- Email: [email protected]
Coupling between the lower and middle atmosphere observed during a very severe cyclonic storm ‘Madi’
Journal of Atmospheric and Solar-Terrestrial Physics xxx (2017) 1–13
Contents lists available at ScienceDirect
Journal of Atmospheric and Solar-Terrestrial Physics journal homepage: www.elsevier.com/locate/jastp
Coupling between the lower and middle atmosphere observed during a very severe cyclonic storm ‘Madi’ H. Hima Bindu a, M. Venkat Ratnam a, *, V. Yesubabu a, T. Narayana Rao a, S. Eswariah b, C.V. Naidu c, S. Vijaya Bhaskara Rao d a
National Atmospheric Research Laboratory [NARL], Gadanki, India Chungnam National University (CNU), Daejeon 305-764, South Korea Department of Meteorology and Oceanography, Andhra University, Visakhapatnam, India d Department of Physics, Sri Venkateswara University, Tirupati, India b c
A R T I C L E I N F O
A B S T R A C T
Keywords: Gravity waves Tides Tropical cyclone Coupling MLT
Synoptic-scale systems like cyclones can generate broad spectrum of waves, which propagate from its source to the middle atmosphere. Coupling between the lower and middle atmosphere over Tirupati (13.6 N, 79.4 E) is studied during a very severe cyclonic storm ‘Madi’ (06–13 December 2013) using Weather Research and Forecast (WRF) model assimilated fields and simultaneous meteor radar observations. Since high temporal and spatial measurements are difficult to obtain during these disturbances, WRF model simulations are obtained by assimilating conventional and satellite observations using 3DVAR technique. The obtained outputs are validated for their consistency in predicting cyclone track and vertical structure by comparing them with independent observations. The good agreement between the assimilated outputs and independent observations prompted us to use the model outputs to investigate the gravity waves (GWs) and tides over Tirupati. GWs with the periods 1–5 h are observed with clear downward phase propagation in the lower stratosphere. These upward propagating waves obtained from the model are also noticed in the meteor radar horizontal wind observations in the MLT region (70–110 km). Interestingly, enhancement in the tidal activity in both the zonal and meridional winds in the mesosphere and lower thermosphere (MLT) region is noticed during the peak cyclonic activity except the suppression of semi-diurnal tide in meridional wind. A very good agreement in the tidal activity is also observed in the horizontal winds in the troposphere and lower stratosphere from the WRF model outputs and ERA5. These results thus provide evidence on the vertical coupling of lower and middle atmosphere induced by the tropical cyclone.
1. Introduction Generation of atmospheric gravity waves (GWs) can be caused by a variety of processes including flow over topography, jet streams, frontal systems, and moist convection. GWs generated by these tropospheric sources may propagate vertically into the stratosphere and mesosphere, contributing to the momentum budget of the middle atmosphere (Fritts and Alexander, 2003). Convection is known to generate atmospheric waves of all scales from planetary-scale (Matsuno, 1966; Holton, 1972; Lindzen, 1974; Salby and Garcia, 1987; Bergman and Salby, 1994; Wheeler and Kiladis, 1999) to small-scale high-frequency gravity waves (Fovell et al., 1992; Dewan et al., 1998; Piani et al., 2000; Choi et al., 2006; Chun and Kim, 2008; Dutta et al., 2009). Both the observational
and theoretical studies suggest convection as the most important source of GWs in the tropics (McLandress et al., 2000; Alexander et al., 2000). A large number of observational studies, dating from as early as 1980's (Larsen et al., 1982), have suggested that short-scale convectively generated GWs do indeed play an important role in the tropical middle atmospheric dynamics. The deep convection during the tropical cyclones (TCs) triggers GWs, which can modulate the dynamics of the middle atmosphere (Fritts and Alexander, 2003) and also influences the large scale atmospheric circulation (Piani et al., 2000; Kim et al., 2003). It is well appreciated that vertically propagating short period GWs have a profound effect on the structure and general circulation of the middle atmosphere. They are also responsible for the coupling between different layers of the atmosphere.
* Corresponding author. E-mail address: [email protected] (M. Venkat Ratnam). https://doi.org/10.1016/j.jastp.2018.01.029 Received 9 October 2017; Received in revised form 24 January 2018; Accepted 27 January 2018 Available online xxxx 1364-6826/© 2018 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Hima Bindu, H., et al., Coupling between the lower and middle atmosphere observed during a very severe cyclonic storm ‘Madi’, Journal of Atmospheric and Solar-Terrestrial Physics (2017), https://doi.org/10.1016/j.jastp.2018.01.029
H. Hima Bindu et al.
Journal of Atmospheric and Solar-Terrestrial Physics xxx (2017) 1–13
successfully reproduces observed inertia-gravity waves in the stratosphere with vertical wavelengths less than 6 km during cyclone. WRF has been widely used in recent times as a regional model for the prediction of TCs and severe weather events in the Indian subcontinent, in particular in Bay of Bengal (BoB) region (Srinivas et al., 2010, 2012; Osuri et al., 2013). Recently, it has been shown that WRF simulations can be effectively used to investigate the Cyclone-generated GWs (Hima Bindu, 2016) over BoB, supporting the previous observational and numerical studies on characteristics of TC-induced GWs (Chane- Ming et al., 2002, 2010) and TC activity in the Upper Troposphere and Lower Stratosphere (UTLS) (Ibrahim et al., 2010). However, to the best of our knowledge so far no study has been made to investigate the coupling between lower and middle atmosphere during the cyclone activity as the top boundary of the WRF inputs are at the best 1 hPa. To overcome this limitation, in the present study we have combined WRF simulations covering troposphere and stratosphere and meteor radar observations to cover the MLT region. The main objective of the present study is to better understand the vertical coupling between lower and middle atmosphere during a very severe cyclonic storm (VSCS) Madi. We have adopted reanalysis approach to generate atmospheric fields using the regional model WRF as the reanalysis can provide a better representation of the atmosphere state through the process of assimilating available observations in downscaled simulations of mesoscale model.
The large-scale dynamics of the mesosphere and lower thermosphere (MLT) are dominated by solar atmospheric tides generated mainly due to asymmetry in the tropospheric water vapour and the ozone in the stratosphere. As cyclones are the zones of very high water vapour, they can generate the tidal activity. Amplitudes of these tides increase with altitude and become comparable to the background winds in the MLT region, where they deposit their momentum to the background winds (Lieberman and Hays, 1994). It is now well understood that the thermal structure and dynamics of the MLT and its variability can be determined to a significant degree by studying the large- and small-scale waves propagating into this region from the lower atmosphere (Fritts and Alexander, 2003; Lieberman et al., 2004; Ratnam et al., 2008) and cyclones are believed to be one of the major sources. The major limitation for the coupling studies between lower and middle atmosphere during the cyclones is the non-availability of suitable data sets. As these processes happen over the oceanic region, only during their landfall that are close to the observational site, where there are continuous observations of major meteorological parameters, may only help in delineating the characteristics of waves and tides. Conventionally global model products are used for this purpose, which often show large deviations from limited observations because of model's coarse resolution and inadequate representation of mesoscale processes. The regional mesoscale model products (Evan et al., 2012) with high spatial resolution and better assimilation of observations are able to simulate GW activity and provide valuable addition to complement the ground-based measurements. There are several works (Kim et al., 2009, 2014; Kim and Chun, 2010a,b; Kim and Chun, 2011) related to typhoon-generated gravity waves using Weather Research and Forecast (WRF) model simulations. Kim et al. (2010a,b) successfully reproduced the stratospheric GWs generated by Typhoon 'Ewiniar' using simulations of Weather Research Forecasting model (WRF) and they showed that WRF can simulate the precise location and structure of the GWs as seen in AIRS observations. Using regional WRF model, Evan et al. (2012) studied the impact of GWs on Quasi-Biennial Oscillation. They also showed that WRF
2. Overview of the tropical cyclone Madi (2013) In this section, we briefly mention the history of the cyclone Madi that occurred in 2013. A low pressure area from South China Sea moved across Malay Peninsula and emerged into south Andaman Sea on 1 December 2013 morning. Moving westwards it lay over southeast BoB on 2 December. Continuing its westwards movement, it lay over southwest BoB off Sri Lanka coast on 3 December. It persisted over the same region and became well developed on 4 December. It further strengthened into a depression in the morning of 6 December over southwest BoB and lay
Fig. 1. Model domains used in ARW for the cyclone simulations. The shading gives the terrain height in meters. Inner triangle shows the location of Gadanki/ Tirupati.
2
H. Hima Bindu et al.
Journal of Atmospheric and Solar-Terrestrial Physics xxx (2017) 1–13
centered at 10 N and 84 E, about 350 km northeast of Tricomalee (Sri Lanka). The depression remained practically stationary, as it lay close to the upper troposphere ridge which ran along 10 N. It led to very slow northward movement afterwards. The deep depression further intensified into a cyclonic storm ‘Madi’ with centre near 10.5 N and 84 E at 0 000 UTC of 7 December. It intensified into a severe cyclonic storm over the same region at 0 900 UTC of 7 December. As it lay slightly to the north of the ridge, the severe cyclonic storm then moved slightly northnortheastward and intensified into a very severe cyclonic storm at 0 600 UTC of 8December near 12.3 N and 84.7 E. It gradually weakened into a cyclonic storm on 10 December at 21UTC and into a depression on 11 December at 1800 UTC and crossed Tamil Nadu coast close to Vedaranyam around 1 330 UTC of 12 December. It then emerged into Palk Strait at 1 500 UTC, moved west southwards and again crossed Tamil Nadu coast near Tondi around 1700 UTC of 12 December. It continued to move west-southwest across south peninsula and weakened further into a well-marked low pressure area over southeast Andaman Sea and adjoining Kerala at 0 000 UTC of 13 December 2013.
