بررسی دﻗﺖ شبیه سازی رواناب ماهانه حوضه دریاچه ارومیه با استفاده از ﻣﺪل ﯾﮏﭘﺎرﭼﻪ AWBM در اﺳﺘﺎن ﮐﺮدﺳﺘﺎن در ایستگاه سنته

Changes in community conditions, population growth, inappropriate and unprincipled use of available water resources and climate change are known as the reasons for the decrease in available water resources in recent years. Hence the need for integrated management of existing water resources is quite obvious. One of the important parameters for sustainable planning and management of water resources is the estimation of river flow. In recent decades and with the increasing development of computer technologies, many rainfall-runoff models have been developed for different purposes, each of which has advantages and disadvantages. Simulation is to understand the relationships governing the process of runoff. Simulation is used when the goal is not to involve the main system or the main system is not available. Simulation itself has been a major issue in terms of runoff forecasting and management in hydrological research. One of the most basic topics in hydrological sciences is understanding and recognizing and understanding the processes of production and flow transfer from input to output of the field. It is necessary to generalize through them those votes that do not have a voice, which is not a sign that it's not a sign, and the ideas of the day that this method of influence came into being. The relationship between precipitation and runoff is of great importance in hydrological studies. Because precipitation data are widely used in flood and runoff forecasts, they can use the available information to fill statistical gaps in runoff data. To achieve this goal, identifying the relationship between runoff and water is of high importance and key. calibrated the evaluation of annual water waste estimation in the watersheds of Khuzestan plain with experimental relations and finally this study showed that the coefficients of Katain method are 2.06, Justin 0.63, Institute of Agricultural Sciences method India is 4.67. These coefficients showed that in three methods, respectively, 90% of the level of trust and in the method of the Institute of Crop Sciences of India, 95% of the level of trust was achieved. evaluated experimental methods for estimating runoff in Band-e-Mandar watershed of Fars province. In this study, 6 experimental methods of runoff calculation including Katain, Indian Agricultural Institute, Justin, World Meteorological Organization, Irrigation Department of India and Khozla were estimated.
Methodology
Lake Urmia catchment area with geographical coordinates of 44 degrees and 7 minutes to 47 degrees and 53 minutes east longitude and 35 degrees and 40 minutes to 38 degrees and 30 minutes north latitude is located in northwestern Iran. The area of this basin is 51876 square kilometers, which is 15 / Covers 3% of the total area of the country. The amount, about 5822 square kilometers is the area of the lake itself, which is related to the height of the lake water and changes with its increase or decrease. The present study will simulate the runoff upstream of the Safakhaneh hydrometric station in the southern part of the catchment area of Lake Urmia in Kurdistan Province. The location of the catchment and 3 hydrometric stations are shown in Figure 0.3. Annual precipitation values and evapotranspiration potential as mentioned in the previous sections, in order to simulate the runoff precipitation process with the AWBM model, precipitation, potential, evaporation, transpiration data for modeling and runoff output from hydrometric stations are needed to calibrate and validate the model. For this purpose, precipitation data of Takab synoptic station were used to model the upstream basin of Safakhaneh station. Although data from the Ministry of Energy rain gauge stations upstream of these basins were also available, on the one hand, because the number of missing data from these stations is high and somehow modeled rainfall using these stations, there was no suitable representative for the desired basin, so we tried to use the data of the nearest synoptic station to this basin.
Statistical criteria for model performance evaluation Nash-Sutcliffe Coefficient (NSE)
The objective function is called a good fit measurement, and the optimal values of the parameters are values that represent the minimum value of the function. In each basin, the value of the objective function depends on the set values of the parameters. The point at which the objective function is minimized for the related parameters is called the optimal point of the parameters. The most common objective function used to calibrate hydrological models is the Nash-Sutcliffe efficiency coefficient. The NSE coefficient is a coefficient that shows the relative difference between the observed and simulated values. In this study, this coefficient has been used to evaluate the simulated results and observational data in the selected statistical period. Shown below.
