COMBINATION OF RAMAN-RAYLEIGH LIDAR FOR HUMIDITY AND TEMPERATURE MEASUREMENT OVER CHUNGLI, TAIWAN

1,2CHIH-WEI CHIANG, 2SUBRATA KUMAR DAS, 2JAN-BAI NEE

1Research Center for Environmental Changes, Academia Sinica, Nankang, Taiwan;
2Department of Physics, National Central University, Chungli, Taiwan

Abstract: The combined Raman-Rayleigh lidar system has been designed for the measurement of various atmospheric parameters at Chungli, Taiwan. This paper presents the detailed methodology for retrieving water mixing ratio, temperature and extinction-to-backscatter ratio of aerosols from Raman-Rayleigh backscattering signals in the lower atmosphere. Two case studies; one with dry condition and the other with humid condition were also discussed to demonstrate the effects of humidity on aerosols swelling.


I. INTRODUCTION

The vertical distribution of water-vapour mixing ratio (WMR) and temperature can be obtained through in-situ measurements with radiosonde. The radiosonde observations are important for remote sensing application, but the disadvantage of it is the limited horizontal and temporal resolutions, which are related to the geographical distribution of the meteorological stations and the launching frequency. There are many passive remote sensing sensors, however most of them have poor vertical resolution or yield only column integrated quantities. Such sensors can be ground or satellite based. Satellite measurements have overcome the limitation in horizontal resolution, but they are still limited in temporal resolution. In order to obtain vertical profiles of water vapour and temperature with good temporal and spatial resolution at a desired site and at variable time, it is needed to develop separate techniques for its retrieval. The use of the active sensors such as lidar has overcome the shortcoming mention above and can measure atmospheric water vapour and temperature profiles with high temporal and vertical resolutions in the order of minutes and tens of metres, respectively. The early work demonstrated the technique of Raman spectroscopy in the measurement of tropospheric water vapour [4,9] and temperature [6, 8].

In this paper, the detailed procedures of a nighttime operating Raman-Rayleigh lidar system for the measurement of high vertical resolution profiles of the WMR, temperature and extinction-to-backscatter ratios in the lower troposphere is reported. The lidar-derived WMR and temperature profiles are compared with profiles measured by radiosonde and the correlation between the WMR, temperature and lidar ratio are also discussed.

II. SYSTEM DESCRIPTION

1. Raman-Rayleigh lidar system

The lidar site is located at National Central University, Chungli (24.58 N; 121.10 E, altitude 167 m, MSL), Taiwan. The lidar system consists of transmitter as an Nd: YAG laser operating at 355 nm and a Newtonian telescope, as a receiver with a diameter 45 cm. The lidar system is pointed vertically and operated at night time. The signals are measured by photomultipliers and recorded continuously by two multi-channel analyzers (MCA). The MCA will receive the signals in accordance with the accumulative time and then assemble in the corresponding channel. The range of backscatter signals can be accurately determined with a vertical and temporal resolution of 24 m and 33 seconds, respectively. The signal eventually transferred into personal computer automatically.

2. Lidar equation

The optical power measured by lidar is proportional to the signal backscattered by particles and molecules present in the atmosphere. The lidar signal can be expressed as Eq. (1).

(1)

where PM and PL are the power received from distance Z and laser output energy, respectively; AT is a constant which accounts for the system optical efficiency, the telescope receiver area, and the photomultiplier tube (PMT) spectral efficiency; βatm and τatm are backscatter coefficient and extinction by atmospheric gases (βatm (z) and τatm (z)) and aerosol particle (βa(z) and τa(z)) for laser wavelength respectively.

III. DATA ANALYSIS

Raman lidar measures a scatter radiation, which is shifted in wavelength relative to the excited laser wavelength. When excited at the wavelength 355 nm, there will be Raman shift for nitrogen and water vapour at 387 nm and 407 nm, respectively. The vertical profiles of lidar ratio, WMR and temperature are then computed by using the signals retrieved from Raman-Rayleigh lidar.

