Plowing and subsequent abandonment of semiarid grasslands in the shortgrass steppe region of North America results in both short- and long-term changes in plant community structure. The traditional Clementsian model of succession in which shortgrasses rapidly dominate the vegetation was modified for these grasslands in the 1970s so that it predicted a prolonged stage characterized by the dominance of the bunchgrass Aristida purpurea, followed by a very slow recovery of shortgrasses after large-scale disturbances. Because neither the Clementsian nor the modified model was supported by results of recent scale-dependent field experiments and simulation analyses, we designed a study to evaluate recovery of shortgrass communities on old fields abandoned for 53 yr in northeastern Colorado, USA. Our objectives were: (1) to compare species composition on abandoned fields with that of adjacent, unplowed areas, (2) to compare vegetation on these fields with predictions from the prevailing conceptual models, and (3) to evaluate the relationship between recovery patterns within fields and distance from the source of propagules at the edge of a field. We reached different conclusions based upon the choice of indicator of recovery. For most cases (9 of 13 fields), relative shortgrass cover did not fit predictions of either the Clementsian model or the modified model. High shortgrass cover on two of the remaining fields was similar to that expected by the Clementsian model, and low shortgrass cover on the remaining two fields was similar to that expected by the modified model. Two fields with high shortgrass cover were dominated by Buchloe dactyloides, a species less resistant to drought and grazing than is Bouteloua gracilis, the dominant species in undisturbed communities. Uniformity in cover of other perennial graminoids and density of perennial forbs and annuals on and off fields indicated that these groups had recovered on most fields. However, differences in similarity of species composition on and off fields indicated that none of the fields had recovered. High variability in recovery of vegetation among fields with similar annual climatic variables and soil textures may be attributed to differences in initial conditions, management practices through time, fine-scale climate, and/or other site characteristics that were not measured in this study. We found the perennial bunchgrass Bouteloua gracilis on all fields sampled, and it dominated basal cover on two fields. Four groups of fields were distinguishable based on the relationship between Bouteloua gracilis cover and distance from the edge with unplowed vegetation: (1) fields with uniformly high cover of Bouteloua gracilis; (2) fields with a decrease in cover with distance, and cover dominated by Bouteloua gracilis; (3) fields with a decrease in cover with distance, and cover dominated by Buchloe dactyloides; and (4) fields with uniformly low cover of Bouteloua gracilis and Buchloe dactyloides, and dominated by other perennial graminoids, indicating that a mid- to late-successional stage had been reached. Our results contrast with the conventional view of shortgrass community response to disturbances, and suggest an alternative view of the recovery process that focuses on interactions between individual plants and their environment to explain recovery patterns that vary in time or space. Accounting for this variability in recovery is critical to the management of these systems, especially under conditions of changing climate and land use.
