Black Sea

Satellite data


The Black Sea as a whole basin was divided into eight sub-regions (Kopelevich et al. 2002b). Six of them (#1-5 and #8) lie in the shelf area (depth less than 200 m), the sub-regions #6 and #7 are deep open parts of the basin (Fig. 15). Six regions in the shelf area were considered separately because they were distinguished with respect to their position, bathymetry and influence of river run-off. Regions #1-3 have depth of less than 50 m; region #1 is under strong influence of water discharge of the Dnepr, Dnestr and Bug rivers, whereas region #2 is influenced mainly by the Danube river. Regions #4 and 5 are the northern and southern outer shelf regions (depth of 50-200 m) of the western part of the Black Sea; the southern, eastern and northeastern shelf regions of the eastern part of the basin were combined and evaluated as a single region (#8).

The general circulation of the Black Sea consists of the strong Rim Current (
Oguz et al. 1993) following the narrow continental slope; the Rim Current limits the water and material transfer from coastal zone to the open Black Sea. The interior of the Rim Current zone is formed by two cyclonic cells occupying the western and eastern halves of the basin; for this reason the western (#6) and eastern (#7) open parts of the Black Sea were considered separately.

Fig_16a Fig_16b

Fig.16 shows a monthly coverage of the sub-regions of the Black by satellite data from SeaWiFS and MODIS-Aqua, given by different hatching. As seen, for all of them the coverage was rather full.


The bio-optical algorithms, presented in this Section, were not changed as compared with the previous issue (
Kopelevich et al. 2011c). The new algorithm is used for calculation of coccolithophore concentration Ncoc (see Bio-optical characteristics of the Black Sea in June).

Chlorophyll concentration

From SeaWiFS data, chlorophyll concentration in the sub-regions 1-5 was calculated with the algorithm developed by specialists from MHI NANU (Suetin et al. 2000, 2001, 2002):

Chl= 1.13 [LNW(510)/LNW(555)]-3.33.                                                         (7)

For the sub-regions 6-8 the algorithm developed by specialists from SIO RAS (Burenkov et al. 2002) was used:

Chl= 0.88 [LNW(510)/LNW555)]-2.24.                                                            (8)

Comparison between the two algorithms has shown that the Chl–SIO values (derived with the algorithm (8)) are higher than the Chl–MHI values (derived with the algorithm (7)) if Chl–SIO < 0.4 mg·m-3. TheChl–SIO values are lower than the Chl–MHI values if Chl–SIO > 0.7 mg·m-3. Within the Chl–SIO range from 0.4 to 0.7 mg·m-3 the Chl values derived by the two algorithms differ by less than 15%.

The both algorithms are assumed to be valid in warm season (May-September). They can result in considerable errors in absolute values of Сhl in cold season. However, it can be assumed that relative changes of chlorophyll concentration are reproduced adequately with the both algorithms, and for this purpose our computations were also carried in cold season.

From MODIS-Aqua data, the algorithm to derive chlorophyll concentration in the sub-regions 1-5 was complicated:

Chl = 0.5 (Chl1 + Chl2),                                                                          (9)

where Chl1 = 1.13 [0.66 RRS(488)/RRS(547) + 0.40]-3.33;

Chl2 = 1.13 [2.35 RRS(531)/RRS(547) – 1.44]-3.33.

For the sub-regions 6-8 the Chl algorithm with MODIS data was similar to (8):

Chl = 0.83 [0.996 RRS(531)/ RRS(547)]-4.36.                                                        (10)

Particle backscattering

For calculation of the particle backscattering coefficient b
bp in the Black Sea from SeaWiFS data, the simplified algorithm described in Section Barents Sea Algorithms was used, but Kd(555) was calculated through the ratio [LNW(490)/LNW(555)] instead of LNW(510)/ LNW(555) in the Barents Sea.

Computations of the particle backscattering coefficient bbp in the Black Sea from MODIS-Aqua data were performed by the same algorithm as with SeaWiFS data but the ratio of LNW (488)/LNW (547) was used instead of LNW(490)/LNW (555).

