Influence of soil seed bank, seed rain, inhibition competition and site disturbance on successional processes within three long-term grazing regimes on the Edwards Plateau, Texas

Abstract

Mechanisms of vegetation dynamics on a shortgrass (Hilaria belangeri) dominated semiarid Texas grassland were investigated. Study sites, which potentially support midgrass communities, were relatively homogeneous, open areas within an heterogeneous woodland matrix. Once transformed into a shortgrass configuration by excessive grazing, they persist as such, and are slow to recover once exogenous disturbance is removed. Three (>37 year) grazing treatments [heavy continuous (HC); moderate deferred rotational (MDR); and exclosure (EXC)] were sampled for seed bank. Absolute seed densities were not different (p>0.05) among treatments, but species composition varied greatly. Later-successional midgrass seed pools were lacking in the HC treatment, and poorly represented in the MDR and EXC treatments. Seed densities varied seasonally (p[less than or equal to]0.01) as did composition, which was attributed to differing persistence of constituent taxa. A high proportion of annual dicot taxa were present in the HC treatment, shifting to a greater proportion of perennial monocot taxa within the EXC treatment. Seed rain in the HC treatment lacked midgrasses but the EXC treatment was rich in those taxa. Detrended correspondence analysis (DCA) axes were highly correlated (r[greater than or equal to]0.65, p[less than or equal to]0.05) with grazing disturbance and seasonal gradients for both seed bank and extant vegetation. Four disturbances [control (C) -- undisturbed vegetation; herbicide (H) -- glyphosate herbicide; herbicide - rake (HR) -- glyphosate herbicide with litter removed; and till (T) -- tilled to 15 cm), each 2.8 m², were superimposed on grazing treatments and seeded with midgrasses. Midgrass recruitment was markedly affected by treatment type. Intact vegetation limited midgrass establishment to <2% of seed sites, whereas disturbed treatments had up to 32% success. Time-site trajectories of species density DCA ordinations for disturbed sites within grazing treatment exhibited distinct rates and patterns of change. The HC recovered most slowly, MDR intermediately and EXC the most rapidly and completely, with trajectories converging on the control. The ungrazed (EXC) site was most resilient to this scale of disturbance. Differential responses of disturbance treatments within grazing treatments occurred. Grazing history significantly affected the biotic community, thus propagule availability. Furthermore, extant vegetation exerted strong competitive inhibition on midgrass establishment within intact stands. These mechanisms greatly reduce the probability of shortgrass communities rapidly returning to a midgrass configuration.