MULTIPLE STABLE STATES AND MODELS OF RIPARIAN VEGETATION SUCCESSION ON THE ANIMAS RIVER, COLORADO

Natural and human disturbances have had fundamentally different effects on ecosystems. Range scientists have proposed a new successional model, a ''state-and-transition model,'' which recognizes the possibility of multiple stable states in vegetation. This paper assesses this model via an investigation of the influence of fluvial processes and landforms on variation in a mosaic of riparian vegetation patches of different ages on a river reach in montane southwestern Colorado. We sampled the composition of herbaceous and shrubby plants in 67 contiguous patches along a 6 km reach of the river, measured 14 environmental variables in each patch, and then analyzed the relationship of the vegetation and environmental gradients using correspondence analysis. The variation in vegetation correlates most highly with the age of the patch, surficial-sediment size, and soil development, which represent different aspects of a gradient in time since last disturbance by floods. A consistent pattern of post-flood succession is not evident in the trends in species richness, mean percent cover, and species composition. The evidence suggests that exceptional floods in 1911 and 1927 altered fundamentally the physical environment on newly exposed bars. As a result, succession slowed and its trajectory may have been redirected toward a new stable state. While this may be a case of multiple stable states in a natural ecosystem, we suggest that such cases are rare in natural ecosystems. We also argue that proponents of state-and-transition models conclude wrongly that these models imply that managers can, or should, choose the desired state based primarily on human uses instead of managing for the potential natural vegetation. That decision is a societal one that should not be implicit in a successional model.