**Radio Science**

# Simultaneous MST radar and radiosonde measurements at Gadanki (13.5°N, 79.2°E) 1. Causative mechanism and characteristics of radar backscatterers at VHF

## Abstract

[1] Simultaneous MST radar and radiosonde observations have been carried out from Gadanki (13.5°N, 79.2°E), a tropical station in India. The radar measurements are used to obtain winds, vertical shear of horizontal winds (S), reflectivity (η), aspect sensitivity and horizontal correlation length (ζ) of the radar backscatterers. The radiosonde measurements of pressure, temperature and humidity are utilized for the computations of Brunt Vaisala frequency (N) and potential refractive index gradient (M). These measurements are used to understand the characteristics of the radar backscatterers in terms of prevailing background atmospheric conditions. Observations show strong vertical shears of horizontal winds in the height range of 16–18 km, at the upper edge of the tropical easterly jet (TEJ) associated winds. Results show that these shears and N^{2} both contribute to enhance the echo power in oblique directions (χ = 10°). The oblique beam observed enhancement of echo power, due to enhanced wind shears, is noticed to be confined to a narrow height range and associated backscatterers are more or less isotropic. At the height where N^{2} (or M_{D}^{2}) contribution to oblique echo power is significant, the radar backscatterers appear to be relatively more anisotropic with the horizontal correlation length (ζ) of 16–20 m. Present observations thus bring out clearly that even at oblique beam with zenith angle χ = 10° contribution of enhanced N^{2} to radar reflectivity (η) from refractivity structure with correlation length of ∼10–20 m can be very significant especially at the height above 17.5 km. These results have to be necessarily taken into account while interpreting the observed radar reflectivity observations.

## 1. Introduction

[2] Observations by VHF radar show enhanced radar echoes at vertical incidence relative to off vertical incidence. This aspect sensitivity of the radar echoes at VHF is due to the fact that radar backscatter from clear air in lower and middle atmosphere arise due to refractive index irregularities caused by turbulence [e.g., *Van Zandt et al.*, 1978; *Gage and Balsley*, 1980] and also due to Fresnel reflection/Scattering from sharp vertical gradients in radio refractive index. The Fresnel reflection/Scattering may arise from layers of radio refractivity with horizontal correlation length as large as the antenna diameter or as large as the first Fresnel zone or even larger. Strong echoes are observed, at the radar beams with small zenith angle, from the stable layers associated with temperature inversion [*Gage et al.*, 1986; *Gage*, 1990; *Jain et al.*, 1994; *Jaya Rao et al.*, 1994]. The radar echoes are observed to be aspect sensitive. This aspect sensitivity of the radar echo depends on the radar parameters viz. frequency, beam width, beam zenith angle and the back ground atmospheric parameters such as atmospheric winds, temperature and humidity.

[3] In recent years there have been considerable efforts to understand the aspect sensitivity of the received radar echoes basically to get an idea of the characteristics of the radar backscatterers and also to understand the effect of echo aspect sensitivity in determining various atmospheric parameters, such as horizontal winds from radar observations using DBS method [*Tsuda et al.*, 1986, 1988; *Hocking et al.*, 1990; *Hooper and Thomas*, 1995; *Jain et al.*, 1997].

[4] More recent work on the aspect sensitivity of the radar echoes suggests that even for radar observations at beam zenith angle χ ∼ 10°, or even at beam zenith angle higher than 10°, there could be significant contribution from enhancement in N^{2} where N is the Brunt Vaisala frequency and parameter N^{2} represents atmospheric stability [e.g., *Tsuda et al.*, 1997; *Hooper and Thomas*, 1998; *Worthington et al.*, 1999]. Most of the studies so far have been confined to mid and high latitudes. There are a few experiments that have been reported from the tropical latitudes.

[5] In a recent paper, *Jain et al.* [2001] used MST radar observations from tropical station Gadanki (13.5°N, 79.2°E) and simultaneous radiosonde observations from nearest station Chennai (13.1°N, 80.2°E), which is located 120 km southeast of the radar site. At some heights, anisotropic Bragg scatterers are associated to the gradient in radio refractive index having horizontal correlation length ∼10–20 m, whereas at other heights these are from Bragg scatterers with correlation length less than radar wavelength, indicating the presence of isotropic structures. Thus, both mechanisms are noted to contribute.