JðxÞ ¼ J b þ J 0 ¼
T 1 1 x xb B1 x xb þ ðy yo ÞT ðE þ FÞ1 ðy yo Þ; 2 2 (1)
where x is the analysis state, xb the background, y the observation space [y ¼ Hx], Hthe observation operator,y o the observation, and B, E and F are the background, observation (instrumental) and representativity error covariance matrices, respectively. Out of the three covariance matrices, the background covariance, B, plays a vital role in finding out the 3DVAR solution (Guo et al., 2005). In the present study, B is estimated using the standard National Meteorological Centre (NMC) method (Parrish and Derber, 1992). First to calculate B, a month long series of WRF 24-h forecasts are generated during November 2011, with each simulation initialized at 0 000 UTC and 1 200 UTC using the ERA-I data for both initial and boundary conditions(6-hourly). The forecast differences between 12 and 24 h, valid over the same time, are used as model perturbations to estimate B using NMC method. The observational data used for assimilation in the study are operationally available quality controlled global observations in BUFR format (Prepbufr) from the NCEP Atmospheric Data Project (ADP) archives along with COSMIC GPS radio occultation (GPSRO) profiles. The Prepbufr data comprise of land surface (Synoptic stations, Metar, and AWS), marine surface (Ship and, buoys), upper-air soundings (pilot balloon and radiosonde) and satellite observations such as wind vectors from scatterometers (QSCAT, OSAT and ASCAT) and atmospheric motion vectors from geostationary satellites. The advantage of this data set is that they can be directly assimilated in the WRFDA system and contains observational error information specific to the observational source. The model is initialized at 0 000 UTC on 06 December, model 6hourly forecast as first guest observations from NCEP ADP and GPSRO are assimilated cyclically till 0 000 UTC of 13 December 2013 in 6-hourly interval using 3-DVAR approach to enhance the quality of the downscaled initial conditions. Model simulations were sampled at every 10 min to extract GW parameters.
3. Data and methodology A high-resolution Advanced Research WRF (ARW) model (Version 3.8; Skamarock et al., 2008) and WRF Data Assimilation (WRFDA) package are used to investigate the vertical coupling of lower and middle atmosphere through high frequency GWs and tides during the cyclone period. Brief description of the model configuration, assimilation methodology and data used for model comparison are given in the following sub-sections. 3.1. WRF model and its configuration The model domain and physics configuration used in this study are adopted from Srinivas et al. (2012) for TC predictions over BoB. The model is configured with two-way interactive nested two domains as shown in Fig. 1 with coarse domain covering a larger region of north Indian Ocean with 18 km grid spacing and the inner domain covering the BoB and its neighbourhood with 6 km grid spacing. A total of 120 vertical levels are used with the model top at 1 hPa. The model physics chosen for both 18 and 6 km domains consists of New Thompson scheme (Thompson et al., 2008) for cloud microphysics, a rapid radiation transfer model (RRTM) for long-wave radiation (Mlawer et al., 1997), a RRTM for shortwave radiation, the Yonsei University non-local diffusion (Hong et al., 2006) for boundary layer turbulence, the new Kain–Fritsch mass flux scheme (Kain, 2004) for cumulus convection and the NOAH scheme for land surface processes (Chen and Dudhia, 2001). No explicit GW parameterization is used to represent the wave breaking. ERA-interim (ERA-I) data from ECMWF are used to initialize the model and boundary conditions are updated at 6-hourly intervals. To improve the representation of land-sea contrasts over coastal regions and to supply the realistic ocean feedback in the model fields, the coarse-resolution sea surface temperature (SST) in the ERA-I data are replaced with time-varying high-resolution SST data obtained from Real-Time Global High-Resolution (RTG-HR) (Gemmill et al., 2007).
3.3. Observational data For the validation of simulated track and intensity parameters from the WRF assimilation, the best track provided by India Meteorological Department (IMD) is used as the reference. We have used independent Gadanki (13.5 N, 79.2 E) GPS radiosonde observations and COSMIC GPSRO (covering troposphere and stratosphere) to compare the model outputs and to check the consistency of the model with the observations. We also used Sri Venkateswara University Meteor Radar (SVUMR) (Rao et al., 2014) observations located at 35 km away from Gadanki in Tirupati to investigate the waves and tides in the MLT region during the cyclone. We have also used hourly analysis fields of ERA5 data available at a horizontal resolution of 31 31 km with 137 vertical levels from the surface up to 0.01 hPa (around 80 km). ERA5 is the fifth generation of atmospheric reanalysis produced at ECMWF, disseminated through Copernicus Climate Change Service. The details of ERA5 reanalysis is partly descript by Weger (2015), and Malardel et al., (2015). 4. Results and discussion
3.2. Assimilation methodology
4.1. Validation of simulated TC Madi track
The three-dimensional variational assimilation (3DVAR) incorporated into WRFDA (Barker et al., 2004) is employed for this study. These variational assimilation methods can efficiently combine observations from heterogeneous platforms with model first guess while satisfying model dynamics and accounting for various uncertainties arising both from observations and model. In brief, 3DVAR is a most computationally efficient method and it can provide an “optimal” estimate of the true atmospheric state produced at any desired analysis time through iterative solution of a cost-function (Ide et al., 1997; Barker et al., 2004).
The simulated track is compared with best track data prepared by IMD and is shown in Fig. 2a. The WRF assimilated track matched well with the IMD observed track both in the time and the position and the movement of the storm throughout the period of the cyclone similar to that reported by Rajasree et al. (2013). The simulated track deviated from the reference track by 30 km before the landfall and coincided with the observed track at landfall. Fig. 2b and c shows the time series of mean sea level pressure (MSLP, hPa) and maximum sustained wind speed (m/s), respectively. The simulated MSLP and wind speed are initially in 3
H. Hima Bindu et al.
Journal of Atmospheric and Solar-Terrestrial Physics xxx (2017) 1–13
Fig. 2. (a)Madi cyclone best track estimates provided by IMD (black line) and the track obtained from WRF 3D-Var assimilation (red line). Time series of (b)mean sea level pressure (hPa) and (c)maximum sustained winds (m/s) obtained from the IMD (black line) and WRF assimations (red line). (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)
UTC are shown in Fig. 3a–c, respectively. Good correspondence, though assimilated outputs show slight underestimation, can be noticed in all the three components. The Root Mean Square Error (RMSE) between the radiosonde and assimilated outputs for such 6 comparisons is shown in Fig. 3d–f. RMSE is close to 0.5 K in the mid and upper troposphere where it reaches as high as 2 K in the lower troposphere and stratosphere. Zonal and meridional wind shows the RMSE of 2–4 m/s except near the tropopause where it reaches as high as 7 m/s (5 m/s in the meridional). Thus, observed parameters matched well with the model outputs except near the tropopause. The FORMOSAT-3/COSMIC (FORMOSA Satellite Mission-3/ Constellation Observing System for Meteorology, Ionosphere, and Climate) data set from the COSMIC Data Analysis and Archive Centre (CDAAC) is obtained to further check the consistency between the assimilated outputs of temperature. Note that COSMIC GPS RO consists of daily soundings of temperature over near Gadanki/Tirupati latitude during 11–12 December 2013. Vertical profiles have high accuracy for temperature (<1 K from 5 to 25 km) but vary in vertical resolution from 100 m at the surface to 1.5 km at 35 km altitude (Pirscher et al., 2010; Anthes, 2011). GPS RO data observed by low-Earth-orbit satellites have been used previously for GWs observations in the lower stratosphere (Tsuda et al., 2000; Liou et al., 2003, 2006; Ratnam et al., 2004). In the present study, profiles of temperature covering complete Indian region (Fig. 4a) (WRF model domain 1), where occultations are available (total 51 occultations are available) are taken for verification with the
good agreement with those reported by IMD during the development and mature stage of the TC, but simulated MSLP underestimates the observed estimate at 18 h and after that it overestimated. Whereas the simulated wind speed is overestimated by 7 m/s for 24 h and then underestimated for 48 h before finally matches well with that of observed during the landfall of the TC Madi. 4.2. Independent verification of the model outputs Assimilated profiles that are obtained from the WRF model outputs represented near realistic features of the background atmosphere and are also used to characterize the GWs (HimaBindu, 2016). During the cyclone Madi period, balloon flights were carried out regularly with high resolution GPS radiosondes (Meisei, Japan, RD-06G) from Gadanki during 06 December to 13 December at around 1730 IST (IST ¼ UTC þ 0530 h). These sondes provided profiles of temperature and horizontal wind up to a mean altitude of ~30 km with ~5 m height resolution. Radiosonde data quality and accuracy are described in Venkat Ratnam et al. (2014) and also used to characterize GWs in the UTLS region over Gadanki (Leena et al., 2012; Pramitha et al., 2015). Temperature and horizontal wind profiles observed by GPS radiosonde launched from Gadanki during the cyclone period (06–13 December 2013) are compared with the assimilated model simulations. A typical comparison between the radiosonde and assimilated profiles of temperature, zonal and meridional winds obtained on 10 December 2013 at around 1 200 4
H. Hima Bindu et al.
Journal of Atmospheric and Solar-Terrestrial Physics xxx (2017) 1–13
Fig. 3. Profiles of (a) temperature, (b) zonal wind and (c) meridional wind observed by GPS radiosonde launched from Gadanki [13.5 N,79.2 E] on 10 December 2013 around 17:30 IST. WRF model assimilated profiles are also superimposed in the respective panels. Root mean square error (RMSE)profiles between GPS radiosonde and assimilated outputs observed in (d)temperature, (e) zonal wind and (f) meridional wind.
