Seed rain in and between vegetation patches in arid Patagonia, Argentina

1 E.E.A. INTA Bariloche. C.C. 277, 8104 – San Carlos de Bariloche, Argentina. 2 Departamento de Agronomía-CERZOS, UNSur, Altos del Palihue, 8000 – Bahía Blanca, Argentina. Adress Correspondence to: Dr. Carlos A. Busso, e-mail: cebusso@criba.edu.ar Recibido/Received 16.V.2007. Aceptado/Accepted 18.VII.2007. Abstract. Seed rain has ecological consequences on plant community structure. This study was conducted in the Monte Austral Neuquino nearby the town of Picún Leufú, Province of Neuquén, Argentina, within an area of 15 km x 15 km (39o 20 32' S; 69o19 30' W). Studies were conducted at two sites, distant 1.5 km from one another. Vegetation is distributed on the soil surface as vegetation patches. Four superficial locations may be found in the soil of these patches: location 1, small dune accumulated under the lee; location 2, found at the edge of the small dune, and below the major grass canopies and tallest shrubs; location 3, located at the edge of the vegetation patch, and location 4 which includes the uncovered corridors among vegetation patches. Seed rain was measured by collecting seeds from traps located in the four study locations. Our working hypotheses were that (1) seeds of the study species arrive to all locations of the vegetation patches and the nude corridors among patches, and (2) the magnitude of the seed rain depends on time of the year. Larrea divaricata showed the greatest values in the corridors among vegetation patches and in the periphery of such patches. A similar pattern was found for S. neaei although this species dispersed seeds further from the mother plant. Atriplex lampa distributed seeds homogeneously in all four locations. Elymus erianthus only contributed 1 seed to the seed rain. Larrea divaricata showed a greater proportion of its total seed rain between December Seed rain in and between vegetation patches in arid Patagonia, Argentina (With 1 Table & 4 Figures)


INTRODUCTION
Seed rain has ecological consequences on plant community structure (Willson & Traveset, 2000).The dynamics of a plant population in environments with vegetation patches depends not only on its survival on a specific phenological stage and growth in different patches, but also of and March in all locations.Seed rain in A. lampa was almost permanent throughout the year while in S. neaei was concentrated in December and January, although it showed some seed rain until April.
Key words: Seed rain, shrub and grass species, vegetation patches, arid Argentina.
Palabras clave: LLuvia de semillas, especies de gramíneas y arbustos, isletas de vegetación, Argentina árida.the degree of the vegetation patch suitability through those stages.The degree of vegetation patch suitability can have an important impact on the quantity and the spatial pattern of recruitments.In addition, site suitability can not be independent of seed arrival because of density-dependent mortality factors.Regarding colonization and community structure, early colonizers can inhibit later colonizations if the first establish so densely that constraint plant growth among them.Thus some aspects of the spatial pattern of plant succession can be related with dispersion.In U.S.A. desertic areas, what appears to be limiting the initiation of primary succession is not seed dispersal but the low permanency of seeds in these habitats (Kollmann & Pirl, 1995;Schupp, 1995;Schupp & Fuentes, 1995;Kollman & Schill, 1996;Aguiar & Sala, 1997;Russell & Schupp, 1998).Aguiar & Sala (1997) emphasize that new plant recruitment depends not only on seedling establishment but also on seed availability which in turn depends on seed production and dispersal.Seed dispersal has two different phases (1) seed movement from the plant to the soil surface, and (2) the subsequent movement of these seeds on that surface.The relationships between wind speed and direction, height of the seed source and seed characteristics determine the position of landing of seeds which are dispersed by wind.The lateral seed movement is mainly controlled by soil roughness and the seed size and shape.It has been recognized as a very significant part of seed dispersal in environments with sparse vegetation.Seeds are captured by uncovered soil only if they are smaller than the size of soil particles or if they possess a morphology and/or appendages which can allow them to anchor in crevices or soil irregularities.
Seeds and seedlings of most populations often interact with individuals of other species.Thus studies at a scale of communities are necessary to evaluate the importance of seed rain in population dynamics and interactions among populations.Peart (1989) emphasized the lack of studies which relate seed rain density to species abundance in the plant community.
The seed flow either into or out of a habitat unit determines the population of that habitat.Seeds which fall on a piece of land can be a function of several variables: (1) the height and origin of the seed source, (2) the seed source concentration, (3) the seed dispersal capacity (i.e., its weight, presence of appendages, etc) and ( 4) the activity of distribution agents (i.e., wind speed and velocity, herbivory and granivory animals, etc.) (Harper, 1977).
Our working hypothesis was that the seeds of the study species arrive to all microenvironments of the vegetation patches and the nude corridors among patches.The magnitude of the seed rain depends on the time of the year.
Objectives included to (1) determine the magnitude of the seed rain of Larrea divaricata, Atriplex lampa, Stipa neaei and Elymus erianthus in the different locations of the vegetation patches (spatial variabi- lity), and (2) characterize its seasonal dynamics (temporal variability).

