The nature reserve Bottsand is situated at the western Baltic coast of Germany, about 20 km to the northeast of Kiel (Fig. 1). The lagoon is confined to the south by the Wendtorf Marina and to the east by a dike that has been built in 1880–1882 and was refurbished in 1982. The northern fringe is a 1.4 km long spit bar. The successive westward progradation of the spit since 1870 has effected an almost complete closure of the lagoon (Schrader, 1990; Knief, 2013). Bogs grew through aggradation at the northern margin of the lagoon, where an extensive salt marsh developed (Wolfram, 1996). A concomitant aggradation of the salt marsh at the foot of the dike has tapered the lagoon to a terminal ditch that runs out in a series of ponds. The northern part of the lagoon is separated from the southern part by a broad shoal that falls dry at low water.

The bottom sediment of the lagoon is a fine to medium sand (Grabert, 1971). The sediment surface is covered by a few centimetres of rotted plant debris in places (Lehmann, 2000). The invertebrate fauna of Bottsand is characterised by dense lugworm colonies and clusters of Mytilus edulis and other molluscs that are common in the Baltic Sea. The plant associations are dominated by macroalgae (Fucus vesiculosus, Zostera marina) (Hammann and Zimmer, 2014). In marginal areas of the lagoon, where the water is only a few centimetres deep, patches of sea rush and saltwort (Bolboschoenus maritimus, Salicornia stricta) are found.

The salinity of lagoonal waters varied between 12 and 18 units. It was markedly lower in the terminal northeastern ditch at 6 to 10 salinity units, mainly due to precipitation and ground water seeping (Lutze, 1968a). The water temperatures in the lagoon were mostly 4 to 5 ∘C higher than the surface water temperature of the adjacent Baltic Sea. A significant environmental parameter for the ecosystems in the lagoon is water level changes, which are driven by wind direction change (Sztobryn et al., 2009). Onshore winds may raise the water level up to 2 m, in extreme cases even to 3 m above standard German reference level (NN), and they may fall to 1.5 m below NN in Kiel Bight (Härdtle, 1984). The tidal range is rather low, at approximately ±0.10 m.

The foraminifera from the Bottsand lagoon and the adjoining salt marsh have been studied since the mid-1960s. Early surveys focused on the general distribution of foraminiferal species in the lagoon and Baltic Sea, as well as their growth, reproduction (Lutze, 1968a), and patchiness (Lutze, 1968b) and the redeposition of tests (Grabert, 1971). The faunal inventory, microenvironments, and reproductive cycles of salt marsh foraminifera were described (Brönnimann et al., 1989; Lehmann, 2000; Lehmann et al., 2006). These studies revealed that the lagoon was populated by four species, of which Elphidium williamsoni was the only rotaliid. The highest population densities were recorded at a two-step brink separating the salt marsh from the lagoon. Miliammina fusca, Ammotium salsum, and Trochamminids were frequent here. The salt marsh fauna was dominated by Trochammina inflata, Haplophragmoides manilaensis, Miliammina fusca, and Tiphotrocha comprimata. Balticammina pseudomacrescens was found at the brink and in the distal Festuca lawn.

The annual monitoring of the foraminiferal fauna at the Bottsand lagoon was initiated in 2003. The aim of the project is to create a complement to the Boknis Eck time series (Bange et al., 2011) because the Bottsand lagoon and the adjacent salt marsh are influenced by both surface water dynamics of the Baltic Sea and predominant seasonal weather conditions. As in any other long-term study, the present investigation is driven by a null hypothesis, i.e. the foraminiferal fauna has not changed with reference to last year or to the last survey.

The Bottsand lagoon water levels showed a low-amplitude daily variability of 0.1–0.3 m and occasionally high waters of 0.6–1.0 m NN lasting several days. Water levels below 0.01 m NN were generally not recorded, resembling an average minimum water depth of 0.1 m close to our sampling site. Hence the terminal ditch of the lagoon never fell dry, which is in agreement with personal observations by the author on unnumbered visits to the area. Corresponding tide gauge records from Kiel Fjord documented low water levels of -0.75 m NN or even less during our measuring periods in 2012 through 2017 (Fig. 2).

The temperature showed a successive decline of water temperatures from 18 ∘C in mid-September to ca. 7 ∘C on average in mid-November 2012 through 2016. The diurnal variability was rather high in September, with 5–7 ∘C, whereas it was only 2–3 ∘C in November. The records of the different years matched well during September and October but fanned out with the beginning of November (Fig. 2).

The salinities showed a much stronger and irregular variability than the temperatures. The inter- and intra-annual salinity variability was low during September in 2012 at 2 to 3 units compared to 2015. Strong oscillations with amplitudes of up to 7 salinity units commenced with the beginning of October during these years. The records of the individual years spread apart by the end of October, effecting a variability between a minimum of 8 and average maximum values of 19 salinity units among the years 2012 to 2015. The salinity curve of 2016, however, showed a broad, extraordinary maximum at which salinities exceeded the average autumn peak value for a period of 13 days. The 2016 curve merged with the salinity records from 2012 to 2015 no earlier than at the beginning of October. The salinity sensor of the data logger failed in 2017 but on-site measurements showed markedly lower values of 10–16 salinity units in September and early October.

