,IRS2)
eas
study
s (see
unt of
ber of
bruary
further
y and
June the number of lakes with quite high values
decreases. By the end of September most of the lakes
are back to low values which are still at a higher level
than in February. This can be interpreted as a general
decrease in bioproduction in fall whereas the minimum
will be reached later during the year. The same
development can be observed between August 1996
and February 1997 representing the minimum of all
datasets.
Another parameter describing the seasonal variation in
bioproduction is the amount of change between the
consecutive dates which is shown in absolute DN
values in table 4. !
Changes Aug18 96 | Feb02 97 | May04 97 | Jun02 97 | Sep01 07
Mean - - - - -
(IRS1,IRS2) | Feb02 97 | May04 97 | Jun02 97 | Sep01 97 | Sep25 97
Mean -4,4 3,4 15 3,0 -2,7
Min -16,0 -46| -14,3 -4 7|. -15,9
Max 6,9 19,4 10,5 22,0 8,6
Stdev 4,6 5,0 4,7 4,7 4,0
Tab. 4: Summary statistics for changes of means
(IRS1, IRS2) between the consecutive
datasets of all core lake areas
30 7 ——$—— Aug-Feb
| 1
2d n ——i—— Feb-May
| p - - - &- - - May-Jun
20 + I= ~Jun-Sepl
|
; 4 rd
Number of lakes
> 12 8 4 0 -4 -8 >
DN of changes calculated from Mean
(IRS1,IRS2)
|
12)
|
|
|
Fig. 6: Frequency distribution of means (IRS1,IRS2)
between consecutive datasets calculated
from means of core lake areas
Each time step (tab. 4) is characterized by a high
variation in behavior of lakes expressed by the wide
range between minima and maxima and the standard
deviation of differences. The maximum amount of
changes was found for the difference between August
1996 and February 1997 which is explained by the
major difference between bioproduction in summer and
winter. The decrease between the beginning and the
end of September does not reach this level. Increasing
values can be observed for spring (February 97 - May
97) and for midsummer (June 97 — September,1, 97)
which is comparable to phases of active bioproduction.
A weak decrease can be observed between May and
June which might be caused by the end of first algal
bloom in some lakes.
The separation of the data into equidistant classes of
change of means (IRS1,IRS2) (see fig. 6) also shows
the variation in behavior of the lakes between the
amounts of the difference for each time step. The peak
at small positive values between June and September,
1 represents the remaining of the majority of lakes at
the reached level of bioproduction. Most of the lakes
show negative values for the difference between
August 96 and February 1997, in contrast between
February and May positive values were calculated for
most of the lakes. For June no clear maximum can be
observed, some lakes show decreasing some other
lakes increasing values. A significant maximum for
negative values characterizes the difference between
September, 1 and September, 25. It can be concluded
that in the study area some of the lakes show little
changes during the year, whereas other lakes
represent the opposite with a general high level of
bioproduction and a big variability. However, the
majority of the lakes shows moderate changes.
It is assumed that the synopsis of multitemporal
information can be achieved by classification of the first
principal component of all mean (IRS1,IRS2) datasets
due to the fact, that variations in the spectral signal are
mainly influenced by chlorophyll-a content.
Table 5 shows the results of the statistical analysis of
the first PC values.
1. PC of all means (IRS1,IRS2) DN
Mean 23,0
Min 9,1
Max 61,7
Standard deviation 9,5
Tab. 5: Statistical parameters of core lake areas for first
principal component
The variability of the first principal component was used
for classification of lakes into equal intervals with
broader classes for minimum and maximum values.
The classification (see fig. 7) can be interpreted as a
spatial distribution of bioproduction of lakes, whereas
high DN values represent high bioproduction and low
DN values show lakes with low bioproduction during
the observed period of time.
Legend:
1.PC Classes
DN intenall
ME 0-10
ME 10-13
Il 13-16
M 16-19
RE 19-22
BEN 22-25
ES 25-25
28-31
[] 31-34
3 > 4
A Extreme shallow lakes
us
Fig. 7: Classification of first PC of all mean (IRS1,IRS2)
by segmentation of core lake areas
5. FACTORS INFLUENCING LAKE PROPERTIES
Influences on lake water properties can be grouped in
parameters related to local lakes conditions and in
Intemational Archives of Photogrammetry and Remote Sensing. Vol. XXXII, Part 7, Budapest, 1998 133
