Peculiarities of vertical atmosphere absorption in the millimeter wave band

Transcription

1 RADIO SCIENCE, VOL. 38, NO. 3, 8043, doi: /2002rs002668, 2003 Peculiarities of vertical atmosphere absorption in the millimeter wave band Nikolay V. Ruzhentsev Institute of Radio Astronomy, National Academy of Sciences, Kharkov, Ukraine Received 11 March 2002; revised 27 September 2002; accepted 2 October 2002; published 8 February [1] One complete year of continuous observations of vertical radiation of the atmosphere in two points of millimeter wave band (94 and 38 GHz) was conducted. The total vertical absorption, liquid and vapor water content of atmosphere, and average (effective) temperature of clouds are restored on the basis of these two-frequency radiometric data. Numerous cases of abnormal differences between experimental and theoretical values of cloud absorption at the two frequencies were observed. An analysis of the collection of direct and indirect features that accompany the abnormal events is carried out. It is shown that the main sources of observed abnormal differences between theoretical and experimental data are snow and ice particles in clouds and, probably, water droplets of large size. The possibility of using two-frequency radiometric data not only for retrieving of liquid and vapor water content of atmosphere but also for determining the effective temperature of clouds and for assessing hydrometeor microstructure is noted. For this purpose, polarization measurements are very useful. INDEX TERMS: 0360 Atmospheric Composition and Structure: Transmission and scattering of radiation; 3360 Meteorology and Atmospheric Dynamics: Remote sensing; 0320 Atmospheric Composition and Structure: Cloud physics and chemistry; 0689 Electromagnetics: Wave propagation (4275) Citation: Ruzhentsev, N. V., Peculiarities of vertical atmosphere absorption in the millimeter wave band, Radio Sci., 38(3), 8043, doi: /2002rs002668, Introduction [2] The problems of determination of the phase structure of cloud canopy and estimation of the cloud moisture content are important for climatology, meteorology, ecology, etc. In this regard, an efficient tool for atmospheric remote sensing, such as a two- or multifrequency radiometer method needs continued development, as well as determining new applications. [3] The normalized absorption of droplet fraction of cloud water (t) can be determined by Debye s theory of dielectric properties of water (or water solutions) as expressed by its time of relaxation (t r ). At millimeter wave band (MM WB) (especially at the short wave side of MM WB) these parameters depend only weakly on the salt concentration or any other substances dissolved in water [Shutko, 1986]. However, their values depend strongly on the water temperature (T o ) and on the frequency of radio signal. We expect that the molecular relaxation of small-dispersed drops is connected not only Copyright 2003 by the American Geophysical Union /03/2002RS with the water viscosity, but also with the forces of intermolecular interactions. In addition, at MM WB we can presuppose the appearance of other (nonthermal) mechanisms of radio wave radiation (absorption). [4] The Van De Hulst equation for a cloud drop absorption is: t ¼ 6p e 1 Im W: ð1þ l e þ 2 where W is the liquid water content; l is the wavelength; e is the complex permittivity of water described by Debye s equation: e ¼ e 0 je 00 ¼ e s e o 1 þ jðl r =lþ þ e o ð2þ Here, e 0 and e 00 are the real and imaginary parts of the water dielectric permittivity; e s and e o are the static and optical dielectric permittivities respectively, which are described by the standard models; l r is the relaxation wavelength. These values are described as: e s ¼ 88; 2 0; 4t; e o ¼ 5:5; l r ¼ 2pc e s þ 2 e o þ 2 t r; MAR 8-1

