19(5)
Chlorophyll-a variability
within Basque coastal
waters and the Bay of
Biscay, between 2005
and 2010, using MODIS
imagery
Stéfani Novoa
Guillem Chust
Yolanda Sagarminaga
Marta Revilla
Javier Franco
Victoriano Valencia
Ángel Borja
Chlorophyll-a variability within Basque coastal waters and the Bay of Biscay using MODIS imagery (2005-2010)
Novoa, S., Chust, G., Sagarminaga, Y., Revilla, M., Franco, J., Valencia, V., Borja, Á., 2012. Chlorophyll-a
variability within Basque coastal waters and the Bay of Biscay, between 2005 and 2010, using MODIS imagery .
Revista de Investigación Marina, AZTI-Tecnalia, 19(5): 92-107
La serie ‘Revista de Investigación Marina’, editada por la Unidad de Investigación Marina de
Tecnalia, cuenta con el siguiente Comité Editorial:
Editor:
Dr. Ángel Borja
Adjunta al Editor: Dña. Mercedes Fernández Monge e Irantzu Zubiaur (coordinación de las
publicaciones)
Comité Editorial:
Dr. Lorenzo Motos
Dr. Adolfo Uriarte
Dr. Michael Collins
Dr. Javier Franco
D. Julien Mader
Dña. Marina Santurtun
D. Victoriano Valencia
Dr. Xabier Irigoien
Dra. Arantza Murillas
Dr. Josu Santiago
La ‘Revista de Investigación Marina’ de Tecnalia edita y publica investigaciones y datos originales resultado
de la Unidad de Investigación Marina de Tecnalia. Las propuestas de publicación deben ser enviadas al
siguiente correo electrónico aborja@azti.es. Un comité de selección revisará las propuestas y sugerirá los
cambios pertinentes antes de su aceptación definitiva.
Edición: 1.ª Julio 2012
© AZTI-Tecnalia
ISSN: 1988-818X
Unidad de Investigación Marina
Internet: www.azti.es
Edita: Unidad de Investigación Marina de Tecnalia
Herrera Kaia, Portualdea
20010 Pasaia
Foto portada: Iñigo Onandia (vista aérea del barco Leporre Anaiak, de Bermeo) © AZTI-Tecnalia.
© AZTI-Tecnalia 2012. Distribución gratuita en formato PDF a través de la web: www.azti.es/RIM
92 | Revista de Investigación Marina, 2012, 19(5)
S. Novoa, G. Chust, Y. Sagarminaga, M. Revilla, J. Franco, V. Valencia and Á. Borja
Chlorophyll-a variability within Basque coastal waters
and the Bay of Biscay, between 2005 and 2010, using
MODIS imagery
Stéfani Novoa1*, Guillem Chust1*, Yolanda Sagarminaga1, Marta Revilla1, Javier
Franco1, Victoriano Valencia1, Ángel Borja1
Abstract
Understanding the response of chlorophyll-a (as a proxy for phytoplankton biomass) to both
anthropogenic pressures and natural factors is important for water quality assessment purposes and
for the management of biological resources. In the Basque coastal area, discharges produced by the
Adour and Nervión rivers (south-eastern part of the Bay of Biscay) have been proven to reach up
to 15-20 km off the coast. The first objective of this study was to describe the spatial and temporal
variability of chlorophyll-a in the Basque coast and the Bay of Biscay, in relation to river discharges
at a daily, seasonal and inter-annual scale using MODIS images acquired between 2005 and 2010. The
second objective was to offer a synoptic description of the spatio-temporal variability of chlorophyll-a
in the entire Bay of Biscay, using multivariate statistical methods and satellite imagery. The results
indicate that seasonal chlorophyll-a cycle is slightly different in coastal areas affected by the Adour
and the Nervión river waters, compared to offshore waters. The spring chlorophyll-a peak in March in
offshore waters shifts to May in the Adour nearby area Nervión. The multivariate statistical analysis
highlights the influence of river discharges in the spatial variability of chlorophyll-a in coastal areas
of the bay. The Spanish and French Basque coastal waters are differentiated in terms of chlorophyll-a
concentrations levels reached, river regimes and morphology of the continental shelf. Statistical and
indicator maps have been created to represent the main components of chlorophyll-a variability in the
area of study. They confirm that, at present, phytoplankton is at good status and eutrophication risk
is low in the Basque coastal waters. These maps may provide water quality indicators in a continuous
spatial distribution in the area and may be used for the selection of water quality stations as a function
of the dynamics of the water masses characterised.
Keywords: Chlorophyll, satellite, MODIS, intra and interanual variability, river plumes, Bay of
Biscay, Basque Country
Introduction
Primary production in temperate areas is subject to a
high spatio-temporal variability (Cebrián and Valiela 1999;
Bode et al. 2005, Muylaert et al. 2006, Gameiro et al. 2010).
Phytoplankton blooms are events of rapid production and
accumulation of phytoplankton biomass that are usually
responses to changing physical driving forces originating in
the ocean (e.g., upwelling), the atmosphere (wind), or in land
(precipitation and river runoff). These drivers have different
timescales of variability; thus, phytoplankton blooms can be
short-term episodic events, recurrent seasonal phenomena,
or sporadic events associated with extraordinary climatic or
hydrologic conditions (Cloern 1996, Cloern 2001). Several
studies have reported the high variability present in oceanic,
continental shelf and coastal areas of the Bay of Biscay
1
AZTI-Tecnalia; Marine Research Division; Herrera Kaia s/n 20110;
Fax: +34 946572555; Tel: +34 943004800
* Corresponding author: snovoa@gmail.com; gchust@azti.es
(Tréguer et al. 1979, García-Soto and Pingree 1998, Lampert
2001, Gohin et al. 2003, Loyer et al. 2006).
River discharges into coastal waters are an important element
in the dynamics of the continental shelf in the Bay of Biscay
(Puillat 2004, Guillaud et al. 2008, Prego et al. 2008). The
plume regions off the river mouths play an important role in the
shelf’s physical, biogeochemical and ecological functioning,
as the river discharge includes nutrient, sendiments, pollutants
and other consitutents in addition to freshwater (Arnau et al.
2004, Wysocki et al. 2006, Weston et al. 2008, Petus et al.
