Geophysical Research Letters

Onset of Indian summer monsoon over Gadanki (13.5°N, 79.2°E): Study using lower atmospheric wind profiler



[1] The onset of Indian summer monsoon (ISM) over Gadanki (13.5°N; 79.2°E) is identified using variations in signal-to-noise ratio (SNR), wind speed, wind direction and vertical velocity at 1.5 km (∼850 hPa) using lower atmospheric wind profiler (LAWP). Strengthening of the low level wind speed attaining 8 ms−1 with directional change from south-easterlies to south-westerlies defines the beginning of the monsoon. Enhancement in SNR few days before with noticeable magnitude of 5–10 dB at the time of onset combined with clear-air vertical velocity reversal from downward to upward few days before onset and persisting during monsoon activity supplements the wind speed criteria in identifying onset. Hydrometeor velocity shows large downward values exceeding more than 1 ms−1 indicating occurrence of rainfall. It is proposed that UHF radar at a location can be used to identify the onset of ISM based on wind speed without considering rainfall separately.

1. Introduction

[2] Monsoon, in general, is defined as seasonal reversal in the wind direction between winter and summer. The Indian summer monsoon (ISM) develops in response to the large thermal gradients between the warm Asian continent to the north and cooler Indian Ocean to the south. The strong southwesterly flow in the lower troposphere brings substantial moisture into Indian sector, which is released as precipitation. Because of the critical importance of monsoon rainfall to agricultural productions and economical conditions, predictions of the monsoon's arrival date are of great interest. But there is no precise definition for the onset of the monsoon. However, it is usually noted by a change of wind direction and conventionally identified in terms of rainfall occurrence. Onset of the summer monsoon has been defined by various methods. Using rain gauge data, Ananthakrishnan and Soman [1988] defined the onset of ISM based on Kerala rainfall during which the rainfall amounts increase to over 15 mm per day. Fasullo and Webster [2003] defined the ISM onset and withdrawal by vertically integrated moisture transport over the Arabian Sea. Flatau et al. [2001] defined double onset or bogus onset of ISM based on a conceptual model. Prasad and Hayashi [2005] studied onset in terms of Zonal asymmetric temperature anomaly between 850 hPa and 200 hPa with NCEP/NCAR reanalysis data. Taniguchi and Koike [2006] defined onset based on wind speed exceeding 8 ms−1 at 850 hPa using NCEP/NCAR reanalysis data.

[3] Several investigators found interesting features during evolution of numerous parameters related to the monsoon activity. Krishnamurti et al. [1981] found that there was an increase in kinetic energy at 850 hPa during the process of onset of summer monsoon over Arabian Sea. Rao et al. [2005] found that the advent of monsoon is different at various locations using operational analysis. Goswami and Ajaya Mohan [2001] studied the intraseasonal oscillations and interannual variability of ISM using NCEP/NCAR reanalysis for the period 1956–1997. Joseph et al. [2006] developed a three step method for defining monsoon onset objectively over Kerala (MOK) using NCEP/NCAR winds, OLR and integrated water vapor with a criteria of area mean wind reaching 6 ms−1 at 600 hPa. But all the studies mentioned utilized either radiosonde data or reanalysis data. Very few studies were carried out using wind profilers. Using VHF radar wind profiler at White Sands Missile Range, Nastrom and Eaton [1993] presented a case study on the onset of seasonal transition from dry conditions of early summer to more humid conditions of late summer over South New Mexico. Using UHF radar Krishnan et al. [2005] reported the signatures of intra-seasonal oscillations (ISOs) and its height-time evolution in the lower atmosphere over Gadanki with UHF multilevel wind observations and NCEP reanalysis data.

[4] Proper understanding of the events that evolve during the onset of monsoon is important for its prediction. A necessary step towards understanding these mechanisms is the detailed observation of the monsoon onset. Since strong low level westerly winds are main features of ISM and basic parameter that is obtained from lower atmospheric wind profiler (LAWP) is wind, an attempt is made to study the onset of Southwest Monsoon at Gadanki extending the criteria as mentioned above given by Taniguchi and Koike [2006]. Although a single station observations alone cannot indicate the onset of large scale monsoon over the sub continent, its arrival at a particular station is crucial in understanding the advancement and activity of the monsoon. This study enables one to use LAWP to observe monsoon progress using the unique feature such as vertical velocity variations during precipitation which cannot be obtained from any other measurements.

