ارزیابی مدل wrf برای پیش‏ بینی دما و رخداد سرمازدگی در حوضۀ آبریز زاینده‏ رود

وقوع مخاطرات جوی، همچون یخبندان و سرمای دیررس بهاره، سالانه، خسارات زیادی در بخش کشاورزی ایجاد می‏کند. برنامه‏ریزی به‏موقع می‏تواند خسارت ناشی از بلایای طبیعی را کاهش دهد. امروزه، با استفاده از مدل‏های پیش‏بینی عددی وضع هوا، می‏توان از خسارت‏های ناشی از آن‏ها جلوگیری کرد. در این تحقیق، برای ارزیابی پیش‏بینی دما توسط مدلwrf  در زمان وقوع سرماهای دیررس بهاره، از یازده ایستگاه هواشناسی واقع در حوضه آبریز زاینده‏رود، با درجه تفکیک افقی یک کیلومتر، شبیه‏سازی شد. سپس، با دو رویکرد نقطه‏ای و منطقه‏ای دماهای شبیه‏سازی‏شده با مقادیر دیدبانی متناظر در پیش‏بینی‏های 24 و 48ساعته دمای سطحی (دومتری) ارزیابی شد. براساس نتایج جذر میانگین مربعات خطا، ضریب تعیین اصلاح‏شده و شاخص میانگین اریبی که برای دمای شبیه‏سازی 24ساعته بهتر از 48ساعته است به‏ترتیب 2.8، 0.88، و 0.48 بود. ارتباط قابل قبولی از لحاظ آماری (ضریب همبستگی) بین متغیر مستقل، که همان داده‏های مدل wrf است، و متغیر وابسته، که همان داده‏های دیدبانی‏شده (واقعی) است، وجود دارد.

Evaluation of WRF Model for Temperature Forecast and frosting occurrence in Zayandeh Rud Basin

Evaluation of WRF Model for Temperature Forecast and frosting occurrence in Zayandeh Rud Basin Abstract Introduction Occurrence of Weather hazards such as freezing, annually late spring frost create much damages in agricultural section. Programming, decision making, suitable action and welltimed can decrease damages of resulting from Weather hazards. Therefore consideration of comprehensive and precision of this phenomenon is necessary. Nowadays by use of numerical weather prediction (NWP) models and more phenomenon’s recognition of Weather hazards can prevent from damages of resulting from those. Verification is process that quality, skill and value of forecast assign via forecast results in comparison with corresponding observed values. In recent years, direct numerical weather prediction model forecasts of near surface parameters often suffer from systematic errors mainly due to the low resolution of the model topography and inaccuracies in the physical parameterization schemes incorporated in the model. On the other hand, verification is a critical component of the development and use of forecasting systems. Ideally, verification should play a role in monitoring the quality of forecasts, providing feedback to developers and forecasters to help improve forecasts, and provide meaningful information to forecast users to apply in their decisionmaking processes. One of the purposes, in this study is to evaluate the performance of the WRF model for Temperature Forecast and frosting occurrence in Zayandeh Rud Basin. Materials and Methods Case of study in this research is in Zayandeh Rud Basin. Zayandeh Rud Basin and research area geographical coordinates has occurred at 50 degrees and 20 minute until 52 degrees and 24 minute eastern longitude and 31 degrees and 12 minute until 33 degrees and 42 minute northern latitude. In this study, for evaluation of temperature forecast by the WRF model on Occurrence time of late spring frost from 11 meteorological stations in Zayandeh Rud Basin with 1 Km horizontal resolution from schemes model (KFMYJ and GDMYJ) were simulated. Then, simulated temperatures and the corresponding observed values were evaluated by two methods such as Point and area in predictions of 24 and 48 hour of surface temperature (2m). For evaluation‌ of function forecast models, indicator function different including: Determination coefficient (R2), Root Mean Square Error(RMSE), Mean Square Error(MSE),‌‌‌Mean Absolute Deviation(MAD), Relative Error(Error), correlation coefficient(R), Mean‌ Bias Error(MBE), Mean Absolute Percentage Error(MAPE), Mean Square Skill Score(MSSS). ، and n is observed values ، forecast values and number of data respectively. Results and Discussion According to the results, Root Mean Square Error, Adjusted Rsquare and Mean Bias Error for 24hour temperature simulation are better than 48 and were about 2.8 , 0.88 and 0.48 respectively. Results of output data 24 with 48 hour indicated that error of 48 hour forecast data from 24 hour is higher. Acceptable relation from the viewpoint of statistical (correlation coefficients and coefficient of determination) exist between independent variable (WRF model values) and dependent variable (observed values) that is significant in the level of 5%. Verification Results of output WRF model was showed for Temperature Forecast in time scale of 24 and 48 hour time of late spring frost from 11 meteorological stations in Zayandeh Rud Basin In more than 80% cases, Forecast Results of phenomenon Occurrence coincides with observed Results that under discussion phenomenon has happened. Correlation between observed temperature and WRF model with 1 Km horizontal resolution than 3 Km horizontal resolution have high accuracy. Substantially by the maps of 850mb the dates of late spring frost occurrence can find out nature of late spring frost that is radiation or advection. For instance existence of cold advection on the maps of 850mb is obvious well. Contour lines and isotherm almost have intersected vertically and a strong cold advection has created. It is worth mentioning whatever condition of Baroclinic (intersection isotherm by contour with good angle) get better and angle of intersection approach the vertical angle, advection is stronger. Per of maps of 850mb (case study) existence of cold advection is obvious. We conclude that occurrence of late spring frost in this paper is advection mostly. Zoning maps of used indicators drew in evaluation of predictions of 24 and 48 hour. Just as infer the stations that have low level rate of model error is less. And output of model is near the truth in proportion to the stations that have high level such as Shahrekord station. This indicate that forecast of surface parameters as two meters temperature to take impression from zone topography. Conclusion Daily temperature was suitable for Forecast in time scale of 24 and 48 hour. Although Forecast 24 hour have high accuracy. The verification scores of model in estimation of quantitative temperature in 24 forecasts are better than 48h forecasts. In addition to accuracy forecast of 2 meters temperature is intense relation in zone topography of case study, someway accuracy of‌ model estimates in plain areas is more than mountainous areas. On the whole, results of this paper indicate that from WRF model forecasts can use well for Temperature Forecast and frosting occurrence.