2 MAR 8-2 RUZHENTSEV: VERTICAL ATMOSPHERE ABSORPTION where t is the water temperature in C. The time of relaxation (t r ) is determined as [Basharinov and Kutuza, 1974]: t r ¼ exp½10ð273=t 0 0; 95ÞŠ10 12 sec ð3þ where T 0 is the water temperature. [5] From equations (1) (3), we get the relation between the values of absorption in liquid water of clouds for pair of wavelengths: t l1 t l2 ¼ l2 2 l 2 1 ðe s þ 2Þ 2 þðl r =l 2 Þ 2 ðe o þ 2Þ 2 ðe s þ 2Þ 2 þðl r =l 1 Þ 2 ðe o þ 2Þ 2 : ð4þ [6] The ratio t l1 /t l2 can give information about the dielectric permittivity of water drops. A possible discrepancy between the experimental data and the theoretical ratio t l1 /t l2 might indicate differences of dielectric properties of water drops from properties described by standard models. In addition, the same deviations could indicate the presence a nonthermal mechanism of emission (absorption). [7] Because, in the small particle limit, the ratio of liquid water absorption for two different frequencies is independent of the integral content and concentration of cloud droplets, there is reason to suppose that the ratio t l1 /t l2 can be a parameter for retrieving the average cloud temperature. Besides, the noted anomalous deviations of the ratio from that obtained by calculations (using meteorological data) can help to determine the phase structures of atmospheric water, as well as to observe the dynamics of processes of cloud origin and their transformation. [8] In this paper, using one complete year of continuous observations of vertical atmosphere radiation (absorption), we compared measured and computed (with Debye s theory) values of the cloud absorption ratio for 94 GHz and 38 GHz frequencies (3 mm and 8 mm wavelength). We also analyzed the data for different synoptic situations. [9] The main experimental data considered in paper were obtained using two modulation radiometers (38 GHz and 94 GHz) with a fluctuation sensitivity dt = 0.25 K. Axis-symmetric directional diagrams of horn antennas of radiometers had identical width q 0.5 =5 and were directed at a zenith angle of 40 during one year of measurements. The calibration of radiometers was conducted with the method of two absorbed loads (with temperature T 1 =78K and T 2 = 300 K) or using a radio-frequency radiation of cloudless regions of atmosphere (with usage the near ground meteo-data and Liebe s method for calculations its radiobrightness temperature -T b ). To protect the antennas from precipitation, the radiometers were placed in a long pipe. Figure 1. Theoretical dependencies of ratio t 3 /t 8 on water temperature. [10] The total absorption of the atmosphere in the vertical direction was evaluated from the measurements of the antenna temperature using the following equation: t total ¼ ln T 0 T ant bt bground = ð1 bþ = T 0 Tc g=secq where: T ant antenna temperature of atmosphere measured by radiometer; b dissipation factor of antenna outside of a main beam of directional diagram; T bground effective temperature of background radiation received by back and side lobes of directional diagram of antennas (experimentally obtained values); T c relict cosmic radiation (2.75 K); q zenith angle of sighting; T 0 average temperature of atmosphere (was determined as T 0 =bt 0 where T 0 temperature of near ground air, and b a coefficient depending on wavelength and season (obtained from season-averaged profiles of meteo-fields). In this work were used b = 0.98 for summer and 0.96 for winter in 3-mm range, and b = 0.95 for summer and 0.93 for winter in 8-mm range). [11] For retrieving atmosphere vapor (Q) and water (W) content values we used a set of equations for 3mm and 8mm ranges: t total l ¼ t oxygen l þ CðlÞQ þ kl; ð T 0 ÞW ð5þ Where C and k are the normalized coefficients of absorption in pairs and in liquid water accordingly. As a result of its solution:

3 RUZHENTSEV: VERTICAL ATMOSPHERE ABSORPTION MAR 8-3 Figure 2. The relations of total absorption values for atmosphere vertical direction at 3 mm and 8 mm wave band. h i W ¼ t 3MM total t 3MM oxygen B t8mm total t 8MM oxygen = ð1 B=AÞk 3mm ð6þ h i Q ¼ t 3MM total t 3MM oxygen A t8mm total t 8MM oxygen = ð1 A=BÞC 3mm ð7þ 3MM where t oxygen value computed with meteo-data, A ¼ k 3MM =k 8MM ¼ t 3MM water =t8mm water ; B ¼ C 3MM =C 8MM ðfor both our frequencies B ¼ 6:8Þ: Results of polarization measurements using the radiometer-polarimeter of 3mm range (which had two balanced channels which were loaded on two identical antennas with vertical and horizontal polarizations) are discussed at the end of the paper. The output signal of this radiometer (with dt = 0.4 K) was proportional to a difference between antenna temperatures measured at two orthogonal polarizations. 2. Results [12] The computed ratio t 3 /t 8 dependence on the water droplet temperature is shown in Figure 1. This plot was obtained using equations (1) (4).

