A new procedure to optimize the selection of groups in a classification tree

Agglomerative cluster analyses encompass many techniques, which have been widely used in various fields of science. In biology, and specifically ecology, datasets are generally highly variable and may contain outliers, which increase the difficulty to identify the number of clusters. Here we present a new criterion to determine statistically the optimal level of partition in a classification tree. The criterion robustness is tested against perturbated data (outliers) using an observation or variable with values randomly generated. The technique, called Random Simulation Test (RST), is tested on (1) the well-known Iris dataset [Fisher, R.A., 1936. The use of multiple measurements in taxonomic problems. Ann. Eugenic. 7, 179–188], (2) simulated data with predetermined numbers of clusters following Milligan and Cooper [Milligan, G.W., Cooper, M.C., 1985. An examination of procedures for determining the number of clusters in a data set. Psychometrika 50, 159–179] and finally (3) is applied on real copepod communities data previously analyzed in Beaugrand et al. [Beaugrand, G., Ibanez, F., Lindley, J.A., Reid, P.C., 2002. Diversity of calanoid copepods in the North Atlantic and adjacent seas: species associations and biogeography. Mar. Ecol. Prog. Ser. 232, 179–195]. The technique is compared to several standard techniques. RST performed generally better than existing algorithms on simulated data and proved to be especially efficient with highly variable datasets.

Ecological Investigations with the Continuous Plankton Recorder: The Copepoda of the Southern North Sea. 1932-37.

I. The monthly changes in the distribution and abundance of the Copepoda in the southern North Sea have been investigated from June 1932 to December 1937 by using the Continuous Plankton Recorder. This was towed at a standard depth of 10 metres by ships sailing on regular lines from Hull to Rotterdam, to Bremen and towards the Skagerrak, and later from London to Esbjerg. 2. The methods are described and those limitations which apply more particularly to the Copepoda are discussed (pp. 175 to 186 and 198 to 203). 3. The first part of the report deals with the Copepoda as a whole, i.e. the total population. The difference between the summer and winter distributions is stressed. The variations in numbers from year to year are found to be considerable and it is suggested that they are sufficiently large to be reflected in the success or failure of the broods of those fish which are at some period of their development dependent upon the Copepoda for food. 4. The second part deals with the data concerning the constituent species or groups of allied species ; a list of these is given on p. 197. 5. The group Paracalanus + Pseudocalanus was by far the most abundant and together with the genera Temora and Acartia was found to be responsible for most of the fluctuations in the population (pp. 205 to 208). 6. The distributions, seasonal and spatial, of the other common forms are described, with the exception of that of Oalantts finmarchicus which is to be the subject of a later report. 7. The recorder results are compared with the findings of the International Council survey from 1902 to 1908; some marked disagreements are discussed (pp. 227 to 232). 8. The appearance of the northern forms Oandacia armata and Metridia lucens during the winters of 1932-33, 1933-34 and 1937 are recorded (pp. 222 to 223) 9. A summarised account of the main seasonal changes in the area is given (pp. 232 to 234) and followed by a brief comparison of the 5½ years investigated.