2. Data and Method of Analysis

2.1. LAWP

[5] We make use of the LAWP data from 20 May to 31 July during 1999 (with one day gap) and 1 May to 31 July during 2000 (with 2 days gap) to monitor the changes before and after the onset. The meteorological parameters used to study the evolution of the monsoon onset over Gadanki are horizontal and vertical winds, SNR from 0.3 km to 4.5 km. More details about LAWP are given by Narayana Rao et al. [2001]. During clear atmospheric conditions, the vertical velocity obtained from LAWP represents the clear air vertical velocity. Since LAWP (1.35 GHz) is very sensitive to precipitation, the vertical velocity thus obtained from zenith beam during precipitation conditions represents hydrometeor fall velocity and will normally show large downward values indicating the presence of rainfall.

2.2. Optical Rain Gauge and Other Supporting Data Sets

[6] To obtain the daily accumulated rainfall, co-located optical rain gauge (ORG) is used. It provides accurate measurement of precipitation with a dynamic range of 0.1–500 mm hr−1 [Narayana Rao et al., 2001]. The Outgoing Long wave Radiation (OLR) data from the National Oceanic and Atmospheric Administration ( is utilized to examine the convective activity over the study region. For comparison purpose an 850 hPa wind speed is also used from NCEP/NCAR reanalysis data set ( (for details, see Kalnay et al. [1996]).

[7] LAWP was operated continuously during each hour giving moments (Power/SNR, Doppler shift and Doppler width) and winds. Winds and SNR are first averaged for each hour and then for each day representing the resulting values for that day. Since LAWP is sensitive to precipitation, the averaging is done for clear and precipitation conditions separately. In the present study precipitation days are taken as the days on which vertical velocity exceeds 1 ms−1 and treated them as precipitation echoes [Narayan Rao et al., 1999]. Remaining days are treated as clear days which may include cloudy days with smaller vertical velocities.

3. Results

[8] The differential heating between Indian sub continent and the ocean surrounding it produces a circulation in the lower troposphere and results in cross equatorial flow blowing into Arabian Sea and developing into low level south westerlies. Strengthening of these westerlies over peninsular India is the important feature at the time of onset [Soman and Kumar, 1993]. This strong westerly flow has an important role in transporting moist air onto the Indian subcontinent [Pearce and Mohanty, 1984]. Normally, the onset of the Indian summer monsoon is around the last week of May over the southern tip of peninsular India i.e. Kerala coast of Arabian Sea.

[9] Before looking for the date of arrival of monsoon over Gadanki, it is important to understand the onset of monsoon over Arabian Sea as it gives us the basis for looking into the subsequent advancement across the country. Figure 1a shows the time series of daily wind speed at 850 hPa over Arabian Sea (5.0°N–15.0°N, 45.0°E–75.0°E) during May–June 1999 using NCEP reanalysis data [Kalnay et al., 1996]. The wind speed shows rapid strengthening of the low level wind exceeding the value of 8 ms−1 on 18 May 1999 which happens to be the onset of ISM over Arabian Sea [Taniguchi and Koike, 2006]. Extending the same criteria, similar analysis is done for the Gadanki region averaged between 12.5°N–15.0°N and 77.5°E–80.0°E. It is observed that the wind speed exceeds 8 ms−1 on 9 June 1999 and maintained for one week. This is the day on which the monsoon has arrived at Gadanki and surrounding regions (also according to India Meteorological Department, IMD). The wind speed over Gadanki from LAWP plotted in Figure 1 is well correlated with the wind speed obtained from NCEP reanalysis. It is interesting to note (Figure 1b) that relatively strong winds called as the low level westerly jet (LLJ) exceeding greater than 15 ms−1 are seen twice over Gadanki around 1.5 km with a break during 27 June to 8 July. The first LLJ coincides with the onset phase of the monsoon followed by a break (zonal wind decreased to less than 5 ms−1) and the second LLJ shows the active period/revival of the monsoon.

Figure 1.

(a) Daily wind speed at 850 hPa over Arabian Sea (5–15°N; 45–75°E) (red line) and Gadanki (12.5–15.0°N; 77.5–80°E) (blue line) during April–July 1999 using NCEP reanalysis data. Daily wind speed observed at 850 hPa by Gadanki LAWP (black line) is also plotted for comparison. The vertical lines shown at 18 May 1999 and 9 June 1999 on the x-axis indicate the onset days over Arabian Sea and Gadanki, respectively. Time-height sections of (b) zonal wind, (c) wind direction, and (d) range corrected SNR obtained from LAWP during 20 May to 31 July 1999. Note the different time scales for Figures 1a and 1b–1d.

