Grass-shrub interactions in a Chihuahuan Desert ecosystem: can vegetation structure predict state-transitions?

State transition from perennial grassland to woody plant domination is synonymous with land degradation and desertification in many arid and semiarid ecosystems. Land managers and other stakeholders could benefit from research designed to illustrate vegetation structural properties (e.g., plant size, density, spatial pattern, life history stage) that might provide tangible metrics that can be used to forecast when transition thresholds are being approached and when ecological sites might be at risk for state transitions. Such knowledge would enable management and restoration efforts to be prioritized and applied at times and in places having the greatest likelihood of averting undesirable transitions and promoting desired transitions. We approached this objective via a combination of linked descriptive and experimental studies along a grassland (Bouteloua eriopoda)-to-shrubland (Prosopis glandulosa) encroachment gradient at the Jornada Basin LTER (JRN). In the descriptive phase, B. eriopoda patch structure (area, height) and biomass was estimated in ninety plots (1 x 1 m) established along the shrub encroachment gradient and the distance to and size (height, canopy dimensions, biomass) of all P. glandulosa individuals within a 5m radius of each plot was quantified. In the subsequent experimental phase of the study, all P. glandulosa shrubs within 5 m of grass plots were killed via foliar herbicide in half (n=45) of the plots and B. eriopoda patch structure and biomass was re-recorded at the end of the subsequent growing season. Total shrub canopy biomass ranged from 1.6 g to 809 g and increased significantly along the grassland-to-shrubland gradient (p < 0.0001). Correspondingly, herbaceous biomass ranged from 7.92 g and 185.7 g and decreased significantly along the gradient (p = 0.0277). Unexpectedly, we found no significant relationship between local P. glandulosa neighborhood structure (e.g. shrub biomass) and B. eriopoda productivity (p = 0.162). This would suggest that shrubs located more proximal or distal than 5 m from target grass patches may be influencing local grass production.  However, B. eriopoda biomass production in plots with P. glandulosa neighbors removed was 30% greater than that in areas where shrubs were left intact (20.24 g m^-2 vs. 14.06 g m^-2; p = 0.0027). Analysis of shrub influence on grass patches at finer scales will ostensibly provide an understanding of the relationship between local B. eriopoda productivity and patch structure and P. glandulosa neighborhood attributes. Additional metrics used in community and population ecology models will be evaluated to assess the demographic and competitive interplay between grass and shrub functional types.