79

ECOM II

Today these techniques are widely used over a wide range of scientifically disciplines

and seem to be particularly appropriate to ecological analysis where the boundaries

between groups such as communities or taxonomic units may be far from sharp.

In ordinary cluster analysis a sample or site either belongs or does not belong to a

particular group or set. We can score membership as a value of 1 and non 

membership as 0. In fuzzy clustering a degree of membership can be assigned as a

value between 0 and 1 so that a value of say 0.8 would indicate a high probability

that the sample belonged to the particular group in question. Thus, for a data set

comprising many samples that can be hypothesized as being divisible into c groups,

each sample has a degree of membership of belonging to each of the c groups.

Fuzzy Grouping offers two main methods of data analysis, first fuzzy c means and

second fuzzy ordination. The first technique is appropriate for data that you suspect

can be divided into c groups but for which you may have no a priori information on

group membership. The second, fuzzy ordination is appropriate when you have

information on possible group membership. For example, the samples might be taken

along a transect up a mountain slope and the altitude can be used as a measure of

likely membership of a sample to the high altitude community group. The program

also undertakes Non Metric Multidimensional Scaling primarily as a means of allowing

the fuzzy clusters to be visualised in a 2 dimensional space. However, the program

can also be used to produce a standard ordination using this powerful method. A final

feature that can be useful is the production of summary statistics on each row

(species) and column (sample).

12.6

Simply Tagging

Simply Tagging

 is a Windows program to estimate population size for both closed

and open populations using mark recapture methods, designed to be used by

undergraduates and allows them to simulate data and explore the accuracy and

reliability of mark recapture methods. Features include data simulation options to

model mark recapture experiments for open and closed populations. The user can

vary sampling intensity, population size, animal behaviour and birth/death rates.

Data can be entered as individual animal capture histories or as summary tables.

Ability to subset data by sex or age.

Wide range of methods available for closed populations including Schnabel census and

methods which assume behavioural differences between individuals.

Undertakes full Jolly Seber or constant survival and probability of capture model.

Plots of population estimates and frequency of capture.

As with all Pisces software, Simply Tagging has an attractive user interface, extensive

help system and good quality graphical output.

12.7

Population Estimation by Removal Sampling

Removal Sampling

 is a Windows 95/98 program to calculate the statistics from

removal trapping experiments.

Copyright 2004  PISCES Conservation Ltd