4 MAR 8-4 RUZHENTSEV: VERTICAL ATMOSPHERE ABSORPTION Figure 3. Variations of measured radio brightness temperature (a) and retrieved values of total liquid water content (b), vapor water content (c) and mean cloud temperature (d) during the passage of a warm atmosphere front. [13] Our statistical analysis of experimental data (49N, 37E) and theoretical results show that during the one year period approximately for 40% of observed frontal zone of clouds (about 65% for warm season of year and about 25% for cold season but always when the near ground temperature was T o >+5 C) there is no essential discrepancy between measured and computed values of ratio t 3 /t 8 (i.e. there is strong correlation between the cloud temperature and their measured ratio t 0 3/t 0 8). [14] Measured values of optical depth in relation t3/t were determined as: t total l t clear l, where t clear l is value measured in cloud-free periods (or value computed with near ground meteo-data). For specific of types of frontal zones, all experimental points are situated on a few (or even on one) lines, as it shown in Figures 2a and 2b. We see that these lines have different slopes and (or) vertical shifts (Figures 2c and 2e) for different zones of cloudiness. We suppose that this difference of slopes is caused by the difference of cloud droplet temperature. It follows from Figure 1 that a change of water temperature from 30 C to+20 C leads to change of ctg a (here a is angle of inclination) from 2 to 6.2, which agrees with our experimental data. We note that separating the cloudy zones with ratio t 0 3/t 0 8 points situated along one line, as well as measuring the angle of this lines inclination, are very useful for retrieving values of water content with (6) and (7) [Ruzhentsev and Kuzmenko, 1998a, 1998b]. [15] The difference of vertical shifts of these rays on neper relative to one another (Figures 2c 2e) might be caused by the difference of droplet size probability distribution functions (PDF) in the different regions of cloudiness and precipitation. The big influence of the shape of droplet size PDF on the attenuation of short millimeter waves under thunderstorm conditions was demonstrated by Furashov et al. [1994] (for the set of PDF s given by Marshal-Palmer, Laws-Parson and Joss et al.). [16] In Figure 3 an example of observation of the passing of an atmosphere front is shown with the results of simultaneous retrievals of W, Q, and mean cloud layer temperature T C The retrieval methods are described by Ruzhentsev and Kuzmenko [1998a, 1998b]. For the retrieval of cloud effective temperature, we used the computed data from Figure 1.

5 RUZHENTSEV: VERTICAL ATMOSPHERE ABSORPTION MAR 8-5 Figure 4. The retrieved values of total atmosphere absorption and ratio t3/t during passage of some of atmosphere fronts. [17] However, we note the absence in our experiments of radiosonde data of measurements of the atmosphere temperature and humidity profiles. On account of this, we could use only approximate values of cloud temperatures in our calculations and comparisons. [18] The cloud temperature estimation was accomplished by using the near ground air temperature measurements with an assumed vertical gradient of air temperature as 6 K/Km for summer and 3 K/Km for winter. [19] Although such an approach has only statistical significance (and for some of particular meteorological situations might lead to errors of cloud temperature up to 10 C), nevertheless it allows us to find cases when t 0 3/t 0 8 ratio had anomalously low (or anomalously high) values relative to the predicted one. [20] Such substantial anomalies were observed regularly during passage of about 60% of atmosphere frontal zones (35% for warm seasons and about 75% for cold seasons). In Figure 4 two typical examples of these situations are shown. [21] We think that there are following possibilities to explain, or even to remove, the discrepancies noted above: 1. to change the parameter t r, obtained by the