Fig. 4. (a) The number of COSMIC GPS Radio Occultations during cyclone ‘Madi’ over Indian region. (b)Profile of temperature (black line) observed by COSMIC GPSRO at 12.8768 N, 79.3743 E on 11 December at 02UTC. Simulated profile of temperature obtained from the 3DVAR assimilation (red line) over Gadanki [13.5 N,79.2 E] is also superimposed. (c)Root mean square error observed between the GPSRO profiles and WRF assimilated outputs. Note that comparisons were made above 10 km due to the ambiguity in GPS RO temperature below 10 km. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)
to use them to identify the GW characteristics over Gadanki/Tirupati latitude during the cyclone ‘Madi’ (06–13 December 2013).
assimilated model outputs. Profiles of temperature obtained from COSMIC GPSRO at 12.8768 N, 79.3743 E on 11 December at 02UTC and the 3DVAR assimilation over Gadanki (13.5 N,79.2 E) shows a clear match between them (Fig. 4b).The RMSE between the COSMIC GPSRO and assimilated outputs shown in Fig. 4c is between 0 and 2.5 K from 10 km altitude to the lower stratosphere. Since dry temperature is assumed, profiles of temperature from COSMIC GPS RO are not valid below 10–12 km. Close match above 10 km (less RMSE) of the assimilated outputs with the independent observations in the UTLS region prompted
4.3. Waves generated by a tropical cyclone ‘Madi’ Different waves are believed to be generated locally with periodicities from minutes to few hours during TCs. For this study, vertical velocity is used rather than the horizontal wind components, as previous studies indicated that it is the best observational parameter for extracting the 5
H. Hima Bindu et al.
Journal of Atmospheric and Solar-Terrestrial Physics xxx (2017) 1–13
Fig. 5. Snapshots of vertical velocities observed at 24 km altitude during a VSCS Madi cyclone on (a) 07 Dec. 2013 at 12:30UTC, (b) 08 Dec. 2013 at 18:30UTC, (c) 09 Dec. 2013 at 01UTC, and (d) 09 Dec. 2013 at 14UTC.
except strong northward wind around 15 km altitude persisting throughout the observational period. There exist strong shears around 15 km in the meridional wind which can generate GWs and propagate upwards. There is a chance of mixing of the waves generated due to convection and wind shear. Vertical wind shows updrafts and downdrafts with some episodic enhancements during the cyclone which probably due to the presence of short period GWs associated with cumulus convection (Sato et al., 1995; Kumar, 2006). Background temperature does not show significant variability. Time series of these winds and temperature are subjected to wavelet analysis in order to know the dominant periodicities of these waves. A typical example of wavelet spectral analysis applied for zonal, meridional and vertical winds at 20 km altitude is shown in Fig. 7a–c, respectively. The presence of high frequency waves having periods ranging from 1 to 5 h in the vertical wind component are observed clearly rather than the zonal and meridional components over Tirupati/Gadanki. It is interesting to notice the existence of the high frequency waves in the vertical wind velocity component, starting from half an hour onwards. The above analysis is made at different altitudes and the features are found to be similar. However, low frequency waves are also observed over Tirupati/ Gadanki at lower altitudes but not at higher altitudes and perhaps they might have been filtered by the background winds.
convection induced GWs (Choi et al., 2006; HimaBindu, 2016). Fig. 5 shows the snapshots of vertical velocities observed during a VSCS Madi cyclone obtained from the 3DVAR at 24 km altitude. Inner triangle within the contour shows the location of Tirupati/Gadanki, cyclone track is shown with dotted line and the centre of the cyclone for a given time shown on the top of each panel is marked as a solid filled circle. A systematic updrafts and downdrafts propagating away from the cyclone centre with the motion of the cyclone are clearly observed from this figure. These updrafts and downdrafts reach the Tirupati/Gadanki location quite often particularly as shown in Fig. 5c and d. As short period waves are believed to be generated locally, strong wave activity is seen almost during the entire period of the cyclone Madi irrespective of its intensity. The spatial extent of these waves at 24 km altitude can be seen up to 1000 km from the cyclone centre and they propagate in all the directions from the centre of the cyclone. Similar features are noticed at different altitudes such as 12, 14, 18, and 20 km during the cyclone period from 06 to 13 Dec. 2013 (Figure not shown). Since the main objective of the present study is to investigate the vertical coupling of the lower and middle atmosphere during the tropical cyclone Madi, first we characterize these waves over Tirupati. The background zonal wind, meridional wind, vertical wind and temperature profiles during the period of the cyclone over Tirupati/Gadanki have been extracted from the WRF assimilated data. Fig. 6a–d shows the time - height section of zonal, meridional, vertical wind velocities and temperature over Tirupati/Gadanki, respectively, from 0.5 to 30 km during 06–13 December 2013 (during the cyclone) obtained from WRF 3DVAR assimilated output. Zonal winds show alternating eastward and westward flow in the troposphere with strong westward around 16 km altitude particularly during latter half of the observations and changes to eastward above 20 km. In general, meridional winds are southward in the troposphere
4.4. Vertical coupling of lower and middle atmosphere during cyclone The winds obtained from WRF 3DVAR assimilated outputs are further subjected to 1–5 h band pass filter.Fig. 8(a–c) shows the time series of 1–5 h band pass filtered zonal, meridional and vertical wind over Tirupati/Gadanki during Madi cyclone. Clear downward phase propagation is noticed with alternating updrafts and downdrafts with magnitudes 6
H. Hima Bindu et al.
Journal of Atmospheric and Solar-Terrestrial Physics xxx (2017) 1–13
Fig. 6. Time – altitude variation of (a) zonal wind, (b)meridional wind, (c)vertical wind and (d)temperature over Gadanki (13.5 N, 79.2 E) during Madi cyclone obtained from WRF 3DVAR assimilated outputs.
Fig. 7. Wavelet analysis of (a) zonal wind (b) meridional wind and (c) vertical wind at 20 km altitude obtained from the 3DVAR assimilation.
range from 50 to 150 km with dominant horizontal wavelengths of 100 km. Since convective sources produce a broad spectrum of waves, the location of an instrument site (Gadanki/Tirupati) relative to the source can heavily influence the spectrum of waves seen. Fig. 8d-f shows the time-altitude section of 1–5 h band pass filtered variations observed
reaching as high as 0.1 m/s. FFT in the vertical domain revealed the vertical wavelengths ranging from 4 to 16 km with dominant vertical wavelengths of 8 km. The horizontal wavelengths during the cyclone over Tirupati/Gadanki are estimated by the dispersion relation using observed periods and vertical wavelengths. The horizontal wave lengths
7
H. Hima Bindu et al.
Journal of Atmospheric and Solar-Terrestrial Physics xxx (2017) 1–13
Fig. 8. Time-altitude section of 1–5 h band pass filtered variations observed in (a) zonal, (b) meridional and (c) vertical winds over Gadanki (13.5 N, 79.2 E) during Madi cyclone obtained from WRF 3DVAR assimilated outputs. (d), (e) and (f) Same as (a), (b) and (c) but over 13.45 N and 88 E.
in zonal, meridional and vertical wind over 13.45 N and 88 E (same as that Gadanki/Tirupati but at different longitude) during Madi cyclone obtained from WRF 3DVAR assimilated outputs. Dominant downward phase propagation is noticed with alternating updrafts and downdrafts with magnitudes reaching as high as 0.1 m/s. All these characteristics suggest these waves are of high frequency GWs. Thus, it is again clear that WRF assimilated outputs can be used to investigate the cyclone generated GWs. The advanced SVU meteor radar located at Tirupati is specially designed to provide continuous high-resolution wind measurements
round the clock for observing both small and large scale oscillations in the MLT region (70–110 km). It is advanced meteor radar due to its high meteor count rate sensitivity, configured by ATRAD, Australia. The SVU MR provides an opportunity to elucidate the vertical and lateral coupling of the MLT region (Rao et al., 2014). Horizontal wind profiles observed by SVU meteor radar during the cyclone period (06–13 Dec. 2013) are subjected to FFT analysis. Waves with periods 1–7 h with dominant period 2–5 h is noticed in both zonal and meridional winds (figure not shown). Fig. 9a-b shows the time series of 1–5 h band pass filtered zonal and meridional winds over Tirupati during Madi cyclone. Clear 8
H. Hima Bindu et al.
Journal of Atmospheric and Solar-Terrestrial Physics xxx (2017) 1–13
Fig. 9. Time-altitude section of 1–5 h band pass filtered variations observed in (a) zonal and (b) meridional winds over Tirupati during Madi cyclone obtained from SVU metoer radar. (c) and (d) same as (a) and (b) but during normal weather condition (without cyclone days).
