Aims: Primary biomass production is a fundamental process for ecosystem functioning. Yet, little is known on the mechanisms driving temporal stability of biomass production in annual plant communities, particularly in communities subjected to highly variable environments and undergoing temporal changes in species composition. We aimed to disentangle the relative importance of biomass production, species diversity, dominance and asynchrony of species fluctuations as drivers of biomass stability in Mediterranean and semiarid annual plant communities.

Location: Mediterranean (N31^o42’; E35^o03’) and Semiarid (N31^o23’; E34^o54’) sites, Israel.

Methods: Aboveground biomass and species abundance were monitored in 15 plots of 250m² per site during eight consecutive years. Relationships between stability drivers and community stability were studied at the regional (between-sites) and local (within-sites) spatial scales.

Results: Community biomass stability (temporal mean/SD) increased from the Semiarid to the Mediterranean site concomitantly with higher biomass production, richness, and evenness, but was not associated with changes in species synchrony. Differences in stability between sites were due to opposite effects of site conditions on the mean and SD of community biomass, leading to higher stability in the Mediterranean site. Within sites, species asynchrony was the key driver of stability at the local spatial-scale. Richness and biomass production affected stability indirectly through asynchrony, but in different ways at each site. At the Mediterranean site, these factors had indirect negative effects on stability by reducing asynchrony, but did not rescind a positive effect of asynchrony on community stability. At the Semiarid site, biomass production had indirect positive effects on stability through asynchrony, while richness had no effect on asynchrony and stability. Stability was not driven by species evenness in either site. 

Conclusions: Our study provides new insights into the complex control of biomass stability in the dynamics of Mediterranean and semiarid annual plant communities, with different mechanisms driving stability across the regional vs. local spatial scales. 

 Study sites 

Study sites were located in contrasting ecosystems: Mediterranean (Matta LTER, N 31^o42’; E 35^o03’) and Semiarid (N 31^o23’; E 34^o54’). The climate in both sites is typical Mediterranean, with hot dry summers and mild wet winters, characterized by high rainfall variability which increases with diminishing rainfall. Average rainfall for the hydrological year (October to September) during the experimental period (seasons 2001-2002 to 2009-2010) was 510 mm (CV 25%, range 413-799 mm) and 222 mm (CV 33%, range 132-377 mm) in the Mediterranean and Semiarid sites, respectively. Soil at the Mediterranean site is Terra Rossa, while light brown Rendzina prevails at the Semiarid site. The length of the growing season is determined by the distribution of rainfall, commencing in October-November and ending in April-May. The landscape in both sites is a sparse, stony shrubland, with the spiny dwarf-shrub Sarcopoterium spinosum (L.) Spach as the dominant shrub. The herbaceous vegetation emerging at the onset of the rainy season in open patches between the shrubs is dominated by annual species, and is a main forage source for livestock. Shrubs contribute much less to biomass production due to their lower cover and slower growth. Both sites were subjected to livestock grazing, mainly sheep and goats, before the onset of the experiment, with higher grazing intensity in the Semiarid site.

2.2- Sampling and data collection

We sampled the herbaceous vegetation at peak biomass – late March in the semiarid site and mid-April in the Mediterranean site, during growth seasons 2001-2002 to 2009-2010, except for season 2004-2005. The 2001–2002 season was not included in the stability analyses to avoid immediate disturbance effects caused by the initial setting the rain manipulation treatments. Ten random samples (20 x 20 cm quadrats) of the herbaceous vegetation were taken in the open patches in each one of the 15 plots. This quadrat size is commonly used to study Mediterranean herbaceous communities (Osem, Kigel and Perevolotski, 2002; Peco, Sanchez and Azcarate, 2006) and is appropriate for the small range at which interactions occur between herbaceous plants in both sites. To avoid edge effects, samples were taken at least 1 m away from plot boundaries. Plant roots were cut just below the soil surface to avoid plant fragmentation and allow counting of individual plants to assess plant density. Plant species were identified and their aboveground biomass per quadrat was weighed after 48 h oven drying (75^oC). Since the growth season is short in both sites, biomass loss by litter production prior to sampling is relatively small. Thus, the average aboveground biomass per unit area at peak season can be taken as a proxy of the actual annual primary production (ANPP) in each site. Temporal variation over years in density and biomass of individual species and in community biomass (i.e. sum of species biomasses) was assessed for each plot, by averaging the 10 samples taken in each growing season. Mean biomass of individual plants at the community level was obtained by dividing quadrat biomass by total number of plants in the quadrat. Species nomenclature follows Feinbrun-Dothan and Danin (1998).