Literature review shows that there are no true annual seagrass species. An annual life cycle was suggested by Kuo et al. [10] for the understudied deep water dioecious Halophila tricostata, but recent work by Chartrand [11] showed that this species overwintered with quiescent rhizomes, although yearly recurring seedling recruitment was important for persistence. Similar life history strategies with vegetative quiescent phases have been revealed for other seagrass species.

Based on available evidence, at least 6 out of 63 seagrass species display a facultative annual life history, with true annual populations, namely Zostera marina, Z. japonica, Halophila decipiens, H. beccarii, Ruppia maritima, and R. spiralis. The trait is polyphyletic, as these species belong to different families (Hydrocharitaceae, Ruppiaceae, and Zosteraceae) [12]. Z. marina is the best-known facultative annual seagrass species. This species occurs in the temperate and tropical northern hemispheres, with annual populations recorded at several locations.

Overall, seed production is five times higher in annual populations compared to conspecific perennial populations. However, populations vary greatly in seed production, and some perennial populations also present high seed outputs; for example, a perennial population of Z. marina in Chesapeake Bay had a potential maximum seed production of 40,000 seeds/m² [13] vs. 100,000 seeds/m² of an annual population in the subtropical Gulf of California [14]. Annual Z. marina plants typically have limited rhizome development and allocate most of the aboveground biomass to reproductive shoots [7]. Such differential allocation to vegetative and reproductive structures has been found for terrestrial angiosperms when comparing annual and perennial congeneric species [4,15,16].

Assuming that there is a trade-off between vegetative (clonal) growth and sexual reproduction [16,17,18] and that sexual reproduction competes with the vegetative functions for necessary resources for plant growth and maintenance, an annual life cycle should only be favored over a perennial cycle when the survivorship of the established plant is lower than that of the seed or seedling. Such unfavorable conditions for vegetative development may recur periodically (often seasonally) or at stochastic intervals in highly unpredictable environments [19].

In seagrasses, such periodically unfavorable conditions may be low temperatures combined with high turbidity, as was found in Zostera japonica (British Columbia [20]) and Ruppia maritima (Baltic Sea [21]). Ruppia spp. may colonize shallow coastal lagoons that are only flooded during part of the year, and annual growth forms are reported to be a response to desiccation. Halophila beccarii forms annual populations as a response to decreased salinities on tidal flats in Malaysia [22]. Additionally, the subtidal delicate and shallow-rooted Halophila decipiens does not have a broad tolerance to salinity or temperature changes and may therefore be susceptible to removal or die-off during winter.

Annual populations of the relatively well-studied Z. marina are encountered in a myriad of situations. Comparing habitats of annual populations with the nearest perennial ones, the first seems to be more stressful than the latter. They experience either desiccation, heat stress, anoxia-related stress, shading stress, or a combination of all these well-known stressors of Z. marina and other seagrass species [23]. Populations are usually annual in the intertidal, where they experience periodic desiccation, but in water-retaining depressions and in moist air intertidal, plants have a perennial live history. Subtidal or submersed annual populations seem to be exposed to higher levels of anoxia compared to those in neighboring populations. Anoxia-related stress includes excessive eutrophication and/or organic matter loading, at times accompanied by lower salinity (as a covariate of enhanced nutrient input from freshwater sources), increased shading, warmer circumstances (decreasing dissolved oxygen and likely enhancing microbial processes leading to anoxia), or muddier sediments (mud is often correlated with organic matter and occurs in areas with less flushing). Anoxia results in the microbial production of sulfide and ammonia, which are toxic to Zostera spp. [24,25]. In addition, tidal or submersed annual populations occur in heat-stressed environments and in light-limited (deep) habitats.

System scale ‘experiments’ in the Southwest Netherlands have shown that annual populations can become perennial within 5 years after a change in environment. Three estuary branches were modified for coastal protection during 1961–1986: one branch was modified into an oligotrophic saline lake [45], one branch was modified into a brackish and eutrophic lake [46], whereas one branch remained intertidal with a modified hydrodynamic regime [47]. Prior to the modifications, the branches were connected, and they all hosted intertidal, annual populations of Z. marina [48]. In the newly formed oligotrophic saline lake, the population became perennial upon submergence within 5 years [41]. However, in the newly formed brackish and eutrophic lake as well as the intertidal branch, the populations continued to be annual ([36,39]). This shift in life history, or absence thereof, after modification of the environment, is evidence that population life history traits can be induced by the environment. When the plants became perennial, they presented lower seed production and a number of flowering shoots, higher belowground biomass, and the vegetative shoots showed vigorous growth earlier in the season than before, when the population was still annual and seasonal timing is earlier, suggesting that rhizomes give the shoots a head start as compared to the seed [36].

Transplantation experiments in NW Europe and in North America confirm that seedlings from annual populations can become perennial plants during the first winter (NW Europe [34], Izembek Lagoon, Alaska [49], although their reproductive effort remains high (NW Europe [34], Willapa bay, Washington [50]). Keddy and Patriquin [7] cultivated seedlings in the laboratory from seeds originating from annual and perennial populations in Nova Scotia and found that 28 out of 29 of the seedlings from the ‘annual’ seeds developed into annual plants and 1 developed into a perennial plant. Vice versa, 26 of 28 seedlings from ‘perennial’ seeds developed into perennial plants, whereas 2 of 28 developed into an annual plant. Thus, the findings of Keddy and Patriquin [7] suggest that annual populations have the potential to produce perennial offspring and vice versa.