fitting in order to evaluate the amplitude and phase of each component. A 3-day data window, sliding forward by 1day each time, is used for extraction of the tide. The resulting amplitude and phases of the tide are attributed to the middle day. Each component of the wind can be expressed in terms of the prevailing and tidal components, as follows:
downward phase propagation is noticed with magnitudes reaching as high as 20 m/s. FFT in the vertical domain revealed the vertical wavelengths ranging from 4 to 16 km with dominant vertical wavelengths of 16 km. The horizontal wavelengths during the cyclone over Tirupati/Gadanki are estimated by the dispersion relation using observed periods and vertical wavelengths. The horizontal wavelengths range from 5 to 150 km with dominant horizontal wavelength of 50 km. All these characteristics suggest that these waves are of high frequency GWs. Since high frequency waves are thought to be fairly ubiquitous in the MLT region, the SVU MR data without cyclone activity has been taken for the similar analysis to check whether these waves are generated by the cyclone or not. The time-altitude section of 1–5 h band pass filtered variations observed in zonal, meridional winds over Tirupati during normal weather (no cyclone activity) obtained from SVU meteor radar shown in Fig. 9(c–d) reveals no significant wave activity during the non-cyclone days. Observed amplitudes are below 10 m/s during non-cyclonic periods. Thus, the upward propagating waves obtained from the model outputs shown in Fig. 8 are again noticed in the SVU MR horizontal wind observations (from 70 to 110 km altitude) which showed a clear propagation of GWs into the mesosphere. Thus, it is clear that the imprints of cyclone generated GWs reach as high as MLT region. As mentioned earlier, the zonal and meridional winds from SVU MR are obtained with 2 km and 1 h resolutions from 70 to 110 km. In order to investigate the behavior of the background tides during cyclone period, the hourly averaged zonal and meridional winds were subjected to Fourier analysis. The diurnal (24 h), semi-diurnal (12 h) and terr-diurnal (8 h) tidal oscillations are clearly noticed (figure not shown). The hourly mean winds recorded by the SVU MR are further subjected to harmonic
AðtÞ ¼ Ao þ Ak cos
2π k ðt ϕÞ Tp
[2]
where t is the local time, A[t] is the parameter [U and V] to be fitted, Ao is the prevailing component, A and φ are the amplitude and phase of the tidal, and Tp ¼ 24 (e.g. Venkateswara Rao et al., 2011). The tidal parameters are individually estimated for U and V. Similar analysis is followed for WRF assimilated outputs to know the tidal activity in the troposphere and stratospheric altitudes. Tidal amplitudes observed in zonal wind and meridional wind from WRF assimilation and SVU Meteor radar during 3–19 Dec. 2013 is shown in Fig. 10. Enhancement in the tidal activity in the MLT region during the peak cyclone activity is clearly noticed in the zonal component of SVU meteor radar. However, semi-diurnal component is suppressed in the meridional component. In the troposphere and lower stratosphere, enhancement in the diurnal and semi-diurnal tides is seen at little later times in both diurnal and semi-diurnal components. No significant terrdiurnal amplitudes are noticed in both zonal and meridional components in the troposphere and lower stratosphere. This shows the clear evidence of vertical coupling between the lower and middle atmosphere during a VSCS Madi even due to tides. However, note that there is a big gap between 30 and 70 km and we do not have any observations at these
9
H. Hima Bindu et al.
Journal of Atmospheric and Solar-Terrestrial Physics xxx (2017) 1–13
Fig. 10. Time-altitude sections of (a) diurnal, (b) semi-diurnal and (c) ter-diurnal tides observed in the zonal component at MLT region using SVUMR. (d)–(f) same as (a)–(c) but observed in the TLS region using WRF assimlated outputs. (g)–(i) same as (a)–(c) but for the meridional component. (j)–(l) same as (d)–(f) but for the meridional component.
run with 2-way nested two domains with horizontal resolutions of 18 km and 6 km. But all the results are analyzed from the high resolution 6 km domain with the temporal resolutions of 10 min. The assimilated model outputs are verified first with the IMD observed track to check the model consistency (Fig. 2). The WRF model outputs are further compared with the independent Gadanki GPS radiosonde temperature and horizontal winds profiles (Fig. 3), which are available during the cyclone period and also with COSMIC GPSRO profiles (Fig. 4). A good comparison is found between the WRF assimilated outputs and the independent observations. Further, the snapshots of vertical wind component at different times in the lower stratosphere from the WRF model showed a clear evidence of GWs over Gadanki region during TC Madi (Fig. 5). The dominant periods of these waves are found to be 1–5 h (Fig. 7) and have the dominant vertical wavelengths of 8 km with the horizontal wavelengths of 100 km. All the three components of winds (zonal, meridional and vertical) showed clear downward phase propagation (upward propagating waves) in the upper troposphere and lower stratosphere during the cyclone (Fig. 8). All these characteristics reveal that the high frequency GWs is generated during the VSCS Madi. These upward propagating waves obtained from the model are again noticed in the SVU MR (Tirupati) horizontal wind observations in the MLT region (70–110 km) (Fig. 9).Thus, evidence of coupling between lower and middle atmosphere exists through waves generated during the VSCS Madi is shown. It can be further confirmed with ray tracing of the gravity waves, however, getting background information which is very much essential for accurate implementation of ray tracing is almost impossible during disturbed events like cyclones. Enhancement in the tidal activity in both zonal and meridional components during the peak stage of cyclone is also noticed both in the MLT region using SVU meteor radar and troposphere and lower stratosphere using WRF assimilated outputs (Fig. 10). Interestingly, suppression of semi-diurnal tide in the meridional component is noticed. No
altitudes. Thus, above mentioned link between lower and middle atmosphere through waves and tides remains as qualitative study. To further look into this aspect, we make use of recently released ERA 5 reanalysis data sets. As ECMWF ERA5 data is quite recent, we have validated the ERA5 data products with the independent GPS radiosonde observations to check the consistency of reanalysis data. Fig. 11 shows the profiles of zonal (top panel), meridional winds (middle panel) and temperature (bottom panel) during Cyclone period (07–13 Dec 2013). A good agreement is observed between ERA5 and radiosonde profiles. With this confidence on ERA5 products, to cover the gap between 30 and 70 km the similar analysis (Tidal extraction) procedure is applied for this (ERA5) data over Gadanki/Tirupati and obtained all the three tidal components for both zonal and meridional winds. Fig. 12 shows the time-altitude sections of diurnal, semi-diurnal and terr-diurnal tidal components in zonal and meridional components during entire December 2013 month covering both cyclone days and non-cyclonic days. More or less similar features observed in the SVU meteor radar i.e, enhancement in diurnal tide and suppression in semi-diurnal tide and no significant amplitudes in the terr-diurnal tide during cyclone period are again noticed even in ERA5 data sets between 40 and 60 km particularly in meridional component. It is clear that significant changes in the tidal activity during cyclone period exist in lower and middle atmosphere when compared to non-cyclonic periods. 5. Summary and conclusions This paper presents the vertical coupling of lower and middle atmosphere by gravity waves (GWs) and tides over Gadanki (13.5 N, 79.2 E)/ Tirupati(13.6 N, 79.4 E) during the evolution of an intense tropical cyclone Madi (2013) using Weather Research and Forecast (WRF) version 3.8.1 model simulations using three dimensional variational assimilation technique and SVU Meteor Radar observations. The model is 10
H. Hima Bindu et al.
Journal of Atmospheric and Solar-Terrestrial Physics xxx (2017) 1–13
Fig. 11. Profiles of Zonal wind (top panel), Meridional wind (middle panel) and Temperature (bottom panel) from ERA5 reanalysis data products and GPS Radiosonde observations over Gadanki during the period of Cyclone ‘Madi’ (07–13 Dec 2013).
Fig. 12. Time-altitude sections of diurnal (left panle), semi-diurnal (middle panel) and ter-diurnal (right panel) tides observed in the ERA-5 reanalysi data sets of zonal (left panels) and meridonal winds (right panles) during December 2013.
Enhancement in the diurnal tide, suppression in the semi-diurnal tide and no significant amplitudes in the terr-diurnal tide during cyclone activity is observed particularly in the meridional wind. There are few interesting observations being obtained from the present study. Strong updrafts and downdrafts surrounding the cyclone are observed that are mainly due to the cyclone generated cloud bands which
significant amplitudes in the terr-diurnal component in the lower atmosphere are noticed though significant amplitudes do exists in the MLT region. Thus, coupling between lower and middle atmosphere through tides is also possible during the cyclone activity. Similar features are also observed (Fig. 12) in the recently released ERA 5 reanalysis data sets covering the gap between WRF outputs and SVU meteor radar.
11
H. Hima Bindu et al.
Journal of Atmospheric and Solar-Terrestrial Physics xxx (2017) 1–13
is the source for the observed GW activity. These GWs propagate in all the directions and those reached over Gadanki/Tirupati are being observed closely. Thus, strong GW activity in the troposphere and lower stratosphere is not throughout the cyclone period but observed as episodic over the Gadanki/Tirupati. Since these high frequency GWs prefer to propagate vertically than in the horizontal direction, enhanced GW activity is observed throughout the cyclone period in the MLT region over Tirupati. Further, stronger amplitudes in the meridional component are observed as the zonal component may be filtered due to strong shears present at 30 km altitude (Fig. 6). Thus, one to one relation between the wave activity in the lower and middle atmosphere may not present mainly due to wave-mean flow interaction. Vertically propagating GWs generated by cyclone (convection) may provide an important contribution to the momentum budget of the middle atmosphere as suggested by Song et al., 2003. Another interesting feature noticed is the enhancement in the tidal activity during the cyclone period which also has source in the water vapour distribution. Tidal amplitudes in the meridional component is almost double than that observed in the zonal component. In continuation to our earlier work (HimaBindu, 2016) and this work, a set of experiments are planned in NARL and SVU covering the atmosphere from surface up to 110 km to investigate more on the coupling between the lower and middle atmosphere during cyclone passage. Future studies will also include detailed description of source mechanisms for convectively generated GWs through severe cyclones.
