Several oil spills have occurred in different marine ecosystems around the world in the last 30 years (Varela et al. 2006; Soto et al. 2014; Duda and Wawruch 2017; Tang et al. 2019). Assessing the effects of these accidents on marine phytoplankton is crucial, considering its basic status in marine food webs and its determining role in material and energy fluxes. Reports on in situ phytoplankton community changes at the early stages of an oil spill and throughout its evolution are rare, even if abundant data from experiments with single species or natural communities are available (Varela et al. 2006; González et al. 2009; Bretherton et al. 2019a, 2020; Putzeys et al. 2022). Microbial communities are rapidly altered after an oil spill incident (Gemmell et al. 2018), so that early monitoring is efficient when in-situ responses to oil pollution are targeted. Therefore, to improve knowledge on the effects of oil spills, it is important to consider rapid and longer responses of in situ phytoplankton following an oil leak. This work presents first field data on the responses of a coastal phytoplankton community to an oil spill from the early stage (the first days) up to a few weeks in the south-western Mediterranean Sea.
The oil leak site was located in the Bay of Bizerte, while the sampling station was located in the Channel of Bizerte (Fig. 1). The wind direction and speed moved the oil slick towards the Bizerte City coast and to the Channel. This was reflected by the presence of PAHs in the sampling station 8 days after the spill. Referring to the WHO 1997 standard (700 ng L− 1), seawater was moderately contaminated (ΣPAHs ≈130 ng L− 1). Some of the measured PAHs (chrysene, fluoranthene, naphthalene, phenanthrene) could have a toxic effect on phytoplankton even at low concentrations (Echeveste et al. 2010a, 2010b; Ben Othman et al. 2017). The PAH water concentrations measured on D08 had certainly been diluted. Therefore, higher levels would have been measured on D01 and D04 if sampled, and phytoplankton was most probably exposed to higher concentrations on D01 and D04 than those measured on D08. PAH levels were much higher in sediment (ΣPAHs = 1222 ng g− 1 dry sediment, Table 1) than in seawater. As a result of their hydrophobic nature, PAHs are known to be sequestered within sediment (Mille et al. 2007; Bradshaw et al. 2012), where it could be degraded partially by bacteria (Acosta-Gonzàlez et al. 2015). Historical data showed that the sediment of Bizerte Channel had been exposed to PAH contamination, but at levels much lower than those recorded after this oil spill (394.1 ng g− 1 dry sediment, Barhoumi et al. 2014).
Impact of oil on the growth of size-fractioned phytoplankton
Phytoplankton can respond rapidly to oil pollution and to the most toxic components of oil – PAHs (Ohwada et al. 2003; González et al. 2009; Gemmell et al. 2018). Our results also showed that phytoplankton displayed significant changes in biomass and abundance following the oil spill. However, the three size fractions exhibited different responses, confirming that cell size is important in determining the reaction of phytoplankton to oil and PAH pollution (Echeveste et al. 2010a, 2010b; Ben Othman et al. 2012; Bretherthon et al. 2020). Moreover, the observed changes in biomass and abundance varied over time after the oil spill. The biomasses and abundances of nano- and micro-phytoplankton significantly decreased on D04 and D08 relatively to D01 (Fig. 3a, b). Short-term negative effects of oil and PAHs on large phytoplankton have been reported, and their toxicity on the photosynthetic activity, growth and biomass of phytoplankton is well documented (Djomo et al. 2004; González et al. 2009; Paul et al. 2013; Ben Othman et al. 2018). PAHs can alter photosystem II functioning and reduce the fluorescence yield in several marine phytoplankton species (Aksmann and Tukaj 2008; Ben Othman et al. 2023). In addition to the toxicity of petroleum hydrocarbons, the presence of an oil slick on the water surface during the first week of sampling (Fig. 1) reduced light penetration and indirectly inhibited phytoplankton growth. Conversely, nearly three weeks after the oil spill (D18), a positive effect was observed on large phytoplankton (> 2 µm), which bloomed and reached high concentrations (2.5-4 105 cells L− 1, Fig. 3a, b). Our results are in agreement with previous studies reporting a reduced phytoplankton concentration in the short term, but phytoplankton outbreaks (particularly for large species such as diatoms) in the long term (Taş et al. 2011; Sheng et al. 2011; Pan et al. 2012; Ozhan et al. 2014). On D18, the oil in the sampled station had been easily dispersed by the hydrological conditions encountered in mid-autumn (wind, water mass movements), and its negative effect on phytoplankton was ultimately reduced. Therefore, the regrowth of nano- and microphytoplankton on D18 could be mainly due to the return of a long-term favourable situation in the station, particularly light conditions. D18 was indeed characterised by the lowest turbidity (Table 2, Fig. 2) and by seawater that recovered the same colour and smell as before the incident. D18 also coincided with the lowest nutrient concentrations (Table 2, Fig. 2), which may reflect their absorption during phytoplankton regrowth. Moreover, nano- and microphytoplankton biomasses and abundances were negatively correlated with all nutrients. Ben Othman et al. (2018) have shown that resumption of phytoplankton growth under PAH contamination coincided with significant change in community composition.
