Seasonal Variations in the Daily Cycles of the Particulate Beam Attenuation Coefficient in the NW Mediterranean Sea

Pierre Gernez, (ACRI-ST and LOV, France)

David Antoine, (LOV, France)

Yannick Huot, (LOV, France)

ABSTRACT

The particulate beam attenuation coefficient at 660 nm (cp) is a descriptor of the particulate matter load. In the open ocean, over large spatial scales, cp varies at first order with the phytoplankton biomass. A second order variability in cp is also observed in response to local changes, such as changes in the mixed layer depth or in the bulk particle composition. Furthermore, a daily variation in cp is associated with the daily cycle in irradiance.

However, generally, in situ investigation of the daily cp variability is limited to a few days. The development of instrumented moorings allows in situ measurements to be performed at high frequency, over long periods (years) and under varying environmental conditions (trophic state, vertical distribution of the water properties, cloudiness, sea state).

A time series of inherent and apparent optical properties (cp, chlorophyll fluorescence, backscattering coefficient, downward and upward irradiances ) is collected at an open ocean mooring (BOUSSOLE project) in the North western Mediterranean Sea (60 km offshore Nice, France). All data are simultaneously recorded every 15 minutes at 4 and 9 m depth. This database allows us to characterize the daily variability in cp, with consideration of the physical forcing and the phytoplankton population composition (as described by monthly HPLC analyses). We show seasonal changes linked to the physical and trophic state of the water.

INTRODUCTION

The measurement of the particulate beam attenuation coefficient (cp) allows the observation of oceanic suspended particles in the size range from 0.1 to 100 m [Stramski and Kiefer, 1991] encompassing a wide variety of living (phototrophes and heterotrophes) as well as non-living (organic detritus and minerals) particles.

A daily cycle in cp, characterized by a diurnal increase and a nocturnal decrease, appears to be a near ubiquitous feature of cp in the world’s oceans. Observations of this feature span many regions of the Pacific [Bishop, et al., 1999; Claustre, et al., 2007; Claustre, et al., 1999; Cullen, et al., 1992; Gardner, et al., 1995; Siegel, et al., 1989] and Atlantic [Gardner, et al., 1993; Marra, 1997; Stramska and Dickey, 1992; Stramska, et al., 1995] oceans as well as the Arabian [Kinkade, et al., 1999] and Mediterranean [Oubelkheir, 2001] Seas.The diurnal increase in the attenuation coefficient has been related to photosynthetic organic production by phytoplankton [Siegel, et al., 1989]. Such an increase is expected whether from an increase in cell size and/or refractive index of the cells [Claustre, et al., 2002; DuRand, et al., 2002; DuRand and Olson, 1998; Stramski and Reynolds, 1993; Stramski, et al., 1995]. Similarly, the nocturnal cpdecrease has been related to the decrease in cell size and refractive index via respiration and cell division. However, in contrast to fluorescence measurement, cp is not only influence by phytoplankton and other organisms are likely contributing to the daily variability in cp. It has been suggested that other organisms may contribute significantly, this was supported by coincident maximum in the daily cp variability and maximum of bacteriochlorophyll[Claustre, et al., 2007]. While the increase in cp is very likely controlled by photosynthetic carbon accumulation, loss processes are not restricted to respiration and include grazing [Cullen, et al., 1992], particles aggregation and sinking. Moreover, diurnal and episodic mixing events contribute to cp variability [Gardner, et al., 1995].

The amplitude of the daily cycle in cp varies widely, from 10% to more than 100% of the morning cp value, but is around 20-30% in most in-situ observations. In addition to the photosynthetic yield of phytoplankton, the daily cp amplitude depends on the proportion of living organisms to non-living particles, the daily cp amplitude decreasing when the proportion of detritus is increasing.The size distribution of the phytoplanktonic community also influences the daily cp amplitude. Laboratory studies have shown that the diurnal increase is more important for picophytoplankton [Claustre, et al., 2002; DuRand and Olson, 1998] than for micro- and nanophytoplankton [DuRand, et al., 2002; Stramski and Reynolds, 1993], confirming in-situ observation[Durand and Olson, 1996]. Daily cp characteristics depend also on the physiological state of the phytoplankton.