Dewan, E., Coauthors, 1998. MSX satellite observations of thunderstorm-generated gravity waves in mid-wave infrared images of the upper stratosphere. Geophys. Res. Lett. 25, 939–942. Dutta, G., Kumar, M.C.A., Kumar, P.V., Ratnam, M.V., Chandrashekar, M., Shibagaki, Y., Salauddin, M., Basha, H.A., 2009. Characteristics of high-frequency gravity waves generated by tropical deep convection: case studies. J. Geophys. Res. 114, D18109 https://doi.org/10.1029/2008JD011332. Evan, S., Alexander, M.J., Dudhia, J., 2012. WRF simulations of convectively generated gravity waves in opposite QBO phases. J. Geophys. Res. 117, D12117 https:// doi.org/10.1029/2011JD017302. Fovell, R., Durran, D., Holton, J.R., 1992. Numerical simulations of convectively generated stratospheric gravity waves. J. Atmos. Sci.49 1427–1442. Fritts, D.C., Alexander, M.J., 2003. Gravity wave dynamics and effects in the middle atmosphere. Rev. Geophys. 41 (1), 1003. https://doi.org/10.1029/2001RG000106. Gemmill, W., Katz, B., Xu, L., 2007. Daily Real-time, Global Sea Surface Temperature High-resolution analysis.NCEP/EMC off. Note 260, 39 pp., Environ. Modell. Cent. Camp Springs, Md. Guo, Y.R., Kusaka, H., Barker, D.M., Kuo, Y.H., Crook, A., 2005. Impact of ground based GPS PW and MM5 3DVar background error statistics on forecast of a convective case. Sci. Online Lett. Atmos 1, 73–76. https://doi.org/10.2151/sola.2005-020. HimaBindu, H., Venkat Ratnam, M., Yesubabu, V., Narayana Rao, T., Amit Kesarkar, Naidu, C.V., 2016. Characteristics of Cyclone generated Gravity waves observed using assimilated WRF model simulations over Bay of Bengal. Atmos. Res. 180, 178–188. Holton, J.R., 1972. Waves in the equatorial stratosphere generated by tropospheric heat sources. J. Atmos. Sci. 29,368–375. Hong, S.Y., Noh, Y., Dudhia, J., 2006. A new vertical diffusion package with explicit treatment of entrainment processes. Mon. Weather Rev. 134, 2318–2341. Ibrahim, C., Chane Ming, F., Barthe, C., Kuleshov, Y., 2010. Diagnosis of tropical cyclone activity through gravity wave energy density in the Southwest Indian Ocean. Geophys. Res. Lett. 37, L09807 https://doi.org/10.1029/2010GL042938. Ide, K., Courtier, P., Ghil, M., Lorenc, A.C., 1997. Unified notation for data assimilation: operational, sequential and variational. J. Math. Soc. Jpn. 75, 181–189. Kain, J.S., 2004. The Kain-Fritsch convective parameterization: an update. J. Appl. Meteor 43, 170–181. Kim, Y.-J., Eckermann, S.D., Chun, H.-Y., 2003. An overview of the past, present and future of gravity-wave drag parameterization for numerical climate and weather prediction models. Survey Article. Atmos.-Ocean 41 (1), 65–98. https://doi.org/ 10.3137/ao.410105. Kim, S.-Y., Chun, H.-Y., Wu, D.L., 2009. A study on stratospheric gravity waves generated by Typhoon Ewiniar: numerical simulations and satellite observations. J. Geophys. Res. 114, D22104 https://doi.org/10.1029/2009JD011971. Kim, S.Y., Chun, H.Y., 2010a. Momentum flux of stratospheric gravity waves generated by Typhoon Ewiniar 2006. Asia-Pac. J. Atmos. Sci. 46, 199–208. Kim, S.-Y., Chun, H.-Y., 2010b. Stratospheric gravity waves generated by Typhoon Saomai (2006): numerical modeling in a moving frame following the typhoon. J. Atmos. Sci. 67, 3617–3636. https://doi.org/10.1175/2010JAS3374.1. Kim, S.-Y., Chun, H.-Y., 2011. Impact of typhoon-generated gravity waves in the typhoon development. Geophys. Res. Lett. 38, L01806 https://doi.org/10.1029/ 2010GL045719. Kim, S.H., Chun, H.-Y., Jang, W., 2014. Horizontal divergence of typhoon-generated gravity waves in the upper troposphere and lower stratosphere (UTLS) and its influence on typhoon evolution. Atmos. Chem. Phys. 14, 3175–3182. https://doi. org/10.5194/acp-14-3175-2014. Kumar, K.K., 2006. VHF radar observations of convectively generated gravity waves: some new insights. Geophys. Res. Lett. 33, L01815 https://doi.org/10.1029/ 2005GL024109. Larsen, M., Swartz, W., Woodman, R., 1982. Gravity-wave generation by thunderstorms observed with a vertically pointing 430 MHz radar. Geophys. Res. Lett. 9, 571–574. Leena, P.P., Ratnam, M.V., Krishna Murthy, B.V., Vijaya Bhaskara Rao, S., 2012. Detection of high frequency gravity waves using high resolution radiosonde observations. J. Atmos. Sol. Terr. Phys 254–259. Lieberman, R.S., Hays, P.B., 1994. An estimate of the momentum deposition in the lower thermosphere by the observed diurnal tide. J. Atmos. Sci. 51, 3094–3105. Lieberman, R.S., Oberheide, J., Hagan, M.E., Remsberg, E.E., Gordley, L.L., 2004. Variability of diurnal tides and planetary waves during November 1978–May 1979. J. Atmos. Sol. Terr. Phys. 66 (6–9), 517–528. https://doi.org/10.1016/ j.jastp.2004.01.006. Lindzen, R.S., 1974. Wave-CISK in the tropics. J. Atmos. Sci. 31, 156–179. Liou, Y.A., Pavelyev, A.G., Huang, C.Y., Igarashi, K., Hocke, K., Yan, S.K., 2003. Analytic method for observation of the gravity waves using radio occultation data. Geophys. Res. Lett. 30, 2021. https://doi.org/10.1029/2003GL017818. Liou, Y.A., Pavelyev, A.G., Wicker, J., Liu, S.F., Pavelyev, A.A., Schmidt, T., Igarashi, K., 2006. Application of GPS radio occultation method for observation of the internal waves in the atmosphere. J. Geophys. Res. 111, D06104 https://doi.org/10.1029/ 2005JD005823. Malardel, S., Wedi, N., Deconinck, W., Diamantakis, M., Kuhnlein, C., Mozdzynski, G., Hamrud, M., Smolarkiewicz, P., 2015. A New Grid for the IFS. ECMWF, p. 6. Newsletter No. 146-Winter 2015/16. Matsuno, T., 1966. Quasi-geostrophic motions in the equatorial area. J. Meteorol. Soc. Jpn. 44, 25–43. McLandress, C., Alexander, M.J., Wu, D.L., 2000. Microwave Limb Sounder observations of gravity waves in the stratosphere: a climatology and interpretation. J. Geophys. Res. 105, 11,947–11,967. https://doi.org/10.1029/2000JD900097.
Acknowledgements The authors wish to thank the Sri Venkateswara University Meteor Radar [SVU MR] for providing necessary data for the present study. We are grateful to the National Atmospheric Research Laboratory [NARL] for providing GPS radiosonde data used in the present study. We also thank the India Meteorological Department [IMD] for the best track data through their website. ERA-Interim data from ECMWF are used to initialize the model and boundary conditions. The prepbufr global observational dataset are obtained from http://rda.ucara.edu. The data used in the present study can be obtained on request. Appendix A. Supplementary data Supplementary data related to this article can be found at https://doi. org/10.1016/j.jastp.2018.01.029. References Alexander, M.J., Beres, J.H., Pfister, L., 2000. Tropical stratospheric gravity wave activity and relationship to clouds. J. Geophys. Res. 105, 22,299–22,309. https://doi.org/ 10.1029/2000JD900326. Anthes, R.A., 2011. Exploring Earth's atmosphere with radio occultation: contributions to weather, climate and space weather. Atmos. Meas. Tech 4, 1077–1103. https:// doi.org/10.5194/amt-4-1077-2011. Barker, D.M., Huang, W., Guo, Y.R., Xiao, Q.N., 2004. A three dimensional (3-D-Var) data assimilation system for use with MM5: implementation and initial results. Mon. Weather Rev. 132, 897–914. Bergman, J.W., Salby, M.L., 1994. Equatorial wave activity derived from fluctuations in observed convection. J. Atmos.Sci 51, 3791–3806. Chane-Ming, F., Roff, G., Robert, L., Leveau, J., 2002. Gravity wave characteristics overTromelin Island during the passage of cyclone Hudah. Geophys. Res. Lett. 29 (6), 1094. https://doi.org/10.1029/2001GL013286. Chane-Ming, F., Chen, Z., Roux, F., 2010. Analysis of gravity-waves produced by intense tropical cyclones. Ann. Geophys. 28, 531–547. https://doi.org/10.5194/angeo-28531-2010. Chen, F., Dudhia, J., 2001. Coupling an advanced land surface/hydrology model with thePenn State/NCAR MM5 modeling system, Part I: model description and implementation. Mon. Weather Rev. 129, 569–585. Choi, Y.G., Lee, S.C., McDonald, A.J., Hooper, D.A., 2006. Wind-profilerobservations of gravity waves produced by convection at mid-latitudes. Atmos. Chem. Phys. 6, 2825–2836. https://doi.org/10.5194/acp-6-2825-2006. Chun, H.-Y., Kim, Y.-H., 2008. Secondary waves generated by breaking of convective gravity waves in the mesosphere and their influence in the wave momentum flux. J. Geophys. Res. 113, D23107 https://doi.org/10.1029/2008JD009792.
