Kaszab Zoltán (szerk.): A Magyar Természettudományi Múzeum évkönyve 75. (Budapest 1983)

Rajczy, M. ; Padisák, J.: DIVDROP analysis - a new method for the interpretation of species importance in diversity changes

Usually the number of species is fairly large in most of cases and only several species have any remarkable effect on diversity. It is not worth having the results printed out for all the species as it would be a waste of paper, so an algorithm was built into the programme to choose the species we are interested in. The method itself is a practical one without any theoretical consideration. Let where / is the serial number of the individual, A H"

jk is the diversity difference of the j-th individual omitting the k-th speices. -— One species might be of importance in following two cases: 1. if the average diversity difference per individual is great enough (the species has an overall effect or a large local effect), i.e. where n is the number of individuals in the data series, c

s is the critical level for this series (in our study c

3 — 0.05 proved to be efficient); 2. if the average diversity difference per individual is small but at least one single value of A H"

jk is great enough, i.e. where c

s is the critical level for the individuals (in our study Cj = 0.05 proved to be efficient). The hydrobiological example Recently, studies on the seasonal succession of phytoplankton communities and analyses on the roles of single species have become some of the most frequented subjects of hydrobiological studies (e. g. LYNCH & SHAPIRO 1981, SOMMER 1981, WALL & BRIAND 1980, WHITING et al. 1978). During an annual cycle phytoplankton communities seldom appear as discrete units. In most cases, mainly in shallow lakes, closely attached communities integrade one with another through time and often exhibit no distinct boundaries between them. A special difficulty is that the changes in abundance of about hundred or more species of algae must usually be studied. That was the case in a day-to-day study carried out in Lake Balaton in the summer of 1978 (PADISÁK 1980). During the thirty-day study 134 species of algae were found in the samples. Most of them may be considered as rare with an average number over the month below 300 ind-1""

1 (frequency measure: 1), 41 species had an average between 300 and 3000; 29 between 3000 and 20,000; 12 between 20,000 and 80,000 and only 2 above 80,000 ind-l"

1 (frequency measures: 2, 3, 4 and 5, respectively). Ordering the species based on their decreasing number seems to be a convenient way to estimate their importance in the community, but it may lead to a considerable loss of information because a heavy storm broke out at the middle of the study period which was followed by characteristic changes in the population dynamics of the most abundant species. The diversity also changed in correlation with that storm. Its changes could not be completely explained as based on the abundance patterns of the species. DIVDROP analysis was more successful in this case. Species having a frequency measure of 1, 2 or 3 had no significant effect on diversity changes, their (nevertheless always positive) effect was under the critical level (c

s = 0.05), except for three species. All the species with a frequency measure of 4 had an average diversity difference above the critical level as well. After removing them from the data-set the diversity decreased except for Nitschia sp. 1., removing its data an increase in diversity was noted in two samples (Fig. 1A). The species mentioned above can be divided into three