The contribution of rewetting to vegetation restoration of degraded peat meadows

Question: What is the contribution of a rise in groundwater level to vegetation restoration of degraded peat meadows compared to abandonment only? Location: Abandoned peat meadows in the central part of The Netherlands. Methods: Comparison of species composition and species abundance of vegetation and seed banks of reference and rewetted peat meadows, using plant trait and seed bank analysis. Results: Vegetation of rewetted meadows shared on average only 27% of their species with the reference meadow, while this was 50% on average for species in the seed bank. Rewetted meadows had a lower total number of species and a lower number of wet grassland and fen species present in the vegetation, but had higher species richness per m², although evenness was not affected. Rewetting increased the dominance of species of fertile and near neutral habitats, but did not result in an increase of species of wet or waterlogged habitats. Re-wetted meadows were dominated by species relying mainly on vegetative reproduction and species with a low average seed longevity compared to the reference meadow. Conclusion: Rewetting was not effective as a restoration measure to increase plant species diversity or the number of wet grassland and fen species in the vegetation. If no additional restoration management is applied, the seed bank will be depleted of seeds of species of wet grassland or fen habitats, further reducing the chances of successful vegetation restoration.     

Restoration of the Cirsio dissecti‐Molinietum in The Netherlands: Can we rely on soil seed banks?

Abstract. Vegetation and soil seed banks of a threatened Atlantic fen meadow community were studied using recent phytosociological records and seedling emergence from soil samples. Similarly managed but differently degraded stands that suffered different levels of species impoverishment were compared. The actual vegetation was related to a set of phytosociological references representing the subassociations of the community. DCA positions of reference relevés from the different subassociations were overlapping, suggesting that in all references many common species occur. Recent records were positioned in‐between the seed bank samples and the references. The soil seed banks of all stands were dominated by ordinary species. Most character species had at most sparse seed banks and no seedlings of locally extinct character species, mentioned in historic floristic records, were detected. In contrast species of pioneer and small‐sedge communities as well as those of heathlands were abundant in the seed banks. Based on the vertical distribution of seeds in the soil layers most fen meadow species were classified into transient or short‐term persistent seed bank types. We concluded that complete restoration of the Cirsio dissecti‐Molinietum without reintroduc‐tion is only likely in stands that were degraded only a few years ago. On the other hand, the presence of viable seeds of Nanocyperion and Parvocaricetea species is promising for the restoration of these communities even after decades. Recreation of pioneer habitats by sod cutting will preserve these species.     

Prospects for fen meadow restoration on severely degraded fens

The majority of fens in Europe have been transformed for agricultural purposes and have disappeared or become degraded. Fen meadows that developed under low-intensity management of fens also have become degraded. In this paper, we consider the available restoration methods, biotic constraints for restoration and new prospects and approaches for the restoration of severely degraded fens. Due to irreversible changes in landscape settings, hydrology, soil and trophic conditions, a full restoration to natural mires is unlikely. Yet, an improvement of the ecosystem functions and revival of biodiversity in degraded fens is possible. A restoration of semi-natural meadows is one of the alternative targets. Important for restoration efforts to succeed are a sufficient reduction of nutrient levels and preventing acidification. In general, a combination of topsoil removal and seed transfer is an effective measure for fen meadow restoration, provided that groundwater seepage can be re-established. There are also several biotic limitations to fen meadow restoration, due to limited propagule availability of target species and the legacy of the former vegetation in form of its soil seed bank and high seed production by unwanted species. Under the present environmental conditions, the re-development of fen meadows on degraded fens will result in species compositions different from those observed in the past and such restoration may require considerable time and effort.     

Plant traits: Their role in the regeneration of alpine plant communities in sub‐arctic Finland

Abstract. Question: How do the relative frequencies of plant traits (clonality, growth form, seed weight, diaspore morphology) vary during the life cycle and how does this affect regeneration? Location: Alpine meadow and heath communities at Kilpisjärvi, sub‐Arctic Finland. Methods: Control plots and three treatments were used to measure relative species abundances for five life cycle stages: standing vegetation, seed rain, seed bank and seedlings emerging in gaps and in closed vegetation. Results: The relative frequencies of plant traits varied between the life cycle stages. The meadows were dominated by weakly clonal herbs, small or intermediate seeds and unappendaged diaspores, while the heaths were dominated by clonal dwarf shrubs, small seeds and fleshy fruits. In the meadows, species with small seeds dominated during the seed rain and in the seedling stage in gaps, while species with intermediate seeds dominated the seed bank and the seedling stage in closed vegetation. Species with unappendaged diaspores dominated throughout the life cycle. In the heaths, seed bank and seedling stage were practically absent. Conclusions: The observed differences in plant trait spectra between life cycle stages indicate that important environmental factors differ among the stages. Small seeds are advantageous for dispersal, whereas intermediate seeds have a greater probability of germinating and establishing in closed vegetation. Appendages facilitate dispersal, whereas unappendaged diaspores favour seed burial. Although the plant growth form spectrum largely reflects environmental constraints during the regeneration cycle, information on seed weight and diaspore morphology improves our knowledge of the relative importance of morphological adaptations of sexual structures in different stages during the life cycle.     

