Abstracts

Diversity of the Fauna

The goal of the work is the application of statistical methods, which are not traditional at the analysis of seasonal and long-term dynamic of birds. With the help of these methods it is possible to examine decreasing and increasing density for the researched period, a periodicity in its change, and the differences in dynamic of birds in different places.

Density of birds was counted on the integrated area of forests and the integrated area of all researched places. Our data we present as stationary stochastic process, and for the analysis of this process we use a method of autocorrelation and spectral decomposing of time series (EXCEL and SPSS programs).

Long-term tendency of small decreasing of density was obtained (p < 0.05). The linear model of a trend can be justified with the help of spectral decomposing of stochastic process. The spectral analysis describes "contribution" of each frequency to a variance. If the process has linear trend, the spectral concentration will have peak at point=0 (Computer biometrics, 1990). For our data the peak in a point =0 was obtained, that indicates availability of linear trend.

But determined component of time series, besides a trend, contains seasonal and cyclical components. From here there is a problem to examine the relationship between the members of the same series, excepting a trend.

Really, an autocorrelation coefficient is calculated as a measure of interrelation between the members of the same time series (Lakin, 1990). For this purpose the members of a series are moved on an adopted time unit, in our case it is one decade, that allows derive series on two variables.

With the help the method of autocorrelation the following outcomes were obtained. In a periodic process of bird dynamic the relationship between the density of birds of current decade and the density of previous decade was obtained for the integrated areas (R=0,39), the greater relationship for the forests (R=0,63). In a periodic process of dynamic of mean number of birds in the different periods of an annual cycle (winter, spring migrations, female, autumn migrations, period before winter) the greatest relationship between the density of birds of the current period and the previous period was obtained (R=0,73).

According the plot of autocorrelation functions, there is a difference between the periodicity of a bird density on the integrated areas and on areas of all forests: if on the first ones the deviation in relation to average is high in the autumn and spring, but on the second ones this deviation is high in summer. Probably, it denotes anthropogenic effect that is less in forests.

We thank V.G. Ivliev for number of bird fond data.

Note. Abstracts are published in author's edition

|Home Page| |English Part|

Go to Home