كاربرد مدل‌هاي عددي پيش‌بيني وضع هوا در بهبود دقت تعيين موقعيت به روش ماهواره‌اي

Application of Numerical Weather Prediction Models to Improve the Accuracy of Satellite Positioning Method

ارائه خدمات ناوبري در تمامي شرايط آب و هوايي به كاربران، همواره به عنوان يكي از مزاياي بسيار برجسته سيستم هاي تعيين موقعيت ماهواره اي به حساب مي آيد. امواج ارسالي از ماهواره هاي‌اين سيستم ها از لايه هاي مختلف جو عبور كرده و‌اين امر منجر به‌ايجاد انكسار در مسير حركت موج و در نهايت تأخير در دريافت امواج مذكور مي گردد.‌اين تأخير به برآوردي ناصحيح از موقعيت گيرنده منجر مي شود. از آنجا كه بيشتر فرآيند هاي جوي در پاييني‌ترين لايه اتمسفر زمين (تروپوسفر) رخ مي­دهد،‌اين لايه، يكي از منابع مهم‌ايجاد خطا در تعيين موقعيت مطلق دقيق (PPP) با سيستم تعيين موقعيت جهاني GPS است. از روش‌هاي مختلفي براي تعامل با‌اين منبع خطا استفاده مي‌شود. در‌اين مقاله از مدل پيش‌بيني جهاني سستامينن كه با داده هاي حاصل از مدل استاندارد اتمسفري حمايت مي شود و روش رديابي اشعه با استفاده از مدل پيش‌بيني عددي وضع هوا ، مدل پيش‌بيني و تحقيقاتي آب و هوا (WRF) جهت برآورد تأخير مايل تروپوسفري استفاده گرديده است. نتايج در سطح تعيين موقعيت مقايسه و بررسي شده است. موقعيت مطلق دقيق يك نقطه پس از حذف خطاي مورد بحث از هر دو روش در يك بازه دوازده روز تعيين گرديد. جهت بررسي نتايج از روش تكرار پذيري استفاده شده است. تكرار پذيري در مؤلفه ارتفاعي نتايج حاصل از تصحيح تأخير تروپوسفر با استفاده از مدل سستامينن و رديابي اشعه به ترتيب 13/3 ميلي متر و 0/98 ميلي‌متر به دست آمده است. كاهش 2/15 ميلي متر در خطاي مربعي متوسط نمايانگر پتانسيل مدل‌هاي عددي پيش بيني وضع هوا جهت دستيابي به برآوردي صحيح‌تر در مقدار تخمين تأخير تروپوسفري با استفاده از روش رديابي اشعه است.

Introduction The remarkable advantage of Global Navigation Satellite Systems (GNSS) is providing navigation service to the user communities, which is independent of the weather condition. The GNSS signals pass through different layers of the Earths atmosphere. Propagating signals are refracted while passing through every layer and therefore, the reception of the signals is delayed. This delay distorts the accuracy of the position which is computed for a receiver. Troposphere, which is the lower most part of the Earths atmosphere, is one of the most important source of bias in this respect. To analyze the impact of tropospheric delay, it is normally divided into dry and wet components. The contribution of the dry and wet parts is reported to be 90 and 10 percent respectively. In contrary to the wet component, the existing models are precise enough to compute the contribution of the dry part on the signal delay. Therefore, analyzing the efficiency of numerical weather models for modeling the tropospheric delay as compared to the existing standard techniques seems to be remarkable. Materials and methods In this paper, three methods are used for computing the tropospheric error. The first method is based on the Saastamoinens global troposphere model. This model is commonly used for computing the tropospheric error. Required input parameters are derived from the standard atmosphere model. In the second and third method, ray tracing is used for correcting the GPS measurements. A Numerical Weather Prediction (NWP) model is used for this purpose. Here, meteorological data measured at the position of the station are used as the input parameters of the model. In the third method, the required surface meteorological data are extracted from a NWP model. The World Research and Forecasting (WRF) model is used for this purpose. The WRF daily forecasts with a horizontal resolution of 0.1 degrees in 25 pressure levels (from 1000 to 50 mill bars) are used together with the GPS carrier phase and code measurements at station TKBN. This permanent GPS station is located at ϕ=36̊ 47ˊ 9.33˝ and λ=50̊ 55ˊ 48.20˝. Sling rate of the GPS measurements is 30 seconds. The observation time interval starts at November 2 and ends at November 13, 2011. Results and discussion Computed tropospheric corrections are applied to the raw measurements above. Daily precise positions of this station are estimated using the corrected data. Repeatability of the point positions is used as a measure for analyzing and comparing the obtained results. The repeatability of the station coordinates in the north component increases from 3.13 mm in the first method to 0.98 mm in the second one. The repeatability of the east and north components also improves by 1.73 mm and 2.11mm respectively when the raw observations are corrected by the tropospheric corrections which are derived from the second method above. This proves the efficiency the WRF numerical weather model for estimating the tropospheric error when the surface meteorological data that is observed at station is used. Conclusion The use of ray tracing with surface data from WRF model in comparison with Saastamoinen model does not lead to improved repeatability coordinates in all three components. The results of this study emphasize the Numerical Weather Prediction models used in this study has been poor to calculate the surface data. Otherwise the results of the ray tracing method based on WRF model with surface meteorological data is observed at station is the best method for Computed tropospheric corrections.