Comparison of soil seed banks of habitats distributed along an altitudinal gradient in northern Iran

In this study we investigated the variations in soil seed banks along an altitudinal gradient in the Alborz mountains, Iran, covering three habitats from lower to upper altitudes: forest, forest-subalpine grassland ecotone and subalpine meadow. In each habitat from 1850 to 2400 m, 20 quadrats were established along four transects, and the above-ground vegetation and the germinable seed banks were determined. Results show that the similarity between seed bank and vegetation was lowest in the ecotone located at intermediate altitudes. Together with the contrasting highest density and species diversity of seeds at these altitudes, the ecotonal role of this habitat was confirmed.We found evidence that lower altitudes could act as storage for seeds of some species growing at higher altitudes; the role of the ecotone was more prominent as a reserve for the meadow plant seeds than the role of the forest as a reserve for seeds of the meadow and ecotone habitats. Soil seed banks, particularly from the ecotone, can be used for restoring vegetation in some degraded sites.     

Assessing soil seed bank persistence in flood‐meadows: The search for reliable traits

Abstract. To assess seed bank persistence of target species in endangered flood‐meadows (alliances Cnidion and Molinion), we investigated established vegetation and soil seed bank of 46 plots for 3 yr and 2 yr, respectively. As traits of seed persistence we calculated various continuous indices that refer to the frequency and abundance of species in above‐ground vegetation and at different soil depths. Furthermore, we tested the significance and soundness of easily observed traits such as maximum seed density per plot and seed attributes (mass, size and shape) as predictors of soil seed bank features. In linear regression, SAI, the seed accumulation index, showed the best agreement (R²= 0.64) with the seed longevity index that was derived from the database by Thompson et al. (1997) for a set of 115 species. The second best predictor (R²= 0.39) of the seed longevity index was maximum seed density per plot in the lower soil layer (5–10 cm). Depth distribution indices and seed attributes showed weaker but still significant relations. The dynamic character of flood‐meadows was reflected by a large proportion of species with a strong tendency to accumulate seeds in the soil relative to their importance in above‐ground vegetation. Most of these species have a ruderal strategy, exploiting gaps after flood disturbances, while the dominants of flood‐meadows tended to have short‐lived seed banks. Compared to other grassland types, a relatively large proportion of rare and endangered target species can be expected to form long‐term persistent seed banks. However, only under marginal conditions that facilitate seed survival in the soil (e.g. fallow) are these persistent seed banks likely to contribute to restoration. We conclude that easily observed traits of persistence such as seed weight, size and shape do not meet the accuracy needed in scientific and practical applications. Thus, there is a crucial demand for further seed bank studies in poorly investigated habitats and of rare species.     

Impact of flooding on potential and realised grassland species richness

In order to reduce flood risk, river management policies advise floodplain restoration and the recreation of water retention areas. These measures may also offer opportunities for the restoration of species-rich floodplain habitats through rewetting and the restoration of flood dynamics. The potential to enhance biodiversity in such flood restoration areas is, however, still subject to debate. In this paper we investigate whether flooding along a small altered lowland river can contribute to the potential and realised species richness of semi-natural meadows. We compare the seed bank and vegetation composition of flooded and non-flooded semi-natural meadows and test the hypothesis that flooding contributes to an input of diaspores into the meadow seed banks, thereby promoting seed density and potential species richness. Furthermore we hypothesise that, where habitat conditions are suitable, flooding leads to a higher realised species richness. Results showed that seed densities in flooded meadows were significantly higher than in non-flooded meadows. The seed banks of flooded meadows also contained a higher proportion of exclusively hydrochorous species. However, the seed bank species richness, as well as the species richness realised in the vegetation did not differ significantly between flooded and non-flooded meadows. Finally, the seed bank and standing vegetation of flooded sites showed larger differences in species composition and Ellenberg nitrogen distribution than non-flooded sites. From these results we conclude that, although flooding does contribute to the density and composition of the seed bank, most imported seeds belong to only a few species. Therefore, it is unlikely that flooding substantially enhances the potential species richness. Furthermore, even if new species are imported as seeds into the seed bank, it seems unlikely that they would be able to establish in the standing vegetation. However, it is unclear which factors impede the establishment of imported species in the vegetation. The implications of our findings for flood meadow restoration are discussed.     

