**4. Discussion**

The present study provides the first information of the spatiotemporal variation in lagoonal meiofaunal assemblages in one of the richest Italian regions in number and extension of lagoons within the Tyrrhenian coast, i.e., the Sardinian Island. We demonstrated significant changes in meiofaunal diversity and community structure in relation to the main environmental gradients which, in the studied lagoon, are driven by salinity and dissolved oxygen concentration in water, sediment organic enrichment, and different degrees of confinement and saprobity. Our results highlight the importance of meiofaunal studies often neglected, by ecologists and policy makers, to further our understanding of the functioning of Mediterranean lagoons.

From an abiotic point of view, the heterogeneity of the Cabras Lagoon can be related to two main interlinked environmental and ecological gradients. The first gradient is due to the marine influence seaward, the second one is related to spatial differences in the trophic status and saprobity condition across the lagoon. In particular, the most confined site C2 differed from the other two investigated sites as being characterized by muddy, organically enriched sediments, and due to its lack of connection and exchange with both continental and marine waters, mostly influencing C1 and C3, respectively. Differently, the northern site C1 was most directly affected by the riverine input of freshwater and had sandy sediments with a very low phytopigment and OM content. This is consistent with the high energy and the low water residence time found at C1, which helps explaining the resuspension and export of fine sediment particles from this site to the central sector of the lagoon [56,57]. Finally, the seaward site C3 located in one of the three creeks connecting the lagoon to the Gulf of Oristano was in terms of sediment features intermediate between C1 and C2. It is worthwhile noting that in this sector of the lagoon man-made structures constructed in proximity of the inlet have caused modifications in the sedimentary regime of the lagoon and the water exchange with the adjacent Gulf [41,56]. Yet, the largest variation in salinity was found at C3 due to a marked drop in salinity in the whole lagoon in the winter of the present study, nullifying temporarily the salinity gradient and further highlighting the high seasonal and interannual environmental variability of the Cabras Lagoon [35,36].

Overall, these results show that the environmental (dis)similarities between the three study sites were constant in the two periods investigated, although summer and winter data-points were most clumped at C1 and more widespread, but still clearly separated one another, at C2 and C3. Thus, temporal variation did not alter significantly the location of the site-points in the multivariate model ordination. This highlighted that the most relevant source of variation in the study area was not due to differences between summer and winter dates, but to differences between sampling sites.

Regarding the meiofaunal richness, our data can be compared only with a few studies because the complete list of meiofaunal taxa is rarely reported. However, an overall high number of richness (i.e., 16 taxa) was recorded in the present study showing levels higher than those documented from both Mediterranean and European transitional water bodies [17–19,22,58,59]. Nematodes were the dominant taxon as frequently reported in lagoon systems worldwide [2,17,22,60,61]. This is likely related to their capacity to colonize the fine (suboxic or anoxic) sediments that generally characterize lagoons [19]. As reported in literature, the second most abundant group is represented by copepods [14,22,58,59]. Copepods are one of the most sensitive taxa to oxygen limitation and therefore they are confined to the oxic sediments [20,48], but they seem to take advantage of the high abundance of the microphytobenthos occurring in lagoon sediments and that are a primary food source for numerous copepod species [24].

The meiofaunal structure assemblage seemed to be very sensitive to the spatial environmental heterogeneity found in the Cabras Lagoon appearing a promising indicator of biotic changes in transitional water bodies. Indeed, the multivariate analysis (nMDS and PERMANOVA) applied on the assemblage structure clearly distinguished the three di fferent sites in line with their environmental features with a partial overlap of the meiofaunal structure of C1 and C2, and a greater separation of C3 (Table 2, Figure 6). The latter site, characterized by "marine conditions", had the highest abundance of all the meiofaunal taxa and in particular, it was distinguished by the greater abundances of Plathelminthes and Oligochaeta. These two taxa are often closely associated with each other and are among the primary components of transitional environment sediments [20,24]. Oligochaeta are regarded as taxa able to adapt to numerous environmental stress [18,48]. Plathelminthes are e ffective predators of many meiobenthic organisms such as copepods that could explain their higher abundance at C3 where the higher meiofaunal densities and copepods was found [14,62].

Instead, the high degree of confinement and consequent trophic load existing at site C2 was marked by a higher presence of Halacaroidea and Ostracoda. Ostracoda are generally recognized as sensitive taxon to environmental perturbations, but adaptive behaviors to numerous natural and anthropogenic environmental (e.g., organic load and trace element contamination) changes have been documented in several species [22,24,63]. In that site, there was also the only record of Pycnogonida that is generally recognized as a marine taxon. Pycnogonida have a few representatives in the meiofauna, but the *Anoplodactylus* genus has some species that ranges in the meiofaunal body size and are also tolerant to salinity variations until values comparable to those found in C2 (11 PSU) [64].

In C1, the site with the lowest salinity, Tanaidacea was one of the discriminating taxon. Noteworthy is that although many euryhaline species from Tanaidacea are found in transitional habitats, most occur only temporarily in these environments, appearing unable to form stable populations there [65]. Furthermore, Ate¸s et al. [66] reported that some species are particularly related to coarse grain size and low content of organic matter that were the conditions that distinguish C1.

The environmental variables that appeared to mainly a ffect the meiofaunal assemblage of the study area were salinity and DO that were the main responsible parameters for the separation of C3 and the other two sites. Salinity gradient is one of the primary factors that influence meiofauna in transitional environments [8,18] along with the oxygen availability that seems to influence all the meiofaunal taxa and not only the oxygen sensitive ones such as copepods (see above references). The quantity and quality of the organic matter (OM, Chl-*<sup>a</sup>*, and phaeopigments) and Wc, which is an indirect indication of the grain size of the substrates, were also important for the meiofaunal distribution as suggested by many authors [8,17,22] and contributed mostly to the separation between the other two sites (i.e., C1 and C2). Instead, temperature did not show a relevant contribution to the ordination of the meiofauna. This issue as well as the lack of significant di fferences in the comparison of the meiofaunal assemblage structure sugges<sup>t</sup> that meiofauna was more a ffected by spatial than temporal variations. This pattern resembled the distribution patterns of the macrozoobenthos observed at the same study sites where spatial di fferences were greater than significant seasonal changes [30].

When the spatiotemporal variation in the univariate measures were statistically studied, the greatest di fferences were observed between the three sites further supporting the results revealed by the structure of the meiofaunal assemblage (Table 4). In particular, PERMANOVA showed significant di fferences of the total meiofaunal abundance, number of taxa, evenness and the Ne/Co ratio. Among them, the Ne/Co ratio showed a very temporal variable trend (i.e., it was higher in C1 in the winter and lower in C2 and C3 in the summer) which likely is why PERMANOVA did not reveal

significant di fferences between periods. However, the variation in the Ne/Co ratio did not highlight the presence of anthropogenic stress being the values lower than the thresholds reported by Ra ffaelli and Mason [52] for stressful conditions. Instead, the highest evenness at C1 was likely related to the coarser grain size of the sediments that generally host a more diversified meiofaunal assemblage [67,68].
