تاثیر دبی چشمه‌ها بر قابلیت شبیه‌سازی رواناب با مدل SWAT در یک حوزه آبخیز کوهستانی با توپوگرافی شدید: واسنجی و اعتبارسنجی با استفاده از الگوریتم پرندگان (PSO)

نوع مقاله: مقاله پژوهشی

نویسندگان

1 دانش آموخته کارشناسی ارشد علوم خاک دانشکده کشاورزی دانشگاه شهرکرد

2 استاد گروه علوم خاک دانشکده کشاورزی دانشگاه شهرکرد

3 استادیار گروه علوم خاک دانشکده کشاورزی دانشگاه ولی عصر رفسنجان

4 استادیار گروه مهندسی آب دانشگاه علوم کشاورزی و منابع طبیعی گرگان

چکیده

در طول چند دهه اخیر استفاده از مدل­های هیدرولوژیکی به منظور شبیه­سازی، درک بهتر فرایندها و صرفه جویی در هزینه­ها در حوضه‌های آبریز بسیار گسترش یافته است. برای شبیه­سازی رواناب در مقیاس حوزه­های آبخیز، مدل­های هیدرولوژیکی گوناگونی از جمله SWAT توسعه داده شده است. در این پژوهش از مدل SWAT برای شبیه­سازی رواناب ماهانه در حوزه آبخیز بازفت استفاده شد و تأثیر دبی چشمه­ها موجود در منطقه بر دقت شبیه­سازی آن در دو اجرا مورد ارزیابی قرار گرفت. نتایج نشان داد که تأثیر دبی چشمه­ها بر جریان پایه، جریان بیشینه و شیب هیدروگراف­ها نسبت به داده­های مشاهداتی بسیار نزدیک­تر از حالتی است که دبی چشمه­ها وارد مدل SWAT نشده است؛  سپس مدل همراه با دبی چشمه­ها برای واسنجی و اعتبارسنجی با استفاده از الگوریتم پرندگان (PSO) انتخاب شد. در بخش آنالیز حساسیت، پارامترهای شماره منحنی (CN2)، زمان تاخیر برای آبخوان (GW_DELAY)، ضریب نفوذ آبخوان (RCHRG_DP)، فاکتور تاخیر دمای توده برف (TIMP) و متوسط کل بارش ماهانه در طول دوره پیش­بینی (PCPMM) بیشترین حساسیت را بر قابلیت شبیه­سازی مدل داشتند. مقادیر شاخص ارزیابی مدل شاملr-factor، p-factor و ضریب تبیین (R2) به ­دست آمده در مرحله واسنجی به ترتیب برابر 01/1، 76 درصد و 79/0 و در مرحله اعتبارسنجی برابر 76/0، 72 درصد و 57/0 بود. بنابراین با توجه به تأثیر چشمگیر دبی چشمه­ها و پارامترهای ورودی مدل بر شبیه­سازی روند تغییرات جریان در منطقه مورد مطالعه، به نظر می­رسد که در نظر گرفتن این عوامل در مدل­سازی با SWAT برای شبیه­سازی رواناب در حوزه­های آبخیز کوهستانی با توپوگرافی شدید، ضروری باشد.

کلیدواژه‌ها


عنوان مقاله [English]

Impact of springs flow on discharge simulation using SWAT model in a highly mountainous and topographic watershed: calibration and validation using PSO algorithm

نویسندگان [English]

  • I. Fazeli Farsani 1
  • M. H. Salehi 2
  • A. A. Besalatpour 3
  • M. R. Farzaneh 4
چکیده [English]

Introduction In order to achieve sustainable management of water resources, integrated knowledge of water resources and modeling is essential, especially in arid and semi-arid regions where water resources have become scarcer with increasing demands from socioeconomic development and population growth. In recent years, utilization of hydrological models has been increased to simulate watershed processes for cost saving purposes. Various hydrological models such as soil and water assessment tool (SWAT) have been developed to simulate runoff in the watersheds. In this study, SWAT was used to simulate monthly runoff in Bazoft watershed and the impact of springs discharge on the simulation accuracy was evaluated.
Materials and Methods Bazoft is one of the watersheds in Karun basin, (31° 37′ to 32° 39′ N and 49° 34′ to 50° 32′ E) located in northern part of the Karun river basin in southwestern Iran. The area of the watershed is 2168 km2. The main river in the watershed is Ab Bazoft which is joined by the Karun River at the outlet of the watershed. The elevation ranges from 880 m in the south of the watershed to 4200 m on Zardkuh Mountain in the north eastern area. Because the topography is very high in the watershed, the rainfall distribution is different, the average of rainfall in the northern part of the watershed is 1400 mm, while the southern part is 500mm. SWAT was used because the model is a continuous time, spatially and semi-distributed and basin-scale model, in which hydrological processes and water quality are coupled with crop growth and agricultural management practices. Input data include digital Elevation Model (DEM), land use, soil type, meteorological and hydrological observed data were provided. After running the model, a sensitivity analysis was done using the one at time method and SUFI-2 program. For calibration and uncertainty analysis in this study, we used the PSO (particle swarm optimization) algorithm. SUFI-2 and PSO are linked to SWAT in the environment of SWAT-CUP software. We used about two-third of the observed data for calibration and the remaining for validation. The simulation period was from 1992 to 2008. The study period was 1998–2008 for calibration and 1992–1997 for validation. The calibration and validation period results were analyzed at monthly time scale. ). The objective function was the Nash– Sutcliffe coefficient. Two indices, the P-factor and the R-factor, are used to quantify the goodness of calibration performance. The P-factor is the percentage of data bracketed by the 95PPU band, and ideally we would like to bracket all measured data, except the outliers, in the 95PPU band, and the R-factor is the average thickness of the 95PPU band divided by the standard deviation of the corresponding measured variable. Theoretically, the value of P-factor ranges between 0 and 100%, while that of R-factor ranges between 0 and infinity. In ideal conditions when the uncertainty model is perfect, P-factor will be 1 and the R-factor will be 0.
Results and Discussion The results showed that the simulated base flow, peak flow, and hydrograph trend by entrance of spring discharge data to the model were more in agreement to the observed runoff data than the model with no spring discharge data. Therefore, the constructed model with the spring discharge data was selected to calibrate the particle swarm optimization (PSO) algorithm. In the sensitivity analysis, the parameters of curve number for moisture condition II (CN2), groundwater delay time (GW_DELAY), deep aquifer percolation fraction (RCHRG_DP), snow pack temperature lag factor (TIMP), the average monthly precipitation during the prediction period (PCPMM), temperature and precipitation parameters and surface runoff lag time coefficient (SURLAG) were the most sensitive parameters in the watershed.
Conclusion The calibration and validation results for the base period (1992-2008) showed that the accuracy of the simulations was satisfactory for the discharge and sediment values. The obtained evaluation criteria r-factor, p-factor, and R2 for the calibration period were 1.01, 76% and 0.79 and for the validation period were 0.76, 72% and 0.57, respectively. Therefore, due to the noticeable effects of spring discharge data and the input parameters on the runoff simulation in the study area, it appears that it is essential to consider these factors for the runoff simulation using SWAT in similar mountainous watersheds with high topography.

کلیدواژه‌ها [English]

  • Discharge simulation
  • Calibration and validation
  • SWAT model
  • PSO algorithm