Seed bank diversity in mesic grasslands in relation to vegetation type,management and site conditions

Questions: 1. Do different management types (i.e. hay meadow, silage meadow, meadow‐pasture, pasture) have different impact on the size and composition of the seed bank of mesic grassland (Arrhenatheretalia)? 2. How strong is the effect of management on the seed bank in relation to above‐ground vegetation, edaphic factors and land‐use history? 3. Are there differences in C‐S‐R plant strategy types and seed longevity under different management regimes? Location: Lahn‐Dill Highlands in central‐western Germany. Methods: Above‐ground vegetation and the soil seed bank of 63 plots (at 21 sites) in mesic grasslands were studied. Differences between management types in quantitative seed bank traits and functional characteristics were tested by ANOVA. The impact of management, above‐ground vegetation, site conditions and land‐use history on seed bank composition were analysed by partial CCA. Results: Management had no significant impact on species richness and density of the seed bank but significantly influenced their floristic composition and functional characteristics. CCA revealed that even after adjustment for soil chemical parameters and above‐ground vegetation management still had significant impact on seed bank composition. ANOVA revealed that silage meadows contained higher proportions of R‐strategy compared to hay meadows. In contrast, in hay meadows and meadow‐pastures proportions of S‐strategy were higher than in silage meadows. Conclusions: The type of grassland management has little impact on quantitative seed bank traits. Management types with a high degree of disturbance lead to an increase of species following a ruderal strategy in the seed bank. Irrespective of management type only a limited proportion of characteristic grassland species is likely to re‐establish from the seed bank after disappearance from above‐ground vegetation.     

Influence of Grazing on Soil Seed Banks Determines the Restoration Potential of Aboveground Vegetation in a Semi‐arid Savanna of Ethiopia

Species composition, number of emerging seedlings, species diversity and functional group of the soil seed banks, and the influence of grazing on the similarity between the soil seed banks and aboveground vegetation, were studied in 2008 and 2009 in a semi‐arid savanna of Ethiopia. We tested whether the availability of persistent seeds in the soil could drive the transition from a degraded system under heavy grazing to healthy vegetation with ample perennial grasses. A total of 77 species emerged from the soil seed bank samples: 21 annual grasses, 12 perennial grasses, 4 herbaceous legumes, 39 forbs, and 1 woody species. Perennial grass species dominated the lightly grazed sites, whereas the heavily grazed sites were dominated by annual forbs. Heavy grazing reduced the number of seeds that can germinate in the seed bank. Species richness in the seed bank was, however, not affected by grazing. With increasing soil depth, the seed density and its species richness declined. There was a higher similarity in species composition between the soil seed bank and aboveground vegetation at the lightly grazed sites compared with the heavily grazed sites. The mean similarity between the seed banks and aboveground vegetation was relatively low, indicating the effect of heavy grazing. Moreover, seeds of perennial grasses were less abundant in the soil seed banks under heavy grazing. We concluded that restoration of grass and woody species from the soil seed banks in the heavily grazed areas could not be successful in semi‐arid savannas of Ethiopia.     

Seed bank offers potential for active restoration of mountain meadows

The nitrogen‐fixing legume Lupinus polyphyllus invaded semi‐natural mountainous grasslands across Europe during the last decades. This invasion resulted in degraded habitats through changes in the structure and function of the mountain meadow vegetation. In our study, we analyzed (1) the effects of increasing cover of L. polyphyllus on the seed bank of mountain meadows, and (2) the potential of the seed bank of these stands for active restoration of mountain meadows in terms of species composition and number. We conducted a seed bank analysis on 84 plots with increasing cover of L. polyphyllus in three mountain‐meadow types of the Rhön Biosphere Reserve, Germany. Seedlings from 119 species germinated from the seed bank samples, including 17 Red List species but only a few seedlings of L. polyphyllus. The species composition of the seed bank matched distinct patterns of the three meadow types, but differed from the species composition of the current aboveground vegetation in a nonmetric multidimensional scaling ordination. While the influence of L. polyphyllus on the current vegetation was visible, no effects on the seed bank were apparent. L. polyphyllus had no influence on total seed density, seed density of typical mountain‐meadow species, or species numbers in the seed bank. Only the seeds of the Red List species were significantly related to the cover of L. polyphyllus. We conclude that the seed bank offers potential for active restoration of species‐rich mountain meadows, but species absent from the seed bank have to be added by other measures.     

