摘要：

The ridge-slough landscape of the Everglades (Florida, USA), is characterized by elevated ridges dominated by sawgrass (Cladium jamaicense) interspersed among deeper sloughs dominated by floating, submerged and emergent macrophytes and calcareous periphyton. Interactions among hydrologie conditions (water depth, hydroperiod), plant composition and production, and respiration are hypothesized to create alternative peat accretion equilibria at the point scale, while spatial interactions among patches create regular pattern at the landscape scale. Despite significant research on these interactions, few studies have examined the hypothesis that the ridge-slough landscape consists of spatially coupled alternative stable patch states, and none has used diagnostic indicators thereof to assess landscape resilience to hydrologie change. Dense random sampling of water depths (a proxy for soil elevation) along a gradient of hydrologie impairment of drained to relatively natural to impounded conditions was used to evaluate four predictions related to this hypothesis: (1) bimodal soil elevation distributions show strong fidelity to community type; (2) positive autocorrelation at short distances with negative values at longer distances; (3) strong anisotropy (diagnostic of flow orientation), and spatial structure (diagnostic of the strength of landscape self-organization); and (4) loss of these features with hydrologie modification. Our results support the hypothesis that soil elevations are strongly bimodal and anisotropic in areas with minimal hydrologie impact, and spatial autocorrelation patterns indicate the operation of scale-dependent feedbacks. These metrics change markedly with hydrologie modification, although with differences between drainage and impoundment. Moreover, changes in landform precede associated changes in vegetation, suggesting their utility as diagnostic indicators of landscape degradation and recovery.

展开