Microalgal pigments and their relation with phytoplankton carbon biomass on the northeastern Mediterranean Sea shore, with special emphasis on nanophytoplankton

Summary Marker pigments are used as a proxy for biomass of distinct phytoplankton classes in different oceanic regions. However, sometimes disagreements are observed between microscopy and accessory-pigment based approaches in distinct regions mainly due to changing environmental factors governing diversity and structure of community composition. In this study, concordance between microscopy and HPLC-CHEMTAX methods were investigated first time in coastal waters of Erdemli, Turkey, in the Levantin Basin of the northeastern Mediterranean Sea by weekly intervals during 2015-2016. According to our results, marker pigment of diatoms, fucoxanthin, which was the most prominent pigment in the study area during most of the year, was a better indicator of diatom abundance than diatom carbon biomass. CHEMTAX derived values of diatom chlorophyll a (Chl a) were not in concert with either abundance or carbon biomass of this group. Contribution of dinoflagellates and cryptophytes to the phytoplankton community was underestimated with pigment based approach. Accessory pigment of cyanophytes, zeaxanthin, was also an important pigment in the samples. Biomass of haptophytes seemed to be overestimated by HPLC-CHEMTAX analysis. In contrast to diatoms, CHEMTAX derived chlorophyll a values of cryptophytes were correlated with abundance of this group but not with alloxanthin. Inclusion of live counts of nanoplanktic cryptophytes, haptophytes and prasinophytes provided a better correlation between microscopy and pigment based results. According to CHEMTAX analysis, nanoplankton and picoplankton constituted ∼55% of Chl a in the region.

Pseudoscourfieldia sp. and distinct cryptophyte species were counted as live. Some of the 27 Chrysochromulina species in live counts could be mixotrophic. 28 Temperature and salinity were measured with a WTW LF330 model conductivity meter. 29 The volume (V) of each cell (100-1000 cells) was calculated by measuring its appropriate   The CHEMTAX 1.95 program (excel version, [59,60] was applied in order to obtain the 3 taxonomic composition of phytoplankton from marker pigments. The 3 input matrices are: (1) the 4 measured concentration of marker pigments and Chl a in the samples; (2) the theoretical input ratios 5 of marker pigments to Chl a for each class of phytoplankton to be quantified; and (3) a ratio limit 6 matrix restricting the iterative adjustments of these ratios operated by CHEMTAX. By using Matrix 1 7 and 2, CHEMTAX iteratively modify the difference between observed and calculated total pigment 8 concentration utilizing changes in pigment ratios.

12
Initial accessory pigment:Chl a ratio matrices were shown in Table 2 The input pigment:Chl a ratio matrices used here included Chl c2 and But but excluded Ddx and 20 Caro, which are shared by many phytoplankton groups [14]. Chl c2 rather than Peri was significantly 21 correlated with Dino-C (p<0.05, r 2 =0.86), thus this pigment was included in the input ratio matrix.

22
For adjusting potential changes in pigment:Chl a ratios, data set was separated to two parts based 23 on high and low C:Chl a ratios as threshold being 8 (see also Appendix 4b, on which whole data set 24 was run in CHEMTAX analysis).

25
The parameters set for the calculations were as follows: ratio limits were set to 500, weighting 26 was 'bounded relative error by pigment', iteration limit = 5000, epsilon limit = 0.0001, initial step size 27 = 25, step ratio = 2, cutoff step = 3000, elements varied = 5, subiterations = 1, weight bound = 30 [59,  Linear regression analysis was performed in order to assess the relationship between random 30 variables (pigment and/or carbon concentrations).    According to annual average pigment concentrations, CHEMTAX assigned Chl a concentrations, 7 phytoplankton carbon biomasses, diatoms were the most important group in the study area (Fig. 3).

8
Although dinoflagellates was the second important group in terms of carbon biomass (Fig. 3d), 9 unambiguous marker pigment of this group, Peri, was among the marker pigments having the 10 minimum concentration. However, Chl c2 could be partially associated with dinoflagellates since 11 there was a positive correlation between Dino-C and Chl c2 (r 2 = 0.86, p<0.05). When CHEMTAX 12 assigned Chl a values of phytoplankton groups and microscopy based carbon data were compared, 13 contribution of haptophytes to the total Chl a seemed to be overestimated while the contribution of 14 cryptophytes was underestimated ( Fig. 3b and d). Carbon biomass of chlorophytes was higher than 15 of prasinophytes due to a few big chlorophyte species such as Halosphaera viridis or some 16 filamentous species (Fig. 3d) Figure 3 Annual average percentage contributions of (a) pigments to total accessory pigments (b) 5 CHEMTAX derived phytoplankton groups to chlorophyll a (c) phytoplankton to total abundance and 6 (d) phytoplankton to total carbon biomass during September 2015-September 2016 at the sampling 7 location.  Table 1.