Restoration of wooded meadows ‐ a comparative analysis along a chronosequence on Öland (Sweden)

Abstract. Wooded meadows on the Baltic Island of Öland result from traditional agricultural management over centuries which has led to a species‐rich vegetation with high species diversity. Today, nearly all of these meadows have been abandoned and became rapidly overgrown by deciduous shrub and tree species forming a closed canopy which resulted in a rapid and strong decrease in species numbers of the herb layer. Recent efforts aim to restore overgrown wooded meadows by cutting single shrubs and trees to open the canopy. However, the effects of abandonment as well as of any restoration management in wooded meadows have rarely been documented until now. Mechanisms driving succession after restoration such as the dispersal potential of the respective species over time and space have not been analysed yet. Therefore, a chronosequence was studied which included a traditionally managed wooded meadow, an overgrown meadow which has been abandoned for more than 100 yr and a meadow which was restored 36 yr ago by cutting and is now grazed. We analysed the soil seed bank of the 3 meadows in comparison with the established vegetation and endozoochorous seed dispersal by cattle and sheep. After abandonment 87% of the typical grassland species vanished from the established vegetation and were replaced by species characteristic of woodland and disturbed grassland communities. The mean number of species decreased from 52 species per plot (4 m²) to 18 species. Mean species number and number of seeds in the seed bank declined significantly from the traditionally managed to the overgrown meadow. Most of the grassland species were assigned to a transient seed bank type while only 1/3 could be classified as having a short‐term persistent seed bank. Thus, restoration of wooded meadows cannot rely on the soil seed bank. Endozoochorous seed dispersal by cattle and sheep was shown for 15% of the species with seed densities per 100 g air dried dung of 737 and 767, respectively. Movement of animals between ancient and restored wooded meadows is recommended since many of the species only occurred in low densities and therefore, will probably not be found in the dung samples.     

Role of soil seed bank along a disturbance gradient in an alpine meadow on the Tibet plateau

We examined the role of the soil seed bank along a grazing disturbance gradient and its relationship with the vegetation of alpine meadows on the Tibet plateau, and discussed the implications for restoration. The seed bank had a high potential for restoration of species-rich vegetation; 62 species were identified in the vegetation and 87 in the seed bank, 39 species being common to both. Mean seed density was 3069–6105 viable seeds m⁻². The density of buried seeds increased significantly with increasing disturbance, indicating that restoration of disturbed areas is not seed limited. Seed density and species richness decreased with depth. The proportion of perennial species decreased with decrease in disturbance both in seed bank and in vegetation. A large portion of species with persistent seeds in the disturbed areas indicate that this seed type can be regarded a strategy of adaptation to current disturbances. Detrended correspondence analysis (DCA) showed significant differences of species composition between seed bank and vegetation, except for the seriously disturbed site. Our results suggest that the establishment of new species in severely disturbed areas is more dependent on the seed bank. By contrast, the restoration in less-disturbed and mature meadows does not rely on seed banks, and the establishment of the vegetation in these communities is more likely to rely on seed dispersal from the standing vegetation and on species with vegetative reproduction.     

Restoration of brook valley meadows in the Netherlands

Until recently, restoration measures in Dutch brook valley meadows consisted of re-introducing traditional management techniques, such as mowing without fertilisation and low-intensity grazing. In the Netherlands, additional measures, such as rewetting and sod cutting, are now carried out on a large scale to combat negative influences of drainage and acidifying influences by atmospheric deposition. An analysis of successful and unsuccessful projects shows that restoration of brook valley meadows is most successful if traditional management techniques are applied in recently abandoned fields that had not been drained or fertilised. Large-scale topsoil removal in former agricultural fields that had been used intensively for several decades is often unsuccessful since seed banks are depleted, while hydrological conditions and seed dispersal mechanisms are sub-optimal. In areas with an organic topsoil, long-term drainage had often led to irreversible changes in chemical and physical properties of the soil. Successful sites were all characterised by a regular discharge of calcareous groundwater provided by local or regional hydrological systems, and, where not very long ago, populations of target species existed. On mineral soils, in particular, sod removal in established nature reserves was a successful measure to increase the number of endangered fen meadow species. It is argued that attempts to restore species-rich meadows should be avoided on former agricultural fields, where pedological processes have led to almost irreversible changes in the soil profile and where soil seed banks have been completely depleted. From a soil conservation point of view, such areas should be exploited as eutrophic wetlands that are regularly flooded.     

