نوع مقاله : مقاله پژوهشی
نویسنده
گروه علوم و مهندسی آب، دانشکده کشاورزی، دانشگاه کردستان، سنندج، ایران
چکیده
کلیدواژهها
موضوعات
عنوان مقاله [English]
نویسنده [English]
In this study, the performance of two empirical and physical snow models at the Zarrineh Obatou synoptic station during the period 1989-2022 was evaluated. To calibrate these models, the Generalized Likelihood Uncertainty Estimation (GLUE) method was employed by selecting 6 parameters and generating 8000 random vectors from the uncertainty domain of these parameters. The models were then run based on these parameters over the period 1989-2022, using RMSE, MBE, and Nash-Sutcliffe coefficient to identify the best simulations (1% of total simulations). This procedure was also used for model validation, with the models being calibrated based on odd years and validating on even years. The results indicated approperiate performance of both models in simulating snow depth at the Zarrineh station, with the physical model demonstrating better overall performance compared to the empirical model. The results also showed that the best performance of both models occurred during the simulation of moderate snow depths, with both models tending to overestimate light snow and underestimate heavy snow. Sensitivity analysis of models indicated that snow melting processes are key processes in both models. Given the limited measured data on snow in Iran and the necessity of using snow models for various purposes such as estimating past snow and projection of future snow in response to climate change, the overall results of this study suggest that the studied models in this research have a good potential for simulating various snow-related variables in Iran and employing them (especially the physical model) is strongly recommended.
کلیدواژهها [English]
EXTENDED ABSTRACT
Snow is one of the most influential components in the hydrological cycle and a determining factor in various climatic, hydrological, and environmental systems. It plays a fundamental role in the exchange of moisture and heat fluxes between the Earth's surface and the atmosphere. In the Earth's climate system, snow is considered as an important indicator in climate change studies due to energy exchanges between the snow surface and the atmosphere. Snow also plays a crucial role in the spatial distribution of water reserves on Earth. Monitoring and observing changes in snow depth and analyzing long-term trends are of great importance. In Iran, very few studies have been conducted on snow depth modeling, leading to a noticeable research gap in this area. To address this research gap in Iran, the present study aims to calibrate and validate two snow models at one of the highest-elevation synoptic stations in Iran.
The Zarrineh-Obatou station in Kurdistan province, with an elevation of 2142m, was selected for this research. In this study, two snow models, both are the sub-models of the CoupModel, were selected. Although both models consider a one-dimensional vertical profile for snow, they differ in complexity. In one model (empirical model), which has a relatively simpler structure, empirical functions dependent on air temperature and solar radiation are used for simulating snowmelt. In the second model (physical model), which has a relatively more complex structure, an energy balance method on the snow surface and within the snowpack is used for snowmelt modeling. The calibration procedure of these models was performed using the Generalized Likelihood Uncertainty Estimation (GLUE) method with the selection of six parameters. After generating 8000 random vectors from the uncertainty domain of these parameters and running the models based on them for the period 1989-2022, RMSE, MBE, and the Nash-Sutcliffe coefficient indices were used to identify the best simulations (1% of all simulations). The same procedure was also used for model validation, with the models being calibrated based on odd years and validating on even years. Sensitivity analysis of the models was also conducted by plotting the cumulative distribution function of the validated parameters and comparing it with the initial uniform distribution of the parameters.
The results show that the GLUE method was able to effectively identify the best simulations among total simulations. Both models demonstrated proper performance in simulating snow depth at the Zarrineh-Obatou station during the period 1989-2022 based on all indices (RMSE, MBE, and Nash-Sutcliffe coefficient). However, the physical model, especially in terms of balancing underestimation and overestimation, exhibited better performance compared to the empirical model. Sensitivity analysis of the models indicated the varying importance of selected parameters, with the models showing higher sensitivity to specific parameters. Parameters related to snowmelt were of high importance in both snow models. Regarding the performance of the models in simulating snow depths ranging from low to high, the results indicated that both models performed best when simulating moderate snow. In simulating light snow, both models tended to overestimate, while in simulating heavy snow, they tended to underestimate.
To improve this issue, in addition to considering the entire study period, it is recommended to directly take the years with heavy and light snow into account as well in calibration process. Furthermore, it is suggested to use the hourly time scale for snow simulation instead of the daily time scale. The results also suggest it is possible to obtain better results from these snow models by measuring other variables like snow density in addition to measuring snow depth, which is the only available snow data in Iran.
The overall results of this research indicate that the studied models have the potential to effectively simulate various snow-related variables, especially the physical snow model. Since general climate models that project future climatic conditions do not directly include snow in their outputs, it is possible to employ the main outputs of these models (air temperature, precipitation, solar radiation, etc.) and utilizing them as the input of the studied snow models in this research to project snow condition in future periods under different climate change scenarios. This will provide a realistic outlook on the variability of snow in future periods, aiding in optimal water resource management in the country. It is worth mentioning that by utilizing these snow models, it is also possible to estimate snow depths during past periods where snow data have not been accurately measured.
The author contributed to the conceptualization of the article and writing of the original and subsequent drafts.
Data available on request from the author.
The author would like to thank the the vice chancellor for research affairs of university of Kurdistan for support of the present study.
The author avoided data fabrication, falsification, plagiarism, and misconduct.
The author declares no conflict of interest.
)