2010). In the presence of a river plume, light and circulation
patterns, stratification and nutrient pathways are significantly
altered. Hence, the ecological processes taking place in the
plume area are influenced by this exchange between the
continental shelf and the river water (Sierra et al. 2002). The
extent of the influence of the river discharge depends mainly
on the river flow regime and on the volume of the discharges
(Cravo et al. 2006), at seasonal and annual scales (Signoret
et al. 2006). Also, the exchange dynamics between river
discharges and coastal waters is affected by atmospheric and
Revista de Investigación Marina, 2012, 19(5) | 93
Chlorophyll-a variability within Basque coastal waters and the Bay of Biscay using MODIS imagery (2005-2010)
marine factors, such as wind and waves (Stumpf et al. 1993;
Gonzalez et al. 2004). The changing levels of river discharges
and the dynamics taking place on the coast affect phytoplankton
productivity and growth, since nutrients play an important
role in primary production (Harding 1994, Lavín et al. 2006,
Domingues et al. 2008).
Understanding the dynamics of chlorophyll-a (chl-a) in
relation to both anthropogenic pressures (e.g. urban waste
water inputs) and natural nutrient inputs is important for
water quality assessment purposes (Gohin et al. 2008), and
also to determine the influence of those inputs on coastal
activities such as fisheries. Hence, the first objective of this
investigation was to examine and describe the variability of
chl-a concentration in surface waters of the Basque coastal
area, using satellite imagery and its possible relationship with
river discharge. The hypothesis is that rain events and river
discharge act as fertilization factors, enhancing phytoplankton
growth and leading to increased biomass (i.e. chl-a) in the
Basque coast. To perform this study, three approaches to analyse
rainfall, river discharges and chl-a patterns were followed:
(i) daily, (ii) seasonal, and (iii) inter-annual variability. The
second objective was to offer a synoptic description of chl-a
spatio-temporal variability in the entire Bay of Biscay, using
multivariate statistical methods.
Materials and methods
In situ dataset
A time series of daily data at inland stations of two rivers,
the Adour (France) and the Nervión (Spain) were acquired
for the 2005-2010 period. In the case of the Nervión River,
rainfall (mm), river discharge (m3.s-1), turbidity (NTU) and
suspended and dissolved matter (mg.m-3) were measured at a
station located on the river. The station selected was the nearest
station to the river mouth without influence of the marine
tides; its name is Abusu. The data was downloaded from the
“Diputación Foral de Bizkaia” web site (www.bizkaia.net/).
In the case of the Adour, only daily river discharge data was
available and the data was downloaded from the DIREN web
site (Directions Régionales de l’Environnement; www.hydro.
eaufrance.fr/). The total river discharge was computed by
combining the contributions from the 6 main streams of the
Adour basin (Figure 1; Petus, 2009).
Satellite data
A total of 1672 OC5 products for the 2005-2010 period
were downloaded from Ifremer’s ftp server (ftp://ftp.ifremer.
fr/ifremer/cersat/products/gridded/ocean-color). The OC5 is an
empirical algorithm, developed and validated by Gohin et al.
(2002, 2008) and Gohin (2011). This algorithm derives chl-a,
turbidity and suspended matter products with look-up tables
applied to standard MODIS/Aqua Level-2 reflectance products
(processed with SeaDAS).
Daily data analysis
First, the relationship between chl-a concentration derived
from satellite images and rainfall events and river discharge
were studied. For this purpose, two OC5 chl-a products
corresponding to two days before and one day after a rainfall
event were downloaded. The rainfall event occurred on the
13th of January 2009 (a time of the year with high precipitation
levels), but no images were available on that day due to the
cloud cover, hence, chl-a products corresponding to the 11th
and the 14th of January 2009 are presented here. Satellite chl-a
concentrations extracted from a 3 x 3 pixel window near the
river mouth was compared to river discharge data. Secondly,
the response of variables measured in situ and variables
derived from satellite images in relation to the rainfall event
was studied. The in situ variables measured (rainfall, river
discharge, turbidity and organic matter) were averaged,
between days -1 and 7, being -1, one day before the rainfall
Figure 1. Study area in the southeastern corner of the Bay of Biscay. The crosses (+) represent river stations, the black circles ( • ) represent the marine
stations.
94 | Revista de Investigación Marina, 2012, 19(5)
S. Novoa, G. Chust, Y. Sagarminaga, M. Revilla, J. Franco, V. Valencia and Á. Borja
event, 0 the day of the event, 1 the day after the event, and so
on. Chl-a, turbidity and suspended matter data, within surface
waters of the river plume area, were acquired using satellite
imagery. Values of the three variables were extracted using the
averaged 3 x 3 pixel window located near the river mouths. The
distance from the river mouth was approximately 1.5 pixels (or
0.9 nautical miles); this distance was necessary to avoid the
effect of the land pixels. The analysis could not be performed
for the Adour due to the lack of rainfall data.
Then, cross-correlation analyses with daily lags were
performed with the different in situ variables obtained for
the Nervión River, i.e. rainfall versus river discharge, organic
matter and turbidity.
Seasonal analysis
Chl-a 90th percentile (P90) and average maps were produced
for the 2005-2010 period (see Novoa et al. 2012, for more
details). This was performed by initially selecting all cloudfree images available for the study area during the 20052010 period. Subsequently, the P90 and average values were
calculated for each pixel of all images selected for that period
using IDL programming. As such, the IDL program calculated
for each pixel the average chl-a value and, for P90 maps, the
value where 90% of the observations were equal or below. This
approach resulted in two types of products, monthly P90 and
mean chl-a maps for the 2005-2010 period. Results obtained
with P90 and mean chl-a maps were compared for coastal
and offshore waters. This comparison was performed between
chl-a concentration values extracted using a 3 x 3 pixel window
positioned near the Adour and Nervión river mouths and at the
offshore reference station.
Spatial analysis
Inter-annual analysis
The long-term trends (years 2005-2010) of sea surface
(0-1 m) chl-a, suspended particulate matter and turbidity in
river plume areas, were analysed along with the variability of
rainfall and river discharge of the rivers at the inland stations.
In situ data from the Littoral Water Quality Monitoring and
Control Network (LQM) of the Basque Country, managed by
the Basque Water Agency, used previously in Revilla et al.
(2009) and in Novoa et al. (2012), was used to perform a time
series analysis of the 2005-2010 period, and was compared to
the same analysis performed with satellite data. The locations
of the Nervión (LN10) and the reference stations are shown in
Figure 1. The 3 x 3 pixel window positioned at those locations
was used to extract the chl-a concentrations.
Time series analysis of several variables measured in situ
and with satellite images was performed. In the case of the
satellite-derived variables, 2-week means were used to avoid
missing values. In the case of in situ data, daily measurements
of river discharge and rainfall were employed for the analysis,
while chl-a in situ measurements used in this analysis were
performed once per trimester. The variability of each variable
was examined by means of the “decompose” procedure in R
language, which extracts the seasonal component of the time
series, and outputs the trend followed by each variable without
this component. Subsequently, the trend values were linearly
regressed against the years, and the significance of the trendline was measured with the p-value returned by the regression.
