lee,et al.  Foraging over Sargassum by Western North Atlantic Seabirds , 2012 .

[10]  Allan W. Stoner,et al.  Pelagic Sargassum: Evidence for a major decrease in biomass , 1983 .

[11]  M. Merino.  Upwelling on the Yucatan Shelf: hydrographic evidence , 1997 .

[12]  Lian Feng,et al.  Sargassum coverage in the northeastern Gulf of Mexico during 2010 from Landsat and airborne observations: Implications for the Deepwater Horizon oil spill impact assessment. , 2016, Marine pollution bulletin.

[13]  Chuanmin Hu,et al.  Mapping macroalgal blooms in the Yellow Sea and East China Sea using HJ-1 and Landsat data: Application of a virtual baseline reflectance height technique , 2016 .

[14]  J. Largier,et al.  The Effects of Wind on Upwelling off Cabo Catoche , 2016, Journal of Coastal Research.

[15]  Chuanmin Hu,et al.  Mapping and quantifying Sargassum distribution and coverage in the Central West Atlantic using MODIS observations , 2016 .

[16]  S. Stapleton,et al.  Sargassum accumulation may spell trouble for nesting sea turtles , 2015 .

[17]  M. Vincx,et al.  Hiding and feeding in floating seaweed: Floating seaweed clumps as possible refuges or feeding grounds for fishes , 2007 .

[18]  Didier Tanré,et al.  Atmospherically resistant vegetation index (ARVI) for EOS-MODIS , 1992, IEEE Trans. Geosci. Remote. Sens..

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

[20]  Song Qin,et al.  High-Resolution Satellite Observations of a New Hazard of Golden Tides Caused by Floating Sargassum in Winter in the Yellow Sea , 2017, IEEE Geoscience and Remote Sensing Letters.

[21]  J. Sheinbaum,et al.  Flow structure and transport in the Yucatan Channel , 2002 .

[22]  Erika Young,et al.  Satellite images suggest a new Sargassum source region in 2011 , 2013 .

[24]  Jim Gower,et al.  Satellite Images Show the Movement of Floating Sargassum in the Gulf of Mexico and Atlantic Ocean , 2008 .

[25]  J. Gower,et al.  Distribution of floating Sargassum in the Gulf of Mexico and the Atlantic Ocean mapped using MERIS , 2011 .

[26]  Adriana Zingone,et al.  Green and golden seaweed tides on the rise , 2013, Nature.

[27]  L. Oey,et al.  A Synopsis of the Circulation in the Gulf of Mexico and on its Continental Margins , 2013 .

[28]  A. Huete.  A soil-adjusted vegetation index (SAVI) , 1988 .

[29]  Andy Liaw,et al.  Classification and Regression by randomForest , 2007 .

[30]  A. Sherman,et al.  Pelagic Sargassum community change over a 40-year period: temporal and spatial variability , 2014, Marine biology.

[31]  B. Witherington,et al.  Young sea turtles of the pelagic Sargassum-dominated drift community: habitat use, population density, and threats , 2012 .

[32]  Xingfa Gu,et al.  Effect of radiometric corrections on NDVI-determined from SPOT-HRV and Landsat-TM data , 1994 .

[33]  P. Mumby,et al.  Remote Sensing Handbook for Tropical Coastal Management , 2000 .

[34]  N. Putman,et al.  Oceanic Habits and Habitats: Caretta caretta , 2013 .

[35]  Robert P. Anderson,et al.  Evaluating predictive models of species’ distributions: criteria for selecting optimal models , 2003 .

[36]  M. Thiel,et al.  The ecology of rafting in the marine environment. I: The floating substrata , 2004 .

[37]  Badih Ghattas,et al.  A review of supervised machine learning algorithms and their applications to ecological data , 2012 .

[38]  Oguz Gungor,et al.  Classification of multispectral images using Random Forest algorithm , 2012 .

[39]  Chuanmin Hu,et al.  Ocean Color Satellites Show Extensive Lines of Floating Sargassum in the Gulf of Mexico , 2006, IEEE Transactions on Geoscience and Remote Sensing.

[40]  L. Breiman.  Random Forests--random Features , 1999 .

[41]  V. Lucieer,et al.  Do marine substrates 'look' and 'sound' the same? Supervised classification of multibeam acoustic data using autonomous underwater vehicle images , 2013 .

[42]  L. Collado-Vides,et al.  Severe impacts of brown tides caused by Sargassum spp. on near-shore Caribbean seagrass communities. , 2017, Marine pollution bulletin.

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

[44]  J. Schell,et al.  Recent Sargassum Inundation Events in the Caribbean: Shipboard Observations Reveal Dominance of a Previously Rare Form , 2015 .

[45]  Chuanmin Hu.  A novel ocean color index to detect floating algae in the global oceans , 2009 .

[46]  Tom Linton,et al.  Development and implementation of Sargassum Early Advisory System (SEAS) , 2013 .