[6] In this paper we report a study on this aspect using a series of radiosonde launches that were carried out from the MST radar site Gadanki (13.5°N, 79.2°E), a tropical station in India, during summer monsoon season from 19 July to 14 August 1999. These measurements are the first of its kind that have been carried out from Gadanki. This was done during the summer monsoon season when the TEJ wind and associated vertical shear of horizontal winds are large. These simultaneous observations are examined to determine the role of atmospheric stability parameter (N^{2}) and sharp vertical shear of horizontal winds (S) in giving rise to enhanced radar backscatter and to determine the characteristics of the radar backscatterers in the troposphere and in the lower stratosphere. In part II of this paper, simultaneous MST radar and radiosonde measurements have been utilized to determine the height profiles of various atmospheric turbulence parameters.

## 2. Observations

[7] Simultaneous MST radar and radiosonde observations have been carried out from Gadanki (13.5°N, 79.2°E) during summer monsoon season from 19 July to 14 August 1999. The Indian MST radar is a sensitive, coherent pulse Doppler radar operating at 53 MHz, with peak power aperture product of 3 × 10^{10} Wm^{2}. The phase array consists of 1024 three element Yagi antennas occupying the geometric area of 130 × 130 m^{2}. Radiation pattern of this system has 3° beam widths with gain of 36 dB and a side lobe level is −20 dB [*Jain et al.*, 1994; *Rao et al.*, 1995]. The radar was operated every day between 16:45 to 17:30 IST in a standard mode. Details of the experiments specification file (ESF) used for these observations are given in Table 1. The radar spectrum data are converted to moments for obtaining the atmospheric parameters using the adaptive method given by *Anandan et al.* [1997]. Radiosonde have been launched every day from radar site at 16:20 IST. These radiosonde carried pressure, temperature and humidity sensors and data were obtained at one minute interval corresponding to a height interval of ∼300 m. The balloon reached height of 24 km in about 1 hour 20 min. from the time of launch.

Parameter | Specifications |
---|---|

- a
E 10y = beam direction 10° east from the zenith in east-west plane. W10y = beam direction 10° west from the zenith in east-west plane. Zy = vertical beam direction formed using east-west plane array. Zx = vertical beam direction formed using north-south plane array. N10x = beam direction 10° north from the zenith in north-south plane. S10x = beam direction 10° south from the zenith in north-south plane.
| |

Pulse width (μs) | 16 |

Inter pulse period (μs) | 1000 |

Coded/Uncoded | Coded (16 baud code, each baud = 1 μs) |

Range resolution | 150 m |

No.of beams | 6(E10y,W10y,Zy,Zx,N10x,S10x )^{a} |

Coherent Integration | 128 |

No. of FFT points | 128 |

Nyquist frequency (Hz) | ±4 |

Doppler resolution (Hz) | 0.06 |

Observational window: | |

Lowest range bin (km) | 3.6 |

Highest range bin (km) | 32 |

Incoherent integration | 1 |

Beam dwell time | ∼16 sec |

STC length (μs) | 40 |

No. of scan cycle | 8 |

## 3. Results and Discussion

[8] Simultaneous MST radar and radiosonde observations from radar site are used to examine the role of vertical potential refractive index gradient (M) and the vertical shear of horizontal winds (S) in giving rise to observed radar backscatter and also to get an idea of the horizontal correlation length (ζ) of the backscatterers. The radiosonde measurements of pressure, temperature and humidity are used to compute potential temperature (θ), atmospheric stability parameter (N^{2}) and the vertical potential refractive index gradient (M). The parameters N^{2} and M are defined as follows:

For the dry atmosphere, where q and q′ can be neglected, M ≅ M_{D}, then

[9] In these expressions, symbols have their standard meaning [e.g., see *Cohn*, 1995]. The radar observations are used to obtain the zonal (u) and meridional (v) component of the horizontal winds and also to compute the vertical shear of horizontal wind (S), i.e.,

[10] Simultaneous radiosonde and radar observations are then used to compute the Richardson number (Ri).

to get an idea of the height range where active turbulence could give rise to refractive index irregularities responsible for enhanced radar reflectivity at these height levels. For the purpose of computing Ri using radiosonde and radar data, the radiosonde data are interpolated using linear interpolation at the height interval of 150 m to match the height levels of measurements by two instruments. The atmospheric stability parameter (N^{2}) and vertical shear of horizontal wind (∂U_{h}/∂Z) are computed corresponding to each radar range gate. For these computations the vertical gradient is determined using a three-point method, thus limiting the height resolution of these parameters to 300 m. The measurement of Ri in the atmosphere is necessary to define Ri over a specified height interval (Δz). As per the *Reiter and Lester* [1969] discussion, Ri is scale-dependent, and typically there is enough structure in the “mean” velocity fluctuation i.e. Ri_{Δz1} < Ri_{Δz2} where Δz1 < Δz2. It means that high-resolution measurements of velocity and temperature are required to observe the small values of Ri that accompany small scale instability and turbulence in the free atmosphere. Keeping the above considerations and the height interval (Δz) of the present measurements in mind, small values of Ri (i.e. Ri ≤ 1) are taken as an indicator of the presence of refractivity structures, associated to mechanical turbulence.