MATERIALS AND METHODS
S St tu ud dy y s si it te e. .This research was conducted in the shrubby steppe of L. divaricata and A. lampa, nearby the town of Picún Leufú, Province of Neuquén, within the Monte Austral Neuquino (MAN).As in other arid and semiarid environments of the world, vegetation at this study site had a scattered distribution and a patchy structure (Aguiar et al., 1992;Aguiar & Sala, 1999).Studies were effected within an area of 15 x 15 km (39º 20-32' S; 69º 19-30' W) and conducted at two sites, distant 1.5 km from one another.Four superficial locations may be found in the soil of these vegetation patches: location 1, small dune accumulated windward which generally presents a litter cover; location 2, it is found at the edge of the small dune, and below the major grass canopy and tallest shrubs; location 3, under the lee and located at the edge of the vegetation patch, and location 4 which are the uncovered corridors among vegetation patches (Fig. 1).In each of the two study sites, four vegetation patches were selected and four locations investigated within each patch: (1) windward, (2) below L. divaricata, (3) under the lee, and (4) in the bare corridors among vegetation patches (Fig. 1).
C Cl li im ma at te e. .In Picún Leufú, average long-term  annual precipitation is 137.2 mm, and average annual precipitation during 1999-2005 was 200 mm.Extreme values of precipitation during 1928-1950 have been 62.7 and 245.3 mm (Morello, 1958).Maximum and minimum precipitations occur during winter and spring, and summer, respectively.Monthly average values for diurnal and nocturnal temperatures are 28 and 15 ºC in January, respectively, and 7 and 2 ºC in July, respectively.Mean annual wind speed is 13 km/h (Servicio Meteorológico Nacional, 1958).Summer is the windy season.November, December and January have the greatest wind speed, with a mean of 17 km/h at 2 m from the soil level.
V Ve eg ge et ta at ti io on n. .The shrubby steppe of the dominant L. divaricata and the co-dominant A. lampa occupies approximately 1,080,000 hectares in the MAN.The average height of this steppe is 1.20 m.It is usually present either in plain areas or in those with very little slope (< 5%) (Movia et al., 1982).Three strata can be found within each of the vegetation patches in this steppe (Movia et al., 1982): (1) L. divaricata, Bougainvillea spinosa (Cav.)Heimerl, Monttea aphylla (Miers) Bentham et Hooker and L. cuneifolia Cav. are present with a mean height of 1.50 m, but which can be reduced to 0.80 m on stony or fine-textured soils, (2) the second stratum varies between 0.70-1.00m and is characterized by the presence of A. lampa (0.5-0.8 m height).This species is present either on sandy mounds together with Prosopis alpataco Philippi, Schinus polygamus (Cav.)Cabr., Lycium chilense Miers S So oi il ls s. .Aridisols dominate the area.There are mostly old structural plains.Landscape is controlled for the horizontal disposition of the subjacent mantle.It includes Typic Paleortids, Petrocalcic Calciortids, Petrocalcic Paleargids and Typic Torriorthents, with shallow and saline phases (Ferrer & Irisarri, 1989).According to the soil classification (Soil Survey Staff, 1975) no water is available during long periods.The annual average precipitation is approximately 140 mm and it does not have a clear seasonal distribution (Italconsult Argentina, 1966).However, a light tendency to the Mediterranean type could be observed, with greater humidity during the winter (Morello, 1958).When temperature is sufficient for plant growth, soil water potentials are < 1.5 MPa most of the time.During the warm season, a continuous period of three of more months with available moisture does not exist (Soil Survey Staff, 1975).In addition, the scarce rain has a low infiltration, especially in the corridors between vegetation patches, and most of it is lost by runoff (Bonvissuto, personal observation, EEA INTA Bariloche, Argentina).S Sa am mp pl li in ng g m me et th ho od ds s. .Wind speed behind the shrub protection in the patch, in the patch interspaces at the soil surface level and at 1.5 m height from the soil surface was measured using a portable anemometer.
The magnitude in which seeds arrive to the four microenvironments (within and between vegetation patches, Fig. 1) was studied, which is a measure of its lateral movement.Seed traps were installed in each of the four locations in undisturbed vegetation patches under field conditions.Traps were constructed using plastic from soft drinks of 10 cm diameter.The upper part of each was inverted and utilized as a funnel.This rested on the rest of the plastic container.Traps were 16.5 cm deep and were buried so that the upper part was at the soil surface level.These traps had small holes at the other extreme which allowed for water drainage.
The total of traps was 1 trap/location x 4 locations/vegetation patch x 4 vegetation patches/site x 2 sites = 32 traps.Traps were installed in 1999 and the number of seeds of each species trapped on them was counted every 15 days during more than a year.The diverse seed or fruit structure and sizes of the study species, which would eventually influence its distribution in space, were described by Correa (1978Correa ( , 1984Correa ( , 1988) ) and can be seen in Fig. 2.