common technique of theoretical evaluations (4) for water, and their realistic values for small drops of water, to agree with the measured data; 2. to explain such discrepancies by the uncertainties in phase state and parameters of hydrometeors that lead to the observed anomalies. [22] We believe that the second way of explanation of observed anomalies is more physical and more fruitful. It is based on the series of features derived from our experimental data as well as from the theoretical results of other authors (where the influence of diffraction by the super big drops and ice crystals on the hydrometeors absorption value at MM WB was considered). In Figure 5 and Figures 2c 2e, we depict the results obtained this way. [23] Actually for most of the records obtained in the warm seasons as well as for many of records obtained in the cold seasons (but with the positive C temperatures of near ground air) it is typical to find good correspond-

6 MAR 8-6 RUZHENTSEV: VERTICAL ATMOSPHERE ABSORPTION Figure 5. The coefficient of scattering at 94 GHz (curves 1 and 3) and 39 GHz (curves 2 and 4) for a monodisperse distribution of 1 gm/m 3 of water content (3 and 4) as well as for a cloud composed by spherical ice particles (1 and 2) as a function of particle diameter. ence between the measured and calculated values of t 3 /t 8 ratio. In addition, there are some regularities concerning the anomalously high values of t 0 3/t 0 8 that were observed mostly in winter or during transition-seasons as well as the cases of the anomalously low values observed in summer season only. [24] A possible cause of winter and transitional-seasons excesses might be the existence of ice fraction of clouds with predominant sizes 0.5 to 2.5 millimeter of ice or snow particles (that often occur in nature which follows from the data of in situ measurements contained, e.g., in the work of Hatanaka et al. [1993]). [25] In this case it is reasonable to turn to the results obtained by Lhermitte [1988] where a theoretical evaluation of extinction (attenuation) values due to the scattering by ice particles at 3 mm WB (using Mie s theory) was carried out. From Figure 5 it follows that the coefficients of scattering at 3 mm WB can reach more then 10 db/km (curve 1). At the same time at 8-mm WB the scattering coefficient increases by less than 3 db/km (if we neglect some small changes of ice dielectric permittivity for the considered interval of frequencies, it is possible to create curve 2 of Figure 5). Thus, the existence of frozen particles of water might cause above mentioned anomalies. Our estimations show that the cloud zones with increased values of vapor humidity do not affect noticeably the extinction coefficient, especially in cold seasons. [26] As we have mentioned above, sometimes in summer there are cases of an anomalous decrease of the ratio t 0 3/t 0 8. It is possible that such a decrease is caused by the super large drops of water with sizes >1.5 mm (Figure 5) or by big ice crystals (3 mm 4 mm), as well as by very large values of water vapor content. [27] For ice particles, this decrease might be caused by the opposite signs of gradients in the change of absorption at 8 mm and 3 mm WB. In summer the cases of high ratio t 3 /t 8 were observed in Cb clouds (not in Sc clouds characterized by a very high vapor water content Q reached 40kG/m 2 ) which are characterized by a high value of liquid water content W (1 2 kg/m 2 and even more) and by a small value of a vapor water content (usually near 0.15 kg/m 2 )[Andreev, 1982]. It is very likely that the presence of ice crystals is the main cause of the phenomena considered here, because of numerous observations [Mazin and Shmerer, 1983] that Figure 6. The examples of recording of synchronous change of radiobrightness temperature (T b ) and polarization difference (T V T H ) of the sky during passing a single cloud (a) and field of cloudiness (b) above the measuring point (a zenith angle of sightings q = 40 ).

