Soil seed banks in restrictive environments: dynamics and response to human disturbance.

Gypsum steppes and saline steppes are restrictive environments, on which factors controlling seedling recruitment and establishment in early stages are little known. Information on the factors controlling recruitment and establishment of the gypsophytes and halophytes will improve our limited knowledge on arid edaphic endemism and provide a basic tool for the management and conservation of protected ecosystems, such as gypsum outcrops and saline steppes. Abandonment of agricultural lands in large areas of the Iberian Mediterranean region during the last decades is recognized like the most important regional global change driver. Our knowledge on secondary old-field succession in Mediterranean environments and its relationship with soil seed bank is extremely scarce and null on restrictive soils, such as gypsum steppes. We sampled a complete set of old-fields on gypsum soils (1 to 60 years since abandonment) in order to test whether re-installation of gypsum vegetation after cropland abandonment requires: (i) soil physical restructuration and (ii) chemical re-adjustment to enable growth and survival of specialized gypsophilous vegetation, and more specifically how time from abandonment drives such environmental change. At the same time, we sampled soil seed bank in order to answer two main questions: 1) How long did specialist species (gypsophytes) take to reappear in seed banks after crop abandonment of restrictive habitats? and 2) which role did the seed bank play in the recovery process of these habitats? In chapter 1, our results showed that secondary old-field succession on semi-arid Mediterranean gypsum soils is controlled by a complex set of factors with tiny overlaps between them. Surprisingly, time since abandonment explains only a tiny fraction of variance (3 %) found in composition. However, soil chemical features independently from time since abandonment are important for explaining differences found in old-field composition. These findings suggest that secondary succession on specialized Mediterranean soils does not follow the widely described amelioration process in which soil features and composition are tightly fitted along time. On the other hand, in chapter 2, we found high seed bank richness (77 species) and middle range seed density values (11,670 seed/m2). Soon in the succession we detected the main constituents of the specialist vegetation typical of gypsum environments, in aboveground and seed bank species composition. We found high correlation between species composition in the soil seed bank and the aboveground vegetation all along the succession process. These results suggest that composition and the density of the soil seed banks were not determined by the time after disturbance. Differences found in soil seed banks of different sites were mainly due to aboveground composition and physical and chemical edaphic parameters. In these restrictive environments specialized plants establish soon in the succession and almost no species replacement happens. Gypsum soils in arid and semi-arid environments impose an important restriction of vascular plants, as it has been reported by several authors. However, factors controlling the recruitment of gypsophytes and gypsovag species are not completely understood. On the other hand, few studies have focused on the impact of seed consumption by ants on the conservation status of endangered plant populations. In chapter 3, we researched the soil seed bank dynamics of Thymus funkii subsp. funkii, a gypsovag shrub endemic to the south-eastern Iberian Peninsula, to determine possible reasons for low recruitment in isolated populations growing on crusted gypsum soils. Recruitment-restriction hypotheses involving main factors in the seed bank dynamics were analysed, and a diagrammatic dynamics model drafted from a population perspective. Viable seed production averaged 362 per plant. Seed shadows resulting from primary dispersal were concentrated under the mother plant. There were two thyme seed predators: coleopteran larvae (pre-dispersal phase) and Messor bouvieri ants (pre- and post-dispersal phases), which predated 16.4% and 3.8% of population yield, respectively. However, ants may simultaneously contribute to the spatial secondary dispersal, since they lost a small fraction of seeds they harvested. Seeds showed innate physiological dormancy to high summer temperatures, although this disappeared rapidly. The response of germination to other temperature × light conditions was wide and rapid. No permanent seed bank was detected in the soil. Seeds buried in the upper soil layer (2 cm) germinated highly and fast. In contrast, a significant fraction of seeds buried deeply (8 cm) remained viable for over one year. Germination in the field accounted for only 1.0% of the seed output, and seed death due to pathogen attack and aging was also low (3.0%). Seedling emergence was confined to autumn and showed a high correlation with total plant cover, irrespective of plant species. Seedling survival was very low. The small spatial correspondence between primary dispersal and seedling emergence patterns may reflect the importance of secondary dispersal in T. funkii seed bank dynamics. Hypotheses and consequences of such horizontal seed movements for conservation management are discussed. In chapter 4, we examined the spatial and temporal patterns in the recruitment of Teucrium libanitis Schreb. (Labiatae), a gypsophyte endemic to the semi-arid south-eastern Iberian Peninsula, from the perspective of soil seed-bank dynamics. Seed production, seed dormancy, soil seed bank, aerial seed bank, seed predation and seedling emergence and survival were monitored for one biological cycle. Several microhabitats were considered to determine safe sites for the Teucrium libanitis recruitment. The gypsophyte T. libanitis forms a transient aerial seed bank and a persistent soil seed bank in gypsum steppe with a strategy well-adapted to a semi-arid environment. The existence of this persistent soil seed bank supports the theory that the recruitment is produced by the occurrence of favourable climatic events. Seeds were redistributed horizontally, accumulating under plant cover, but seedling survival was not dependent on these different microhabitats. Recruitment was not limited by predation or by seed death by aging and pathogen attack. At least, in chapter 5, we analyzed the fate of seeds removed by ants, seed supply, seedling survival, and the ability to form persistent seed bank in the soil by the critically endangered population of the non-myrmechorous perennial halophyte Helianthemum polygonoides in order to determine to what extent seed removal by ants represents a real bottleneck for recruitment and thus a threat for long-term population viability. In the absence of ants (Messor bouvieri and Aphaenogaster dulcineae), plants produced on average 969 viable seeds. Apparently the impact of seed-harvester ants was dramatic: 75.5% of reproductive plants were affected by pre-dispersal predation and the primary seed shadow was reduced up to 93%. However, M. bouvieri lost 7% of fruits (5.5% of seeds removed) by dropping on trunk trails, mostly under plant cover, the most propitious microenvironment for recruitment. In addition, a small fraction of seeds were mistakenly rejected in refuse piles, although here recruitment was extremely low. A significant fraction of seeds (¿ 40%) persisted viable in the soil for over two years. In addition H. polygonoides had ability to form short-term persistent seed banks, in spite of the intense ant seed removal. In general, seedling recruitment was very low, denoting the scarcity of safe sites in the habitat. Those data overall corroborates that seed removal by ants, although highly intense, does not compromise the viability of perennial-plant populations, because seed supply and seed reserve in the soil are enough for exploiting current and future safe sites. A population viability analysis demonstrated that other threats affecting the survival of reproductive H. polygonoides plants would constitute the real risk for the conservation of that critically endangered species