[11] The radar observations of the mean signal to noise ratio (S/N) for four oblique beams are also used to compute the radar reflectivity (η) [*Ghosh et al.*, 2000]

where (S/N)_{o} represents the mean signal-to-noise ratio (SNR) for oblique beams.

[12] Substituting the values of various radar constants (see Tables 1 and 2), the following expression is obtained for the η:

where P_{t} is the peak radar-transmitted power in watts and the same has been used here according to its value for each day of observation.

Symbols | Parameter | Values |
---|---|---|

λ | Radar wavelength | 5.66 m |

Δr | Range resolution | 150 m |

A_{e} | Effective antenna area | 1.2 × 10^{4} m^{2} |

K_{B} | Boltzman's constant | 1.38 × 10^{−23} J/K |

B_{N} | Receiver bandwidth (effective) | 1.7 × 10^{6} Hz |

α_{r} | Receiver path loss | 4.4 dB |

α_{t} | Transmitter path loss | 2.65 dB |

T_{C} | Cosmic noise temperature | 6000 K |

T_{r} | Receiver noise temperature | 607 K |

N_{B} | Number of bauds for coded pulse | 16 |

N_{C} | Number of coherent integration | 128 (Table 1) |

r | Range of backscatter echo in meters |

[13] Radar average SNR for oblique and vertical beams is used to compute aspect sensitivity. Here aspect sensitivity implies the difference of vertical beam and oblique beam SNR in dB. Aspect sensitivity is inturn used to compute θ_{s} that is e^{−1} half width of the polar diagram of the received backscatterer [*Hocking*, 1986]. The parameter θ_{s} is characteristic of the scatterer. For small beam pointing angle, as in case of Indian MST radar, θ_{s} give a direct measurement of the horizontal correlation length (ζ) of the radar backscatter [*Hocking et al.*, 1990] and the same is given by

where θ_{s} is expressed in degrees. For Indian MST radar the effective aperture (A_{e}) is 1.2 × 10^{4} m^{2} and the same corresponds to an effective antenna diameter (D_{eff}) of ∼110 m. The Fresnel effect is significant when the horizontal correlation length ζ ≥ 0.29 D [*Gage*, 1990]. Therefore, for Indian MST radar observations, Fresnel effects become significant for ζ ≥ 32 m which corresponds to θ_{s} ≤ 2.7°.

[14] Figure 1 shows the height profile of the received backscattered echo power for 21 July 1999. The large fluctuations in echo power with height indicate that the scatterers are not uniformly distributed, but they are in discrete layers. The present series of observation have been carried out using the radar resolution of 150 m. The radar received echo power even with this range resolution shows sharp well-defined peaks. This confirms that such peaks are arising due to layers of reflectivity, which are comparable or thinner than the range resolution. However, a close examination of Figure 1 shows that such layers are quasi periodic in height with the separation in range ∼300 to 700 m with the mean interval of 500 m. Thus making the layers of enhanced reflectivity clearly distinguishable. It must be mentioned here that many workers have reported such sharp layers of enhanced radar reflectivity [e.g., *Rottger and Schmidt*, 1979; *Barat*, 1982; *Sato and Woodman*, 1982; *Jaya Rao et al.*, 1994; *Ruster et al.*, 1998; *Jain et al.*, 2001].

[15] Figure 2 shows the typical mean height profile of winds speed and vertical shear of horizontal wind on one day of observation. The horizontal bars shows the standard deviations over 45 min during which the sets of wind observations are taken. From this figure, strong wind shears are clearly observable at the upper edge of the jet stream. Shears are relatively weaker near the bottom of the jet stream. An arrow indicates the height of tropopause, given by the radiosonde. It would now be interesting to look for the contribution of such wind shears to the observed radar reflectivity (η).