It is intriguing that the seedlings of the reviewed annual populations produce reproductive shoots very early in development; in other words, they are “programmed for scenescence” several months later. Secondly, it is intriguing that they, nevertheless, may shift to a perennial life history when the environment becomes more favorable for vegetative survival in critical periods.

During early growth, the seedlings of annual Zostera marina plants may not receive any indications from their environment that they will encounter adverse conditions for perennial growth later in the season, and the rapid development of generative shoots and early scenescence are perhaps “programmed”. Chartrand [11] found indications for such programming in deep water annual populations of Halophila decipiens in tropical Australia. However, it is also possible that a more stressful environment may already manifest early in the season and induce lower productivity/respiration ratios in the seedlings. This lower P/R ratio may induce the plant to invest more resources into sexual reproduction, which is also suggested by a review of the effect of disturbance on sexual reproduction in seagrasses by Cabaço and Santos [51], and supported by later studies, for example, showing relations between sexual reproductive effort and temperature [52,53,54], but see [55], desiccation [43,56], nutrients [57,58], mechanical disturbance [59], and high salinity [60].

Population genetic studies in NW Europe [61] and in San Francisco Bay US [62] suggest a lack of genetic differentiation between annual and perennial populations, as well as high rates of gene flow between them, although genetic diversity is generally larger in the annual than in perennial populations [63]. Muñoz-Salazar and coworkers [64] found significant genetic differentiation between perennial Z. marina populations from the Pacific coast and annual ones in the Gulf of California. This genetic divergence may be explained by the different life histories (annual vs. perennial), but it could also have been generated by limited gene flow between the two regions, as the tropical waters and current patterns of the southern Gulf of California have presented a barrier to gene flow and migration since the end of the Pleistocene. Oetjen and coworkers [65], using a genome scanning approach (using SNP and microsatellite markers), found some indications of selection between the subtidal perennial and intertidal annual populations in NW Europe. Divergent selections between the types of populations were detected at six loci, of which three were linked to genes involved in osmoregulation, water balance, and sexual reproduction (seed maturation). Selection could be enhanced by the different timing of the flowering initiation, even if annual populations are located in the immediate proximity of perennial populations via reproductive isolation [18,66].

The above suggests that the annual vs. perennial life cycles in facultative annual Z. marina (and possibly the other facultative annual seagrass species) may be reversible, involving tradeoffs between vegetative and generative functions. Genetic evidence of such inflection of tradeoff was, for example, found in the terrestrial annual Arabidopsis thaliana. Modulation of the activities of only three genes influenced the indeterminacy of meristems and longevity of the plants, resulting in a growth form with the increasing development of vegetative buds, higher longevity, and extensive woodiness, indicative of perennial plants [67]. In the two terrestrial facultative annuals described in the literature, Erythrante guttata and Oryza sativa, possible genetic mechanisms for such reversibility between life histories have been investigated. Friedman and coworkers [3], when identifying phenotypic and genetic trade-offs between flowering and vegetative growth in E. guttata, found that differential responses to photoperiod and vernalization (the induction of a plant’s flowering process by exposure to the prolonged cold of winter) of plants from annual or perennial populations involved quantitative trait loci (QTL) and differential gene expression. QTL was also found to influence resource allocation in annual and perennial populations of rice Oryza [68,69].

In general, gene expression may be involved in frequent and precocious flowering. Perennial plants require reprogramming of some meristems to start the production of reproductive organs. Overexpression of the Flowering Locus (FT) gene from A. thaliana resulted in precocious flower development independent of photoperiod [70]. In the same plant, it was found that micro RNAs are involved in gene expression; miRNA 172 (miR172) caused early flowering through disruption of the downregulation of floral repression genes [71]. Interspecies gene transfer between perennial Arabis alpina and A. thaliana, showed a perennial and an annual signaling pathway to flowering, involving Squamosa promotor binding protein-like 15 (SPL15) and FL pathways, respectively [72]. The functional overlap between the pathways may enable flexible responses to shifting environments, as well as life history variation [72]. In general, from an evolutionary perspective, life history traits are among the most labile trait syndromes in flowering plants and annuality has evolved convergently in different lineages of flowering plants, though mechanisms underlying transitions are still unclear [16].

Chartrand [11] found that the general condition of the seagrass plants of the deep-water annual population of the seagrass H. decipiens declined before the light levels fell below the critical threshold for growth, from which she suggests that senescence and sexual reproduction were programmed. She observed shifts in hormones involved in these processes similar to shifts previously reported in terrestrial plants [4]. Up- and downregulation of corresponding areas could be confirmed with metabolomic profile analysis. Such changes in metabolomic expression may be heritable (epigenetic); epigenetic changes may last through cell divisions for the duration of the plant’s life and may also last for multiple generations, even though they do not involve changes in the underlying DNA sequence of the organism [73]. In short, annual life cycles in facultative annual species seem to be induced by the environment (for example, by low P/R ratios) or (epi-) genetic programming.