A striking coherence of transient, short-term water level and salinity maxima was recognised at the beginning of the salinity rise in September 2016, following the lunar semi-diurnal M2 tidal constituent of 12 h and 25 min. The northern part of the lagoon was thereby flushed with salt-rich water during every flood, which was, however, diluted again only in part during the next low tide, probably by less-saline ground water seeping (Fig. 3).

There were 12 foraminiferal species recorded in the present study. They comprise nine arenaceous species and three calcareous species of the genus Elphidium (Table 1). Miliammina fusca, Ammotium salsum, Elphidium williamsoni, and Trochammina inflata were common to frequent with proportions varying between 1 and 57 % of the living fauna, even though individual species were occasionally not found (Fig. 4). Arenoparella mexicana depicted more constant proportions ranging from 6 to 47 %, mostly varying between 10 and 36 %. All these species have been found in lagoonal or terminal ditch faunas in 1966 and 1996 as well (Lutze, 1968a; Lehmann, 2000). The proportions of Miliammina fusca and Elphidium williamsoni were much higher in the 1960s than in the early 2000s. Elphidium incertum suddenly appeared in 2016 and made up 89 % of the entire fauna, thereby outnumbering all other species. It returned to much lower proportions in 2017. E. incertum has not been found in the lagoon before. In total, the six ranked species comprise on average 87.4 % of the living fauna.

Among the rare species, Elphidium gunteri and Reophax moniliforme had also not been recorded in the Bottsand lagoon before 2003. They recurrently showed up during certain years since then and were, with two exceptions, not found again in the following year. A similar stochastic occurrence pattern is recognised in Labrospira jeffreysii.

The foraminiferal population densities varied between 3 and 67 specimens per 10 cm3 surface sediment, and they steadily increased during the years 2003 to 2015. The population densities suddenly increased to 922 specimens per 10 cm3 in 2016, of which E. incertum accounted for 817 specimens; the afore-mentioned trend continued for the other species (Fig. 5). The standing stock of E. incertum returned to 3 and 7 specimens per 10 cm3 in 2017, levels that compare well to other, less frequent species in these samples. The Fisher's alpha diversity index ranged from 1.1 to 3.9 and mostly varied around 2, a value that is common for marginal marine foraminiferal assemblages (Table 1).

A long-term assessment of the benthic foraminiferal faunas of the Bottsand lagoon is hampered by the fact that the previous studies focused on different locations and used a different methodology. In 1966, the lagoonal bottom was inhabited by Elphidium williamsoni (Cribrononion articulatum of authors) and Miliammina fusca only, both showing profound seasonal variations. At the vegetation boundary or in the terminal ditch of the lagoon, Ammotium salsum, Jadammina macrescens, Trochammina inflata, Tiphotrocha comprimata, and Haplophragmoides wilberti (H. bonplandi of authors) were common to frequent (Lutze, 1968a). In 1996, Ammotium salsum spread and populated the entire lagoon and Ovammina opaca was frequent as well (Lehmann, 2000). The latter species may not have been recognised because the samples were picked dry in the 1960s. However, the bottom of the terminal ditch was not sampled in 1996 but the brink before the vegetation boundary was. Balticammina pseudomacrescens, Psammosphaera sp., and Arenoparella mexicana were recorded here as well and showed a small-scale lateral variability. The latter two species were not found in 1966, and Ammotium salsum was missing here in 1996. The trend of increasing species richness continued in the early 2000s as described above, in particular with the recurrent recruitment of two more Elphidium species, Elphidium incertum, and the repopulation of the terminal ditch by Ammotium salsum. However, the interannual fluctuations of dominant species' proportions and population densities showed no coherence with meteorological records, as for instance seasonal or annual average temperatures or sunshine hours. The conclusion of Lutze (1968a), that these parameters exerted a positive influence on the faunas, is therefore not corroborated by the data from this study. The overall trend of increasing species richness and the interannual variation in species' abundances may not be related to changes in substrate properties because the surface sediment was always a crumbly to semi-fluid, organic-rich, sandy mud at the monitoring site.

Elphidium incertum is commonly found in boreal to arctic seas of Europe where it occurs in intertidal to subtidal environments around the British Isles and in marginal seas and fjords below the halocline (e.g. Horton and Edwards, 2006; Polyak et al., 2002; Nordberg et al., 2017). The salinities prevailing at the sampling sites on British intertidal flats were open marine, and they were not lower than 25 units in the southern Kara Sea, which assigns E. incertum as a stenohaline organism (Darling et al., 2016). In the western Baltic Sea, however, E. incertum was dwelling in sandy muds at depths between 16 and 25 m that were bathed by inflowing Kattegat water of 20 to 24 salinity units (Lutze, 1974; Schönfeld and Numberger, 2007; Polovodova et al., 2009). The species has only been occasionally recorded at depths shallower than 13 m, and at presumably even lower salinities, thus challenging the above-mentioned threshold value.

