Table 1.
Summary of the satellite datasets for the different types of harmful algal blooms (HABs) in the Red Sea.
Fig 1.
Bathymetry and the position of reported harmful algal blooms (HABs) in the Red Sea.
(a) Bathymetry of the Red Sea (acquired from the General Bathymetric Chart of the Oceans (GEBCO_2014 Grid, version 20150318, http://www.gebco.net)). (b) Map showing the in situ sampling locations associated with the different Harmful Algal Blooms (HABs) in the Red Sea; Ost: Ostreopsis sp.; HA: H. akashiwo; KF: K. foliaceum; TE: T. erythraeum; NS: N. scintillans/miliaris (We note that no in situ data were collected during the occurrence of C. polykrikoides blooms in the Red Sea, although their cyst life-stages were identified along the Red Sea, off the south-western coast of Saudi Arabia [16]).
Table 2.
Summary of the different types of harmful algal blooms (HABs) reported by various studies in the Red Sea.
Fig 2.
Summary of the steps involved in the development of our model.
Flow chart showing the major steps involved in the detection and classification of different Harmful Algal Blooms (HABs) in the Red Sea.
Fig 3.
Examples of Rrs derived from daily MODIS-Aqua data for the different HAB species in the Red Sea waters.
(a) H. akashiwo (b) Ostreopsis sp. (c) K. foliaceum (d) N. scintillans/miliaris (e) C. polykrikoides (f) T. erythraeum.
Table 3.
Different sets of nequations for each HAB species based on the band-ratios/difference algorithm and difference in their derivative spectra.
Fig 4.
Examples of second-order derivative Rrs from daily MODIS-Aqua data for the different HABs in the Red Sea waters.
(a) H. akashiwo (b) Ostreopsis sp. (c) K. foliaceum (d) N. scintillans/miliaris (e) C. polykrikoides (f) T. erythraeum.
Fig 5.
Distinct cluster of different HAB species (based on training datasets) in the Red Sea waters.
Ost- Ostreopsis sp.; NS- N. scintillans/miliaris; TE- T. erythraeum; KF-K. foliaceum; HA-H. akashiwo; CP- C. polykrikoides.
Fig 6.
Remotely sensing Ostreopsis sp. blooms in the central Red Sea and comparison with in situ measurements.
(a) and (b) False color composite image generated using the NIR, green and blue reflectances (i.e., R = Rrs (748), G = Rrs (547), B = Rrs (443)). (c) Satellite-derived Chl-a using the ABI algorithm. (d) and (e) MODIS-Aqua-derived maps of the Ostreopsis sp. blooms in the central Red Sea on 27th February 2012 [In (e) the red and white dots denote the presence and absence of HABs, respectively, as detected using satellite data. Also, the yellow circles and yellow triangles indicate the in situ sample points and highlight the presence and absence of Ostreopsis sp. blooms, respectively]. (f) Variations in Ostreopsis sp. cell counts recorded from the different sampling sites in the central Red Sea on 27th February 2012 [red squares and red circles denote satellite detection of HAB presence and absence, respectively].
Fig 7.
Remotely sensing H. akashiwo blooms in the central Red Sea and comparison with in situ measurements.
(a) and (b) False color composite image generated using the NIR, green and blue reflectances (i.e., R = Rrs (748), G = Rrs (547), B = Rrs (443)). (c) Satellite-derived Chl-a using ABI algorithm. (d) and (e) MODIS-Aqua-derived maps of the H. akashiwo blooms in the Red Sea on 27th May 2010 [In (e) the red and white dots denote the presence and absence of HABs, respectively, as detected using satellite data. Also, the yellow circle and yellow triangle indicates the in situ sample points and highlight the presence and absence of H. akashiwo blooms, respectively]. (f) Variations in H. akashiwo cell counts recorded from the two different sampling sites in the Red Sea off the southern coast of Saudi Arabia on 27th May 2010 [red square and red circle denotes the satellite detection of HAB presence and absence, respectively].
Fig 8.
Remotely sensing K. foliaceum blooms in the southern Red Sea and comparison with in situ measurements.
(a) and (b) False color composite image generated using the NIR, green and blue reflectances (i.e., R = Rrs (748), G = Rrs (547), B = Rrs (443)). (c) Satellite-derived Chl-a using ABI algorithm. (d) and (e) MODIS-Aqua-derived maps of the K. foliaceum blooms in the Red Sea on 8th May 2013 [In (e) the red and white dots denote the presence and absence of HABs, respectively, as detected using satellite data. Also, the yellow circle and yellow triangles indicate the in situ sample points and highlight the presence and absence of K. foliaceum blooms, respectively]. (f) Variations in K. foliaceum cell counts recorded from the four different sampling stations in the coastal waters of Al Salif, southern Red Sea on 8th May 2013 [red squares and red circles denote satellite detection of HAB presence and absence, respectively].
Fig 9.
Remotely sensing T. erythraeum blooms in the southern Red Sea and comparison with in situ measurements.
(a) and (b) False color composite image generated using the NIR, green and blue reflectances (i.e., R = Rrs (748), G = Rrs (547), B = Rrs (443)). (c) Satellite-derived Chl-a using ABI algorithm. (d) and (e) MODIS-Aqua-derived maps of the T. erythraeum blooms in the Red Sea on 27th December 2012 [In (e) the red and white dots denote the presence and absence of HABs, respectively, as detected using satellite data. Also, the yellow circle and yellow triangles indicate the in situ sample points and highlight the presence and absence of T. erythraeum blooms, respectively]. (f) Variations in T. erythraeum cell counts recorded from the four different sampling stations in the Yemeni coastal waters off Red Sea, near Al Hodeida City on 27th December 2102 [red squares and red circles denote satellite detection of HAB presence and absence, respectively].
Fig 10.
Remotely sensing N. scintillans/miliaris blooms in the southern Red Sea and comparison with in situ measurements.
(a) and (b) False color composite image generated using the NIR, green and blue reflectances (i.e., R = Rrs (748), G = Rrs (547), B = Rrs (443)). (c) Satellite-derived Chl-a using ABI algorithm. (d) and (e) MODIS-Aqua-derived maps of the N. scintillans/miliaris blooms in the Red Sea on 3rd March 2009 [In (e) the red and white dots denote the presence and absence of HABs, respectively, as detected using satellite data. Also, the yellow circles and yellow triangles indicate the in situ sample points and highlight the presence and absence of N. scintillans/miliaris blooms, respectively]. (f) Variations in N. scintillans/miliaris cell counts recorded from the five different stations in the coastal waters of Yemen, southern Red Sea during 3rd March 2009 [red squares and circles denote satellite detection of HAB presence and absence, respectively].
Table 4.
Statistical accuracy assessment results for detecting the presence and absence of different HAB species in the Red Sea watersa.
Fig 11.
Remotely sensing Chl-a concentrations associated with different HAB species in the Red Sea and comparison with in situ phytoplankton cell counts.
(a) Variations in total cell counts of harmful species and their satellite-derived Chl-a concentrations observed in the Red Sea. (b) Variations in the area density of remotely sensed HABs in the Red Sea. TE- Trichodesmium erythraeum; Ost- Ostreopsis sp.; NS- Noctiluca scintillans; KF-Kryptoperidinium foliaceum; HA-Heterosigma akashiwo.