NSE=1-(∑_i▒(Q_(m,i)-Q_s )_i^2 )/(∑_i▒(Q_(m,i)-(Q_m ) ̅ )^2 )
Observational flow rate Q_s: Indicates computational flow Q_ (m, i): Observed flow rates during the simulation period And all three are in terms of m ^ 3⁄s. The performance range of the NSE evaluation index in the simulations performed by the model is given in Table (3-2). Rainfall-runoff simulation In order to study climatic parameters on runoff, it is necessary to use rainfall-runoff models. In this study, support vector machine regression (SVR), gene expression programming (GEP) and IHACRES (IHACRES) were used to generate monthly runoff. All AWBMs provided (3) are all computers whose functionality is similar to that of a valid computer..
The AWBM model uses surface storage capacities (C3, C2, C1) with areas (A3, A2, A1) to simulate runoff levels, and the water function of each storage surface is independent of the others in daily time steps (or Hourly) is calculated. The water balance equation of each surface is such that precipitation is added to the surface reserve and evaporation and transpiration are reduced. The equation of water balance in case n is the number of reserves in the basin is as follows:
Store_(n+1)=Store_n+Rain-Evan (n=1,2,3)
Where, zero is considered when the storage moisture content is negative, but if the storage moisture is more than the reservoir capacity, the excess moisture is converted to runoff and the storage moisture remains equal to the reservoir capacity .In the model, it is assumed that the two main sources of surface runoff and base water are runoff. This model has three surface storage capacities (C_3, C_2, C_1) and 3 levels corresponding to surface storage capacities (A_3, A_2, A_1) and an average storage capacity (C_ave), the relationships between these components are:
A_1=0⁄133 A_2=0⁄433 A_3=0⁄433
C_1=0⁄01 C_ave/A_1 C_2=0⁄33 C_ave/A_2 C_3=0⁄66 C_ave/A_3
Calibrate this model with AWBM2002 subroutine. Initially, this subroutine is considered a C_ave and using BFI and K obtained from the NBFLOW subroutine and with the help of relations 2 and 3, the values of C3, C2, C1 with hypothetical levels A1 = 0⁄133, Gets A2 = 0⁄433, A3 = 0⁄433 and finally corrects these three tablets. C_ave is first calculated from the assumed values of 5, 50, 100, 150 and 300, which are considered by default in the model, in such a way that the runoff obtained and the actual runoff obtained from the following equation have a low The most differences are:
Act=e1A1+e2A2+e3A3
The amount of actual monthly runoff e_n: Monthly computational runoff from each of the stored weapons A_n: The level of each storage capacity.
Conclusion
Parameter A2 is the same as A1, except that the optimal range of variation is 0-0.3. The sensitivity of the objective function values to the values of the BFI parameter is completely opposite to parameters A1 and A2, so that the closer the value of this parameter is to one, the more optimal the objective function values are. The sensitivity of the model to the parameters C1, C2 and C3, however, is low in a wide range of their variation. Changing the KBase and KSurf parameters will also result in better values of the target function in the range of 0.8 to 1.
The amount of rainfall recorded at the stations is not a good indicator of rainfall in the basin. Although in the AWBM model, it is possible to change the amount of input data with the Data Scaling tool, and from this tool, the results were obtained by defining the appropriate scale, but due to limitations, this scaling could not calculate the precipitation values in the months with peak flow. Increase in such a way that the simulated flow approaches the peak flow. On the other hand, because all three sub-basins are mountainous and there is a large difference in altitude in the basin, so considerable rainfall at higher altitudes is obvious, and since other than the Data Scaling tool, another tool to approximate rainfall values There is no entrance to the realities of the region. In some hydrological models such as SWAT, for hydrological simulation in mountainous areas, it is possible to define altitude bands and also to define the altitude gradient of precipitation, and this causes the precipitation modeled at altitudes far from the precipitation recorded in More stations to be considered. Overall, the performance of the AWBM model in precipitation-runoff simulation of the upstream basin station in the calibration period with a Nash-Sutcliffe index value greater than 0.7 is very good.