12
H. Hima Bindu et al.
Journal of Atmospheric and Solar-Terrestrial Physics xxx (2017) 1–13 Sato, K., Hashiguchi, H., Fukao, S., 1995. Gravity waves and turbulence associated with the cumulus convection observed with the UHF/VHF clear-air Doppler radars. J. Geophys. Res. 100, 7111–7119. Skamarock, W.C., Klemp, J.B., Dudhia, J., Gill, D.O., Barker, D.M., Dudha, M.G., Huang, X., Wang, W., Powers, Y., 2008. A Description of the Advanced Research WRF Ver.30.NCARTechnicalNote, NCAR/TN-475CSTR, Mesoscale and Microscale Meteorology Davison. National Centre for Atmospheric Research, Boulder, CO, USA, p. 113. Song, I.S., Chun, H.Y., Lane, T.P., 2003. Generation mechanisms of convectively forced internal gravity waves and their propagation to the stratosphere. J. Atmos. Sci. 60, 1960–1980. Srinivas, C.V., Yesubabu, V., Venkatesan, R., Ramarkrishna, S.S.V.S., 2010. Impact of assimilation of conventional and satellite meteorological observations on the numerical simulation of a Bay of Bengal Tropical Cyclone of Nov 2008 near Tamilnadu using WRF model. Meteorol. Atmos. Phys. 110, 19–44. Srinivas, C.V., Yesubabu, Hari Prasad, K.B.R.R.V., Venkatraman, B., Ramakrishna, S.S.V.S., 2012. Numerical simulation of Cyclonic Storms FANOOS, NARGIS with assimilation of conventional and satellite observations using 3-D-VAR. Nat. Hazards 63 (2), 867–889. Thompson, G., Field, P.R., Rasmussen, R.M., Hall, W.D., 2008. Explicit forecasts of winter precipitation using an improved bulk microphysics scheme. Part II: implementation of a new snow parameterization. Mon. Weather Rev. 163, 5095–5114. Tsuda, T., Nishida, M., Rocken, C., Ware, R.H., 2000. A global morphology of gravity wave activity in the stratosphere revealed by the GPS occultation data (GPS/MET). J. Geophys. Res. 105, 7257–7273. Venkat Ratnam, M., Pravallika, N., Ravindra babu, S., Ghouse Basha,Pramitha, M., Krishna Murthy, B.V., 2014. Assessment of GPS radiosonde descent data. Atmospheric Measurement Techniques 7, 1011–1025. Venkateswara Rao, N., Tsuda, T., Gurubaran, S., Miyoshi, Y., Fujiwara, H., 2011. On the occurrence and variability of the terdiurnal tide in the equatorial mesosphere and lower thermosphere and a comparison with the Kyushu-GCMJ. Geophys. Rep. 116, D02117 https://doi.org/10.1029/2010JD014529. Weger, I., 2015. In: “Preliminary Comparison of ECMWF ERA-interim and ERA-5 Reanalyses” on Core-Climax Workshop, Brussels, January 15-16, 2015, and ECMWF – Computing and Forecast System, ICAS 2015, Annecy, France, September 2015. Wheeler, M., Kiladis, G.N., 1999. Convectively coupled equatorial waves: analysis of clouds and temperature in the wavenumber– frequency domain. J. Atmos. Sci. 56, 374–399.
Mlawer, E.J., Taubman, S.J., Brown, P.D., Iacono, M.J., Clough, S.A., 1997. Radiative transfer for inhomogeneous atmosphere: RRTM, a validated correlated-k model for the long wave. J. Geophys. Res. 102, 16663–16682. Osuri, K.K., Mohanty, U.C., Routray, A., Mohapatra, M., 2013. Real-time track prediction of tropical cyclones over the north Indian Ocean using the ARW model. J. Appl. Meteorol. Climatol 52, 2476–2492. https://doi.org/10.1175/JAMC-D-12-0313.1. Parrish, D.F., Derber, J.C., 1992. The national meteorological Centre's spectral statistical interpolation analysis system. Mon. Weather Rev. 120, 1747–1763. Piani, C., Durran, D., Alexander, M.J., Holton, J.R., 2000. A numerical study of three dimensional GWs waves triggered by deep tropical convection and their role in the dynamics of the QBO. J. Atmos. Sci. 57, 3689–3702. Pirscher, B., Foelsche, U., Borsche, M., Kirchengast, G., Kuo, Y.H., 2010. Analysis of migrating diurnal tides detected in FORMOSAT-3/COSMIC temperature data. J. Geophys. Res. 115, D14108 https://doi.org/10.1029/2009JD013008. Pramitha, M., Venkat Ratnam, M., Leena, P.P., Krishna Murthy, B.V., Vijaya Bhaskar Rao, S., 2015. Identification of Inertia Gravity Wave sources observed in the troposphere and the lower stratosphere over a tropical station Gadanki. Atmos. Res. 176–177 [2016] 202–211. Rajasree, V.P.M., Kesarkar, A.P., Bhate, J.N., Umakanth, U., Singh, V., Varma, T.H., 2013. 2016. Appraisal of recent theories to understand cyclogenesis pathways of tropical cyclone Madi. J. Geophys. Res. Atmos. 121, 8949–8982. https://doi.org/10.1002/ 2016JD025188. Rao, S.V.B., Eswaraiah, S., VenkatRatnam, M., Kosalendra, E., Kishore Kumar, K., Sathish Kumar, S., Patil, P.T., Gurubaran, S., 2014. Advanced meteor radar installed at Tirupati: system details and comparison with different radars. J. Geophys. Res. Atmos. 119, 11,893–11,904. https://doi.org/10.1002/2014JD021781. Ratnam, M.V., Kishore Kumar, G., Krishna Murthy, B.V., Patra, A.K., Jagannadha Rao, V.V.M., Vijaya Bhaskar Rao, S., Kishore Kumar, K., Ramkumar, G., 2008. Longterm variability of the low latitude mesospheric SAO and QBO and their relation with stratospheric QBO. Geophys. Res. Lett. 35, L21809 https://doi.org/10.1029/ 2008GL035390. Ratnam, M.V., Tetzlaff, G., Jacobi, C., 2004. Global and seasonal variations of stratospheric gravity wave activity deduced from the CHAMP/GPS satellite. J. Atmos. Sci. 61 (13), 1610–1620. Salby, M.L., Garcia, R.R., 1987. Transient response to localized episodic heating in thetropics. Part I: excitation and short-time near-field behaviour. J. Atmos. Sci. 44,458–498.
13
Contents lists available at ScienceDirect
Journal of Atmospheric and Solar-Terrestrial Physics journal homepage: www.elsevier.com/locate/jastp
Coupling between the lower and middle atmosphere observed during a very severe cyclonic storm ‘Madi’ H. Hima Bindu a, M. Venkat Ratnam a, *, V. Yesubabu a, T. Narayana Rao a, S. Eswariah b, C.V. Naidu c, S. Vijaya Bhaskara Rao d a
National Atmospheric Research Laboratory [NARL], Gadanki, India Chungnam National University (CNU), Daejeon 305-764, South Korea Department of Meteorology and Oceanography, Andhra University, Visakhapatnam, India d Department of Physics, Sri Venkateswara University, Tirupati, India b c
A R T I C L E I N F O
A B S T R A C T
Keywords: Gravity waves Tides Tropical cyclone Coupling MLT
Synoptic-scale systems like cyclones can generate broad spectrum of waves, which propagate from its source to the middle atmosphere. Coupling between the lower and middle atmosphere over Tirupati (13.6 N, 79.4 E) is studied during a very severe cyclonic storm ‘Madi’ (06–13 December 2013) using Weather Research and Forecast (WRF) model assimilated fields and simultaneous meteor radar observations. Since high temporal and spatial measurements are difficult to obtain during these disturbances, WRF model simulations are obtained by assimilating conventional and satellite observations using 3DVAR technique. The obtained outputs are validated for their consistency in predicting cyclone track and vertical structure by comparing them with independent observations. The good agreement between the assimilated outputs and independent observations prompted us to use the model outputs to investigate the gravity waves (GWs) and tides over Tirupati. GWs with the periods 1–5 h are observed with clear downward phase propagation in the lower stratosphere. These upward propagating waves obtained from the model are also noticed in the meteor radar horizontal wind observations in the MLT region (70–110 km). Interestingly, enhancement in the tidal activity in both the zonal and meridional winds in the mesosphere and lower thermosphere (MLT) region is noticed during the peak cyclonic activity except the suppression of semi-diurnal tide in meridional wind. A very good agreement in the tidal activity is also observed in the horizontal winds in the troposphere and lower stratosphere from the WRF model outputs and ERA5. These results thus provide evidence on the vertical coupling of lower and middle atmosphere induced by the tropical cyclone.
1. Introduction Generation of atmospheric gravity waves (GWs) can be caused by a variety of processes including flow over topography, jet streams, frontal systems, and moist convection. GWs generated by these tropospheric sources may propagate vertically into the stratosphere and mesosphere, contributing to the momentum budget of the middle atmosphere (Fritts and Alexander, 2003). Convection is known to generate atmospheric waves of all scales from planetary-scale (Matsuno, 1966; Holton, 1972; Lindzen, 1974; Salby and Garcia, 1987; Bergman and Salby, 1994; Wheeler and Kiladis, 1999) to small-scale high-frequency gravity waves (Fovell et al., 1992; Dewan et al., 1998; Piani et al., 2000; Choi et al., 2006; Chun and Kim, 2008; Dutta et al., 2009). Both the observational
and theoretical studies suggest convection as the most important source of GWs in the tropics (McLandress et al., 2000; Alexander et al., 2000). A large number of observational studies, dating from as early as 1980's (Larsen et al., 1982), have suggested that short-scale convectively generated GWs do indeed play an important role in the tropical middle atmospheric dynamics. The deep convection during the tropical cyclones (TCs) triggers GWs, which can modulate the dynamics of the middle atmosphere (Fritts and Alexander, 2003) and also influences the large scale atmospheric circulation (Piani et al., 2000; Kim et al., 2003). It is well appreciated that vertically propagating short period GWs have a profound effect on the structure and general circulation of the middle atmosphere. They are also responsible for the coupling between different layers of the atmosphere.