[10] The progression of monsoon can be understood using OLR, which can be used as proxy for tropical deep convection. Figure 2a shows the latitudinal average of OLR (10°N–15°N) plotted day wise against longitude ranging from 60°E to 90°E. The probability of rainfall with 0.7–0.8 is associated with OLR values around 220 Wm−2 [Soman and Kumar, 1993], hence OLR less than 220 Wm−2 is considered as regions of convective activity. The deep convection started in the Arabian Sea around 18 May 1999 and further moved to Gadanki region around 9 June and maintained for few days showing the monsoon arrival. The observed un-weighted weekly accumulated rainfall using rain gauges over the region of Andhra Pradesh surrounding Gadanki is depicted in Figure 2b. Since daily point rainfall data may be highly variable, a weekly rainfall has been considered here. It is found that the increase in rainfall is sharp during the week of the arrival of monsoon and persisted for the next few days. The time series of wind speed reveals that the monsoon has arrived Gadanki on 9 June and is active for about 18 days. Thereafter a small break from 27 June to 8 July is observed, though it is active over Arabian Sea for nearly 12 days and again monsoon revived over the study region consistent with NCEP/NCAR reanalysis data set shown in Figure 1.

Figure 2.

(a) Longitude-time section of outgoing long-wave radiation (OLR) averaged between 10°N and 15°N latitudes observed from 20 May 1999 to 31 July 1999. The vertical line along 79.2°E indicates the longitude of the Gadanki region. OLR values less than 220 W/m2 indicates regions of tropical deep convection. (b) Weekly accumulated rainfall surrounding Gadanki region. (c) Same as Figure 2a but for 1 May 2000 to 31 July 2000.

[11] To identify the onset of monsoon over Gadanki, the features observed by LAWP with the evolution of the monsoon are presented. Figures 3a–3d shows the time series of wind speed and wind direction, SNR during clear conditions, clear air vertical velocity and hydrometeor fall velocity and rain fall during 1999 monsoon season around 1.5 km. The data are averaged for 3 range bins around 1.5 km for consistency corresponding to 850 hPa. The wind speed during the month of May is small and steady with magnitudes around 5 ms−1. At the beginning of June, the wind speed starts increasing and shows large enhancement exceeding 8 ms−1 on 9 June 1999 indicating the arrival of monsoon. The monsoon activity is maintained for few weeks up to 26 June 1999. Thereafter the wind speed starts decreasing showing break in monsoon and again monsoon becomes active around 9 July. The wind direction around 1.5 km (right vertical axis) shows change in magnitude exceeding 180° on the monsoon arrival date and persisted during the active period reaching maximum up to 270° showing the wind direction reversal from south easterlies to south westerlies. Although wind speed shows magnitudes exceeding 8 ms−1 on some days (for example, around 4 June), wind direction remains southeasterly. But, both wind speed and direction show enhancement on 9 June satisfying the criteria of 8 ms−1 and southwesterly direction. During active period, these southwesterlies are dominant throughout the altitude range (Figure 1c). The time series of SNR obtained from zenith beam during clear conditions (Figure 3b) shows enhancement just few days before the onset of monsoon (4 June) and continues to increase during the active monsoon and decrease during break monsoon. Appreciable changes in range corrected SNR (r2 × SNR) are seen mostly below 2.5 km during the monsoon period (Figure 1d). On normal days the r2 × SNR is between −5 dB to −2 dB below 2.5 km. During active period the r2 × SNR increases ranging from 5 dB to 10 dB. Overall an increase of nearly 10 dB is observed in SNR during monsoon. Interestingly around 27 May, a shift in wind direction (∼250°) indicating southwesterlies and enhancement in SNR of 10–15 dB is observed giving a sign of monsoon arrival. But these changes do not seem to be persistent for next few days. Moreover, wind speed shows no appreciable change in magnitude and remains below 8 ms−1. Therefore these changes can not be attributed to monsoon arrival.

Figure 3.

Time series of (a) wind speed and wind direction, (b) signal-to-noise ratio (SNR), (c) clear air vertical velocity (blue line), and hydrometeor fall velocity (red line, axis is on the right side) observed using LAWP averaged for 3 range bins around 1.5 km, and (d) rain fall from co-located optical rain gauge (ORG) during May–July 1999. The arrow on the x-axis in Figure 3a at 9 June 1999 represents the onset day based on inference deduced from IMD weather report on progress of monsoon.

