Initial results

In the following sections, sample results are provided essentially to illustrate the site characteristics.

The figure below is a time series of the temperature over the first 200 meters of the water column, from July 2001 to December 2005. The well-known winter mixing and spring stratification clearly appear, as well as the increase of the surface temperature from year to year. The summer of 2003 was one of the warmer recorded over the last decades in northern Europe, and the signature of this unusual climate is evident here in the temperature record. Another noticeable feature appears in winter of 2005, when the temperature did not reached the usual 12.9 degrees minimum, which is in principle a recurrent characteristics of this area where deep mixing occurs in winter, forming by this way the deep waters of the western Mediterranean basin.

Time series (July 2001 – June 2005) of the potential temperature over the 0-200 meters depth.

Back to top

Pigment samples for HPLC analysis have been collected at about 12 depths between the surface and 200 meters at the BOUSSOLE site since July 2001. Surface samples at 5 and 10 meters depth were collected in triplicate. During certain surveys only surface samples were taken. In the following section, results from these analyses over 2 years from July 2001 to July 2003 are briefly described.

Total chlorophyll a seasonal evolution
During the winter periods the highest values for TChla have been observed, particularly for the month of February, when maximum values at the surface reached 1.34 (in 2002) and 2.44 mg.m-3 (in 2003, Figure below). In February 2002 the water column appeared to have undergone strong vertical mixing just before the survey, with Tchla a concentrations averaging 0.845 mg.m-3 (± 5%), down to 130 meters depth.

March and April are characterized by the spring bloom in surface waters, with TChla values reaching 1.23 mg.m-3. In March 2003, a deep chlorophyll maximum (DCM) already began to form at 50 m.

During summertime, surface TChla concentrations remain low (0.1 to 0.2 mg.m-3). However, this trend is not reflected for deep samples (below 10 m) where important development of phytoplankton takes place during the oligotrophic conditions of this Mediterranean site. The thickness of the DCM at 50 m tends to decrease towards the end of the summer period.

Autumn marks the transition between stratified summer conditions and winter turbulence of the water column, with the progressive increase of surface TChla concentrations.

TChla variations between July 2001 and April 2004 for depths < 5 meters (stars) and for depths > 5 meters and < 10 meters (circles). Each point is the average of  all measurements taken during a cruise, and the vertical bars are the standard deviations within each cruise.

Back to top

TChla (diamonds; solid line) and Divinyl Chlorophyll a (circles, dotted line) 0-200 meters vertical profiles, and for the dates indicated.

Changes in community structure
Seven carotenoids were partitioned into three phytoplankton size classes and normalized by the sum of the 7 pigments (Vidussi et al, 2001):

The f igure below presents the variations of these 3 classes between July 2001 and July 2003 at the surface (between 0 and 10 m). Generally, for surface waters, microphytoplankton (characterized by diatoms and dinoflagellates) and nanophytoplankton (characterized by prymnesiophytes and crysophytes) tend to vary in opposition with picophytoplankton (characterized by cyanobacteria at this site). The first are not a dominant class over the studied period, with minimum values in autumn. Nevertheless, at depth microphytoplankton can reach important proportions relative to the other two classes, particularly in spring. Nanophytoplankton is a dominant class in surface waters over most of the year, except during the autumn when picophytoplankton tends to take over. However, the DCM is dominated by nanophytoplankton over the whole summer-autumn period.

The presence of divinyl chlorophyll a, an indicator of prochlorophytes, has generally been observed during autumn at the DCM and even in surface waters during winter convection.

As for the concentrations in chlorophyll a degradation products (Chlorophyllid a and Phaeophorbid a), they were most often below the detection limits of the instrument.

Time series of average abundances of pico- (diamonds), nano- (circles) and micro (triangles) phytoplankton for surface samples (depth < 10 m) between July 2001 and April 2004. Each point is the average over all measurements taken during a cruise, and the vertical bars are the standard deviations within each cruise.

Back to top

The figure below (panels a, b and c) show the July 2001- August 2004 time series of the reflectance ratio R(l)/R(560), for l = 443, 490 and 510. The next two figures show, for the same time period, the diffuse attenuation coefficients of the upper layers for the wavelengths 412, 443, 490, 510, 560 and 670 nm. Finally, and the reflectances (R = Eu(0-)/Ed(0-)) for the same wavelengths.

