Use of Visual Cover Assessments as Quantitative Estimators of Some British Woodland Taxa

Abstract

SUMMARY (1) Ten observers made visual estimates of plant cover on a series of 4 m2, 50 m2 and 200 m2 quadrats in a woodland. (2) Mean cover values from the ten observers corresponded reasonably closely with point quadrat estimates of cover for most species in the 4 m2 quadrats. (3) For all taxa, significant differences occurred between the estimates of different observers at all quadrat sizes. Using an observer drawn at random, 90% confidence intervals were in the range + 10-20% cover. The corresponding range for an individual repeating an estimate on the same quadrat was + 5-15% cover. (4) Variability between observers was usually lowest when estimating broad-leaved species and highest with fine-leaved species and bryophytes. (5) Observers differed in the consistency with which they tended to under- or over-estimate cover in relation both to species and to quadrat size; this consistency was not correlated with experience. (6) The biases of different observers for a particular species and quadrat size were sufficiently consistent that the use of mean bias correction factors brought about clear gains in the precision of cover estimates for most species.