A permutation resampling method was performed to test the
significance of the trend-lines. The time series was permuted
1000 times, and the observed slope from the regression
between the trend and the time series was then compared to
the permuted slopes. The null hypothesis established that the
observed slope did not show a significant change over the
years. If the observed value is found to be outside 95% quantile
range of the histogram, then the null hypothesis is refuted. For
more information on this procedure see Davison and Hinkley
(1997).
Cluster and Principal Component Analysis
Figure 2. Polygons used to extract mean chl-a concentrations at different
distances from the river Nervión and Adour mouths (MODIS
Terra, 17/05/2004).
The evolution of chl-a concentration in the river plume
areas with increasing distance from the coast, was examined
using the monthly P90 chl-a maps calculated with images
downloaded between 2005 and 2010. As such, mean values of
chl-a at different distances from the coast were extracted in the
form of polygons, as shown in Figure 2.
Two multivariate statistical analyses were performed over
the entire Bay of Biscay, an unsupervised classification and a
Principal Component Analysis (PCA), to synoptically describe
the variability of chl-a concentration throughout the seasons
and from coasts to open waters. The entire bay was considered
for these analyses to examine the spatial variability of the
Basque coast area in relation to the entire bay.
The unsupervised cluster classification of the chl-a OC5
products was performed with the k-means routine from the ENVI
program. All the products available for the 2005-2010 period
were included in this analysis. This routine initially computes
class means evenly distributed in the data space. Three classes
were selected for this classification. Then, it iteratively clusters
the pixels into the nearest class using a minimum distance
technique. Every single iteration recalculates class means and
reclassifies pixels with respect to the new means. This process
Revista de Investigación Marina, 2012, 19(5) | 95
Trophic studies of small pelagic fish in the Bay of Biscay: methodological aspects
continues until the maximum number of iterations is reached.
For more information see Tou and Gonzalez (1974).
The PCA was performed using the ENVI program as well.
All the OC5 chl-a products available for the 2005-2010 period
for the entire Bay of Biscay were used for this purpose. The
PCA produces uncorrelated output bands, to segregate noise
components, and to reduce the dimensionality of data sets. This
routine finds several sets of orthogonal axes that have their
origin at the data mean and that are rotated so the data variance
is maximized. It produces principal component bands that
are linear combinations of the original spectral bands and are
uncorrelated. The first PC band contains the largest percentage
of data variance and the second PC band contains the second
largest data variance, and so on. The percent of total variance
can be determined from the eigenvalues, which are the measure
of variance in a PCA. The set of weights applied to band values
to obtain a principal component are called eigenvectors. To
obtain more information on this procedure see Richards and
Jia (2006).
In this study, the results provided by the PCA offer a
visualization of the similarities and differences among pixels
in relation to the temporal evolution of the chl-a concentration
averages. This analysis also provides a visualization of the
global connections between the chl-a pixels of the entire Bay
of Biscay. In this case, the PCA variables are the averaged chl-a
months, which are 70 in total and the number of observations
used in the analysis are the total number of pixels in the area
of study, hence 244097 points of observation (the land is not
included). Monthly composites were chosen for this analysis
because the PCA principles state that missing values should be
avoided. Some areas show missing values during large periods
of time (more than 10 days in some cases), due to clouds or
image artifacts. The 70 monthly averaged composites rarely
provided missing values for the Bay of Biscay area.
Results
Daily variability
Images before and after the rainfall event, which occurred
on the 13th of January 2009, showed an increase of chl-a in the
coastal area one day after the event (Figure 3). In the case of the
Nervión River, there was an increase of river discharge on the
same day of the rainfall event and the maximum river discharge
was reached after one day (Figure 3).
Figure 3. MODIS daily image showing chl-a distribution (estimated with the OC5 algorithm) before (a) and after (b) a rainfall and discharge event occurred
on the 13th of January 2009. (c) High chl-a concentration (dashed lines) was observed one day after the rainfall event, and on the same day of the
river discharge increase was recorded at the Nervión river station.
96 | Revista de Investigación Marina, 2012, 19(5)
S. Novoa, G. Chust, Y. Sagarminaga, M. Revilla, J. Franco, V. Valencia and Á. Borja
Figure 4. Diagram showing the evolution of the different variables measured in situ (rainfall, river discharge, organic matter, and turbidity) at the river
Nervión station and with satellite imagery at the river plume (chl-a, suspended matter and turbidty). The values shown are an average of 13 rainfall
events. Some parameter values are transformed for display purposes.
Figure 4 shows that, in the case of the Nervión river, the
averaged turbidity and organic matter measured in situ, present
a maximum increase one day after the event. This analysis
showed that chl-a, suspended matter and turbidity estimated
with satellite imagery reached a maximum concentration
2 days (on average) after a rainfall event i.e. average of 13
rainfall events.
The cross-correlation between rainfall and river discharge
is significant on the same day of the event and on the seven
following days. Rainfall and turbidity measured in situ are
also significantly correlated at day lags between 0 and 7, while
organic matter is significatively correlated with rainfall only
on the day of the event (Figure 5).
Seasonal variability
Figure 5. Cross-correlations between in situ parameters. (a) River discharge
with rainfall, (b) Turbidity with rainfall and (c) Organic matter
with rainfall.
The monthly P90 and mean images for the 2005-2010 period
showed similar seasonal patterns for both the offshore and the
coastal areas, with some exceptions (Figures 6, 7, 8). In offshore
waters, two main chl-a concentration peaks are observed, one
of them is observed during spring (March-April) and the other
in November. However, the P90 and the mean values extracted
from the images show different magnitudes. The P90 chl-a
levels are greater in November than in March, while the mean
data shows the opposite, higher chl-a concentrations in March
compared to November (Figure 8). The summer months in
offshore areas show the lowest levels of chl-a concentration.
In coastal waters, seasonal differences are observed in
different areas (Figures 6 and 7). In the area of the Adour river
plume, the maximum chl-a concentration is observed in May,
for both the P90 and the mean products (Figures 6, 7, 8). A
peak is observed with the P90 chl-a values in February that is
observed with less magnitude on the chl-a mean figure (Figure
8). In the area of the Nervión plume (Figure 8), there are two
Revista de Investigación Marina, 2012, 19(5) | 97
Chlorophyll-a variability within Basque coastal waters and the Bay of Biscay using MODIS imagery (2005-2010)
Figure 6. Monthly P90 chl-a maps for the 2005-2010 period.