Effects of disturbance intensity on seasonal dynamics of alpine meadow soil seed banks on the Tibetan Plateau

Background and Aims

Information on soil seed bank processes is crucial for understanding vegetation dynamics. Despite the documented importance of soil seed banks in many ecosystems, their role is not fully understood in some sensitive habitats, such as the alpine meadows of the Tibetan Plateau.

Methods

We studied the seasonal dynamics of the germinable soil seed bank under four disturbance intensities in an alpine meadow on the Tibetan Plateau as well as seed size distribution relative to disturbance intensity. Composition of the seed bank was compared with that of the standing vegetation.

Results

Density of buried seeds increased with disturbance intensity, but species richness and species diversity decreased. Seed density and species richness of the seed bank varied seasonally in all layers (0–2, 2–7, 7–12 cm) and the whole (0–12 cm). The species composition of seed bank was not significantly influenced by season. There was no trend in seed size distribution as disturbance increased. Seasonal seed bank turnover rates increased with increase in disturbance. The result of the NMDS showed that species composition of seed bank and vegetation exhibited a fairly uniform pattern in each season.

Conclusions

Although as a whole the species composition of the vegetation and seed bank showed a relatively low degree of similarity in each season, similarity was highest in the most disturbed habitat. There was no alteration in species composition of seed bank regardless of disturbance intensity, but seed density decreased as disturbance increased. Disturbances in alpine plant communities might increase persistence of regeneration niches. Regeneration from the seed bank together with vegetative reproduction contributed to aboveground vegetation in highly disturbed habitats. Clonal species played an important role in regeneration of vegetation in slightly disturbed areas, where there was little contribution of ruderals from soil seed banks.     

Seed rain and its relationship with above-ground vegetation of degraded Kobresia meadows

Seed rain is a crucial element in vegetation regeneration, but has been rarely studied in high altitude regions, particularly degraded Kobresia meadow. Weed infestation is a distinctive feature of pasture degradation in Kobresia meadows on the Tibetan plateau, the ecological mechanism of which is closely related with vegetation’s seed rain. In this paper we assess the effect of vegetation degradation on seed rain and consider its implication for restoration of degraded Kobresia meadows in the headwater area of Yellow river, through analysis of seed species composition, number of seeds landing per m² of soil surface, and their relationship with above ground vegetation. Vegetation degradation had an impact on the species composition and numbers of seeds in seed rain and their relationship with above-ground vegetation. Within the un-degraded meadow, which provided a closed vegetation cover, 35 % of the seed rain was of sedge and gramineae species. However, within the degraded meadows, as the extent of degradation increased, so the total number of seeds m^?2 increased, with those derived from sedge and gramineae species forming a declining proportion of the total. Degradation of Kobresia meadow on the Tibetan plateau is exacerbated by the seed input of weed species (such as Oxytropis ochrocephala, Carum carvi, Aconitum pendulum, Pedicularis kansuensis in this study). Therefore, a major priority for the restoration of such degraded meadows should be the elimination of these weeds from the above ground vegetation by human intervention.     

The potential of soil seed banks of a eucalypt wetland forest to aid restoration

Soil seed banks can play an important role in the regeneration of wetland vegetation. However, their potential role in the restoration of degraded wetland forests is less certain. I surveyed the soil seed bank and extant floras of four sites across a eucalypt wetland forest of variable vegetation condition. At each site, the extant vegetation was surveyed within two 5 × 5 m² quadrats, each from which five composite soil seed bank samples were collected. Across the four sites, 57 (including 18 exotic) species were identified in the extant vegetation, while from the seed bank samples 6379 seedlings emerged from 80 taxa, 33 of which were exotic species. The soil seed bank was dominated by native and exotic monocots, and contained very few seeds of wetland tree or shrub species. Overall, the similarity between the extant and seed bank floras was very low (~24 %). Soil seed banks are likely to be of limited use in the restoration of degraded wetland forests, because the dominant species in such systems—woody and clonal plants—are typically absent from the soil seed bank. Wetland soil seed banks may contribute to the maintenance and diversity of understorey vegetation, however, they may also act as a source of exotic plant invasions, particularly when a wetland is degraded.     

