School of Botany - Theses
Permanent URI for this collection
Search Results
1 - 1 of 1
-
ItemEffective survey methods for detecting plants(University of Melbourne, 2001)Plant survey methods widely used in Australia to compile species lists and conduct baseline surveys are evaluated. In general flora surveys use one of two methods, a fixed- area quadrat or a wandering search. An untested assertion exists in the literature that suggests search sampling is better than quadrat sampling for detecting species. In order to test this claim empirically, an area of sclerophyll forest in southern Australia was sampled intensively using systematic quadrat sampling and search sampling. The number of species detected within quadrat samples was compared to the number detected in search samples. The pattern of species accumulation with time is modeled according to three types of botanical experience: the novice, average and expert searcher. The models are then used to predict the number of species gained with further sampling time. The likelihood of quadrat sampling to detect the same species within search samples is investigated. The expectation of life-form bias in the collections of individual searchers is also analysed. The results suggest that the number of samples taken should be guided by a consideration of how confidently the survey should detect the presence of relatively scarce species. A survey of botanists and biologists confirmed that individuals are poor judges of sample size and subjective estimates should be avoided. The results also suggest that in many circumstances, fixed-area quadrat sampling is preferable to search sampling. Measures of the adequacy of vegetation surveys are suggested.