In the mixed C₃/C₄ grassland of the southern Great Plains, USA, the invasive woody legume, honey mesquite (Prosopis glandulosa), affects grass production and composition differently beneath the canopy (subcanopy) than in spaces between trees (intercanopy) due in part to the dominant presence of C₃ Texas wintergrass (Nassella leucotricha) beneath the mesquite canopy and soil enrichment from N-fixation by mesquite. This arrangement, unlike most Prosopis systems worldwide that have C₄ grass or C₃ subshrub understories, uniquely affects grass production spatially and seasonally during mesquite expansion and possibly after anthropogenic removal of mesquite. We compared herbaceous and soil N responses in subcanopy and intercanopy microsites during the first 2 years following a root-killing herbicide mesquite treatment. Perennial grass (PGR) and total herbaceous (THB) production were greater in treated than untreated intercanopy and subcanopy microsites at 1-yr post-treatment, with Texas wintergrass comprising the largest portion of PGR. In year 2, PGR production declined in both treated microsites with no differences between treatments. However, THB production remained greater in treated than untreated microsites due mainly to increased annual forb production that supplanted PGR production from year 1. Increased annual forb production in treated microsites in year 2 was likely due to high rainfall in fall of year 1 that stimulated forb seed germination, increased light from the loss of shading by mesquite, and soil inorganic N that increased from year 1 to year 2. Pre-treatment spatial heterogeneity of herbaceous composition and soil N, caused by mesquite, affected post-treatment patterns of herbaceous production. The unexpected replacement of PGR by annual forbs in year 2 revealed that grass forage production following brush control can deviate markedly from predicted models under certain conditions.  

Data collected by field sampling and lab analysis