98 | Revista de Investigación Marina, 2012, 19(5)
S. Novoa, G. Chust, Y. Sagarminaga, M. Revilla, J. Franco, V. Valencia and Á. Borja
Figure 7. Monthly mean chl-a maps for the 2005-2010 period.
Revista de Investigación Marina, 2012, 19(5) | 99
Chlorophyll-a variability within Basque coastal waters and the Bay of Biscay using MODIS imagery (2005-2010)
Figure 8. P90 Chl-a monthly values (a) and means (b) for the 2005-2010
period at 3 different locations: The Nervión river plume, the
Adour river plume and the offshore reference station.
Figure 9. Monthly mean, maximum and minimum chl-a concentrations,
and monthly mean river discharge and standard deviations, at
the Nervión (a) and Adour (b) rivers, for the 2005-2010 period.
major chl-a peaks, one in March and one in November. Both
the P90 and the mean values of the pixels located at the Nervión
plume show this pattern and the chl-a concentrations are higher
in November than in March. The patterns of the Nervión river
discharges and chl-a levels observed in the river plume area
coincide, except for the chl-a concentration increase observed
in June (Figure 9). The variability of the river discharges is
also greater during the spring and the autumn months. In the
case of the Adour, the patterns are slightly different, as the
highest river discharges occurred in April-May, and the highest
mean chl-a concentration values at the Adour river plume area
are observed in May (Figure 9). These are also the periods of
greatest river discharge variability (Figure 9).
The chl-a seasonal pattern slightly changes with distance
from the coast (Figure 10). Chl-a 90th percentile between 1.2
and 11 nautical miles from the Nervión river mouth, exhibits
two major peaks, one in spring and another in November. The
magnitude of the chl-a concentration decreases with increasing
distance. However, two particularities can be observed. First,
at 1.2 miles from the coast, there is a peak in June that is not
observed at a farther distance. Second, the peak occurring
in November is of greater magnitude compared to the peak
occurring in March at 1.2 miles, but this is not observed at
larger distances. In the case of the Adour, the month with the
highest chl-a concentration is May, up until a distance of 11
nautical miles; at this distance, March is the month with the
highest chl-a concentration.
100 | Revista de Investigación Marina, 2012, 19(5)
S. Novoa, G. Chust, Y. Sagarminaga, M. Revilla, J. Franco, V. Valencia and Á. Borja
Figure 10. P90 Seasonal variation of chorophyll-a with increasing distance from the coast, from 1.2 nautical miles until 11 nautical miles, at the Nervión
(a) and the Adour (b) river plumes
Inter-annual variability
Chl-a concentration calculated from monthly means show
considerable inter-annual variation in magnitude and duration
of peak events. In general, the chl-a variability coincided with
the river discharge events during the 2005-2010 period. A peak
in spring, followed by another peak of lower magnitude in
autumn was observed, while the lowest values were observed
during the summer months. Chl-a concentration is generally
higher in plume waters than at the offshore station, which is not
usually influenced by river discharges (Figure 11).
The variation of chl-a concentration in coastal water usually
coincides with the long duration of river discharges (Figures
11, 12), and do not necessarily coincide with their magnitude.
The highest mean chl-a concentration in the Adour plume area
was 6.5 mg.m-3 (June, 2007), followed by 4.6 mg.m-3 (May,
2009). The lowest mean chl-a concentration in the Adour river
plume was 0.5 mg.m-3 and occurred in August 2008. The mean
annual chl-a concentration reached 2.4 mg.m-3 in 2007 and 1.3
mg.m-3 in 2008. The levels of the other years varied between
those two values.
The highest mean chl-a concentration in the Nervión plume
area was 3.0 mg.m-3 (November, 2008), followed by 2.9 mg.m-3
(October, 2009). The lowest monthly mean chl-a concentration
was 0.3 mg.m-3 and occurred in July 2009. The annual mean
chl-a concentration was the highest in 2005, with 1.3 mg.m-3,
while the lowest annual mean chl-a concentration was 1.0
mg.m-3 and was observed in 2008. In general, the variability
of river discharges coincided with the rainfall variability at the
Nervión inland station (Figures 12 and 13).
The highest monthly mean chl-a concentration found at the
offshore location was 1.5 mg.m-3 (March, 2006), followed by
1.3 mg.m-3 (April, 2008). The lowest mean chl-a concentration
at the offshore location was of 0.2 and occurred in September
2006. The maximum annual chlorophyll mean value was
reached in 2007 as well, and was of 0.7 mg.m-3; the minimum
mean value was observed in 2008, and was of 0.5 mg.m-3.
Revista de Investigación Marina, 2012, 19(5) | 101
Chlorophyll-a variability within Basque coastal waters and the Bay of Biscay using MODIS imagery (2005-2010)
Figure 12. Time series of daily river discharges at the Nervión (a) and the
Adour (b) river stations.
Figure 11. Time series of two-week mean chl-a concentration, at the
offshore station (a), the Nervión (b) and the Adour (c) river
plumes.
Figure 13. Time series of daily precipitation (rainfall) values at the
Nervión river station.
Table 1. Time series analysis of several variables measured in situ and with satellite imagery. The values provided are the slopes of trend analysis. The
asterisk (*) means that the bootstrap analysis proved the trend is significant.
slope
chl-a
(in situ)
Nervión
-0.0220*
Adour
Offshore
-0.0010*
chl-a
(satellite)
River discharge
Rainfall
Turbidity (satellite)
-0.0027*
0.0012*
0.0002*
0.0086*
-0.0020*
0.067*
0.0017*
102 | Revista de Investigación Marina, 2012, 19(5)
0.0103
0.0141
S. Novoa, G. Chust, Y. Sagarminaga, M. Revilla, J. Franco, V. Valencia and Á. Borja
The bootstrap analysis revealed a significant decrease in
chl-a at the Nervión river plume region, measured with satellite
and in situ data between 2005 and 2009 (Table 1). The in situ
data resulted in a steeper slope than the satellite data (-0.0220
vs. -0.0027). The chl-a concentration measured in the Adour
river plume showed as well, a significant decrease during the
same period. The river discharge and rainfall measured at the
inland river stations, increased significantly. The turbidity
measured with satellite imagery significantly increased over
the same period of time in plume and offshore waters.
Unsupervised classification and principal component
analysis
Figure 15. Unsupervised K-means classification performed with 1672
chl-a OC5 products. The black area is the land and was not
considered in the classification. Class 1 is shown in yellow,
class 2 in blue and class 3 in green.
Figure 14. Linear regression between suspended matter and chl-a
estimated with satellite imagery, for the Nervión (a) and
the Adour (b) river plumes. The data are shown on a
logarithmic scale for display purposes.