On these figures, the open diamonds and the black dots are for the in situ data (several points per cruises plus the standard deviation around the mean of these points), the black triangles are for the HPLC-determined chlorophylla concentration, the stars are for the AOPs (either R, ratio of R at two wavelengths or Kd) as computed through a model that is fed with the in situ chlorophylla concentration, and the dotted curve is the chlorophylla concentration that would be derived through satellite algorithms using the in situ reflectances ratio. The difference between the big black dots and the stars, therefore, represents the anomaly of the measured AOPs as compared to what is predicted by standard (global) bio-optical models, considering the in situ chlorophylla concentration.

Similarly, the difference between the stars and the dotted curve represents the error in the chlorophyll-a concentration that is obtained when satellite algorithms, which were derived from global data sets and are therefore based on average global relationships between reflectance ratios and the chlorophylla concentration, are applied to waters with anomalous optical properties.

One striking feature is the large anomaly of the blue-to-green ratio in summer of 2001, with lower-than-expected blue-to-green ratios, as already identified and quantified in Claustre et al. (2002) for the year 1999 (PROSOPE cruise). The reason for that anomaly has been attributed to the presence in the water of desert dust particles, absorbing in the blue and scattering in the green. This anomaly is however less important during the summer of 2002, and it is vanishing during the summer of 2003. The transient character of this anomaly has to be confirmed (for instance with the up coming data for the summer of 2004). The transient character of this anomaly has to be confirmed.

Examining the diffuse attenuation coefficients reveals that their values in the summer of 2001 are largely above the modeled values across the entire spectrum, at least from 412-560nm, and that this difference is disappearing in summer of 2003. Because is largely determined by the absorption properties of the medium, these observations mean that an excess absorption exists at all wavelengths in the July to September period in 2001.

The same feature does not appear in the reflectances, which are lower than the modeled values in the blue, again during summer of 2001, yet greater than the modeled values in the green. Because is, to the first order, proportional to backscattering and inversely proportional to absorption, these observation confirm the presence of an excess backscattering in the green part of the spectrum. Again, the transient nature of this anomaly remains to be explained.

Another anomaly is observed in February 2003, which is now in the opposite direction as compared to the summer anomaly, with larger-than-expected blue-green ratios. During the February 2003 cruise, the chlorophylla concentration was as large as 2mg m-3. The reason for this behavior might be in the low scattering coefficients that are typical of a fresh phytoplankton bloom, where the proportion of large healthy cells characterized by a low backscattering efficiency is high, and the contribution of small detritus is low.

Back to top

Times series (July 2001 – August 2004) of the reflectance ratio R(443)/R(560). The open diamonds are for the in situ data (several points per cruises plus the standard deviation around the mean of these points), and the black circles are the average of these data over each cruise. The black triangles are for the HPLC-determined chlorophyll-a concentration, the stars are for the reflectance ratio as computed through a model that is fed with the in situ chlorophyll-a concentration (Morel and Maritorena, 2001), and the dotted curve is the chlorophyll-a concentration that would be derived through “satellite algorithms” using the in situ reflectances ratio.

Times series (July 2001 – August 2004) of the diffuse attenuation coefficients Kd(412), Kd (490) and Kd (560). The open diamonds are for the in situ data (several points per cruises plus the standard deviation around the mean of these points), and the black circles are the average of these data over each cruise. The black triangles are for the HPLC-determined chlorophyll-a concentration, the stars are for Kd as computed through a model that is fed with the in situ chlorophyll-a concentration (Morel and Maritorena, 2001).

Back to top

Times series (July 2001 – August 2004) of the irradiance reflectances R(412), R(490) and R(560). The open diamonds are for the in situ data (several points per cruises plus the standard deviation around the mean of these points), and the black circles are the average of these data over each cruise. The black triangles are for the HPLC-determined chlorophyll-a concentration, the stars are for R as computed through a model that is fed with the in situ chlorophyll-a concentration (Morel and Maritorena, 2001).