1. Aerosol extinction-to-backscatter ratio

The basic lidar equation for the inelastic-backscatter signal can be written as Eq. (2):


           (2)


where the definitions of parameter are similar with Eq. (1), the subscript R stands for Raman signal; is the molecule number density; is the range-independent differential Raman cross-section for the backward direction, where is the scattering cross-section, p is the scattering angle and W is the solid angle; the τatm(R) is the extinction by atmospheric gases (τair(R) ) and aerosol particle extinction (τa(R) ) for the Raman wavelengths, which is written as Eq. (3).


                     (3)


where and are the volume extinction coefficient for gas and aerosol particle, respectively.

The particle extinction coefficient can be obtained from Ansmann et al. (1990) by means of Eq. (2) which is shown in the following:


(4)


where particle extinction is assumed to be proportional to, is constant which depends upon the size and composition of the particles. In this study, the value κ is taken as 1.

From two lidar equations Eq. (1) and Eq. (2), a solution for the particle backscatter coefficient can be obtained and is shown in Eq. (5).


               (5)


The vertical profile of the lidar ratio can be obtained by using Eqs. (4) and (5) as,

                    (6)

2. Water-vapour mixing ratio (WMR)

The WMR is considered as the ratio of mass of water vapour to the mass of dry air in a given volume. With respect to height, Whiteman et al. [13] expressed WMR as Eq. (7).

             (7)

Where and are the system calibration constant and the transmission correction function for the WMR measurement respectively. and are the backscatter Raman signals of water vapour and nitrogen respectively from the atmosphere.

Therefore, WMR () can be calculated by knowing the parameter , and utilizing the ratio of water vapour to nitrogen Raman backscatter signal () measured by lidar.

a. The transmission correction function

The respective extinction coefficient for the Raman shifted wavelength for water vapour and nitrogen signal is the sum of extinction coefficients, scattered by air molecules and aerosols. For air molecules, the atmospheric transmissions differ primarily as dependence, due to Rayleigh scattering. And in case of aerosols, an additional correction is needed and this term can be derived by using the extinction coefficient profile derived from Eq. (4). This will provide the aerosols correction for the differential atmospheric transmission.

b. The system calibration constant Cw

The system calibration constant can be retrieved by using relative humidity (RH). In this paper, the RH data has been considered from radiosonde measurement. From Saucier et al. (1989), the saturation vapour pressure can be shown as Eq. (8).

                            (8)

where is saturation vapour pressure (hPa); a = 7.5 ; b = 237.3 ; is 6.11 hPa (saturation vapour pressure at 0oC); t is temperature (oC).

The saturation vapour pressure can retrieve the saturation WMR [5], which can be written as Eq. (9).

                  (9)

   where Ws(z) the saturation water mixing ratio (g/kg) is, P(z) is the pressure at height z.

The three parameters viz RH, saturation WMR, and WMR are related with a parametric equation represented as Eq. (10).

            (10)

where the symbols have the usual meaning cited above. Thus, the WMR can be derived from radiosonde based on Eqs. (9) and (10). Finally in order to estimate the system calibration constant Cw, we converge the measured profiles of WMR from lidar (by using Eq. (7)) to that obtained from radiosonde data at the reference point.

3. Temperature algorithm

The retrieval of vertical profiles of atmospheric temperature is based on the assumption of the hydrostatic equilibrium and the ideal gas law [7]. Given a reference temperature T(zref) and an atmospheric transmission, the atmospheric temperature (T) can be expressed as follows:


 

(11)


where M is the atmospheric molecular weight; R is the molar gas constant; g is the acceleration due to gravity; is the concentration (number density) of air molecules and aerosol.

Since lidar backscatter signals are proportional to , the lidar equation (Eq. (1)) can be rewritten as:


(12)

where AT is constant.

Thus the Eq.(11) can be rewritten as Eq. (13):

   (13)


However, in the lower troposphere the aerosols are dominating factor and its effect cant be neglected especially for the height below 5 km over the study site [3]. Thus to obtain accurate vertical profiles of atmospheric temperature in the lower troposphere, the iterative method employed by Chen et al. [2], needs the correction and must be modified by considering the corrected backscatter signals for the aerosol as well as molecular transmission, shown as Eq. (14):


 

                 (14)


where the definition of symbols is same with above mention, the aerosol transmission can be derived from Eq. (4) while the molecular transmission can de calculated by using Eq. (15) .

(15)

where is the molecular Rayleigh backscattering cross section.