The logarithmic regression between the suspended matter
and the chl-a concentrations (Figure 15) did not show a very
high determination coefficients in both river plumes (Nervión
r2=0.13; Adour r2=0.12).
The classification map showed three distinct surface water
classes in the Bay of Biscay according to chl-a concentration
(Figure 15). The first class (in yellow) includes the French
part of the continental shelf influenced by river discharges,
where chl-a concentrations are the highest. The Adour river
plume area is included in this class. The second class (in
blue) corresponds to an area of transition between river
influenced areas and oceanic waters. This class has probably
lower chl-a concentrations than the first class and it includes
the region inside the continental shelf beyond the influenced
areas by the French rivers. Northern Spain’s coastal waters
belong to the second classes well except for Galician waters.
The third class (in green) would correspond to oceanic
waters, where chl-a concentration is lower in general.
This class includes the area beyond the continental shelf.
However, some regions in the centre of the Bay located over
deep water (>2000 m) and beyond the shelf, belong to class
2, indicated similar chl-a levels than transitional waters
(Class 2, in blue). These higher concentrations are probably
reached during bloom periods.
The first PCA band explained 57.3% of the total variance of
the data and shows prevalent positive values in coastal areas,
which are characterised by higher chl-a concentration (Figure
16).
Revista de Investigación Marina, 2012, 19(5) | 103
Chlorophyll-a variability within Basque coastal waters and the Bay of Biscay using MODIS imagery (2005-2010)
Figure 16. Spatial projection of the first Principal Component band. The
lowest eigenvalues are shown in dark purple and the highest
in red; the blue, green and yellow colours are used for the
eigenvalues in between.
Discussion and conclusions
This study described the spatial and temporal variability
of chl-a in the Basque coast and the SE Bay of Biscay. The
main conclusions yielded by this study are that river discharges
closely affect the variability of chl-a concentration levels in
coastal areas at a daily, seasonal and inter-annual scale. In
general, an increase of chl-a concentration was observed 1 or
2 days after a rainfall event, in Basque coastal areas near the
Adour and the Nervión river plumes. Such a rapid increase of
chl-a after an input of nutrients has not been reported previously
in the area, where other studies found longer reactive periods.
In particular García-Soto et al. (1990) reported an increase of
primary productivity 4-5 days after a rainfall event in a coastal
embayment near the Nervión estuary, and Alvarez et al. (2009)
suggested a delay of 7 to 9 to detect primary productivity
after an increase in nutrient concentrations from upwelling
nutrients in the Cantabrian coast. Possible explanations for
these differences that would need a more detailed study may
be that the increase in chl-a concentration, observed with
satellite images shortly after the rainfall event, is not produced
by phytoplankton proliferation caused by nutrient inputs,
but to an effect of the storms. Winds (turbulence) and waves
(bottom shear stress), usually more intense during winter
storms, have major effects on the particles in suspension and
on the dynamics of coastal waters and the river plumes (Huret
et al. 2007). These disturbances are producing a resuspension
and an advection of particulate material (e.g. microalgae and
sediments) of marine, benthic, or estuarine origin (Abreu et
al. 2009). In fact, benthic diatoms were found in some water
104 | Revista de Investigación Marina, 2012, 19(5)
samples collected in coastal areas during the LQM survey (Borja
et al. 2007-2010). Furthermore, the presence of higher levels of
dissolved and suspended particulate inorganic material, due to
resuspension or to river plume advection, could be also causing
an overestimation of chl-a concentration by satellite imagery
(Lahet et al. 2001, Bowers et al. 2009).
The seasonal variability patterns of chl-a concentration are
different for the Adour and the Nervión river plumes, and differ
as well from the pattern observed in offshore waters. In the
case of the Adour, the highest chl-a concentration levels are
observed during late autumn and spring, but mainly between
April and June, the months of the greatest river floods driven
by its pluvio-nival regime. In the case of the Nervión, the cycle
is similar to offshore waters, where blooms are observed in
spring and late autumn-early winter, except for a secondary
peak detected in June. The increase of chl-a concentration
during spring is due to the higher concentration of nutrients
present produced by river discharges, but also to the increase
of light availaibility during this season (Orive et al. 2004,
Butrón et al. 2009)
Although the spatial influence of Nervión river is rather
constrained (1.2 nautical miles from estuary), the Adour
discharge has an influence up to 8 nm from coast, which derives
in important phytoplankton growth in the marine areas under
river discharge influence. These findings were also provided
by other studies in the area (Retailleau et al. 2009, Lampert et
al. 2002, Loyer et al. 2006).
The inter-annual analysis also shows a significant decrease
of chl-a concentration (estimated with satellite data) between
2005 and 2010 in coastal areas (near the Nervión and the Adour
river mouths) and a significant increase in offshore waters. The
negative tendency in coastal waters, detected both by satellite
and in situ data, could be related to the pollution control
programmes set up in Nervión river in 1993 (García-Barcina
et al. 2006), and from 1975 in the Adour river (Tudesque et
al. 2008) that have considerably improved the water quality
of these estuaries. Another reason for the chl-a decrease
during the study period, could be the increase of suspended
material related to the increase in river discharges, which could
have caused a decrease of water clarity and thus, affected
phytoplankton growth. In fact, other authors (Iriarte and Purdie
2004; Butrón et al. 2009) have addressed that light limitation
due to turbidity and/or low residence time due to higher river
flows, can limit the phytoplankton growth.”
Nevertheless, in offshore waters the time series analysis
performed with in situ and satellite chl-a estimates are
contradictory: in situ measurements show a significant
decrease of surface chl-a between 2005 and 2010, whilst
satellite data estimate a significant increase, in the same
location. Interestingly, although a negative chl-a tendency
was found by Revilla et al. (2010a) in surface waters with in
situ measurements, they also reported (Revilla et al. 2011) an
increase in chl-a concentration at greater depths in the photic
layer. Thus a possible explanation of this contradiction could
lay on the fact than under clear water conditions, which is the
case for this reference station, satellite estimates are integrating
S. Novoa, G. Chust, Y. Sagarminaga, M. Revilla, J. Franco, V. Valencia and Á. Borja
the part of the water column where an increase in chl-a occurs.
In this case, chl-a satellite estimates would coincide with chl-a
in situ estimates.