The microhabitat of E. incertum is shallow endobenthic. It lives underneath the immediate sediment surface and above the redox boundary (Wefer, 1976). The abundance maximum is centred in the uppermost 5–6 mm of the oxidised near-surface layer (Lutze, 1974). The species was found at the sediment surface only when the redox boundary had moved upwards during hypoxic periods (Wefer, 1976). Therefore, and even under pore water deoxygenation, the E. incertum abundance maximum was entirely covered by the 0–1 cm sample interval at the Bottsand monitoring site.

The size of E. incertum from the Baltic Sea was ca. 600 to 800 µm, whereas specimens from the Bottsand lagoon were only ca. 125 to 250 µm in diameter and thus much smaller (Plate 1). Populations of E. incertum with consistently smaller specimens of 350 to 550 µm test diameter were described from the margins of the Bornholm and Gotland basins in the central Baltic and denominated as Cribrononion incertum subsp. A by Lutze (1965). The chambers were slightly inflated and the umbilicus was depressed and covered with small pustules (Plate 1). Morphotypes with a similar size (360 µm) have been recorded from 19 m water depth at Wismar Bight, southern Baltic Sea (Frenzel et al., 2009), and were identified as Haynesina germanica. The specimens from Bottsand, however, were lacking diagnostic features of H. germanica, in particular the collar of pustules at the base of the apertural face of the last chamber (Austin et al., 2005), the smooth, transparent chamber wall, the widened sutures close to the umbilicus, and the tendency of evolute, oblique coiling. Instead, their chambers are slightly inflated and the outline is lobate (Plate 1). These features are also characteristic for the subarctic Haynesina orbicularis or the small-sized Elphidium albiumbilicatum. The latter is common in the Baltic Sea as well. However, the field of pustules densely covering the umbilicus is quite large compared to the individuals from Bottsand and extends on the flanks of both sides of the sutures, whereas pustules follow a thin seam along both sides of the narrow incised sutures of the Bottsand specimens. The pustules merge, thicken, and form one to three bundles bridging the sutures between later chambers, which is a diagnostic feature discriminating E. incertum from other Elphidium species (Lutze, 1965). Finally, a small, lobate E. incertum with a pustulous umbilicus from the Kara Sea has been described, where it is common in estuarine settings under the influence of major rivers (Polyak et al., 2002). The occurrences from Bottsand, Wismar Bight, and the central Baltic provide corroborating evidence and suggest that this morphotype has to be considered as an ecophenotype rather than as a subspecies or even a different taxon.

Bottsand is a nature reserve and strictly closed to human activities like fisheries or ship traffic that could convey alien species to this area. However, the lagoon is flushed by sporadically high waters, often associated with strong northerly gales, which may well have stirred up foraminiferal propagules by deep, grounding waves and carried them into the lagoon in suspension. This has probably happened more than once a year since the late 19th century, when the lagoon started to form (Schrader, 1990). The question remains of why E. incertum or other Baltic deep water taxa have never been found at the Bottsand lagoon before even though hundreds of samples were examined since the 1960s. The salinity records provide a persuasive explanation: values of more than 19.9 units have never been recorded (Table 1). They were created in early September 2016 by enhanced evaporation due the high temperatures and regular inflows from the adjacent Baltic Sea during high tides, which effected, in combination, a successive salt concentration in the terminal ditch (Fig. 3). It is conceivable that a period of 13 days around the lower salinity limit of this species and high temperatures could have triggered the propagules to grow.

One may argue that E. incertum had already occurred earlier, that the tests were not preserved, and the species had therefore been missed. Indeed, only a few, corroded E. williamsoni specimens were observed among the unstained tests in the samples, which were not systematically investigated in the present study. Nonetheless, Lutze (1974) noted that E. incertum was less frequent in the dead assemblage compared to its proportion of the living fauna in Kiel Bight. The turnover rate and test production was considered to be substantially lower than in other species. If this is applicable to the Bottsand lagoon too, a sporadic occurrence of a few living specimens between 1996 and 2003 would hardly have left any record in the surface sediments.

The infaunal E. incertum requires high concentrations of food particles, even though it is not strictly bloom feeding, and it is less tolerant to extended hypoxic conditions (Gustafsson and Nordberg, 1999; Polovodova et al., 2009). Both preconditions are given in the organic detritus surface layer and shallow water depth of the ditch. As such, the population dynamics with a huge increase in standing stock was similar to any other first colonising invasive species, even though the constant population densities of the other species suggest that E. incertum either used a different food source or microhabitat than the other, indigenous foraminiferal species, or that the carrying capacity of the habitat has not been used at all. The fact that the diversity did not decline substantially with the bloom of the newcomer corroborates this conclusion (Fig. 5). Furthermore, the faunal data suggested a high resilience of E. incertum to periods of low salinity as recorded in late 2017. Whether the species will continue to survive in the coming years remains to be tested by continuing the monitoring programme.