Yellow substance absorption

In the Black Sea it is possible to retrieve the values of the yellow substance absorption coefficient ag(440). The version of a semianalytic algorithm described in Burenkov et al. (2001b) was used. It is difficult to estimate errors of the derived values of ag because they depend strongly on errors of the atmospheric correction. Of course, the cases with negative values of LNWwere rejected out, but we could not check the errors in LNW which were not so pronounced.

Results and discussion

The mean monthly distributions of chlorophyll concentration and the particle backscattering coefficient 
bbp in the Black Sea from January 1998 to December 2012 are presented on the color maps; the mean monthly distributions of the yellow substance absorption coefficient ag for these years are given on these maps only for warm season (May-September). As seen, the northwestern part of the sea exposed to the influence of the runoff of the Dnieper, Dniester, Bug, and Danube rivers (regions 1 and 2) is distinguished by enhanced values of all of the considered characteristics. A clearly manifested impact of the Caucasian and Turkish rivers is restricted by a narrow near-shore band along the eastern and southern coasts of the sea.

The influence of the river runoff from the northwestern coast mostly extends southward along the coast; the transverse propagation is significantly weaker. The latter is mainly implemented by the mesoscale eddies permanently observed in the Black Sea.

Fig_17a Fig_17b  Fig_18a Fig_18b

Fig. 171819 show a variability of the monthly means of the bio-optical characteristics in different regions. A similarity between seasonal changes of all characteristics is observed in the regions #1-3 being under strong influence of the river run-off. The result is easy to understand: rivers bring to a coastal zone both particulate and dissolved matter as well as nutrients which stimulate primary production and increasing chlorophyll concentration.

Fig_19a Fig_19b

The variations of Chl values in the regions #6, 7 are weakly connected with the ones in the western regions; they correlate with each other (#6 and #7), and Chl in the region #7 correlates with Chl in the region #8. The latter can be explained by similarity of natural conditions governing by variability of primary production and chlorophyll concentration in the central and eastern regions of the Black Sea.

According to
Demidov (2001), who analyzed seasonal variability of chlorophyll concentration in the Black Sea since 1960 to 1997, there were three maxima observed in its seasonal changes: in winter-spring (January-March), summer (June-August) and autumn (October-November). The satellite data, presented in Fig. 17, are in agreement with that, but they show that number of seasonal maxima and the months with the maxima can be changed. In the region # 2 the principal maximum was most frequently observed in June, but they also occurred in other months (for example, in January in 2005). In the open Black Sea they were observed in October-December with exception of 2001 in June (# 6) and in May (# 7). Magnitudes of the Chl principal maxima and minima in the open sea, excluding 2001, were rather stable.

It is a surprising result that the 
bbp variations in the open parts and in the coastal regions are almost synchronous. The enhanced values of bbp covering the whole basin are observed repeatedly in June; this problem is discussed in section Bio-optical characteristics of the Black Sea in June.

Fig_20a Fig_20b

A variability of the monthly means of SST is shown in Fig.20. As it is demonstrated in Burenkov et al. (2011a), changes in winter SST influence an intensity of spring phytoplankton bloom and the June maximum of the particle backscattering coefficient from sea surface temperature.

Inter-annual variability

The seasonal (for chlorophyll concentration and the yellow substance absorption coefficient) and annual (for the particle backscattering coefficient) mean values with their standard deviations in different regions of the Black Sea are given in 
Table 3.

As seen, the mean values of chlorophyll concentration almost in all sub-regions were minimal in 2003 and 2011; they were highest in most regions in 2001. As for the particle backscattering coefficient, the minimal mean annual values in the most sub-regions were in 2011.

Bio-optical characteristics of the Black Sea in June

Field and satellite studies, carried out in the north-eastern part of the Black Sea in 2004-2012, showed that the Chl values and spatial distributions of bio-optical parameters in June strongly differed from other seasons. There were two hypotheses to explain this phenomenon: (i) the particulate matter from the river runoff spreading to the whole basin across the boundary Rim Current by mesoscale eddies and via turbulent exchange; (ii) coccolithophore bloom covering the whole basin (Cokacar et al. 2001, 2004). A lot of concurrent field measurements of water-leaving radiances  LW and seawater constituents (total suspended matter TSM, chlorophyll-a Chl and coccolithophore concentrations Ncoc) were made. That allowed to develop regional algorithms for June (using data of ocean color scanner SeaWiFS). Following empirical relationships were obtained and used for computations in our previous CD (Kopelevich et al. 2011c):

Chl = 0.73 [LWN (510)/LWN (555)]-2.73                                        (11)
n=81, r2=0.54;

TSM = 72.4 bbp +0.11,                                                          (12)
n=48, r2=0.25;

Ncoc=768 bbp1.55,                                                                (13)
n=48, r2=0.54,

where n is number of subsatellite points, r2 is coefficient of determination, Ncoc is coccolithophore concentrations in 106 cells/l, bbp(550) is measured in m-6 and TSM in mg/l.