Smaller phytoplankton species are traditionally reported to be more sensitive than large ones to pollutants like oil compounds because of their higher surface-to-volume ratio (Echeveste et al. 2011; Ben Othman et al. 2012), and significant decreases in their abundance and growth are generally reported after PAH exposure (Kottuparambil and Augusti 2018; Ashok et Augusti 2022). On the contrary, the biomass and abundance of picophytoplankton was stimulated after the oil pollution event of Bizerte, even when an oil film covered the surface water of the station (on D04 and D08, Fig. 3). Our results are in agreement with other studies showing proliferation of small phytoplankton after an oil spill incident (González et al. 2009; Huang et al. 2011; Kottuparambil et al. 2023). Quigg et al (2021) argued that the dominance of small phytoplankton following an oil spill is related to their higher growth rate than that of large cells. However, the increase of picophytoplankton could result from the release from grazing by sensitive grazers like heterotrophic nanoflagellates (Johansson et al. 1980; Almeda et al. 2018; Gemmell et al. 2018). Additionally, correlations between picophytoplankton and abiotic factors (nutrients) were absent, suggesting that pico-sized biomass may be more controlled by biotic factors such as zooplankton grazing. Concomitantly with our work, Chkili et al. (submitted) investigated the protozooplanktonic community and its grazing rate, and reported a net decrease in grazing mortality rate of picophytoplankton starting from D08, with the absence of their potential heterotrophic nanoflagellate grazers.
The different responses of small and large phytoplankton to the oil spill were followed by a shift in the phytoplankton size structure from dominant micro-sized cells on D01 (61% of Chl a) to pico-sized cells on D18 (∼60% of Chl a). This change could influence the grazer community and ultimately the food web structure.
Impact of oil on the phytoplankton community composition
The oil spill had a clear effect on the community composition of nano- and microphytoplankton, which changed differentially in the short and long terms (Figs. 4, 5, 6).
Short-term impact. In the short term, the compositions of the two communities showed high dissimilarity on D01 relative to D04 and D08 (Table S1). During this short period, the presence of petroleum slick in the station (Fig. 1) was accompanied by a decrease in the abundance and biomass of nano- and micro-sized cells, but some species maintained their growth. This indicates a growth inhibition of sensitive species and a possible selection of pollutant-tolerant species following the oil spill in the short term.
