Comparison of field methods in vegetation monitoring

Abstract

For the purpose of vegetation monitoring, three methods for estimating plant cover, namely visual-estimate (VE), point-frequency (PF) and subplot-frequency (SF), were compared with regard to accuracy, precision, sensitivity and time consumption. The cover estimates by the three methods were compared with ‘true’ values measured on photographs using an image analysis technique. VE had the highest accuracy, precision and sensitivity. PF systematically overestimated cover. The greatest drawback of PF was the low sensitivity to species detection. It failed to detect 22–30% of the species in two forests and a bog. SF has high precision and sensitivity but low accuracy. Time consumption on one quadrat was about 6 min for VE and SF, 8.5 min for PF. Inter- and intra-person error with the three methods depends on plants. Inter-person error with VE was slightly greater for small and wide-spread plants, especially mosses, than for other life-forms. For wide-spread plants, intra-person error was slightly greater with PF than with the other methods. VE and PF are strongly correlated and convertible, but SF is neither correlated nor convertible with the other methods.