Chapter 13 . Cross-Sectional Estimation of Chlorophyll a in Lakes 253
various times. Further, mass loads of pollutants entering the reservoir, for example,
nitrogen and phosphorus, are highly seasonal. The data show a persistent seasonal
metalimnetic oxygen minima, and high spatial heterogeneity with respect to Secchi
depth, nutrient concentration levels, turbidity and chlorophyll-a (Soyupak et al.
1998). The data utilized in this study covers the years 1991-1993 (see data of
Yemioen et al. (1994) and compilation by Karul (1998b) and (1999a)).
Mogan and Eymir lakes are small and shallow lakes located near Ankara in
Central Anatolia. The total watershed of the lakes is about 970 km 2 at the oudet of
Eymir. Both of lakes are formed by the inundation of the main canal through
deposition of alluvial material. Hence, the lakes are analogous to wide riverbeds.
The database that was utilized for Mogan and Eymir Lakes during this study
covers the years 1993-1995 (Altonbilek et al. 1995) and was compiled by Karul
(1999a). Mogan Lake is situated between the northern latitudes of 39° 28' and 39°
53' and eastern longitudes of 32°30' and 33°00'. Mogan Lake is a shallow eutrophie
lake with seasonally dense growth of macrophytes and wide seasonal and diurnal
variations in dissolved oxygen concentrations. The main inflow enters the lake
through a swamp area and the outflow gives rise to a canal, wh ich enters Eymir Lake.
Hydrologieally, Lake Eymir is fed by the outflow from Lake Mogan as weil as
Koolak<;:o Creek, a small tributary confluence near the lake oudet. On the bases of
OECD Lake classification criteria (Vollenweider and Kerekes 1981), Lake Eymir
is eutrophie lake.
126, 31 and 34 valid data sets were obtained from Keban Dam Reservoir
(Yemioen et al. 1994), Mogan Lake and Eymir Lake (Altonbilek et al., 1995),
respectively. Each data set for KDR included P0 4 N0 3 , Alkalinity, Suspended
Solids, pH, Temperature, Electrical conductivity,' Dissolved Oxygen, Secchi
depth, densities of Daphnia species only, and bulk densities of species belonging
to Cladocera and Copepoda as input and Chlorophyll-a as output. Data sets for
Mongan and Emir Lakes include Total Phosphorus, N0 3 , NH 3 , Suspended Solids,
Temperature, Electrical conductivity, pH, Turbidity, Secchi depth as input and
Chlorophyll-a as output.
13.3
Methodology
Data sets available for each water body were collected over several months from
the upper 1 m (euphotic zone) of pre-selected stations for each lake. Since the
available data was not appropriate to form times series a steady-state approach
was adopted. Temporal component was eliminated, since the prime goal of the
study was to develop neural network tools for chlorophyll-a concentrations
utilizing the data related to governing environmental parameters. Eliminating
temporal component was something new when compared to almost the earlier
related works (Recknagel et al. 1997; Yabunaka et al. 1997; Karul et al. 1998a).
various times. Further, mass loads of pollutants entering the reservoir, for example,
nitrogen and phosphorus, are highly seasonal. The data show a persistent seasonal
metalimnetic oxygen minima, and high spatial heterogeneity with respect to Secchi
depth, nutrient concentration levels, turbidity and chlorophyll-a (Soyupak et al.
1998). The data utilized in this study covers the years 1991-1993 (see data of
Yemioen et al. (1994) and compilation by Karul (1998b) and (1999a)).
Mogan and Eymir lakes are small and shallow lakes located near Ankara in
Central Anatolia. The total watershed of the lakes is about 970 km 2 at the oudet of
Eymir. Both of lakes are formed by the inundation of the main canal through
deposition of alluvial material. Hence, the lakes are analogous to wide riverbeds.
The database that was utilized for Mogan and Eymir Lakes during this study
covers the years 1993-1995 (Altonbilek et al. 1995) and was compiled by Karul
(1999a). Mogan Lake is situated between the northern latitudes of 39° 28' and 39°
53' and eastern longitudes of 32°30' and 33°00'. Mogan Lake is a shallow eutrophie
lake with seasonally dense growth of macrophytes and wide seasonal and diurnal
variations in dissolved oxygen concentrations. The main inflow enters the lake
through a swamp area and the outflow gives rise to a canal, wh ich enters Eymir Lake.
Hydrologieally, Lake Eymir is fed by the outflow from Lake Mogan as weil as
Koolak<;:o Creek, a small tributary confluence near the lake oudet. On the bases of
OECD Lake classification criteria (Vollenweider and Kerekes 1981), Lake Eymir
is eutrophie lake.
126, 31 and 34 valid data sets were obtained from Keban Dam Reservoir
(Yemioen et al. 1994), Mogan Lake and Eymir Lake (Altonbilek et al., 1995),
respectively. Each data set for KDR included P0 4 N0 3 , Alkalinity, Suspended
Solids, pH, Temperature, Electrical conductivity,' Dissolved Oxygen, Secchi
depth, densities of Daphnia species only, and bulk densities of species belonging
to Cladocera and Copepoda as input and Chlorophyll-a as output. Data sets for
Mongan and Emir Lakes include Total Phosphorus, N0 3 , NH 3 , Suspended Solids,
Temperature, Electrical conductivity, pH, Turbidity, Secchi depth as input and
Chlorophyll-a as output.
13.3
Methodology
Data sets available for each water body were collected over several months from
the upper 1 m (euphotic zone) of pre-selected stations for each lake. Since the
available data was not appropriate to form times series a steady-state approach
was adopted. Temporal component was eliminated, since the prime goal of the
study was to develop neural network tools for chlorophyll-a concentrations
utilizing the data related to governing environmental parameters. Eliminating
temporal component was something new when compared to almost the earlier
related works (Recknagel et al. 1997; Yabunaka et al. 1997; Karul et al. 1998a).