have been observed in situ, using a satellite method similar to that employed by
another study (Smyth et al. 2004). We examine these results in relation to physical
processes and climatic indices.
18.2 Methods
Blooms of E. huxleyi were identified in global AVHRR and SeaWiFS imagery as
described in the following sections.
18.2.1 AVHRR Imagery
Daily global, 1/4
o resolution remote sensing reflectances (R rs ), derived from
Rayleigh-corrected radiances of the AVHRR visible and near-infrared channels,
were acquired from the Clouds from AVHRR Extended (CLAVR-X) Project
(Heidinger et al. 2002). Input radiances from the different AVHRR platforms had
been intercalibrated to ensure a reliable and comparable time series of R rs . CLAVR-X
was developed specifically to separate the cloud and sea-surface signals. To separate
the effect of atmospheric contamination, R rs was estimated using a method for
detecting E. huxleyi blooms developed by Smyth et al. (2004):
R rs ¼ ðR1 À R2Þ=ðexpððÀ0:057=2Þ Â p1Þ
where R1 and R2 are Rayleigh-corrected channel 1 (0.580–0.680 μm) and channel 2
(0.725–1.1 μm) reflectances and pl is the optical path length, a function of the
satellite and solar zenith angle. The factor of À0.057 is the calculated value of the
Rayleigh optical depth in channel 1 for a pl of unity and a standard atmosphere.
18.2.2 SeaWiFS Imagery
SeaWiFS Global Area Coverage Level 1a data were acquired from the SeaWiFS
Project and Distributed Active Archive Center, processed to Level 2 radiances
while retaining the coccolith-flagged pixels normally masked during standard
processing, and spatially binned to 9-km resolution. SeaWiFS classification of
pixels into E. huxleyi bloom and non-bloom was based on 8-day mean normalized
water-leaving radiances using the supervised, multispectral classification scheme of
Iglesias-Rodriguez et al. (2002). The spectral signature of E. huxleyi blooms was
empirically ascertained to distinguish them from other oceanic conditions and has
been well validated.
280
S. Schollaert Uz et al.
another study (Smyth et al. 2004). We examine these results in relation to physical
processes and climatic indices.
18.2 Methods
Blooms of E. huxleyi were identified in global AVHRR and SeaWiFS imagery as
described in the following sections.
18.2.1 AVHRR Imagery
Daily global, 1/4
o resolution remote sensing reflectances (R rs ), derived from
Rayleigh-corrected radiances of the AVHRR visible and near-infrared channels,
were acquired from the Clouds from AVHRR Extended (CLAVR-X) Project
(Heidinger et al. 2002). Input radiances from the different AVHRR platforms had
been intercalibrated to ensure a reliable and comparable time series of R rs . CLAVR-X
was developed specifically to separate the cloud and sea-surface signals. To separate
the effect of atmospheric contamination, R rs was estimated using a method for
detecting E. huxleyi blooms developed by Smyth et al. (2004):
R rs ¼ ðR1 À R2Þ=ðexpððÀ0:057=2Þ Â p1Þ
where R1 and R2 are Rayleigh-corrected channel 1 (0.580–0.680 μm) and channel 2
(0.725–1.1 μm) reflectances and pl is the optical path length, a function of the
satellite and solar zenith angle. The factor of À0.057 is the calculated value of the
Rayleigh optical depth in channel 1 for a pl of unity and a standard atmosphere.
18.2.2 SeaWiFS Imagery
SeaWiFS Global Area Coverage Level 1a data were acquired from the SeaWiFS
Project and Distributed Active Archive Center, processed to Level 2 radiances
while retaining the coccolith-flagged pixels normally masked during standard
processing, and spatially binned to 9-km resolution. SeaWiFS classification of
pixels into E. huxleyi bloom and non-bloom was based on 8-day mean normalized
water-leaving radiances using the supervised, multispectral classification scheme of
Iglesias-Rodriguez et al. (2002). The spectral signature of E. huxleyi blooms was
empirically ascertained to distinguish them from other oceanic conditions and has
been well validated.
280
S. Schollaert Uz et al.