Note that Chl algorithm for June differs from the algorithm described in 4.2. This can be associated with difference in phytoplankton composition in June and autumn.

The algorithm to derive chlorophyll concentration in June from MODIS-Aqua data uses the ratio of 

Chl = 0.89 [0.996RRS(531)/RRS(547)]-6.0.                                          (14)

In this issue the Chl and TSM concentrations were also calculated with (11)-(12) and (13)-(14), but a new algorithm was developed for Ncoc. The absorption coefficient of yellow substance ag was also calculated.

The obvious drawback of the previous algorithm for Ncoc was the non-calcite related backscattering is not accounted for. An adequate model for the particle backscattering coefficient bbp in the Black Sea must include not only the coccolithophore concentration Ncoc but a parameter characterizing the river runoff. For development of satellite algorithm, the second parameter should be also derived from satellite data. A new idea of evaluation of the “non-coccolithophore” component of bbp (denote it as bb_riv) was based on an assumption that the values of ag and bb_river are correlated with each other, and bb_riv can be represented through ag. This assumption was supported by results of our field studies (Kravchishina et al. 2013, Kopelevich et al. 2013a).

On this basis, a new model has been developed and it takes the form:

bbp - bbp_bg = Kcoc Ncoc + Kriv(ag – ag_bg),                                        (15)

where bbp_bg and ag_bg are the background values of bbp and ag, derived from satellite data over 2003-2010 as their lowest monthly means; bbp_bg = 0.0025 м-1, ag_bg = 0.047м-1; the coefficients Kcox and Kriv were determined from the statistical characteristics of the Ncoc values from direct measurements of Ncoc and of bbp , ag calculated from data of our floating spectroradiometer (Artemiev et al. 2000) measured in 2004-2008 (Kopelevich et al., 2013b); Kcoc=2.74 ·10-3, Kriv=0.157.

As seen from (15), for calculation of Ncoc we need to know bbp and ag. For calculation them, values of the remote sensing reflectance RRS(488) and RRS(555), derived from satellite data, were used as the input parameters.

For bbp and ag the low-parametric models were used with taking into account the spectral dependences of the backscattering by coccolithophore and river particles (Kopelevich et al., 2013a,b).

Finally, we obtained a linear system of two equations with two unknowns ag and Ncoc and RRS(488) and RRS(555) as the input parameters

f1(λi)ag + f2(λi)Ncoc = g, i=1,2                                                       (16)

The results of our calculation are presented as color maps and Figures 21-24. The color maps are given for all area of the Black Sea, even though the algorithms were developed with data for the north-eastern part only. So it should be kept in mind that for other sub-regions the data presented are not reliable.

Fig_21   Fig_22

Fig. 21, 22, 23, 24 show a variability of the June mean concentrations of ChlTSM, Ncoc and ag in the eastern subregions # 7, 8 from 1998 to 2012.

Fig_23   Fig_24

One can see that there is no correlation between inter-annual variability of Chl  and  TSM, whereas the TSM  and Ncoc changes are very close to each other. But there were years with no marked coccolithophore blooms (2003, 2010), but the high values of bbp were also observed in those years (Fig.18b); according to our model, they should be attributed to the bb_riv contribution.

The coccolithophore blooms are more changeable in comparison with the river runoff which is more stable; it is well seen from comparison between Fig.
23 and 24.

The highest values of
Ncoc in both sub-regions were observed in 2012. Burenkov et al. (2011a) demonstrated that the inter-annual changes of intensity of coccolithophore blooms in June can be linked to the winter sea surface temperature, and they gave some explanation of that. It should be noted that the highest value of Ncoc in 2012 also corresponded to cold winter, though the lowest SST value was observed in February 2006.