2003 13th Int. Crimean Conference “Microwave & Telecommunication Technology” (CriMiCo’2003). 8-12 September, Sevastopol, Crimea, Ukraine © 2003: CriMiCo’2003 Organizing Committee; Weber Co. ISBN: 966-7968-26-X. IEEE Catalog Number: 03EX697 68 PATH LOSS DUE TO RAIN FADING AND PRECIPITATION IN 26 GHZ LMDS SYSTEMS: CONSIDERATION OF IMPLEMENTATION IN TURKEY Sercan Uslu and İbrahim Tekin Faculty of Engineering and Natural Sciences, Sabancı University Orhanlı – Tuzla, 34956 İstanbul, Turkey E-mail: sercanuslu@su.sabanciuniv.edu, tekin@sabanciuniv.edu Abstract - This paper explains the rainfall attenuation pre- diction model recommended in ITU-R PN.838 and investigates the variability of 26 GHz LMDS radio channel obscured by rain attenuation in seven representative cities for the climate of tur- key. Model is applied both with recommended input values presented in ITU-R PN.837-1 and values obtained by process- ing long term monthly precipitation data for Adana, Ankara, Antalya, Erzurum, Istanbul, Izmir, and Samsun. Results for both applications are presented and the necessity of using location- specific monthly climate statistics for the LMDS system design is addressed. I. Introduction Availability of radio link and the system range defines the reliability of Local Multipoint Distribution Services (LMDS) as in any radio system. Since system range and availability are highly constrained by climate characteris- tics of region of operation, path loss caused by precipita- tion in that region is one of the key issues that must be considered for the LMDS system design process. Pre- cipitation based path loss of the signal is an issue which is easy to understand but difficult to predict. Under good weather conditions precipitation based path loss is not evident whereas during heavy rainfalls, signals are ab- sorbed by the water within the rain and this absorption causes signal attenuation. Atmospheric disturbances such as thunderstorms also distort the signal and cause unacceptable number of errors [1]. Although the difficulty is obvious for prediction, it is possible to make some estimation. By means of the standardized method of calculation for the precipitation based path loss in Inter- national Telecommunications Union – Radio Communi- cations sector (ITU-R) documents, we can calculate the effect of precipitation in the region of operation for the former years via historical precipitation data and make predictions for the possible future effects. Taking this as a starting point, this paper investigates the variability of LMDS radio channel obscured by pre- cipitation in seven cities of Turkey; Adana, Ankara, Anta- lya, Erzurum, Istanbul, Izmir, and Samsun, which are representative for typical precipitation regions of the country. The underlying reason of investigating variability in seven cities, which fall into two rain climatic zones specified by ITU-R recommendations, is the necessity of using location-specific data due to the high value of es- timated attenuation standard deviation (19%) within a rain climatic zone [2]. II. Overview of Turkey’s Climate and Regional Rainfall Characteristics In general, when the geographic location is taken into consideration, Turkey is under the effect of Mediterra- nean climate. But significant differences in climatic con- ditions and therefore in precipitation characteristics are observable between diverse regions of the country as a result of irregular topography. As given in Figure 1, in the driest regions (the Central Anatolia), annual rainfall is less than 300 millimeters whereas in the wettest regions (the Black Sea coastal region) annual rainfall can reach up to 2,200 millimeters, and the region receives rainfall throughout the year [3]. In the Eastern region of Anatolia, terrestrial climate is seen as a result of its topography and annual precipitation in this region averages about 500-800 millimeters with actual amounts determined by elevation. Anatolian Pla- teau, experiencing a steppe climate, rainfall is low and usually in the form of snow whereas annual precipitation in the Aegean and Mediterranean coastal areas varies from 580 to 1,300 millimeters, almost all the time in the form of rain [3]. Figure 2 represents monthly temperature and precipitation values for Turkey over long term. III. Propagation Environment Path losses relative to free-space in line-of-sight (LOS) radio systems are classified under eight groups which are listed in Table I. Among these, the most sig- nificant factor on signal attenuation in the bands from about 1 GHz to 40 GHz is precipitation – rain in specific. Precipitation is a wide term that includes cloud, fog, hail, rain, and snow. Although signal degradation due to hail can also be important, as its occurrence is relatively un- common when compared to rain in the selected regions, its affects are ignored in this research. Other forms of precipitation – snow, hail, and fog are also not consid- ered since they are not affective in frequencies between 1 GHz to 40 GHz due to their drier nature when com- pared to rain [1]. The only form that could be considered as wet is snow but since precipitation data used classi- fies this form under rain, it has already been covered. In general, for radio communication systems operat- ing above 10 GHz signal attenuation due to rainfall is the main component of the path loss. Specifically, it is rain drops that cause attenuation effects in signal like ab- sorption and scattering. So the rain attenuation is de- pendent on shape and size distribution of the rain drops as well as on temperature, angle and velocity of rainfall, and rain rate [4]. IV. Estimation of Rain Fading and Precipi- tation Effects Modeling ITU-R recommendations PN.838 and PN.837 are about specific rain attenuation and characterization of precipitation for propagation modeling. In ITU-R PN.838, for frequencies under 40 GHz and path lengths shorter than 60 km [1], attenuation originating from rainfall is defined by γR and modeled as follows: km dB in R k R / α γ ⋅ = (1) where R is the rainfall rate in mm/hour with different val- ues for each climate zone [5]. k and α are frequency dependent coefficients for linear or horizontal polariza- tion [4]. These frequency-dependent coefficients are calculated by