Most in-situ studies have been conducted from ships. Therefore, observations of the daily cp variability are temporally limited to a few days. The development of instrumented moorings and autonomous profiling devices has increased the resolution and the duration of the observations [Bishop, et al., 2002; Dickey, et al., 2006; Dickey, et al., 1991]. In the North Western Mediterranean Sea, at an oceanic site, the BOUSSOLE mooring allows continuous recording of the surface optical variability since September 2003 [Antoine, et al., 2006; Antoine, et al., 2008a]. It provides a comprehensive data set of optical measurements, including cp, chlorophyll fluorescence and particulate backscattering bbp.

The site is characterized by a strong seasonality in the environmental and trophic conditions [Antoine, et al., 2006; Bosc, et al., 2004; D'Ortenzio, et al., 2005; Marty and Chiaverini, 2002; Morel and André, 1991]. Variability in cp has been previously reported as the results of changes in particle abundance and composition[Oubelkheir, 2001]. The timeseries of cprecordedat 4 and 9m in 2006 and 2007presents an important variability at two different time-scales: at the seasonal (Figure 1) and at the daily (Figure 2) time-scale. The goal of this study is to describe the features of the daily cp variability as it varies during the course of the seasons.

DATA AND METHOD

All data are collected within the frame of the BOUSSOLE project, which consists of an instrumented mooring and monthly visits to the site. Data collection started in July 2001. Information about the project is available on the web site A general presentation of the project [Antoine, et al., 2006] as well as a description of the mooring itself [Antoine, et al., 2008a] have been published. All the data used in this study have been collected between the 1stJanuary 2006 and the 31stDecember 2007. Underwater instruments are fixed at two depths, whose nominal values are 4 and 9m.

Measurement site

The BOUSSOLE site is located in the Ligurian Sea, one of the basins of the North Western Mediterranean Sea, at 60 km offshore from the coast. Water depth at this open ocean site is 2440 m, and waters are permanently of the Case 1 type[Antoine, et al., 2008b], following the definition of [Morel and Prieur, 1977]. Ocean currents are usually weak (< 20 cm.s-1), because the selected position is in the central area of the cyclonic circulation that characterizes the Ligurian Sea [Millot, 1999].This is important for the study of temporal variability at the daily scale; most of the variability does not originated from changes in water masses.

Monthly cruises measurements

The density () is measured with a Seabird SBE 911 plus ® CTD. The mixed layer depth (zm) is computed using a = 0.125 kg.m-3 criterion. The averaged value of zm is computed for each cruise.

The beam attenuation coefficient (c) is measured using a 25 cm pathlength WetLabs C-star ® transmissiometer (660nm). Influence of particle and dissolved material absorption is negligible at this wavelength. Therefore, the particulate beam attenuation coefficient (cp) is equal to c minus the contribution of pure sea water(cw).The factory calibrated value of cw is 0.364m-1. In practice however, in order to correct for uncertainties in calibration, cw is determined for each profile as the attenuation value between 350 and 400 m[Loisel and Morel, 1998].

Automatic buoy measurements

A comprehensive dataset of inherent and apparent optical properties is collected on the BOUSSOLE buoy. In this extended abstract, only measurements of cpand Photosynthetic Available Radiation (PAR) are presented. All measurements are performed every 15 minutes.

cp is measured with a 25 cm pathlength WetLabs C-star ® transmissiometer (660nm), equipped with copper plates to prevent from biofouling contamination. Buoy cp measurements are adjusted to monthly cruise values in order to correct for instrumental drift. Downward irradiance is measured above the water and underwater at seven wavelengths (442, 490, 510, 554, 560, 665 and 683 nm) using Satlantics OCI-200 ® radiometers. An approximate value of PARis computed by discrete integration over the seven measured wavelengths. Nominal depth (zN) and water temperature (T) are measured witha Seabird SBE 37SI ® CTD.

Data selection

Buoy data are measured from a fixed position. The buoy is mostly stable, but it may move (being tilted or plunged) under strong currents. zNis usually at 10m but may increase up to 20m ore more. During2006 and 2007, 383 complete days have been collected. Days when zNwas deeper than 11m have been removed from the analysis. Moreover, a visual inspection of each day reveals that unwanted variations in cp occurred in consequence of occasional changes in zNand/or T as the results of buoy plunges and/or of the intrusion of water masses with different particles load. From the 383 days of the collection period, 274 days (i.e. 72%) have been selected for the analysis. This represents a dataset of about 26000 cp measurements.