### 3.1. Typical Height Profiles of Various Observed Parameters

[16] Figures 3 to 5 show the height profiles of N^{2}, wind speed, square of vertical shears of horizontal wind (S^{2}), relative humidity (RH), Richardson number (Ri), logM^{2} and logM_{D}^{2}, logη, aspect sensitivity and horizontal correlation length (ζ) for three different days of observations. The averaging time for the height profiles of wind speed, wind shear and η is 45 min. An attempt is made to identify the contributions of enhanced S^{2}, N^{2} and enhanced relative humidity to the enhancement of η. Thick solid lines are drawn in Figures 3 to 5 to represent the case of the height levels of enhanced η that could be associated to the enhancement of one of these parameters as discussed below.

#### 3.1.1. Case I: The Contribution of Enhanced S^{2}

[17] The solid lines A and D in Figure 3 and line B in Figure 5 represent the case where shears are high compared to N^{2} and Ri is low (Ri ≤ 1). It can be noted that in all these cases peak reflectivity (η) and low aspect sensitivity and horizontal correlation length (ζ) appears at a height slightly lower (150–300 m) than the height of minimum Ri. The enhanced η in these cases is predominantly due to refractive index irregularities associated to shear generated turbulence. This is consistent with low aspect sensitivity and small ζ (<10 m) and suggests that the back scatterers are more or less isotropic in nature. One of the reasons that the peak in η and minimum in ζ appears at height somewhat lower than the height of peak shear. This could be due to low frequency gravity waves as discussed by *Tsuda et al.* [1985] to explain their observations from the Arecibo radar. It is also apparent, from Figures 3 to 5, that the contribution of enhanced S^{2} to the observed radar reflectivity is confined to a narrow height range.

#### 3.1.2. Case II: The Contribution Due to Enhanced N^{2} and Humidity (RH)

[18] At some height levels, enhanced humidity and N^{2} are also observed to contribute to η at oblique beam (χ = 10°). Solid line A in Figures 4 and 5 represents cases where enhanced humidity contribute to enhanced η. The lines C and D in Figure 5 represent the cases where enhanced N^{2} (or M_{D}^{2}) contributes to enhance η. For all these cases, Ri are large (Ri ≥ 1), aspect sensitivity is high and horizontal correlation length is large (ζ ≥ 10 m).

[19] It is also evident from Figures 3 to 5 that enhanced values of η, above the altitude of 17.5 km, are generally due to enhanced N^{2} (or M_{D}^{2}). At these heights, Ri and ζ are large. This suggests that enhanced N^{2} is important contributor to enhance η above 17.5 km.

#### 3.1.3. Case III: The Contribution Due to Enhanced N^{2} and S^{2}

[20] The cases where enhancement in N^{2} and S^{2} both contribute to enhancement in η are represented by height range between lines B and C in Figure 3, height range between lines B and C, and D and E in Figure 4. Out of these three cases in η, two cases of enhanced S^{2} appear just above the peak of N^{2} (see lines B and C in Figure 3 and lines D and E in Figure 4). For the third case, enhanced S^{2} appears just below N^{2} (see lines B and C in Figure 4). It can also be noticed from the three cases, discussed here, that where enhancement in η is due to enhancement in S^{2}, Ri ≤ 1, aspect sensitivity is low and values of ζ are small (≤10 m). It should be again mentioned here that in these cases also height of peak η and low values of ζ appear at slightly lower height of peak shear and low Ri as already discussed for case I. It can also be noted, from the three cases under discussion, that at height where enhancement in η is due to N^{2} the same corresponds to large values of Ri, high aspect sensitivity and relatively large values of ζ are observed.

[21] In summary it can be said that results presented in Figures 3 to 5 show that enhanced S^{2}, N^{2} and relative humidity (RH) at different height levels contribute to enhanced η even at oblique beam at large as 10°. Sometimes the contributions of enhanced N^{2} and S^{2} appear together i.e. one on the top of the other. The layer of enhanced humidity contributes at lower heights (≤10 km). At heights above 17.5 km, enhancement in η appears to be mainly due to enhancement in N^{2}.

### 3.2. Height-Date Contour Maps for Various Parameters

[22] So far, observations for three individual days are examined. It is necessary to examine the observation taken over the whole month in detail to get a clear picture of the contribution of S^{2} and N^{2}. For this purpose, height-date contour maps of various parameters such as wind speed, S^{2}, N^{2}, logη, aspect sensitivity and ζ are drawn in Figure 6. The solid line on the contour maps represents the height of tropopause determined from radiosonde observations. The gaps in this line are due to nonavailability of radiosonde data for these days.

