PREDICTING BUFFELGRASS SURVIVAL ACROSS A GEOGRAPHICAL AND ENVIRONMENTAL GRADIENT

This research was designed to identify relationships between T-4464 bu ffelgrass (Cenchrus ciliaris L.) survival and climatic and soil charac teristics. At 167 buffelgrass seeding sites in North America we collec ted climatic and soils data where the grass: 1) persisted over time an d increased in area covered (spreads), 2) persisted over time but does not increase in area covered (persists), and 3) declined over time an d all plants eventually died (dies). At 30 sites in Kenya we collected climatic and soils data in the area where T-4464 seed was originally collected. Only total soil nitrogen and organic carbon differed among survival regimes. Total soil nitrogen and organic carbon concentration s were least where buffelgrass spreads, intermediate where the grass p ersists and greatest where the grass dies. To predict buffelgrass surv ival among the 3 survival regimes, and between areas where the grass s preads or dies, we used discriminant function analyses. A model includ ing organic carbon, total soil nitrogen, sand, clay, potassium and cat ion exchange capacity correctly classified 78% (r2 = 0.8) of the seedi ng sites in the 3 survival regimes. A model including sand, total soil nitrogen, calcium, mean minimum temperature in the coldest month, ann ual precipitation and winter precipitation correctly classified 88% (r 2 = 0.8) of the seeding sites between spreads and dies. Survival regim e selection prior to brush control, seedbed preparation and sowing wil l reduce planting failure probabilities, soil erosion and economic los ses, and enhance long-term beef production.