Data processing to characterize the daily variability

For each solar day, the median is computed. Moreover, the sunrise, sunset and next sunrise cpvalues are computed as the median of cp values +/30 min around the times of sunrise, sunset and next sunrise. These quantities are used to derive the daytime gross, nocturnal loss and net daily variation (respectively Gcp, Lcp and Ncp, after [Claustre, et al., 2007; Siegel, et al., 1989]). Gcp, Lcp and Ncp are expressed in % in defined by:

RESULTS

Seasonal variability

The cp seasonal variability follows those of the trophic regime (Figure 1). In winters, low chlorophyll concentrations [Antoine, et al., 2006; Marty and Chiaverini, 2002]and low cp values are associated with deep mixing [D'Ortenzio, et al., 2005]. The development of a vernal bloom leads to an increase in the chlorophyll concentration and progressive stratification. After the bloom, the situation evolves toward oligotrophy, with a stratified and nutrient depleted surface layer above the deep chlorophyll maximum. Surface waters present low cp values. Oligotrophy lasts during summer. Episodic breakings of the stratification, associated with strong winds, may occur. In autumn, the erosion of the thermocline precedesthe situation of deepmixing in winter.

Figure 1. Timeseries of cp(9 m) and zm in 2006 and 2007.

[Blue thick line is the monthly cruise-averaged zm. In February 2006, zm is deeper than 400 m. Black thick line is, whereas grey line show all cp measurements.]

Both years 2006 and 2007 presents the same seasonal features, but their amplitude is totally different (Figure 1). In 2006, cp spanned two orders of magnitude between the winter and the bloom, varying from 0.01 to 1 m-1. In 2007, cp spanned only one order of magnitude between the winter and the end of the bloom, varying from 0.1 to 1 m-1. The cp value during the winter 2006 is exceptionally low. It is likely the result from a vigorous mixing that “washed-up” the water column from the surface to the bottom. Depth profiles of cp were constant from 0 to 200m (data not shown). On the contrary, during the winter 2007, a negative gradient from 0.1 m-1 in surface to 0.01 m-1 at 200 m was observed. During the winter 2007, the mixed layer depth never got deeper that 300 m. During the summer, in 2006 and in 2007, cp values are similar and remained around 0.1 m-1.

Daily variability

In a first part, three examples of four days cptime series are presented during the different seasons: winter, spring and summer. In a second part, the time series of the daytime gross(Gcp) and night-time loss(Lcp) variations is presented.

Figure 2. Examples of cpduring winter (top), spring (middle) and summer (bottom).

[ is in dashed line and PAR is in red. Grey rectangles indicate night-time].

Examples of daily variability atdifferent seasons

During the four days selected as examples for winter, remains stable, around 0.015 m-1. Gcp varies between 20 and 30 % (not shown graphically), with an exception at 70 % on the 9th February 2006 (Figure 2, top). The shape of the cycles is not regular and not synchronised with sunset and sunrise. During the four days representatives of the bloom in 2007 (Figure 2, middle), increases by 300 % (from 0.15 to 0.6 m-1) between the 4th and the 8th April. Daily cycles are superimposed over the general trend, with Gcp values increasing from 70 % the 4th April to 140 % the 8th April. During the four days representatives of the summer, remains stable, around 0.09 m-1(Figure 2, bottom). Gcp varies between 6 and 16 %.

Some preliminary conclusions can be drawn from these examples. First, a daily cp cycle is observed whatever the season. Second, its amplitude seems to vary with the season. Third, its shape is not regular and is not systematically synchronised with sunrise and sunset.

Time series of Gcp and Lcp

Gcp shows an important variability (Figure 3). Seasonal features are observed. In winter, Gcp varies between 0 and 30%. Gcp increases up to 70 % during the bloom. After the bloom,Gcp decreases down to 20% and remains at a low level during summer.

In order to compare Gcp with Lcp, it is useful to consider the variations of Ncp, which is equal to Gcp minus Lcp.Ncpvarieslike the first derivate of. Ncp is zero whenis stable and Ncp is positive (respectively negative) when increases (respectively decreases). Concomitant positive Ncp and increasing values are observed during the development of the bloom, whereas negative Ncp and decreasing values are observed during the bloom collapses. In winter and in summer, is fluctuating and Ncp is alternatively positive or negative.

The blooms in 2006 and 2007 are not similar. In 2006, the bloom starts the 15th March and lasts almost two months. It develops from the 15th to the 5th April and extends until the 5th May.Discrepancies between Gcp, Lcp andoccur. In particular, at the end of April, Gcp and Lcp started to collapse one week before . After an apogee, the bloom definitively ends when Ncp became negative after the 5th May. Such a phenomena (i.e. anticipated decrease in Gcp and Lcp) previously occurred between the 1st and 15th April, with the difference that Ncp remained positive. It was immediately followed by a strong parallel increase in Gcp and Lcp that maintainshigh values of .