   The iterative method used by Chen et al., [2] for retrieving the atmospheric transmission () consists of three steps. In the first step, we consider the ideal atmosphere by setting the atmospheric transmission factor as 1 for all heights bins. The second step is to calculate the atmospheric concentration () by considering the step one and making the use of Eq. (12). In the third and final step, atmospheric concentration is used to calculate a new atmospheric transmission by considering Eq. (14) and (15). The steps two and three are repeated until both atmospheric concentration and atmospheric transmission profiles converge respectively.

IV. MEASUREMENT AND RESULTS

1. Comparison with radiosonde measurements

In this paper, to demonstrate the methodology described in section 3 we have considered two cases for study: one when dry condition (RH<60 %) was dominant and other when humid condition (RH>60 %) was dominant. Both the cases have been studied for lower troposphere below 4 km. The measurements were carried on Dec. 24, 2003 (dry condition) and May 12, 2004 (humid condition) without substantial change of the meteorological condition. Fig. 1a shows the comparison of Raman lidar-derived (solid line) and radiosonde-derived WMR (circled symbol) while Fig. 1b shows the comparison between temperatures measured from lidar (solid line) and radiosonde (circled symbol) for Dec. 24, 2003. The horizontal bars represent the standard deviation from the mean value during the observation period. Fig. 2a and 2b is same as Fig. 1a and 1b but shown when humid condition was dominant and the measurement was made on May 12, 2004. The value of RH is also shown in Fig.1a and Fig.2a for reference to know about the dry and wet condition of the environment.

In both the cases, the comparison shows a good agreement between the WMR derived from lidar and radiosonde in the wet condition, however in the dry condition some difference has been found. In Fig. 1a, the difference for WMR measurements occurred at higher height (> ~2.2 km) which is mainly related to less water vapour content, and can cause the uncertainties in both lidar and radiosonde measurement. This effect is more deliberate when the RH value decreases more and the atmosphere becomes drier. However, when the RH >60%, a fairly good correlation has been found between both the measurement of WMR but at ~1km WMR derived from lidar deviates from radiosonde measurement which can be due to the measurement error of radiosonde at very lower altitude. Similarly also for Fig. 2a, good correlation is found among both the measurement of WMR when RH> 60% i.e. wet condition. But discrepancy appears under the drier condition.

The temperature profiles derived from lidar and radiosonde shows an excellent correlation in the both the cases. The measurement errors are within the range of ~1K. The discrepancy between the lidar and radiosonde measurement of temperature can also due to the fact that lidar observation is done for one hour while the radiosonde is the single data observation.


Figure 1. a) Comparison of water vapour mixing ratio measured by lidar (-) and radiosonde (), the cross symbol (+) shows the relative humidity. b) Temperature measurements by using by lidar (-) and radiosonde (). Both are simultaneous observation on Dec. 24, 2003. The horizontal bars in the Fig. 2b represent the standard deviation from the mean profile.

Figure 2. Same as Fig. 1a and 1b respectively but for May 12, 2004

Figure 3. (a) and (b) show the backscattering ratio profiles of aerosols measured on Dec. 24, 2003 and May 12, 2004, respectively. The horizontal bars represent the standard deviation magnitude of the backscattering ratio during the period of measurement.

Figure 4. (a) and (b) show aerosol extinction-to-backscatter ratio on Dec. 24, 2003 and May 12, 2004, respectively. The error bars show the variation of results during the measurements.

2. Aerosols and humidity analysis

Figs. 3a and 3b show the aerosol backscattering ratio profiles for Dec. 24, 2003 and May 12, 2004, respectively. The detailed calculation of aerosol backscattering ratio profile and the definition of aerosols over Chungli, Taiwan has been reported in [3].

