Predicting plant traits and functional types response to grazing in an alpine shrub meadow on the Qinghai-Tibet Plateau

Abstract

The identification of easily measured plant functional types (PFTs) that consistently predict grazing response would be a major advance. The responses to grazing of individual traits and PFTs were analyzed along a grazing gradient in an alpine shrub meadow on the Qinghai-Tibet Plateau, China. Three response types were identified; grazing increaser (GI), grazing decreaser (GD), and neutral (NE) for both traits and PFTs. Seven traits were measured: plant height, economic group, cotyledon type, plant inclination, growth form, life cycle, and vegetative structure. The first five were significantly affected by grazing. Ordinal regressions for grazing response of the seven traits showed that the best single predictors of response were growth form (including the attributes “Scattered”, “Bunched” or “Closely Bunched”), and plant inclination (“Rosette”, “Prostrate”, or “Erect”), followed by economic group (“Shrub”, “Grass”, “Sedge”, “Legume”, “Forb”, or “Harmful”) and plant height (“Tall”, “Medium”, or “Small”). Within the four optimal traits, the summed dominance ratio (SDR) of small plants, forbs, rosette and bunched plants, invariably increased, while that of tall plants, shrubs, grasses, and erect plants decreased, when grazing pressure was enhanced. Canonical correspondence analysis (CCA) identified eleven explanatory PFTs based on 195 defined PFTs, by combining the different attributes of the four optimal traits. Among explanatory PFTs, the most valuable in predicting the community response to grazing were Tall×Shrub×Erect×Scattered and Small×Forb×Rosette, as these have the closest connections with grazing disturbance and include fewer species. Species richness, diversity, and community evenness, did not differ among grazing treatments because turnover occurred in component species and their relative abundances along the grazing gradient. We have demonstrated that a minimum set of PFTs resulting from optimal individual traits can provide consistent prediction of community responses to grazing in this region. This approach provides a more accurate indicator of change within a changing environment than do univariate measures of species diversity. We hope to provide a link between management practices and vegetation structure, forming a basis for future, large scale, plant trait comparisons.