As regards microphytoplankton, diatoms were the main species on D01 and remained dominant on D04 and D08 despite the presence of oil in the station. The ability of diatoms to resist oil spills is documented. The presence of the silica frustule can protect them from the lethal effect of oil and confer them a better chance to survive than the naked phytoplankton cells do (Ikavalko 2005; Varela et al. 2006; Hallare et al. 2011; Ozhan et al. 2014; Parsons et al. 2015). Genes involved in the formation of the silica shell of diatoms (e.g., sil3) are downregulated after exposure to PAH mixtures (Bopp and Lettieri 2007). However, other researchers showed that diatoms are more affected by oil than phytoflagellates and dinoflagellates (Mishamandani et al. 2015; Taş et al. 2011; Fiori et al. 2016), and the response of coastal diatoms to oil is clearly size-dependent. Small diatoms (< 20 mm) are apparently stimulated by oil, whereas large diatoms (> 20 µm) are negatively affected by high oil concentrations (González et al. 2009). Furthermore, when Bretherton et al. (2020) assessed the role of different physiological traits in the response of various phytoplankton taxa to oil, cell size was most important in determining the biomass response to oil. In our study, the abundance of centric diatoms decreased from D01 to D08, dominated by small (mean length/width ratio 35 µm / 3 µm) Leptocylindrus species (Fig. 4c), at the expense of pennate diatoms whose contribution increased gradually to form almost all microphytoplankton on D08 (∼90%; Fig. 4a). Large Pseudo-nitzschia and Nitzschia species (length/width ratio ranges 45–108 µm / 3–13 µm) were the main taxa on D01 and D04, and Nitzschia spp. formed 91% of pennate diatoms on D08 (Fig. 4b). Our results are in line with those of Bretherton et al. (2019a) showing the dominance of the Pseudo-nitzschia genus after 3 days of oil exposure. The tolerance of pennate diatoms to toxic petroleum products may be explained genetically, since they are known to be more tolerant than other diatoms to oil compounds (Ozhan and Bargu 2014; Bretherton et al. 2019b). Furthermore, Melliti Ben Garali et al. (2021) provided the first evidence that two Pseudo-nitzschia species isolated from the Channel and Lagoon of Bizerte could tolerate a 15-PAH mixture by enhancing their biovolume, and were even able to bioconcentrate PAHs and degrade them, probably in synergy with their associated bacteria. Phytoplankton (including diatoms) can release extracellular polymeric substances (EPS) during environmental stress, e.g., oil exposure (Sun et al. 2018; Kamalanathan et al. 2019). The bacteria associated to phytoplankton can influence EPS production, and phytoplankton can in turn alter the bacterial community (Kamalanathan et al. 2019). Therefore, the phytoplankton-bacteria association can undoubtedly play an important role in phytoplankton resistance to oil. A benthic diatom and its associated bacteria isolated from a station close to Bizerte Channel have been found efficient in removing benzo(a)pyrene and fluoranthene (Kahla et al. 2021). All these findings clearly suggest that Nitzschia and Pseudo-nitzschia species from Bizerte, where previous PAH contamination has been reported (Lafabrie et al. 2013; Bancon-Montigny et al. 2019), have developed tolerance to PAHs through adaptation mechanisms that make them physiologically able to resist oil pollution. The trophic mode might also influence the sensitivity of phytoplankton to contaminants, including oil compounds. Conversely to autotrophic species, mixotrophic or heterotrophic diatoms can rely on sources of organic carbon to grow and to decrease their dependence on photosynthesis, so that they can better withstand contaminant toxicity (Debenest et al. 2009; Larras et al. 2012; Brethertron et al. 2021). Interestingly, heterotrophy has been reported for some Peudo-nitzschia species (Mengelt and Prezelin 2005; Loureiro et al. 2009), including species from our study site (Melliti Ben Garali et al. 2016). Our results suggest that diatoms are resistant to hydrocarbons and that certain species – e.g., members of the Pseudo-nitzschia and Nitzschia genera – may grow under oil contamination conditions. This is of particular interest given the fact that several species within these genera are known to produce the neurotoxin domoïc acid, including strains from the Bizerte waters (Bouchouicha Smida et al. 2014a; Sakka Halili et al. 2016). Blooms of Pseudo-nitzschia and Nitzschia are a known fact in the Lagoon and Channel of Bizerte (Sahraoui et al. 2012; Bouchouicha Smida et al. 2014b; Melliti Ben Garali et al. 2020). Oil hydrocarbon pollution can promote the growth of these diatoms and could increase the frequency of toxic blooms in these coastal waters, which are chronically prone to this kind of pollution.