An unsupervised classification has resulted in three types of
water bodies (Adour, Nervión and Oceanic) that converge in a
relatively small area. The differences between these areas are
mainly due to average chl-a concentrations, to the influence of
river discharges and to the morphology of the coastal shelf and
slope. Interestingly, the station used as reference of offshore
waters, not affected by river discharges (Revilla et al. 2010a;
b), belongs to the same class as the coastal areas of the Spanish
Basque coastal waters. Therefore, this confirms that, at present,
phytoplankton is at good status and eutrophication risk is low
in the Basque coastal waters (Revilla et al. 2009; Garmendia
et al. 2011).
The PCA resulted in a first band explaining 57% of chl-a
variability located near coastal areas, where highest chl-a
concentration levels are found, and which are related with
river discharges areas and the morphology of the continental
shelf, confirming the results found with the unsupervised
classification.
The findings presented in this study reveal the value of
remote sensing as a valuable tool for this type of studies, as
it considerably improves the spatial and temporal sampling
coverage needed to detect high frequency variability in primary
production processes, which at present cannot be afforded by
in situ monitoring programmes. Nevertheless, its applicability
is limited to the study of upper ocean layers and estimations
in turbid coastal waters may show inaccuracies if proper
validated algorithms are not used. Thus, a complementation
between both methodologies may at present largely improve
the study and monitoring of primary production processes in
coastal areas.
A next step for this study would be to deeper analyse the
processes and factors affecting the variability of chl-a, and
continue to improve the estimation of chl-a in coastal areas
with satellite imagery.
Acknowledgements
This research and data was supported by the Basque
Water Agency (URA) through the project “La clorofila,
elemento clave”. Funds for Aquitania–Euskadi cooperation
(Basque Government), and by the European Structural Funds
Programme INTERREG IVa (LOREA project: Littoral, Ocean
and River in Euskadi-Aquitaine). This study was supported
also by the project WISER (Water bodies in Europe: Integrative
Systems to assess Ecological status and Recovery), funded
by the European Union under the 7th Framework Program,
Theme 6 (Environment including Climate Change) (contract
No. 226273), www.wiser.eu. Thanks are given to the NASA
Goddard Earth Science Distributed Active Archive Centre,
for the MODIS Level IB and Level II images. S. Novoa has
benefited from a PhD Scholarship granted by the Fundación
de Centros Tecnológicos Iñaki Goenaga. The authors wish to
acknowledge as well the contribution of Dr. Rubén Roa for his
help with the time series analysis and of Maialen Garmendia
for her thorough revision.
This is contribution no. 585 from AZTI-Tecnalia (Marine
Research Division).
References
Abreu, P. C., M. Bergesch, L. A. Proença, C. A. E. Garcia, and C.
Odebrecht. 2009. Short- and Long-Term Chlorophyll a Variability
in the Shallow Microtidal Patos Lagoon Estuary, Southern Brazil.
Estuaries and Coasts 33: 554-569.
Alvarez, E., E. Nogueira, J. L. Acuna, M. Lopez-Alvarez, and J. A. Sostres.
2009. Short-term dynamics of late-winter phytoplankton blooms in
a temperate ecosystem (Central Cantabrian Sea, Southern Bay of
Biscay). Journal of Plankton Research 31: 601-617, doi:10.1093/
plankt/fbp012.
Arnau, P., C. Liquete, and M. Canals. 2004. River mouth plume events and
their dispersal in the Northwestern Mediterranean Sea. Oceanography
17:23-31.
Bode, A., M.T. Álvarez-Ossorio, N. González, J. Lorenzo, C. Rodríguez,
M. Varela, and M.M. Varela. 2005. Seasonal variability of plankton
blooms in the Ria de Ferrol (NW Spain): II. Plankton abundance,
composition and biomass. Estuarine, Coastal and Shelf Science 63:
285-300.
Borja, A., A. B. Josefson, A. Miles, I. Muxika, F. Olsgard, G. Phillips, J. G.
Rodríguez, and B. Rygg. 2007. An approach to the intercalibration of
benthic ecological status assessment in the North Atlantic ecoregion,
according to the European Water Framework Directive. Marine
Pollution Bulletin 55: 42-52.
Borja, A., J. Bald, J. Franco, J. Larreta, I. Muxika, M. Revilla, J. G.
Rodríguez, O. Solaun, A. Uriarte, and V. Valencia. 2009. Using
multiple ecosystem components, in assessing ecological status in
Spanish (Basque Country) Atlantic marine waters. Marine pollution
bulletin 59: 54-64.
Borja, A., I. Galparsoro, X. Irigoien, A. Iriondo, I. Menchaca, I. Muxika, M.
Pascual, I. Quincoces, M. Revilla, J. G. Rodríguez, M. Santurtún, O.
Solaun, A. Uriarte, V. Valencia, and I. Zorita. 2011. Implementation of
the European Marine Strategy Framework Directive: a methodological
approach for the assessment of environmental status, from the Basque
Country (Bay of Biscay). Marine Pollution Bulletin 62: 889-904.
Borja A., J. Bald, M. J. Belzunce, J. Franco, J. M. Garmendia, J. Larreta,
I. Muxika, M. Revilla, G. Rodríguez, O. Solaun, A. Uriarte, and V.
Valencia. 2007. Agencia Vasca del agua. Red de seguimiento del
estado ecológico de las aguas de transición y costeras de la comunidad
del País Vasco: masa de agua costera Matxitxako-Getaria. (www.
uragentzia.euskadi.net).
Borja A., J. Bald, M. J. Belzunce, J. Franco, J. M. Garmendia, J. Larreta,
I. Muxika, M. Revilla, G. Rodríguez, O. Solaun, A. Uriarte, I. Zorita,
and V. Valencia. 2008. Agencia Vasca del agua. Red de seguimiento
del estado ecológico de las aguas de transición y costeras de la
comunidad del País Vasco: masa de agua costera Matxitxako-Getaria.
(www.uragentzia.euskadi.net).
Borja A., J. Bald, M. J. Belzunce, J. Franco, J. M. Garmendia, J. Larreta,
I. Muxika, M. Revilla, G. Rodríguez, O. Solaun, A. Uriarte, and I.
Zorita, 2009. Agencia vasca del agua. Red de seguimiento del estado
ecológico de las aguas de transición y costeras de la comunidad
del País Vasco: masa de agua costera Matxitxako-Getaria. (www.
uragentzia.euskadi.net).
Borja, A., J. Bald, M. J. Belzunce, J. Franco, J. M. Garmendia, J. Larreta,
I. Muxika, M. Revilla, J. G. Rodríguez, O. Solaun, A. Uriarte, V.
Valencia, I. Zorita, I. Adarraga, F. Aguirrezabala, I. Cruz, A. Laza,
M. A. Marquiegui, J. Martínez, E. Orive, J. Mª Ruiz, S. Seoane,
J.C. Sola, and A. Manzanos. 2010. Agencia Vasca del agua. Red de
seguimiento del estado ecológico de las aguas de transición y costeras
de la comunidad del País Vasco: masa de agua costera Matxitxako-
Revista de Investigación Marina, 2012, 19(5) | 105
Chlorophyll-a variability within Basque coastal waters and the Bay of Biscay using MODIS imagery (2005-2010)
Getaria. (www.uragentzia.euskadi.net).