In 2007, the bloom starts earlier, around the 15th February. Until the 10th March, the situation is similar to that in 2007: between the 15th February and the 5th March, Gcp and Lcp increase in parallel, with Gcpsignificantly greater than Lcp. Between the 5th and the 10th March, is still increasing but Gcp and Lcp decrease.The bloom is abruptly interrupted from the 15th to 20th March. Gcp is still greater than Lcp, but both values are negative. The bloom then develops and lasts untilthe 15th April, when Ncp becomes negative. As in 2006, the collapse of the bloom is anticipated by a decrease in Gcp and Lcp.

Figure 3. Time series of Gcp and Lcp in 2006 (top) and in 2007 (bottom).

[ is in thick black line. Gcp and Lcp are the thin grey lines. The area between Gcp and Lcphas been colored. When filled in grey(respectively in black), itrepresents the days when Ncp (i.e.Gcp minus Lcp) is positive (respectively negative). Note that Gcp and Lcp have been averaged using a three days running mean.]

DISCUSSION AND PERSPECTIVES

The development of the bloom is associated with the particular situation where Gcpis significantly over Lcp, as it happened in mid-March 2006 and in the beginning of April 2007. During such a phase, the daytime photosynthetic carbon fixation (i.e. “new” production) is maximal, whereas night-time losses are probably reduced to cell respiration. On the contrary, the end of the bloom is associated with the situation where Lcp is over Gcp. The reasons why night-time losses became more important than daytime growth is still unknown.

In spite of higher PAR values, Gcp is minimal in summer. This is somehow surprising, because greatest daily cp amplitude has been reported for pico- and nanoparticles, typical of oligotrophic situations. At this time of the year however, the contribution of detritus to cp is expected to be maximal [Oubelkheir, 2001], which could explain why Gcpappears so low.

All bio-optical mechanisms that influence cp variability are not accessible by the routine automatic measurements performed at the BOUSSOLE site. However, futureanalyses of the interactions between night-time fluorescence and cptime series willhelpthe interpretation.

REFERENCES

Antoine, D., et al. (2006), BOUSSOLE: A Joint CNRS-INSU, ESA, CNES and NASA Ocean Color Calibration and Validation Activity, Technical Memorandum, 59 pp, NASA.

Antoine, D., et al. (2008a), The "BOUSSOLE" buoy. A new transparent-to-swell taut mooring dedicated to marine optics: design, tests and performance at sea, J. Atmos. Ocean. Technol.

Antoine, D., et al. (2008b), Assessment of uncertainty in the ocean reflectance determined by three satellite ocean color sensors (MERIS, SeaWiFS and MODIS-A) at an offshore site in the Mediterranean Sea (BOUSSOLE project), J. Geophys. Res., 113.

Bishop, J. K. B., et al. (1999), Spatial and temporal variability of POC in the northeast Subarctic Pacific, Deep-Sea Research Part Ii-Topical Studies In Oceanography, 46, 2699-2733.

Bishop, J. K. B., et al. (2002), Robotic observations of dust storm enhancement of carbon biomass in the North Pacific, Science, 298, 817-821.

Bosc, E., et al. (2004), Seasonal and interannual variability in algal biomass and primary production in the Mediterranean Sea, as derived from four years of SeaWiFS observations, Global Biogeochemical Cycles, 18, doi:10.1029/2003GB002034.

Claustre, H., et al. (2002), Diel variations in Prochlorococcus optical properties, Limnology And Oceanography, 47, 1637-1647.

Claustre, H., et al. (2007), Gross community production and metabolic balance in the South Pacific Gyre, using a non intrusive bio-optical method, Biogeosciences, 5, 463-474.

Claustre, H., et al. (1999), Variability in particle attenuation and chlorophyll fluorescence in the tropical Pacific: Scales, patterns, and biogeochemical implications, Journal Of Geophysical Research-Oceans, 104, 3401-3422.

Cullen, J. J., et al. (1992), Photosynthetic Characteristics And Estimated Growth-Rates Indicate Grazing Is The Proximate Control Of Primary Production In The Equatorial Pacific, Journal Of Geophysical Research-Oceans, 97, 639-654.

D'Ortenzio, F., et al. (2005), Seasonal variability of the mixed layer depth in the Mediterranean Sea as derived from in situ profiles, Geophysical Research Letters, 32.