As regards nanophytoplankton, cryptophyceae (dominated by H. fusiformis and H. marina) were the main taxa from D01 to D08 (Fig. 5). This is not in agreement with the observation of Brussaard et al. (2016), who reported the disappearance of this phytoplankton group after short-term exposure to oil. The major short-term change within nanophytoplankton observed in our study was the decrease of chrysophyceae (mainly M. pusilla) and prymnesiophyceae (mainly Imantonia rotunda) (Fig. 5). M. pusilla has been found sensitive to oil (Bretherton et al. 2020), or not (Brussaard et al. 2016), and a beneficial effect of oil exposure on chrysophyceae has even been found (Finkel et al. 2020). However, prymnesiophyceae can be a dominant group after an oil spill (Brussaard et al. 2016). In fact, the sensitivity of phytoplankton to oil is largely species dependent, since toxic oil compounds can affect different cellular and molecular targets (Ozhan et al, 2014; Kamalanathan et al. 2021, 2022). Furthermore, species can have differential core metabolism plasticity to counteract oil-induced oxidative stress and can modulate their tolerance to oil exposure (Kamalanathan et al. 2021, 2022).
Long-term impact. The community compositions of micro- and nanophytoplankton on D18 greatly differed from those of the other days (Table S1, Figs. 4, 5, 6). On D18, the wide regrowth of both size fractions (Fig. 3) coincided with significant changes in the community composition, as speculated. Centric diatoms (microphytoplankton) were more abundant after 18 days and formed ∼40% of the community, although the community structure did not change that much since Leptocylindrus spp. remained the dominant taxon (Fig. 4c). Centric diatoms have been reported to be faster growing species than pennate diatoms after a long post-spill period (Verlecar et al. 2006; Hallare et al. 2011). In the present study, pennate diatoms remained an important component of microphytoplankton, but the community changed deeply: A. glacialis had a low contribution from D01 to D08 (8–11%), but proliferated (2 105 cells L− 1) and became dominant on D18 (87%) (Fig. 4b). Gallo et al. (2016) reported a dominance of A. glacialis under increased levels of CO2 in seawater, while Sahu et al. (2022) found that its development can be supported by salinity and a high nitrogen concentration. The return of favourable environmental factors in our study site in the longer term may be responsible for A. glacialis development. The species abundance was negatively correlated with the N and Si concentrations, indicating their utilization during the A. glacialis increase.
As regards nanophytoplankton, the Chaetoceros spp. bloom (1.3 105 cells L− 1) was the most significant change observed after 18 days (Fig. 5). The growth of these red tide organisms was recently found stimulated after 10 and 14 days of oil exposure (Lv et al. 2023). Nutrients – phosphorous in particular – are considered the main regulating factors of Chaetoceros species growth (Shevchenko et al. 2006; Silkin et al. 2011). During our study, phosphorous seemed to support the proliferation of Chaetoceros spp., since the abundance of these organisms was negatively correlated with P concentrations (rs = -0.699, P < 0.05).
Species succession after an oil spill might also be related to different sensitivity levels to hydrocarbons (Sargian et al. 2007; Huang et al. 2011; Hemmer et al. 2011). Dominant species after several days of oil exposure can be considered more sensitive than species growing in the presence of high amounts of oil. In our study, A. glacialis and Chaetoceros spp, which can be sensitive to oil pollutants, dominated on D18 and took the place of Nitzschia and Pseudonitzschia species, which can resist to hydrocarbons as discussed above. Another reason might be that A. glacialis and Chaetoceros spp could outcompete and eliminate other species via allelopathic mechanisms, and become more competitive for resources (nutrients, light) and hence grow better. Allelopathy plays a significant role in the dominance, succession, and formation of natural diatom communities (Leflaive and Ten Hage, 2009; Zhang et al. 2019). Additionally, the difference in the diatom community composition after long-term oil pollution could be partially due to indirect trophic interactions because herbivorous zooplankton is sensitive to oil compounds (Hjorth et al. 2007; González et al. 2009; Almeda et al. 2014; Van Dinh et al. 2019). For one reason or another, our results clearly show that the oil spill was followed by a profound change in the phytoplankton community composition after 18 days. This modification was accompanied by significantly lower diversity indexes (S, H’ and J, Fig. 7), which reinforces the fact that oil spillage has deep effects on the structure of primary producers. Huang et al. (2010) also observed a drop in biodiversity and in the number and evenness of phytoplankton species after 15 days of crude oil exposure during different seasons.