12
Average Chl a concentration during the sampling was 1.19 ± 1.5 µg L -1 (n=50). Chl a values were 13 <1 µg L -1 within half of the samples. When the whole dataset was taken into account, there was not 14 any correlation between carbon and Chl a concentrations (p>0.05) during the sampling dates. 15 However, when the samples were separated as high (8) and low (<8) C:Chl a ratios, a significant 16 correlation in the high and low C:Chl a samples (r 2 =0.6, p<0.05, n=11 and r 2 =0.5, p<0.05, n=39, 1 such as light, temperature, species composition and nutrients control the variation of this ratio in 2 different time periods, it is better to divide the dataset in timeseries studies. Average C:Chl a ratios 3 were 32 ± 23 and 3 ± 1.6 in the high and low C:Chl a samples, respectively. The highest Chl a 4 concentration observed on 3 February 2016, did not correspond to the highest carbon biomass during 5 the sampling period (Fig. 4). Despite relatively low carbon biomass on 3 February 2016, Chl a 6 reached to the highest level. High carbon values seen between January and May were mainly due to 7 diatoms. The lowest phytoplankton abundance, carbon biomass and Chl a concentrations were 8 observed in December-early January. Increase in Chl a concentration on 1 July 2016 was due to 9 cyanophytes inferred from the rise in Zea concentration. growth phases of this species on these dates, respectively. (Fig. 6).

23
Intraclass variations of Fuco:Diat-C were prominent during January, April and other months.

24
Minimum values of Fuco:Diat-C were observed in January and April when big sized diatom species 25 Trieres mobiliensis and Proboscia alata were dominant respectively (Fig. 5, 6). The highest Fuco:Diat- Diat-C or Diat-A (Table 3). Dividing diatoms to two distinct groups in CHEMTAX input ratio did not 2 provide a better correlation between Diat-Chl a and Diat-C or Diat-A (Appendix 3, supplemantary 3 material). Table 3 Relationship between Group-A-Group-C and marker pigments-Group-Chl a values in the 6 samples having high (8) and low (<8) C:Chl a ratios. Cyanophytes were not correlated with Zea or 7 DVChl a since picoplanktic cyanophytes were not counted.    (Table 3, Fig. 6). During bloom concentration of the prasinophyte Pyramimonas spp. on 2 10 September 2015, Lut concentration was relatively high, while Chl b could not be detected (Fig. 6).  Zeaxanthin has demonstrated that cyanophytes increased at the end of May and remained high until 10 the beginning of August (Fig. 6). Zea was an important component of accessory pigments composing 11 14% of annual average accessory pigments. CHEMTAX allocated Cyano-Chl a constituted 25% of 12 average Chl a.

13
DVChl a containing cyanophyte Prochlorococcus sp. was separately shown as Phrochloro-Chl 14 a in CHEMTAX analysis. They reached peak values during winter-spring period (Fig. 6).   (Table 3). High number of diatom species most probably having 31 distinct marker pigment:carbon ratios (pls see section 3.5.1 in results) and variations in their abundance 32 related to changing nutrient, light and temperature levels seem to complicate estimation of diatom 33 carbon biomass via CHEMTAX analysis in the study region. Low fucoxanthin content of big-sized 1 diatoms could easily be inferred from differences in peak levels of diatom carbon biomass and 2 fucoxanthin maxima during the study especially in April (Figs. 6, Appendix 4a, c).

3
Second important group in terms of CHEMTAX analysis following diatoms was cyanophytes 4 especially during warm period (Fig. 3b, Appendix 4b). Although carbon biomass of haptophytes was 5 low, they were among major components of pigments and CHEMTAX derived Chl a following 6 cyanophytes. Similar to diatoms, marker pigment of haptophytes, rather than group specific Chl a, 7 were correlated with carbon biomass of this group. This could also be attributed to species diversity of 8 haptophytes. The highest pigment diversity was found among haptophytes represented by five with carbon biomasses obtained from microscopy for these groups (Table 3). However, carbon 12 biomass of cryptophytes appeared to be underestimated with HPLC-CHEMTAX based approach.

13
Similar to the present study, CHEMTAX seemed to underestimate cryptophyte biomass but

32
In this study, average contribution of Cyano-Chl a attributed by CHEMTAX to the total 33 accessory pigments was estimated as 28% (Prochloro-Chl a + Cyano-Chl a, Fig. 3b). In the Adriatic The highest Chl a values of cyanobacteria derived by CHEMTAX were observed during July and 2 August in the present study which was consistent with flowcytometer results in a nearby location [29].