[23] The following points may be noted from Figure 6:

- There are strong shears present in the height range of 16–18 km (see panel b) and these shears appear at the upper edge of TEJ observed at these latitudes during summer monsoon season as seen from panel (a).
- The contour maps of log η show a minimum in the height range of 14–16 km and a secondary maximum between 16–18 km (panel d). An intercomparison of contour maps of η and S
^{2}shows close correspondence between enhanced S^{2}and η in this height range. - An examination of N
^{2}contour maps shows a clear enhancement in the height range of 17–20 km. It is also apparent that enhancements in N^{2}and S^{2}some times appears very close to each other i.e., one on the top of the other. Comparing the contours of η and N^{2}, it is evident that enhanced N^{2}also contributes to enhancement in η, especially above the height of 17.5 km. - An examination of the contour of aspect sensitivity and horizontal correlation length (ζ) shows high aspect sensitivity and large values of ζ (≥10 m) above 17.5 km. Values of ζ as high as 16–20 m are also observed frequently. These large values of ζ indicate the presence of more anisotropic backscatterers at these height levels, as compared to those present at lower heights.

[24] The height range near the tropical tropopause during the summer monsoon season is dominated by strong shear associated to TEJ winds as seen from Figure 6 (panel a). This height range, as shown by *Jain et al.* [2001], also contains multiple stable layers (see Figure 6 also). Therefore, the height range around the tropical tropopause, say between 16–20 km, is ideal for detailed examination of contributions of S^{2} and N^{2} to enhancement in η. With this objective, height-date contour maps for the parameters S^{2}, N^{2} and η are drawn for the height range of 14–20 km. These are shown in Figure 7, and panels (a), (b) and (c) of this figure show the height-date contour maps of S^{2}, N^{2} and η respectively. The solid line shown in panel (c) of this figure represents the height of occurrence of peak vertical shears as noted from the radar observations.

[25] The following points emerge from the close examination of Figure 7:

- There is a close correspondence between the enhanced values of η and height of enhanced shears confirming that shears do play an important role in generating refractive index irregularities, which are associated to mechanical turbulence.
- Enhanced N
^{2}can also be observed to contribute to enhanced layer of η, especially above the height of 17.5 km. However, at the height above 17.5 km, also occasionally layers of enhanced shears are observed to contribute to enhanced η.

[26] Figures 6 and 7 thus give a clear idea of the relative contribution of enhanced S^{2} and N^{2} to enhanced radar reflectivity on a day-to-day basis.

## 4. Conclusions

- Simultaneous radar and radiosonde observations show that enhanced vertical shear in horizontal wind, enhanced N
^{2}and layers of enhanced humidity all contribute to the enhanced η at various height levels depending upon the background atmospheric conditions. - Contribution of humidity layer can be seen at lower heights i.e. below 10 km.
- Contribution of shear (S
^{2}) to enhanced echo intensity is noticed to be confined to a narrow height range. At these heights the backscatterers are more or less isotropic in nature. - As expected, the radar backscatterers are noted to be relatively more anisotropic at the heights where enhancement in N
^{2}(or M_{D}^{2}) contribute η. Aspect sensitivity is generally high at the heights above 17.5 km. In this height range, radar backscatterers with horizontal correlation length (ζ) as large as 16–20 m are observed. - It is also evident from the present set of observations that contribution of N
^{2}to radar reflectivity is significant at some heights even at zenith angle χ ∼ 10°. This contribution appears to be dominant above the height of 17.5 km. This result has significant implication in interpreting the radar reflectivity observations. High aspect sensitivity, as observed above 17.5 km, could introduce a bias in determination of the horizontal winds. This particular aspect needs to be examined separately.

## Acknowledgments

[27] The National MST Radar Facility (NMRF) is set up jointly by the Council of Scientific and Industrial Research (CSIR), Defense Research and Development Organization (DRDO), Department of Electronics, Environment, Science and Technology and Space, Government of India with Department of Space as a nodal agency. The NMRF is operated by the Department of Space, Government of India, with partial support from CSIR. We wish to thank the staff of the National MST Radar Facility for the collection of data used in this paper. We would like to thank India Meteorological Department (IMD) for supporting the radiosonde observation campaign carried out from NMRF, Gadanki, during July–August 1999. We would also like to thank the anonymous reviewers, whose suggestions have resulted in substantial improvement of the paper.