Figs. 4a and 4b show the lidar ratios of aerosols derived from lidar for Dec. 24, 2003 and May 12, 2004, respectively. Comparing Fig.1a and Fig.4a, show that smaller lidar ratio (< 35 sr) under higher RH (>70%) and the increase in lidar ratio with decrease of RH. From Fig.2a and Fig.4b, it was observed that the lidar ratio is distributed around 30-50 sr below 2 km, while the RH greater than 70%. The observations also reveal the maximum RH (~85%) is corresponding with the minimum lidar ratio (~32 sr) and showing the tendency of increasing lidar ratio with decrease in RH. Both studied cases indicates that under drier environment (RH<60), the lidar ratio varied within 50-90 sr. The distribution of lidar ratio in different height with RH indicates the change in aerosol type or the effects of RH on aerosol swelling. This result supported that aerosols begin to swell when the RH increases above ~70% [12]. The profiles of the aerosol backscattering ratio for both cases (Fig. 3a and 3b) showed similar tendency with their WMR profiles (Fig.1a and Fig.2a) and indicates the correlation of aerosols with humidity. This will cause the hygroscopic aerosols to grow which will increase the scattering cross section of the aerosols and thus resulting in the higher backscattering ratio.

V. SUMMARY

A coordinated Raman-Rayleigh lidar system has been successfully developed and applied to measure the vertical distribution of atmospheric water vapour, temperature and the extinction-to-backscatter ratio of aerosols over Chungli, Taiwan. The methodology was discussed in detailed for the retrieval process of the above mention atmospheric parameter. The water vapour and temperature profiles measured by lidar yields good agreement with radiosonde, but the former have the advantage of better vertical and temporal resolution. By using the Raman signal the lidar ratio is studied which is one of the important optical parameter that can yield information about the physical nature of aerosols. The quantification of this ratio has been demonstrated for the growth of aerosols as a function of relative humidity. The profiles of water vapour are similar to their backscattering profiles which indicate that the hydrophilic aerosols can produce higher backscattering.

In any event, those simultaneous intercomparisons provide a clear example of the quantitative capabilities of the Raman-Rayleigh lidar technique for profiling atmospheric water vapour, temperature and aerosols.

REFERENCES

1. Ansmann A, Riebesell M, Weitkamp C., 1990. Measurement of atmospheric aerosol extinction profiles with a Raman lidar. Optics Letters, 15: 746-748.

2. Chen W.N., Tsao C.C., Nee J.B., 2004. Rayleigh lidar temperature measure-ments in the upper troposphere and lower stratosphere. J. of Atm. and Solar-Terr.l Physics, 66: 39-49.

3. Chiang C.W., Chen W.M., Liang W.A., Das S.K., and Nee J.B., 2007. Optical properties of tropospheric aerosols based on measurements of lidar, sun-photometer, and visibility at Chung-Li (25N, 121E). Atmosph. Env., 41: 4128-4137.

4. Cooney J.A., 1970. Remote measurements of atmospheric water vapour profiles using the Raman component of laser backscatter. Journal of Applied Meteorology, 9,182-184.

5. Emanuel K.A., 1994. Moist thermodynamic processes: Atmospheric convection. Oxford Univ. press, New York, pp.107-117.

6. Evans K.D., Melfi S.H., Ferrare R.A., Whiteman D.N., 1997. Upper tropospheric temperature measurements with the use of a Raman lidar. Applied Optics, 36: 2594-2602.

7. Hauchecorne A., Chanin M.L., 1980. Density and temperature profiles obtained by lidar between 35 and 70 km. Geoph. Res. Letters, 7: 565-568.

8. Keckhut P., Chanin M.L., Hauchecorne A., 1990. Stratosphere temperature measurement using Raman lidar. Applied Optics, 29: 5182-5186.

9. Melfi S.H., Lawrence J.D. Jr., McCormick M.P., 1969. Observation of Raman scattering by water vapour in the atmosphere. Applied Phys. Letters, 15: 295-297.

10. Nee J.B., Thulasiraman S., Chen W.N., Venkat Ratnam M., Narayana Rao D., 2002. Middle atmospheric temperature structure over two tropical locations, Chung-Li (25N, 121E) and Gadanki (13.5N, 79.2E). J. of Atm. and Solar-Terr. Phys., 64: 13111319.

11. Saucier W.J., 1989. Principles of meteorological analysis. Dover, 9.

12. Tang I.N., 1996. Chemical and size effects of hygroscopic aerosols on light scattering coefficients. J. of Geoph. Res., 101: 19245-19250.

13. Whiteman D.N., 2003. Examination of the traditional Raman lidar technique. II: Evaluating the ratios for water vapour and aerosols. Applied Optics, 42: 2593-2608.