Back to top

The figure below shows the seasonal and interannual evolution of the diffuse attenuation coefficient for wavelengths 412, 490 and 560 nm. The time series was established from 3 years of measurements carried out monthly (approximately) from the ships.

This parameter is largely determined by the absorption coefficient, and it is thus a good indicator of the concentration of phytoplankton and associated colored dissolved substances. The seasonal cycle appears very clearly, in particular in 2003 and 2004: it starts with an intense mixing of water in winter (maximum in February, with almost homogeneous properties and weak ), continues with the start of the spring bloom in March-April, then evolves gradually to the oligotrophy, which is maximum around September, and which is characterized by a maximum of attenuation at about 50 meters, corresponding to the deep chlorophyll maximum (DCM). For chlorophyll concentrations lower than 0.1mgm at the surface at this period, the global statistical relationships (cf Morel and Berthon, 1989) predict a DCM at about 100 meters. Its development at a lower depth at the BOUSSOLE site is due to the situation of this site in the center of a cyclonic circulation, generating a ôdomeö structure maintained by significant vertical advection.

The interannual variability observed here for seems directly related to the changes of the physical framework. Indeed, the depth of the DCM seems to increase by 2001 to 2004, at the same time as the winter mixing seems to intensify and the spring bloom being more intense. The oligotrophic layer of water characterized by low values of in summer is deeper in 2003 than in 2001 and 2002. An intense mixing in winter on the one hand brings more nutrients to the surface, supporting a stronger bloom, and, on the other hand, more efficiently eliminates the colored dissolved substances accumulated at the surface during summer (phenomenon already shown for dissolved organic carbon; Copin-Mont‰gut and Avril, 1993; Avril, 2002). This could explain the changes observed in the "anomalies" of the marine reflectance (cf. above).

However, the interannual variability of the spring bloom is somewhat distorted here by an insufficient sampling in spring 2002 and 2003 (bad weather). Consequently, it will probably be necessary to rebuild part of the record by using the vertical profiles of the pigments and the values derived from the satellites, coupled with the relationships Chl - that will be established from the existing data.

Time series of the vertical distribution of the diffuse attenuation coefficient at 412 (top), 490 (middle) and 560 (bottom) nm (derived from SPMR’ Ed vertical profiles). Red bars indicate the time of the ship sampling. Units are m-1.

Back to top

Time series (2003) of the absorption coefficients at 440 nm for total particles (circles), phytoplankton (stars) and detritus (triangles). The absorption coefficient of phytoplankton at 676 nm is also shown (diamonds). Top : linear scale. Bottom : log scale, in order to zoom on the summer values.

Back to top

The f igure below shows examples of the absorption spectra by particles, either for the total particles or for the phytoplankton or detritus only. The dominant role of phytoplankton in the absorption budget during the spring bloom clearly appears, as well as the progressive increase of that part of absorption that is due to detritus in summer. The exact cause of the background absorption in the UV that clearly appears when absorption decreases at other wavelengths (see the summer spectrum) remains to be identified.

Spectra of the absorption coefficient of particles at 5 meters (total : solid curve, phytoplankton : dotted curve, and detritus : dashed curve), as determined on filtered samples (method described in section 7.2). Samples collected in 2003, for the dates indicated.

The figure below shows examples of the total attenuation, scattering and absorption coefficients for the surface (0-10m) layer. One can just mention the clear identification of the absorption features on the scattering spectrum, and the change in its slope from spring to summer (increasing slope in summer corresponding to the large contribution of detritus as compared to phytoplankton themselves).

Spectra of the total attenuation (circles), absorption (diamonds) and scattering (stars) coefficients, at the four months indicated, and for the surface layer (mean from the surface to a 10 meters depth). Values determined in situ with the AC9.

The figure below illustrates the richness of the information provided by the IOPs, when compared to modeled values. What is shown here is that a careful analysis of these properties on the one hand confirm what is seen on the AOPs (closure), and, on the other hand, provides clues for the understanding of the causes of variation in these AOPs versus IOPs relationships.

Total scattering coefficient at 550 nm as a function of the chlorophyll concentration. Black diamonds : values determined in situ (AC9). Solid and dashed lines: b(550) as computed from Loisel and Morel (1998), and a factor of two around the model values.

Back to top