Case Studies - Coastal Lagoons: Ecosystem Processes and Modeling for Sustainable Use and Development - page 458

9.5.1.2.2.2 Hydrodynamic and Water Quality Modeling
of the Lagoon System
The ongoing modeling study for the Koycegiz-Dalyan Lagoon system has a seven-
step approach. The first step includes seasonal data analysis. In the second step, the
southern part of the system (Dalyan channels and lakes) was divided into computa-
tional boxes, which were later used as model boxes of finite difference computational
elements. In the third step, the U.S. EPA's one-dimensional finite difference hydro-
dynamic model DYNHYD5 33 and the Lung O'Connor method 34 were used to estimate
the flow distribution and velocities in the channels of the Dalyan Lagoon. In the
fourth step, a mass balance approach was used to verify the results obtained in the
previous step. A modified version of the U.S. EPA's WASP5 35 model was used. In
the sixth step, simple hydraulic calculations were made for the northern part of the
system (Koycegiz Lake) to estimate inflow and mixing effects. In the seventh step,
the CE-QUAL-W2 36 model was applied to used Koycegiz Lake and Dalyan Channel
and lakes for hydrodynamic simulations.
The data analysis step used salinity data obtained from field studies, long-term
daily inflow data gathered from the State Hydraulic Works and State Electric Works,
and water-level elevation data of the Koycegiz Lake gathered from the State Hydrau-
lic Works. The field data indicated that the system is under a saline water influence
that changes spatially and temporally. The hydrodynamic behavior of the lagoon
system is explained in detail in Section 9.5.1.1.1.
As shown in Figure 9.5.9, following the data analysis of the Dalyan Channel
network, the system was divided into 62 computational elements for preliminary
modeling studies.
Data analysis results concluded that a two-dimensional (longitudinal and verti-
cal), dynamic, laterally averaging, numerical model (x-z model) is required to sim-
ulate the flow in the Dalyan Channel networks. Instead of running a two-dimensional
dynamic x-z model and conducting the whole calibration work for that model, a
simpler technique was used. The U.S. EPA's one-dimensional hydrodynamic model
DYNHYD5 was run for a computational network. DYNHYD5 calculates the flow
rates and depths in a node-link computational network. The results were used to
estimate the net flow (upper layer flow plus lower layer flow) carrying capacity of
each channel. Upper and lower layer flows were estimated with a modified version
of the Lung O'Connor method which was modified by Ertürk 37 to handle the
nonlinear vertical salinity gradients.
The flows estimated were checked with a salinity mass balance. Mass balance and
transport equations had to be solved for each control volume in the Dalyan Channel
network. The one-dimensional computational network has 49 control volumes. As there
is a two-layer flow, the number of control volumes increases to 98. The EUTRO module
in the WASP5 model was used to solve the mass balance equations. The estimated flow
rates were used as flow input and the control volumes were used as WASP junctions.
The standard version of WASP5 allows the user to simulate 50 control volumes. Because
of some other program limitations related to number of flows, the source code of WASP5
has been modified and recompiled. The verification was made with the field data from
the August 1999 cruise, when a high influence of saline water was observed. The results
for flow velocities calculated for summer are illustrated in Figure 9.5.10.