Sampling efficiency for species composition assessments using the wheel-point method in a semi-arid savanna

A pilot survey to determine the sampling efficiency of the wheel-point method, using the nearest plant method, to assess species composition (using replicate similarity related to sampling intensity, and total sampling time) was conducted on three plot sizes (20 x 20m, 30 x 30m, 40 x 40m) at two sites in a semi-arid savanna. Total sampling time included the time to demarcate a plot. A 40 x 40m plot had the highest precision and repeatability (number of point observations required to achieve a stated asymptotic value), and was followed by a 20 x 20m plot. These relationships were consistent across sites. A 40 x 40m plot recorded the highest sampling times due to the time required to demarcate a plot. A 20 x 20m plot was the most efficient. A model to assist in making decisions regarding a suitable plot and sample size, and for planning and budgeting of vegetation monitoring programs was developed. Future surveys should include a Monte-Carlo data simulation as an adjunct to data collection. Keywords: grass layer; plot size; point methods; replicate similarity; sample size African Journal of Range & Forage Science 2001, 18(2&3): 93-10