Bowers, D. G., D. Evans, D. N. Thomas, K. Ellis, and P. J. L. B. Williams.
2009. Interpreting the colour of an estuary. Estuarine, Coastal and
Shelf Science 59: 13-20.
Brière, C. 2005. Etude de l’hydrodynamique d’une une zone côtière
anthropisée : l’embouchure de l’Adour et les plages adjacentes
d’Anglet.
Butrón A., A. Iriarte, and J. Madariaga. 2009. Size-fractionated
phytoplankton biomass, primary production and respiration in the
Nervión-Ibaizabal estuary: A comparison with other nearshore and
estuarine ecosystems from the Bay of Biscay. Continental Shelf
Research 29: 1088-1102
Cebrián, J. and I. Valiela. 1999. Seasonal patterns in phytoplankton biomass
in coastal ecosystems. Journal of Plankton Research 21:429-444.
Cloern, J. E. 1996. Phytoplankton bloom in coastal ecosystems: A review
with some general lessons from sustained investigation of San
Francisco Bay, California. Reviews of Geophysics 34: 127-168.
Cloern, J.E. 2001. Our evolving conceptual model of the coastal
eutrophication problem. Marine Ecology Progress Series 210: 223253
Cravo, A., M. Madureira, H. Felicia, F. Rita, and M. Bebianno. 2006.
Impact of outflow from the Guadiana River on the distribution of
suspended particulate matter and nutrients in the adjacent coastal
zone. Estuarine, Coastal and Shelf Science 70: 63-75.
Domingues, R.B., A. Barbosa, and H. Galvão. 2008. Constraints on the
use of phytoplankton as a biological quality element within the Water
Framework Directive in Portuguese waters. Marine Pollution Bulletin
56: 1389-1395.
Gameiro, C., and V. Brotas. 2010. Patterns of Phytoplankton Variability in
the Tagus Estuary (Portugal). Estuaries and Coasts 33: 311-323.
García-Barcina, J. Ma, J. A. González-Oreja, and A. De la Sota. 2006.
Assessing the improvement of the Bilbao estuary water quality in
response to pollution abatement measures. Water Research 40: 951960.
García-Soto, C., and R. D. Pingree. 1998. Late autumn distribution
and seasonality of chlorophyll a at the shelf-break/ slope region of
the Armorican and Celtic shelf. Journal of the Marine Biological
Association of the U.K 78: 17- 33.
García-Soto, C., I. de Madariaga, F. Villate, and E. Orive. 1990. Day-to-day
variability in the plankton community of a coastal shallow embayment
in response to changes in river runoff and water turbulence. Estuarine,
Coastal and Shelf Science 31: 217-229.
Garmendia, M., M. Revilla , J. Bald, J.Franco, A. Laza-Martínez, E.
Orive, S. Seoane, V. Valencia, and Á. Borja. 2011. Phytoplankton
communities and biomass size structure (fractionated chlorophyll
“a”), along trophic gradients of the Basque coast (northern Spain).
Biogeochemistry 106(2): 243-263.
Gohin, F. 2011. Joint use of satellite and in-situ data for coastal monitoring.
Ocean Science Discussions 8: 955-998.
Gohin, F., J. N. Druon, and L. Lampert. 2002. A five channel chlorophyll
concentration algorithm applied to SeaWiFS data processed by
SeaDAS in coastal waters. International Journal of Remote Sensing
23: 1639-1661.
Gohin, F., L. Lampert, J. F. Guillaud, A. Herbland, and E. Nézan. 2003.
Satellite and in situ observations of a late winter phytoplankton
bloom in the northern Bay of Biscay. Continental Shelf Research 23:
1117−1141.
Gohin, F., B. Salquin, H. Oger-Jeanneret, L. Lozac’h, L. Lampert, A.
Lefebvre, P. Riou, and F. Bruchon. 2008. Towards a better assessment
of the ecological status of coastal waters using satellite-derived
chlorophyll-a concentrations. Remote Sensing of Environment 112:
3329-3340.
Gonzalez, M., A. Uriarte, A. Fontán, J. Mader, and P. Gyssels. 2004.
Marine dynamics, In Oceanography and Marine Environment of the
Basque Country. Elsevier Oceanography Series.
Guillaud, J. F., A. Aminot, D. Delmas, F. Gohin, M. Lunven, C. Labry, and
A. Herbland. 2008. Seasonal variation of riverine nutrient inputs in the
106 | Revista de Investigación Marina, 2012, 19(5)
northern Bay of Biscay (France), and patterns of marine phytoplankton
response. Journal of Marine Systems 72: 309 - 319.
Harding, L. 1994. Long term trends in the distribution of phytoplankton
in Chesapeake Bay: roles of light, nutrients and streamflow. Marine
Ecology Progress Series 104: 267-291.
Huret M., F. Gohin, D. Delmas, M. Lunven and V. Garc¸ 2007. Use of
SeaWiFS data for light availability and parameter estimation of
phytoplankton production model of the Bay of Biscay.Journal of
Marine Systems, 65: 509-531.
Iriarte A. and D.A. Purdie. 2004. Factors controlling the timing of major
spring bloom events in a UK south coast estuary. Estuarine, Coastal
and Shelf Science 61: 679-690.
Lahet, F., S. Ouillon, and P. Forget. 2001. Colour classification of coastal
waters of the Ebro river plume from spectral reflectances. International
Journal of Remote Sensing 22: 1639-1664.
Lampert, L. 2001. Dynamique saisonnière et variabilité pygmentaire
des populations phytoplanctoniques dans l’Atlantique Nord (Golfe
de Gascogne). Tesis Doctoral. Université de Bretagne Occidentale,
France.
Lampert, L., B. Quéguiner, T. Labasque, A. Pichon, and N. Lebreton.
2002. Spatial variability of phytoplankton composition and biomass
on the eastern continental shelf of the Bay of Biscay (northeast Atlantic Ocean). Evidence for a bloom of Emiliania huxleyi
(Prymnesiophyceae) in spring 1998. Continental Shelf Research 22:
1225-1247.
