Assessment of Short Term Flood Operation Strategies Using Numerical Weather Prediction Data in YUVACΙK DAM Reservoir, Turkey

Yuvacik Dam Reservoir Basin, located in the Marmara region of Turkey with 248 km2 drainage area, has steep topography, mild and rainy climate thus induces high flood potential with fast flow response, especially to early spring and fall precipitation events. Moreover, the basin provides considerable snowmelt contribution to the streamflow during melt season since the elevation ranges between 80 - 1548 m. The long term strategies are based on supplying annual demand of 142 hm3 water despite a relatively small reservoir capacity of 51 hm3. This situation makes short term release decisions as the challenging task regarding the constrained downstream safe channel capacity especially in times of floods. Providing the demand of 1.5 million populated city of Kocaeli is the highest priority issue in terms of reservoir management but risk optimization is also required due to flood regulation. Although, the spillway capacity is 1560 m3/s, the maximum amount of water to be released is set as 100 m3/s by the regional water authority taking into consideration the downstream channel capacity which passes through industrial region of the city. The reservoir is a controlled one and it is possible to hold back the 15 hm3 additional water by keeping the gates closed. Flood regulation is set to achieve the maximum possible flood attenuation by using the full flood-control zone capacity in the reservoir before making releases in excess of the downstream safe-channel capacity. However, the operators still need to exceed flood regulation zones to take precautions for drought summer periods in order to supply water without any shortage that increases the risk in times of flood. Regarding to this circumstances, a hydrological model integrated reservoir modeling system, is applied to account for the physical behavior of the system. Hence, this reservoir modeling is carried out to analyze both previous decisions and also the future scenarios as a decision support tool for operators. In the first step, a hydrological model with an embedded snow module is used to establish a rainfall-runoff relationship to calculate the inflow into the dam reservoir. The basin is divided into four sub-basins, along with the three elevation zones for each subbasin. Hydro-meteorological data are collected via 11 automated stations in and around the basin and a semi-distributed rainfall-runoff model, HEC-HMS, is calibrated for sub-basins. Then, HEC-ResSim is used to create simulation alternatives of reservoir system according to user defined guide curves and rules based on internal and/or external variables. The decision support modeling scenarios are tested with Numerical Weather Prediction Mesoscale Model 5 (MM5) daily total precipitation and daily average temperature data. Predicted precipitation and temperature data are compared with ground observations to examine the consistency. Predicted inflows computed by HEC-HMS are used as main forcing inputs into HEC-ResSim for the short term operation of reservoir during the flood events.