What processes maintain the vegetation mosaics in arid ecosystems? A modelling study of plant interaction balance and seed dispersal in Patagonia.

CIPRIOTTI, P,A, M.R. AGUIAR, T.WIEGAND, J.PARUELO

Creta, Grecia,

Congreso; 52nd International Symposium of the International Association for Vegetation Science; 2009

Vegetation in most arid ecosystems is organized as two-phase mosaics, where high-cover patches alternate within a matrix of low or null plant cover. However, there is little knowledge about the processes that maintain this structure and how spatial organization feedbacks on these processes. We studied how the balance of biotic interactions between shrubs and grasses (i.e., underground competition and aerial facilitation) and seed dispersal patterns control the long-term coexistence of dominant life-forms and the maintenance of the mosaic. We used a spatially-explicit and individual-based model (DINVEG) that was tailored to describe plant community dynamics of Patagonian steppes dominated by perennial grasses and shrubs with different life histories. DINVEG simulates the spatial and temporal dynamics of six shrub and grass species at plant-level in a 0.25 ha plot with a spatial grain of 400 cm2 and a monthly time-step. To simulate vegetation dynamics in a realistic way, most individual processes (e.g. emergence, growth, survival, seed output) are controlled by the soil water content, the main limiting resource in arid ecosystems. First, we calibrated DINVEG to test three hypotheses on roots’ distribution of shrubs and grasses across soil profile: no overlapping (Walter hypothesis), partial overlapping, and full overlapping. Second, we used the calibrated parameters to run DINVEG in a factorial arrangement under two levels of facilitation effects from shrubs to grass recruitment: null and 3-fold; and two seed dispersal patterns: random and around vegetated patches. Shrub-grass roots overlapping had the major impacts on the coexistence of both life-forms and the mosaic maintenance. Contrary to the Walter hypothesis, coexistence levels resembling field abundance data were only achieved by partial root overlap of shrubs and grasses. Facilitation and seed dispersal pattern had interacting effects on life-form abundance and mosaic structure. Higher facilitation levels and seed dispersal around vegetated patches reduced shrub abundance and cover of high-cover patches at similar values to field data. Moreover, shrubs showed lower levels of aggregation under higher facilitation levels and seed dispersal around vegetated patches prevented. Spatial associations between grasses and shrubs persisted even with null facilitation effect, although they reach maximum intensities under higher facilitation effect.