SELECTING KENTUCKY BLUEGRASS CULTIVARS

A computer-aided learning process has been previously developed to ide ntify unique turfgrass cultivar performance levels under specific grow ing conditions, However, it was not determined if this process was sim ilar to the routine assessments made by turfgrass experts, A survey wa s sent to 191 turfgrass experts to determine turf cultivar recommendat ion consistency, to validate a set of computer developed rules of asso ciation, and to determine the environmental and management characteris tics the experts considered most influential on turfgrass cultivar per formance. Seventy completed surveys, mostly from respondents who had w orked professionally for at least 10 yr at Land Grant universities, we re returned. In the first section of the survey, experts showed a high degree of consistency in recommending Kentucky bluegrass (Poa pratens is L.) cultivars for specific settings. In the second section of the s urvey, there was inconsistent agreement among survey respondents and m achine learning output when presented with rules associating cultivar performance and turf settings. Finally, when asked to indicate which e nvironment or management parameters experts felt most influence perfor mance of specific Kentucky bluegrass cultivars, survey respondents sel ected mowing height, nitrogen fertilization levels, and irrigation, Th e computer-aided learning process also identified mowing height and ni trogen fertilization levels as influencing cultivar performance. The m achine learning process also selected soil pH and average monthly air temperature as influential, Thus, there was not overall agreement betw een survey respondents and machine learning output.