7 RUZHENTSEV: VERTICAL ATMOSPHERE ABSORPTION MAR 8-7 show a very little relative portion of super large size drops in their size PDF. [28] The plausibility of the explanations given above increases if we take into account that such summer anomalies were observed mainly into the thunderstorm clouds. As a rule, the large-size ice or hail formations are presented in the higher layers of these clouds [Mazin and Shmerer, 1983]. Moreover, our recent cycle of polarization observations of cloud canopy have confirmed the data contained in the work of Vivekanandan et al. [1993] about polarization differences between big size snow and ice particles. Such polarization differences with data of our measurements can reach 18 K for zenith angles of 40. Furthermore we observed the polarization differences at 3mm WB when the radio brightness at vertical polarization exceeded the horizontal one and vice versa (Figure 6). These differences of signs of depolarization, as it is shown by the theoretical calculations by Osharin and Troitsky [2000], could characterize the microstructure of cloud ice particles (if the ice particles are spherical then T V > T H, but if these particles have the plain plate form, then T H > T V ). [29] All these arguments lead us to prefer the hypothesis of the influence of large sized particle of frozen water, on the observed anomalies of cloud absorption ratio for different frequencies at MM WB (although hypothetically the same effect might take place in special case when the PDF of droplet sizes is concentrated near a few millimeter value). 3. Conclusions [30] As a result of analysis of one complete year of radiometric data it was concluded: 1. the existence of a temperature dependence t 0 3/t 0 8 was observed; this leads to the possibility of determining cloud mean temperature from the radiometric measured ratio t 0 3/t 0 8; 2. the existence of the numerous cases of abnormal measured ratio t 0 3/t 0 8 that exceed the expected theoretical value, as well as the more rare cases of its anomalous low value; 3. confirmation of the key role of a large-sized ice fraction of cloud water in abnormal excess of the expected values t 3 /t 8 (that can be used for determination the phase structure and the type of clouds); 4. indirect confirmation the possibility of determining some cloud crystalline zones by radiometric polarization measurements. References Andreev, G. A., Earth covers heat emission at millimeter waves (in Russian), Zarub. Radioelectron., N12, 3 39, Basharinov, A. E., and B. G. Kutuza, Determination of temperature dependence of time relaxation of water molecules in clouds as well as determination of possibility of evaluating affective temperature of clouds from radiometer measurements (in Russian), Izv. Vyssh. Uchebn. Zaved Radiofiz., 17(N1), 52 57, Furashov, N. I., V. Y. Katkov, and B. A. Sverdlov, Hysteresis effects in attenuation/rain rate relationships at the shorter millimeter wave lengths, in Proceedings of the International Conference on Millimeter and Sub-Millimeter Waves and Applications, San Diego, California, pp , Int. Soc. for Opt. Eng., Bellingham, Wash., Hatanaka, M., Y. Ohta, H. Takeya, I. Sugioka, A. Nshitsuji, and M. Wada, An evaluation method of snow particle size distribution functions from VTR image for meteorological radar observations, paper presented at International Symposium IGARSS 93, Tokyo, Inst. of Electr. and Electron. Eng., New York, Lhermitte, R. M., Cloud and precipitation remote sensing at 94 GHz, IEEE Trans. Geosci. Remote Sens., 26(N3), , Mazin, L. P., and S. M. Shmerer, Clouds: Their Structure and Formation (in Russian), 280 pp., Gidrometeoizdat, Leningrad, Osharin, A. M., and A. V. Troitsky, Polarization of the thermal radiation of the cloudy atmosphere in millimeter wavelength band, in Proceedings of the International Conference on Mathematical Methods in Electromagnetic Theory (Sept , 2000, Kharkov, Ukraine), vol. 1, pp , Inst. of Electr. and Electron. Eng., New York, Ruzhentsev, N. V., Y. A. Kuzmenko, An annual motion of atmosphere water content from integral absorption radiometer observation, in Proceedings of the URSI Commission-F Open Symposium on Climatic Parameters in Radio Wave Propagation Prediction (April 1998, Ottawa, Canada), pp , Inst. of Electr. and Electron. Eng., New York, 1998a. Ruzhentsev, N. V., and Y. A. Kuzmenko, Some results of atmosphere water motion observation from annual radiometer data, in Proceedings of the URSI Commission-F International Triennial Open Symposium on Wave Propagation and Remote Sensing (22 25 September 1998, Aveiro, Portugal), pp , Inst. of Electr. and Electron. Eng., New York, 1998b. Shutko, A. M., Microwave Radiometer Measurements of Water Surfaces and Bare Soils (inrussian), 190pp.,Nauka, Moscow, Vivekanandan, J., J. Turk, F. S. Marzano, A. Magnai, R. W. Spenser, R. E. Hood, et al., Active and passive remote sensing of precipitation over ocean surfaces, paper presented at International Symposium IGARSS 93, Tokyo, Inst. of Electr. and Electron. Eng., New York, N. V. Ruzhentsev, Institute of Radio Astronomy, National Academy of Sciences, Kharkov, Ukraine. (ruzh@rian.kharkov. ua)