Lavín, A., Valdes, L., Sánchez, F., Abaunza, P., Forest, A., Boucher, J.,
Lazure, P., and A.M. Jegou. 2006. The Bay of Biscay: The encountering
of the ocean and the shelf. In: Robinson, A.R., Brink, K.H. (Eds.), The
Sea, Volume 14. Harvard Univ. Press, Cambridge, MA, pp. 933-1001.
Lihan, T., M. A. Mustapha, S. A. Rahim, S. Saitoh, and K. Iida. 2011.
Influence of River Plume Variability of Chlorophyll a Concentration
using Satellite Images. Journal of Applied Sciences 11: 484-493.
Loyer, S., L. Lampert, A. Menesguen, and P. Cann. 2006. Seasonal
evolution of the nutrient pattern on Biscay Bay continental shelf over
the years 1999-2000. Scientia Marina 70: 31-46.
Muylaert, K., R. Gonzales, M. Franck, M. Lionard, C. Van der Zee,
A. Cattrijsse, K. Sabbe, L. Chou, and W. Vyverman. 2006. Spatial
variation in phytoplankton dynamics in the Belgian coastal zone of the
North Sea studied by microscopy, HPLC-CHEMTAX and underway
fluorescence recordings. Journal of Sea Research 55: 253-265.
Novoa, S., Chust, G., Sagarminaga, Y., Revilla, M., Borja, A. & Franco, J.
(2012) Water quality assessment using satellite-derived chlorophyll-a
within the European directives, in the southeastern Bay of Biscay.
Marine Pollution Bulletin 64:739-750
Prego, R., P. Boi, and A. Cobelo-garcía. 2008. The contribution of total
suspended solids to the Bay of Biscay by Cantabrian Rivers (northern
coast of the Iberian Peninsula). Journal of Marine Systems 72: 342349.
Orive, E., J. Franco, J. Madariaga, M. Revilla. 2004. Bacterioplankton
and phytoplankton communities. In A. Borja and M. Collins (eds.),
Oceanography and Marine Environment of the Basque Country.
Elsevier Oceanography Series.
Puillat, I. 2004. Hydrographical variability on the French continental shelf
in the Bay of Biscay, during the 1990s. Continental Shelf Research
24: 1143-1163.
Petus C., Chust G., Gohin F., Doxaran D., Froidefond J.-M., Sagarminaga
Y. (2010). Estimating turbidity and total suspended matter in the Adour
River plume (South Bay of Biscay) using MODIS 250-m imagery.
Continental Shelf Research, 30, 379–392.
Retailleau, R., H. Howa, R. Schiebel, F. Lombard, F. Eynaud, S. Schmidt,
F. Jorissen, and L. Labeyrie. 2009. Planktonic foraminifera production
along an offshore–onshore transect in the south-eastern Bay of Biscay.
Continental Shelf Research 29: 1123-1135.
Revilla, M., J. Franco, J. Bald, Á. Borja, A. Laza, S. Seoane, and V.
Valencia. 2009. Assessment of the phytoplankton ecological status in
the Basque coast (northern Spain) according to the European Water
Framework Directive. Journal of Sea Research 61: 60-67.
S. Novoa, G. Chust, Y. Sagarminaga, M. Revilla, J. Franco, V. Valencia and Á. Borja
Revilla, M., Á. Borja, A. Fontán, J. Franco, M. González, and V.
Valencia. 2010a. A two-decade record of surface chlorophyll “a” and
temperature in offshore waters of the Basque country (south-eastern
Bay of Biscay). Revista de Investigación Marina 17(2): 13-20.
Revilla, M., J. Franco, M. Garmendia, and Á. Borja. 2010b. A new method
for phytoplankton quality assessment in the Basque estuaries (northern
Spain), within the European Water Framework Directive. Revista de
Investigación Marina 17: 149-164.
Revilla, M., Á. Borja, A. Fontán, M. González, and V. Valencia. 2011.
Time-series of sea surface temperature, nutrients and chlorophyll-a in
the south-eastern Bay of Biscay: in situ measurements along a landocean transect. ICES/NAFO Decadal Symposium.
Richards, J. A., and X. Jia. 2006. Remote Sensing Digital Image Analysis,
Springer., doi:10.1016/S0169-555X(01)00164-7
Sierra, J. P., A. Sánchez-Arcilla, J. G. D. Rı́o, J. Flos, E. Movellán, C.
Mösso, R. Martı́nez, M. Rodilla, S. Falco, and I. Romero. 2002. Spatial
distribution of nutrients in the Ebro estuary and plume. Continental
Shelf Research 22: 361-378.
Signoret, M., M. Monrealgomez, J. Aldeco, and D. Salasdeleon. 2006.
Hydrography, oxygen saturation, suspended particulate matter, and
chlorophyll-a fluorescence in an oceanic region under freshwater
influence. Estuarine, Coastal and Shelf Science 69: 153-164.
Stumpf, R. P., G. Gelfenbaum, and J. R. Pennock. 1993. Wind and tidal
forcing of a buoyant plume, Mobile Bay, Alabama. Continental Shelf
Research 13: 1281-1301.
Tou, J. T., and R. C. González. 1974. Pattern recognition principles. Image
Rochester NY, 377pp.
Tréguer, P., P. Le Corre, and J. R. Grall. 1979. The seasonal variations
of nutrients in the upper waters of the Bay of Biscay region and
their relation to phytoplankton growth. Deep Sea Research Part A.
Oceanographic Research Papers 26: 1121-1152.
Tudesque, L., M. Gevrey, G. Grenouillet, and S. Lek. 2008. Long-term
changes in water physicochemistry in the Adour-Garonne hydrographic
network during the last three decades. Water research 42: 732-42.
Weston, J., Ratle, F., & Collobert, R. 2008. Deep learning via semisupervised embedding. International Conference on Machine
Learning.
Wysocki, L. A., T. S. Bianchi, R. T. Powell, and N. Reuss. 2006.
Spatial variability in the coupling of organic carbon, nutrients, and
phytoplankton pigments in surface waters and sediments of the
Mississippi River plume. Estuarine, Coastal and Shelf Science 69:
47-63.
Revista de Investigación Marina, 2012, 19(5) | 107
www.azti.es
Txatxarramendi ugartea z/g
48395 Sukarrieta (Bizkaia)
Tel.: +34 94 602 94 00
Fax: +34 94 657 25 55
Herrera Kaia, Portualdea z/g
20110 Pasaia (Gipuzkoa)
Tel.: +34 943 00 48 00
Fax: +34 94 657 25 55
Parque Tecnológico de Bizkaia
Astondo bidea. Edificio 609.
48160 Derio (Bizkaia)
Tel.: +34 94 657 40 00
Fax: +34 94 657 25 55