10 Ocean Colour Remote Sensing of Harmful Algal Blooms in the Benguela System
197
biomass cases (Fig. 10.3c), the generation of the effective diameter fails to retrieve
a reasonable estimate (around 30 μm for mono-specific blooms) (Fig. 10.3d). In
high biomass cases however, such as for the bloom shown in Fig. 10.3e, where more
reasonable estimate of effective diameter can be retrieved (Fig. 10.3f), a ratio between
bands 620 and 550 nm can be used to further identify the presence of Myrionecta
rubra (Fig. 10.3h) from a background of mixed diatom blooms. These examples
show the complexities of detecting accessory pigment-based optical variability with
multi-spectral ocean colour data, in particular the very strong effects of biomass on
the useable reflectance signal.
10.4.3 Tracking the Development of a Ceratium Balechii Bloom
and Consequent Anoxia Using the MPH Algorithm
In the autumn of 2009 the development and ultimate decay of an exceptional bloom
of the dinoflagellate Ceratium balechii lead to an anoxia-induced mass mortality of
macrofauna in St Helena Bay (Pitcher and Probyn 2011). Initial build-up of the bloom
was evident in mid-February from chlorophyll concentrations derived by application
of the MPH algorithm to MERIS data (Fig. 10.4). The bloom was first observed to
the north of St Helena Bay in the nearshore region of the southern Namaqua shelf
and was on occasions shown to extend in a narrow band over a distance of 100 km
(Fig. 10.4; 14 March 2009). Under diminished upwelling activity during late summer
and early autumn inshore counter currents resulted in the southward progression of
the bloom and its entrainment into the Bay (Fig. 10.4; 11 April 2009). In early May
the bloom was shown to have accumulated in the shallow, southern reaches of the Bay
(Fig. 10.4; 1 May 2009). Degradation of the bloom occurred in these shallow waters
where subthermocline nutrients, necessary for bloom maintenance, are inaccessible.
The exceptional organic loading of the system as afforded by the bloom resulted
in anoxia through the entire water column. Large fish and lobster mortalities were
consequently observed off the Berg River estuary and adjacent beaches on the 5 May
2009.
Termination of the boom was associated with persistent downwelling conditions
resulting in advection of the bloom from St Helena Bay shortly after the anoxic
event (Fig. 10.4; 4 May 2009). Surveillance of these blooms as depicted in Fig. 10.4
enables continual assessment by fisheries and coastal managers of the risk posed by
these phenomena to marine resources and the coastal environment.
The case study demonstrates the ability of relatively simple algorithms, such
as the MPH, to effectively track very high biomass blooms. It should further be
noted that several of the images (± 40 %) suffered from high sun-glint, which would
prohibit employing EAP-type algorithms that use the full spectrum of water leaving
reflectance data after aerosol and Rayleigh atmospheric correction. The MPH is less
susceptible to glint effects, as it uses only a small set of green to NIR wavebands and
does not employ an aerosol correction (Matthews et al. 2012): this allows increased
frequency of image utility at the event scale, a significant advantage for operational
application.
197
biomass cases (Fig. 10.3c), the generation of the effective diameter fails to retrieve
a reasonable estimate (around 30 μm for mono-specific blooms) (Fig. 10.3d). In
high biomass cases however, such as for the bloom shown in Fig. 10.3e, where more
reasonable estimate of effective diameter can be retrieved (Fig. 10.3f), a ratio between
bands 620 and 550 nm can be used to further identify the presence of Myrionecta
rubra (Fig. 10.3h) from a background of mixed diatom blooms. These examples
show the complexities of detecting accessory pigment-based optical variability with
multi-spectral ocean colour data, in particular the very strong effects of biomass on
the useable reflectance signal.
10.4.3 Tracking the Development of a Ceratium Balechii Bloom
and Consequent Anoxia Using the MPH Algorithm
In the autumn of 2009 the development and ultimate decay of an exceptional bloom
of the dinoflagellate Ceratium balechii lead to an anoxia-induced mass mortality of
macrofauna in St Helena Bay (Pitcher and Probyn 2011). Initial build-up of the bloom
was evident in mid-February from chlorophyll concentrations derived by application
of the MPH algorithm to MERIS data (Fig. 10.4). The bloom was first observed to
the north of St Helena Bay in the nearshore region of the southern Namaqua shelf
and was on occasions shown to extend in a narrow band over a distance of 100 km
(Fig. 10.4; 14 March 2009). Under diminished upwelling activity during late summer
and early autumn inshore counter currents resulted in the southward progression of
the bloom and its entrainment into the Bay (Fig. 10.4; 11 April 2009). In early May
the bloom was shown to have accumulated in the shallow, southern reaches of the Bay
(Fig. 10.4; 1 May 2009). Degradation of the bloom occurred in these shallow waters
where subthermocline nutrients, necessary for bloom maintenance, are inaccessible.
The exceptional organic loading of the system as afforded by the bloom resulted
in anoxia through the entire water column. Large fish and lobster mortalities were
consequently observed off the Berg River estuary and adjacent beaches on the 5 May
2009.
Termination of the boom was associated with persistent downwelling conditions
resulting in advection of the bloom from St Helena Bay shortly after the anoxic
event (Fig. 10.4; 4 May 2009). Surveillance of these blooms as depicted in Fig. 10.4
enables continual assessment by fisheries and coastal managers of the risk posed by
these phenomena to marine resources and the coastal environment.
The case study demonstrates the ability of relatively simple algorithms, such
as the MPH, to effectively track very high biomass blooms. It should further be
noted that several of the images (± 40 %) suffered from high sun-glint, which would
prohibit employing EAP-type algorithms that use the full spectrum of water leaving
reflectance data after aerosol and Rayleigh atmospheric correction. The MPH is less
susceptible to glint effects, as it uses only a small set of green to NIR wavebands and
does not employ an aerosol correction (Matthews et al. 2012): this allows increased
frequency of image utility at the event scale, a significant advantage for operational
application.