A hierarchical, multi-scale, management-responsive model of Mediterranean vegetation dynamics

Abstract

Ecosystems of the Mediterranean basin are characterized by a heterogeneous and dynamic landscape mosaic of vegetation formations. This landscape has been shaped over millenia by disturbances associated with agropastoral land use: clearing, grazing and burning, and by regeneration processes of the natural woody vegetation. The ability to predict the effects of management decisions on the structure and composition of the vegetation is essential for present and future land management. To improve this ability we developed a hierarchical multi-scale, management-responsive model of vegetation dynamics. The model was initially developed at two hierarchically nested spatial scales: “cell” (1 m 2) and “site” (400 m 2). The cell model was conceptually derived from a “States and transitions” framework. Vegetation state in a cell was defined by the identity, age and height of the dominant plant functional type. Biological realism was enhanced by defining transition probabilities not as constants but as functions of variables that represent biological processes: reproduction, colonization, expansion, replacement, aging, mortality, grazing, regrowth after fire, and dispersal. Up-scaling to the site was essentially done by averaging: vegetation state of a site was defined by frequencies of cells in different states. However, there was a feedback from site to cell by seed production and dispersal functions. The model was parameterized and implemented for a study area in Mediterranean woody vegetation (garrigue) in Israel. The biological components and processes represented in the model are sufficiently general to allow adaptation to similar vegetation in other Mediterranean regions. A series of simulation experiments over 100 years was carried out under different management scenarios defined by combinations of initial vegetation cover (cleared and control plots), cattle and goat grazing intensity and fire frequency. The results highlight the following trends: • In most scenarios, initial cover has an effect even after 100 years. • In the absence of any disturbance, vegetation becomes dominated by the two tallest woody functional types. • Under extremely intense multiple disturbances (fire + goat grazing), the vegetation becomes dominated by Herbaceous plants. • Scenarios of intermediate disturbance regimes often result in heterogeneous vegetation composition. • Under most of the management scenarios examined, vegetation composition had not reached a steady state (or steady cycle of fire-regrowth) even after 100 years. The model can be a useful tool for land managers by eliminating obviously undesirable scenarios and focusing on a limited set of desirable or acceptable scenarios.