* Corresponding author. E-mail address: [email protected] (M. Venkat Ratnam). https://doi.org/10.1016/j.jastp.2018.01.029 Received 9 October 2017; Received in revised form 24 January 2018; Accepted 27 January 2018 Available online xxxx 1364-6826/© 2018 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Hima Bindu, H., et al., Coupling between the lower and middle atmosphere observed during a very severe cyclonic storm ‘Madi’, Journal of Atmospheric and Solar-Terrestrial Physics (2017), https://doi.org/10.1016/j.jastp.2018.01.029
H. Hima Bindu et al.
Journal of Atmospheric and Solar-Terrestrial Physics xxx (2017) 1–13
successfully reproduces observed inertia-gravity waves in the stratosphere with vertical wavelengths less than 6 km during cyclone. WRF has been widely used in recent times as a regional model for the prediction of TCs and severe weather events in the Indian subcontinent, in particular in Bay of Bengal (BoB) region (Srinivas et al., 2010, 2012; Osuri et al., 2013). Recently, it has been shown that WRF simulations can be effectively used to investigate the Cyclone-generated GWs (Hima Bindu, 2016) over BoB, supporting the previous observational and numerical studies on characteristics of TC-induced GWs (Chane- Ming et al., 2002, 2010) and TC activity in the Upper Troposphere and Lower Stratosphere (UTLS) (Ibrahim et al., 2010). However, to the best of our knowledge so far no study has been made to investigate the coupling between lower and middle atmosphere during the cyclone activity as the top boundary of the WRF inputs are at the best 1 hPa. To overcome this limitation, in the present study we have combined WRF simulations covering troposphere and stratosphere and meteor radar observations to cover the MLT region. The main objective of the present study is to better understand the vertical coupling between lower and middle atmosphere during a very severe cyclonic storm (VSCS) Madi. We have adopted reanalysis approach to generate atmospheric fields using the regional model WRF as the reanalysis can provide a better representation of the atmosphere state through the process of assimilating available observations in downscaled simulations of mesoscale model.
The large-scale dynamics of the mesosphere and lower thermosphere (MLT) are dominated by solar atmospheric tides generated mainly due to asymmetry in the tropospheric water vapour and the ozone in the stratosphere. As cyclones are the zones of very high water vapour, they can generate the tidal activity. Amplitudes of these tides increase with altitude and become comparable to the background winds in the MLT region, where they deposit their momentum to the background winds (Lieberman and Hays, 1994). It is now well understood that the thermal structure and dynamics of the MLT and its variability can be determined to a significant degree by studying the large- and small-scale waves propagating into this region from the lower atmosphere (Fritts and Alexander, 2003; Lieberman et al., 2004; Ratnam et al., 2008) and cyclones are believed to be one of the major sources. The major limitation for the coupling studies between lower and middle atmosphere during the cyclones is the non-availability of suitable data sets. As these processes happen over the oceanic region, only during their landfall that are close to the observational site, where there are continuous observations of major meteorological parameters, may only help in delineating the characteristics of waves and tides. Conventionally global model products are used for this purpose, which often show large deviations from limited observations because of model's coarse resolution and inadequate representation of mesoscale processes. The regional mesoscale model products (Evan et al., 2012) with high spatial resolution and better assimilation of observations are able to simulate GW activity and provide valuable addition to complement the ground-based measurements. There are several works (Kim et al., 2009, 2014; Kim and Chun, 2010a,b; Kim and Chun, 2011) related to typhoon-generated gravity waves using Weather Research and Forecast (WRF) model simulations. Kim et al. (2010a,b) successfully reproduced the stratospheric GWs generated by Typhoon 'Ewiniar' using simulations of Weather Research Forecasting model (WRF) and they showed that WRF can simulate the precise location and structure of the GWs as seen in AIRS observations. Using regional WRF model, Evan et al. (2012) studied the impact of GWs on Quasi-Biennial Oscillation. They also showed that WRF
2. Overview of the tropical cyclone Madi (2013) In this section, we briefly mention the history of the cyclone Madi that occurred in 2013. A low pressure area from South China Sea moved across Malay Peninsula and emerged into south Andaman Sea on 1 December 2013 morning. Moving westwards it lay over southeast BoB on 2 December. Continuing its westwards movement, it lay over southwest BoB off Sri Lanka coast on 3 December. It persisted over the same region and became well developed on 4 December. It further strengthened into a depression in the morning of 6 December over southwest BoB and lay
Fig. 1. Model domains used in ARW for the cyclone simulations. The shading gives the terrain height in meters. Inner triangle shows the location of Gadanki/ Tirupati.
2
H. Hima Bindu et al.
Journal of Atmospheric and Solar-Terrestrial Physics xxx (2017) 1–13
centered at 10 N and 84 E, about 350 km northeast of Tricomalee (Sri Lanka). The depression remained practically stationary, as it lay close to the upper troposphere ridge which ran along 10 N. It led to very slow northward movement afterwards. The deep depression further intensified into a cyclonic storm ‘Madi’ with centre near 10.5 N and 84 E at 0 000 UTC of 7 December. It intensified into a severe cyclonic storm over the same region at 0 900 UTC of 7 December. As it lay slightly to the north of the ridge, the severe cyclonic storm then moved slightly northnortheastward and intensified into a very severe cyclonic storm at 0 600 UTC of 8December near 12.3 N and 84.7 E. It gradually weakened into a cyclonic storm on 10 December at 21UTC and into a depression on 11 December at 1800 UTC and crossed Tamil Nadu coast close to Vedaranyam around 1 330 UTC of 12 December. It then emerged into Palk Strait at 1 500 UTC, moved west southwards and again crossed Tamil Nadu coast near Tondi around 1700 UTC of 12 December. It continued to move west-southwest across south peninsula and weakened further into a well-marked low pressure area over southeast Andaman Sea and adjoining Kerala at 0 000 UTC of 13 December 2013.
JðxÞ ¼ J b þ J 0 ¼
T 1 1 x xb B1 x xb þ ðy yo ÞT ðE þ FÞ1 ðy yo Þ; 2 2 (1)
where x is the analysis state, xb the background, y the observation space [y ¼ Hx], Hthe observation operator,y o the observation, and B, E and F are the background, observation (instrumental) and representativity error covariance matrices, respectively. Out of the three covariance matrices, the background covariance, B, plays a vital role in finding out the 3DVAR solution (Guo et al., 2005). In the present study, B is estimated using the standard National Meteorological Centre (NMC) method (Parrish and Derber, 1992). First to calculate B, a month long series of WRF 24-h forecasts are generated during November 2011, with each simulation initialized at 0 000 UTC and 1 200 UTC using the ERA-I data for both initial and boundary conditions(6-hourly). The forecast differences between 12 and 24 h, valid over the same time, are used as model perturbations to estimate B using NMC method. The observational data used for assimilation in the study are operationally available quality controlled global observations in BUFR format (Prepbufr) from the NCEP Atmospheric Data Project (ADP) archives along with COSMIC GPS radio occultation (GPSRO) profiles. The Prepbufr data comprise of land surface (Synoptic stations, Metar, and AWS), marine surface (Ship and, buoys), upper-air soundings (pilot balloon and radiosonde) and satellite observations such as wind vectors from scatterometers (QSCAT, OSAT and ASCAT) and atmospheric motion vectors from geostationary satellites. The advantage of this data set is that they can be directly assimilated in the WRFDA system and contains observational error information specific to the observational source. The model is initialized at 0 000 UTC on 06 December, model 6hourly forecast as first guest observations from NCEP ADP and GPSRO are assimilated cyclically till 0 000 UTC of 13 December 2013 in 6-hourly interval using 3-DVAR approach to enhance the quality of the downscaled initial conditions. Model simulations were sampled at every 10 min to extract GW parameters.
3. Data and methodology A high-resolution Advanced Research WRF (ARW) model (Version 3.8; Skamarock et al., 2008) and WRF Data Assimilation (WRFDA) package are used to investigate the vertical coupling of lower and middle atmosphere through high frequency GWs and tides during the cyclone period. Brief description of the model configuration, assimilation methodology and data used for model comparison are given in the following sub-sections. 3.1. WRF model and its configuration The model domain and physics configuration used in this study are adopted from Srinivas et al. (2012) for TC predictions over BoB. The model is configured with two-way interactive nested two domains as shown in Fig. 1 with coarse domain covering a larger region of north Indian Ocean with 18 km grid spacing and the inner domain covering the BoB and its neighbourhood with 6 km grid spacing. A total of 120 vertical levels are used with the model top at 1 hPa. The model physics chosen for both 18 and 6 km domains consists of New Thompson scheme (Thompson et al., 2008) for cloud microphysics, a rapid radiation transfer model (RRTM) for long-wave radiation (Mlawer et al., 1997), a RRTM for shortwave radiation, the Yonsei University non-local diffusion (Hong et al., 2006) for boundary layer turbulence, the new Kain–Fritsch mass flux scheme (Kain, 2004) for cumulus convection and the NOAH scheme for land surface processes (Chen and Dudhia, 2001). No explicit GW parameterization is used to represent the wave breaking. ERA-interim (ERA-I) data from ECMWF are used to initialize the model and boundary conditions are updated at 6-hourly intervals. To improve the representation of land-sea contrasts over coastal regions and to supply the realistic ocean feedback in the model fields, the coarse-resolution sea surface temperature (SST) in the ERA-I data are replaced with time-varying high-resolution SST data obtained from Real-Time Global High-Resolution (RTG-HR) (Gemmill et al., 2007).
3.3. Observational data For the validation of simulated track and intensity parameters from the WRF assimilation, the best track provided by India Meteorological Department (IMD) is used as the reference. We have used independent Gadanki (13.5 N, 79.2 E) GPS radiosonde observations and COSMIC GPSRO (covering troposphere and stratosphere) to compare the model outputs and to check the consistency of the model with the observations. We also used Sri Venkateswara University Meteor Radar (SVUMR) (Rao et al., 2014) observations located at 35 km away from Gadanki in Tirupati to investigate the waves and tides in the MLT region during the cyclone. We have also used hourly analysis fields of ERA5 data available at a horizontal resolution of 31 31 km with 137 vertical levels from the surface up to 0.01 hPa (around 80 km). ERA5 is the fifth generation of atmospheric reanalysis produced at ECMWF, disseminated through Copernicus Climate Change Service. The details of ERA5 reanalysis is partly descript by Weger (2015), and Malardel et al., (2015). 4. Results and discussion
3.2. Assimilation methodology
4.1. Validation of simulated TC Madi track
The three-dimensional variational assimilation (3DVAR) incorporated into WRFDA (Barker et al., 2004) is employed for this study. These variational assimilation methods can efficiently combine observations from heterogeneous platforms with model first guess while satisfying model dynamics and accounting for various uncertainties arising both from observations and model. In brief, 3DVAR is a most computationally efficient method and it can provide an “optimal” estimate of the true atmospheric state produced at any desired analysis time through iterative solution of a cost-function (Ide et al., 1997; Barker et al., 2004).
