Effects of disturbance regimes on honey mesquite (Prosopis glandulosa) cover and spatial pattern, and associated impacts on herbaceous production

Abstract

Herbaceous production and spatial pattern of vegetation were studied in a mesquite savanna in north Texas using a combination of field investigation and remote sensing techniques. Color infrared imagery was classified for analysis of spatial attributes of woody and non-woody components of the savanna including percent cover, mean patch size, patch density, mean shape index, edge density and mean distance to the nearest neighbor. Lacunarity analysis was used to provide an integrative measure of landscape pattern at multiple scales. Comparisons were made between untreated control areas and areas treated with top-killing (triclopyr or 2,4,5-T) or root-killing (clopyralid or triclopyr + clopyralid) herbicides, and between the two treatment types. Significant differences in spatial pattern of vegetation were found between the control and the treatments, as well as between the two treatment types. Three sets of analysis were used to explore the relationship between abundance and pattern of mesquite trees and perennial grass production in interstitial spaces between mesquite. Production data was collected for three functional groups of grasses in a previous study, by clipping herbaceous growth in grazing exclosure cages. First, attributes of four mesquite plants adjacent to cages were measured and their relationship to herbaceous production was analyzed using regression analysis. Results showed that distance to canopy edge, height and canopy diameter of mesquite were significantly related to herbaceous production. Second, production data were compared with percent woody cover in four different neighborhood sizes around each cage in order explore the scale of influence of mesquite on interstitial production. Significant negative relationships were found between percent cover and herbaceous production for all functional groups, with inconsistent results at varying scales. For the third set of analyses, two landscape metrics (inverse distance index and proximity index) were considered to examine a more spatially explicit relationship between the amount and configuration of woody cover and herbaceous production in interstitial spaces. Results indicate that these two metrics do not accurately describe the relationship between spatial pattern of mesquite and interstitial herbaceous production.