Seed distribution constrains the dynamics of the Patagonian steppe

Ecology, Jan, 1997 by Martin R. Aguiar, Osvaldo E. Sala

INTRODUCTION

Vegetation of arid and semiarid environments is sparse and has a patchy structure (Charley and West 1975, Fowler 1984, Schlesinger et al. 1990, Aguiar et al. 1992). Low water availability accounts for the generally low plant cover, but a mechanistic model of the origin, maintenance, and dynamics of patches can only be derived from a thorough understanding of the distribution in space and time of seeds as well as of safe sites for seedling establishment (Harper 1977). Many studies have partially addressed demographic processes such as seed dispersal, seed movement over the soil, seed survival, emergence, and seedling survival. However, few studies link seed dispersal and movement of seeds over the soil surface with the demographic consequences of the final location of seeds (Schupp 1995, Schupp and Fuentes 1995). The spatial distribution of seeds modifies the location of vegetation patches, which, in turn, influences the community by modulating the strength of facilitation and competition (Fowler 1988, Franco and Nobel 1988, Aguiar et al. 1992, Aguiar and Sala 1994). Vegetation patches also constrain ecosystem functioning by determining the spatial pattern of soil organic matter, soil texture, nutrient cycling, and water dynamics (Charley and West 1975, Hook et al. 1991, Gutierrez et al. 1993, Aguiar and Sala 1994, Lauenroth et al. 1996).

The Patagonian steppe in Argentina is composed mainly of tussock grasses 0.2 m in height and hemispheric shrubs 0.6 m in height. Tussock grasses and shrubs are arranged in two kinds of structural patches, one formed by scattered tussocks interspersed with bare-soil areas and the other made up of shrubs each tightly surrounded by a dense ring of grasses (Soriano et al. 1994). The structure of the vegetation creates microsites of different suitability for seedling establishment. In bare-soil areas (exposed microsites), wind speed is five times higher and evaporative demand two times higher than near shrubs with a ring of grasses (protected microsites) (Soriano and Sala 1986). Our previous studies showed that emergence of grass seedlings was equal in the two types of microsites, but survival was three times higher in exposed than in protected microsites. This was the consequence of higher root competition near shrubs with a dense ring of grasses, which overshadowed the protection effect (Aguiar et al. 1992, Aguiar and Sala 1994).

Recruitment of new individuals depends not only on establishment, but also on availability of seeds, which, in turn, depends on seed production and dispersal. Seed dispersal has two phases (Watkinson 1978, Chambers and MacMahon 1994): the air transport of seeds until their landing on the soil surface, and the subsequent movement of seeds over the soil surface. The relationships among wind velocity and direction, height of the seed source, and seed characteristics (e.g., biomass, morphology) determine the landing position of wind-dispersed seeds (Green 1983). Lateral movement of seeds is mainly controlled by surface rugosity and seed shape and size (Matlack 1989, Chambers et al. 1991). Lateral movement has been recognized as a very significant part of dispersal in environments with sparse vegetation (Nelson and Chew 1977, Watkinson 1978, Reichman 1984, van Tooren 1988, Chambers et al. 1991).

Some models of community dynamics assume either no seed limitation or a random or lottery process selecting which species lands first in each patch (Fagerstrom 1988, Coffin and Lauenroth 1990, van Hulst 1992). These models were usually generated for grasslands or forests with high plant cover. We suggest that arid and semiarid communities need a different model because of the importance of spatial heterogeneity and the large proportion of bare-soil areas to total cover. This paper contributes to the empirical basis for developing such a new model.

Our previous studies demonstrated how the balance between facilitation and competition determines the fine spatial distribution of the probabilities of grass seedling establishment in the Patagonian steppe (Aguiar et al. 1992, Aguiar and Sala 1994). However, the spatial distribution of seeds, which is the other part of the equation of recruitment of new individuals, has not been examined. The objectives of this study were: (1) to assess the spatial pattern and dynamics of seed availability in this community and the major factors controlling them; and (2) to combine information on seed patterns with that on seedling establishment to determine recruitment patterns of new individuals in different microsites. We evaluated the spatial distribution of seeds and its seasonal dynamics by following the movements of a cohort of seeds from dispersal until germination. We used this information in a simple model to assess recruitment of new individuals in the different microsites of the steppe.

METHODS

Site description and study species

Our work was conducted in a grass-shrub steppe representative of the Occidental District of Patagonia (Golluscio et al. 1982). The site was located in the Rio Mayo Experimental Site (45 [degrees] 25 [minutes] S, 70 [degrees] 20 [minutes] W) in southwestern Chubut, Argentina. Mean precipitation is 168 mm. Daily mean temperature ranges between 14 [degrees] C in January and 2 [degrees] C in July. Precipitation is concentrated between May and November (winter and early spring), and strong dry winds blow (15 km/h annual mean) mostly from the west.