Wetland drying indirectly influences plant community and seed bank diversity through soil pH

High altitude wetlands on the Tibetan Plateau have been shrinking due to anthropogenic disturbances and global climate change. However, the few studies that have been conducted on wetlands are inconclusive about the effect of soil moisture on seed banks and potential of seed banks in wetlands with different levels of soil moisture for regeneration of dried wetlands. We investigated seed banks and plant communities along a soil moisture gradient. A structural equation model was used to analyze the direct and indirect effects of soil moisture on seed banks, as well as the relationship between plant communities and seed banks. Although soil moisture had no direct effects on seed bank richness and density and indirect effects on seed banks through plant community, it had indirect effects on the seed bank through soil pH. Soil moisture also did not have direct effects on plant community richness, but it had indirect effects through soil pH. Plant community composition changed with soil moisture, but aboveground plant abundance and seed banks composition did not change. Low similarity exists between plant community and seed banks for all wetlands, and similarity decreased along the moisture gradient. The key factor determining plant community diversity was soil pH, while seed bank diversity was mainly affected by soil pH and plant community diversity with wetland drying. Although potential for regenerating the plant community from the seed bank decreased with an increase in soil moisture, drained wetlands still have enough residual seeds for successful restoration of species-rich alpine meadows.     

Hidden biodiversity in the Arctic – a study of soil seed banks at Disko Island,Qeqertarsuaq, West Greenland

Seed dispersal is a key process in plant community dynamics, and soil seed banks represent seed dispersal in time rather than in space. Despite their potential importance, seed bank dynamics in the Arctic are poorly understood. We investigated soil seed banks and corresponding plant community composition in three contrasting vegetation types in West Greenland, viz. dwarf shrub heaths, herb slopes and fell‐fields. Through germination testing, 31 species were documented in soil seed banks. All of these were herbaceous, while no dwarf‐shrub species, which represents the dominating growth form in the above‐ground vegetation, were emerging from the seed bank. Consequently, across vegetation types, the lowest similarity between seed bank and above‐ground vegetation was found in dwarf shrub heath. Nine plant species were exclusively found in seed bank, all of which were non‐clonal forbs. Seed bank size (total number of seeds) and species richness seemed to increase with the level of natural disturbance. Additionally, we examined the effect of different experimental light and temperature conditions on the quantity and diversity of germinating seeds. The difference in diversity in vegetation and seed bank at the species level will impact population dynamics, regeneration of vegetation after disturbances and its potential to respond to climate change.     

Size traits and site conditions determine changes in seed bank structure caused by grazing exclusion in semiarid annual plant communities

1. Contrasting patterns of change in the seed bank of natural grasslands are frequently found in response to grazing by domestic herbivores. Here, we studied the hypotheses that a) patterns of change in seed bank density and composition in response to grazing depend on spatial variation in resource availability and productivity, and b) that variation among species in patterns of seed bank response to grazing is linked to differences in species size traits (i.e. size of plant, dispersal unit and seed). 2. Effects of sheep grazing exclusion on the seed bank were followed during five years in a semiarid Mediterranean annual plant community in Israel. Seed bank density and composition were measured in autumn, before the rainy season, inside and outside fenced exclosures in four neighboring topographic sites differing in vegetation characteristics, soil resources and primary productivity: Wadi (dry stream terraces, high productive site), Hilltop, South‐ and North‐facing slopes (less productive sites). 3. Topographic sites differed in seed density (range ca 2500–18000 seed m^?2) and in seed bank response to grazing exclusion. Fencing increased seed density by 78, 51 and 18% in the Wadi, South‐ and North‐facing slopes, respectively, but had no effect in the Hilltop. At the species level, grazing exclusion interacted with site conditions in determining species seed bank density, with larger or opposite changes in the high productive Wadi compared to the other less productive sites. 4. Changes in seed bank structure after grazing exclusion were strongly related to species size traits. Grazing exclusion favored species with large size traits in all sites, while seed density of tiny species decreased strongly in the high productive Wadi. Species with medium and small size traits showed lesser or no responses. 5. The size of plants, dispersal units and seeds were strongly correlated to each other, thus confounding the evaluation of the relative importance of each trait in the response of species to grazing and site conditions. We propose that the relative importance of plant size vs seed size in the response to grazing changes with productivity level.