RESULTS
There was times when wind speed was greater than 30 km/h at 1.5 m height from the soil surface, and it reached values close to 25 km/h at the soil surface in the interspaces between vegetation patches (Fig. 3).Under the shrubs, values were from 0 to 5 km/h throughout the year.
The total amount of seeds trapped in all sectors of the vegetation patches varied with the study species (Table 1).Larrea divaricata showed the greatest values in the corridors among vegetation patches (location 4) and in the periphery of such patches (location 3).A similar pattern was found for S.neaei although this species dispersed seeds further from the mother plant.Atriplex lampa distributed seeds homogeneously in all 4 locations.Elymus erianthus only contributed 1 seed to the seed rain.(Correa, 1978(Correa, , 1984(Correa, and 1988)).
Table 1.Total of seeds fallen per location at the field in 8 vegetation patches from March 1999 to June 2000 in the study species T Ta ab bl la a 1 1. .Número total de semillas caídas en cada sector en el campo en 8 parches de vegetación desde Marzo 1999 hasta Junio 2000 en las especies en estudio.

DISCUSSION
There was an important seed rain to location 4: 46% of L. divaricata seeds, 55% of the seeds of A. lampa, 45% of the seeds of S. neaei and only one seed of E. erianthus.However, seeded seeds in that location produced numerous seedlings, but none of them survived until the time of harvesting which was conducted after the summer drought (Bonvissuto, 2005).Many seeds of all 4 species arrived to location 4 (Table 1), but there were relatively few in the soil seed bank (Bonvissuto, 2005).
Seeds do not necessarily stay in the places they fall while seed rain is occurring.Lateral movement has been recognized as a significant part of secondary seed dispersal in environments with scarce vegetation (Reichman, 1984).Wind speed and direction are important in secondary seed dispersal.Wind can reach 25 km/h in some cases in the corridors among vegetation patches (Fig. 3).Looking at this figure, we can realize than in the corridors will very likely be more lateral seed movement than in the microenvironments located behind the vegetation patches as a result of wind intensity differences among locations.Wind speed and direction combined with height of the seed source and seed characteristics (i.e., biomass, morphology) to determine the position of seed landing of seeds dispersed by wind (Green, 1983).When precipitation is intense it can also contribute to secondary seed dispersal through runoff (Bonvissuto, personal observation, E.E.A. INTA Bariloche, Río Negro, Argentina).
The total seed rain contributed from all 4 species (L.divaricata, A. lampa, S. neaei and E. erianthus) increased at increasing distance from below the shrub canopies (location 2) (Table 1).The study species demons-trated its dispersion capacity.This is because it was abundant the amount of their seeds which arrived to the most distant locations (corridors among vegetation patches), very likely as a result of the appendages in their fruits.L.divaricata has strong and firm radial trichomes in all seed directions (Fig. 2); this implies that little wind is necessary for making seeds to roll (www.fs.fed.us/database/feis/plants…nical_and_ecological_characteristics.html, 2000).Atriplex lampa presents light bractlets with paper consistency (Fig. 2) which can collaborate in wind dispersal (Osmond et al., 1980).Stipa neaei has a large, beardless awn and very likely is dispersed by wind (www.botany.utoronto.ca/courses/BOT307/307prairie3.html,2005).Time from seed maturity is an important component which influences seed rain.This is because there are seeds which are released as soon as they mature while others need a post-maturation time in the plant before seed dispersal will occur.Seeds of S. neaei, for example, are released from the plant as soon as they mature.Maximum seed rain in this species was during December and January.It is also the case of L. divaricata, which maximum seed dispersal occurred between December and March.On the other hand, A. lampa showed an homogeneous seed rain during the whole year.In this case, seeds were kept in the plant and then released either at once or slowly throughout a prolonged period.
Numerous ecological factors can contribute to seed dispersal (Willson & Traveset, 2000).Ideally, seed maturation should adjust to dispersal agent availability and germination conditions.Limitations to this ideal can derive from natural selection which contributes to avoid (1) seed predators, (2) changes in flowering time and (3) changes in the amount of time required for fruit maturation.In addition, there is environmental variation regarding the timing of (1) maturation and (2) activity of the dispersal agents in any given area.This ultimately will affect rates and quality of seed dispersal.