[12] Figure 3c shows the vertical velocity during clear and precipitation conditions. The vertical velocity is calculated by averaging hourly vertical velocities during clear and precipitation periods separately during the 24 hour period of profiler data. As mentioned earlier, since LAWP is sensitive to precipitation, the Doppler spectrum of the zenith beam shows large positive Doppler shifts during the precipitation depending on the rainfall intensity. Therefore the large downward vertical velocities exceeding more than 1 ms−1 observed from LAWP represent the fall velocity of hydrometeors indicating the presence of rainfall indirectly. This can be clearly seen by comparing hydrometeor fall velocity with the occurrence of rainfall observed by co-located ORG (Figure 3d). Another notable feature is that, the hydrometeor fall velocity and rainfall occurrence coincide with the vigour of the monsoon with hydrometeor velocity reaching up to −5 to −6 ms−1. Thus, Doppler velocity from LAWP plays an important role in the study of monsoon activity without looking into the conventional rainfall parameter. Although rainfall intensity is important to study the monsoon activity, the present study utilizes the relation between Doppler velocity and rainfall qualitatively to identify the arrival of monsoon. The clear air vertical velocity shows a reversal in the direction from downward to upward sometime attaining magnitudes up to 0.8 ms−1. This reversal occurred few days before the onset day similar to that of SNR. Therefore, simultaneous and persistent variations in wind speed, direction, SNR and vertical velocity facilitate us in identifying monsoon onset.

[13] Figure 4 shows the time series of wind speed, SNR during clear conditions, vertical velocity and rainfall during 2000 monsoon season. Similar to the case of 1999, the wind speed shows enhancement at the time of onset on 31 May with southwesterly direction. The increase in SNR and vertical velocity reversal take place at the time of the onset of monsoon. The hydrometeor fall velocity and rainfall show high frequency of occurrence. This gives us an indication that monsoon is active for a longer period with few short breaks. On the other hand, a break in monsoon activity with longer period is seen during 1999. The mechanisms behind the observed features are explained in the following section.

Figure 4.

Same as Figure 3 but for the year 2000 case. The onset day is on 31 May 2000.

4. Discussion

[14] During the evolution of summer monsoon, the low level westerly wind increases giving rise to an increase in kinetic energy over the Arabian Sea [Krishnamurti et al., 1981]. This low level jet is related to onset and active phases of monsoon as noted earlier in this study. Further this jet is also linked to ISOs during monsoon period [Krishnan et al., 2005]. This wind transports moisture on to the continent [Pearce and Mohanty, 1984]. It is well known that the wind profiler gets the echoes from the refractive index gradients due to turbulent fluctuations. These fluctuations are closely coupled with humidity [Tsuda et al., 1988] in the lower troposphere. Thus, SNR observed from the profiler gives the indirect observation of moisture presence. During the progress of monsoon, more moisture is transported giving rise to increased refractive index fluctuations hence increase in the SNR. It is already noted that the increase in SNR took place few days before the onset of monsoon over Gadanki. This might be attributed to the build up of moisture a few days before the onset [Soman and Kumar, 1993]. This increase in the SNR at the time of onset is observed at all levels in the lower troposphere below 2 km.

[15] The clear air vertical velocity shows a transition from mean descent to ascent few days before or at the time of the onset of monsoon. This can be attributed to moisture convergence and rapid enhancement in the low level winds. The large downward vertical velocities exceeding 1 ms−1 indicate the occurrence of rainfall as the monsoon progresses over Gadanki. This can be visualized by comparing with daily accumulated rainfall measured by ORG (Figures 3 and 4).

5. Conclusions

[16] In this study, a systematic change in low level wind speed, wind direction, SNR, vertical velocity and rainfall are studied and the onset/arrival of monsoon over Gadanki was investigated during 1999 and 2000 using LAWP. Enhancement of the low level south westerly winds attaining 8 ms−1 at 1.5 km (850 hPa) with persistency during the next few days indicates the beginning of monsoon season over Gadanki. The strong low level monsoon winds are associated with convergence during strengthening of monsoon causing upward vertical velocity during onset and its progress. Hydrometeor velocity shows large downward velocities indicating occurrence of monsoon rainfall. Enhancement in SNR (zenith beam) indicates increased reflectivity which is thought to be due to increased moisture carried by monsoon winds. Thus, it is shown that changes occurring during the different phases of monsoon can be monitored using ultra-high frequency (UHF) radar at a particular station although monsoon activity is a synoptic phenomenon.


[17] We are grateful to NARL, Gadanki, IMD, ACAS (ISRO), and NOAA-CIRES for providing necessary data for this study. V. V. M. J. Rao is thankful to The Commissioner, Department of Technical Education, Government of A.P, Hyderabad, for permitting him to carry out this research. We sincerely thank T. Tsuda, Kyoto University for his suggestions in improving the present study. We are thankful to the anonymous reviewers for their useful suggestions.