The simulated track is compared with best track data prepared by IMD and is shown in Fig. 2a. The WRF assimilated track matched well with the IMD observed track both in the time and the position and the movement of the storm throughout the period of the cyclone similar to that reported by Rajasree et al. (2013). The simulated track deviated from the reference track by 30 km before the landfall and coincided with the observed track at landfall. Fig. 2b and c shows the time series of mean sea level pressure (MSLP, hPa) and maximum sustained wind speed (m/s), respectively. The simulated MSLP and wind speed are initially in 3
H. Hima Bindu et al.
Journal of Atmospheric and Solar-Terrestrial Physics xxx (2017) 1–13
Fig. 2. (a)Madi cyclone best track estimates provided by IMD (black line) and the track obtained from WRF 3D-Var assimilation (red line). Time series of (b)mean sea level pressure (hPa) and (c)maximum sustained winds (m/s) obtained from the IMD (black line) and WRF assimations (red line). (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)
UTC are shown in Fig. 3a–c, respectively. Good correspondence, though assimilated outputs show slight underestimation, can be noticed in all the three components. The Root Mean Square Error (RMSE) between the radiosonde and assimilated outputs for such 6 comparisons is shown in Fig. 3d–f. RMSE is close to 0.5 K in the mid and upper troposphere where it reaches as high as 2 K in the lower troposphere and stratosphere. Zonal and meridional wind shows the RMSE of 2–4 m/s except near the tropopause where it reaches as high as 7 m/s (5 m/s in the meridional). Thus, observed parameters matched well with the model outputs except near the tropopause. The FORMOSAT-3/COSMIC (FORMOSA Satellite Mission-3/ Constellation Observing System for Meteorology, Ionosphere, and Climate) data set from the COSMIC Data Analysis and Archive Centre (CDAAC) is obtained to further check the consistency between the assimilated outputs of temperature. Note that COSMIC GPS RO consists of daily soundings of temperature over near Gadanki/Tirupati latitude during 11–12 December 2013. Vertical profiles have high accuracy for temperature (<1 K from 5 to 25 km) but vary in vertical resolution from 100 m at the surface to 1.5 km at 35 km altitude (Pirscher et al., 2010; Anthes, 2011). GPS RO data observed by low-Earth-orbit satellites have been used previously for GWs observations in the lower stratosphere (Tsuda et al., 2000; Liou et al., 2003, 2006; Ratnam et al., 2004). In the present study, profiles of temperature covering complete Indian region (Fig. 4a) (WRF model domain 1), where occultations are available (total 51 occultations are available) are taken for verification with the
good agreement with those reported by IMD during the development and mature stage of the TC, but simulated MSLP underestimates the observed estimate at 18 h and after that it overestimated. Whereas the simulated wind speed is overestimated by 7 m/s for 24 h and then underestimated for 48 h before finally matches well with that of observed during the landfall of the TC Madi. 4.2. Independent verification of the model outputs Assimilated profiles that are obtained from the WRF model outputs represented near realistic features of the background atmosphere and are also used to characterize the GWs (HimaBindu, 2016). During the cyclone Madi period, balloon flights were carried out regularly with high resolution GPS radiosondes (Meisei, Japan, RD-06G) from Gadanki during 06 December to 13 December at around 1730 IST (IST ¼ UTC þ 0530 h). These sondes provided profiles of temperature and horizontal wind up to a mean altitude of ~30 km with ~5 m height resolution. Radiosonde data quality and accuracy are described in Venkat Ratnam et al. (2014) and also used to characterize GWs in the UTLS region over Gadanki (Leena et al., 2012; Pramitha et al., 2015). Temperature and horizontal wind profiles observed by GPS radiosonde launched from Gadanki during the cyclone period (06–13 December 2013) are compared with the assimilated model simulations. A typical comparison between the radiosonde and assimilated profiles of temperature, zonal and meridional winds obtained on 10 December 2013 at around 1 200 4
H. Hima Bindu et al.
Journal of Atmospheric and Solar-Terrestrial Physics xxx (2017) 1–13
Fig. 3. Profiles of (a) temperature, (b) zonal wind and (c) meridional wind observed by GPS radiosonde launched from Gadanki [13.5 N,79.2 E] on 10 December 2013 around 17:30 IST. WRF model assimilated profiles are also superimposed in the respective panels. Root mean square error (RMSE)profiles between GPS radiosonde and assimilated outputs observed in (d)temperature, (e) zonal wind and (f) meridional wind.
Fig. 4. (a) The number of COSMIC GPS Radio Occultations during cyclone ‘Madi’ over Indian region. (b)Profile of temperature (black line) observed by COSMIC GPSRO at 12.8768 N, 79.3743 E on 11 December at 02UTC. Simulated profile of temperature obtained from the 3DVAR assimilation (red line) over Gadanki [13.5 N,79.2 E] is also superimposed. (c)Root mean square error observed between the GPSRO profiles and WRF assimilated outputs. Note that comparisons were made above 10 km due to the ambiguity in GPS RO temperature below 10 km. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)
to use them to identify the GW characteristics over Gadanki/Tirupati latitude during the cyclone ‘Madi’ (06–13 December 2013).
assimilated model outputs. Profiles of temperature obtained from COSMIC GPSRO at 12.8768 N, 79.3743 E on 11 December at 02UTC and the 3DVAR assimilation over Gadanki (13.5 N,79.2 E) shows a clear match between them (Fig. 4b).The RMSE between the COSMIC GPSRO and assimilated outputs shown in Fig. 4c is between 0 and 2.5 K from 10 km altitude to the lower stratosphere. Since dry temperature is assumed, profiles of temperature from COSMIC GPS RO are not valid below 10–12 km. Close match above 10 km (less RMSE) of the assimilated outputs with the independent observations in the UTLS region prompted
4.3. Waves generated by a tropical cyclone ‘Madi’ Different waves are believed to be generated locally with periodicities from minutes to few hours during TCs. For this study, vertical velocity is used rather than the horizontal wind components, as previous studies indicated that it is the best observational parameter for extracting the 5
H. Hima Bindu et al.
Journal of Atmospheric and Solar-Terrestrial Physics xxx (2017) 1–13
Fig. 5. Snapshots of vertical velocities observed at 24 km altitude during a VSCS Madi cyclone on (a) 07 Dec. 2013 at 12:30UTC, (b) 08 Dec. 2013 at 18:30UTC, (c) 09 Dec. 2013 at 01UTC, and (d) 09 Dec. 2013 at 14UTC.
except strong northward wind around 15 km altitude persisting throughout the observational period. There exist strong shears around 15 km in the meridional wind which can generate GWs and propagate upwards. There is a chance of mixing of the waves generated due to convection and wind shear. Vertical wind shows updrafts and downdrafts with some episodic enhancements during the cyclone which probably due to the presence of short period GWs associated with cumulus convection (Sato et al., 1995; Kumar, 2006). Background temperature does not show significant variability. Time series of these winds and temperature are subjected to wavelet analysis in order to know the dominant periodicities of these waves. A typical example of wavelet spectral analysis applied for zonal, meridional and vertical winds at 20 km altitude is shown in Fig. 7a–c, respectively. The presence of high frequency waves having periods ranging from 1 to 5 h in the vertical wind component are observed clearly rather than the zonal and meridional components over Tirupati/Gadanki. It is interesting to notice the existence of the high frequency waves in the vertical wind velocity component, starting from half an hour onwards. The above analysis is made at different altitudes and the features are found to be similar. However, low frequency waves are also observed over Tirupati/ Gadanki at lower altitudes but not at higher altitudes and perhaps they might have been filtered by the background winds.
convection induced GWs (Choi et al., 2006; HimaBindu, 2016). Fig. 5 shows the snapshots of vertical velocities observed during a VSCS Madi cyclone obtained from the 3DVAR at 24 km altitude. Inner triangle within the contour shows the location of Tirupati/Gadanki, cyclone track is shown with dotted line and the centre of the cyclone for a given time shown on the top of each panel is marked as a solid filled circle. A systematic updrafts and downdrafts propagating away from the cyclone centre with the motion of the cyclone are clearly observed from this figure. These updrafts and downdrafts reach the Tirupati/Gadanki location quite often particularly as shown in Fig. 5c and d. As short period waves are believed to be generated locally, strong wave activity is seen almost during the entire period of the cyclone Madi irrespective of its intensity. The spatial extent of these waves at 24 km altitude can be seen up to 1000 km from the cyclone centre and they propagate in all the directions from the centre of the cyclone. Similar features are noticed at different altitudes such as 12, 14, 18, and 20 km during the cyclone period from 06 to 13 Dec. 2013 (Figure not shown). Since the main objective of the present study is to investigate the vertical coupling of the lower and middle atmosphere during the tropical cyclone Madi, first we characterize these waves over Tirupati. The background zonal wind, meridional wind, vertical wind and temperature profiles during the period of the cyclone over Tirupati/Gadanki have been extracted from the WRF assimilated data. Fig. 6a–d shows the time - height section of zonal, meridional, vertical wind velocities and temperature over Tirupati/Gadanki, respectively, from 0.5 to 30 km during 06–13 December 2013 (during the cyclone) obtained from WRF 3DVAR assimilated output. Zonal winds show alternating eastward and westward flow in the troposphere with strong westward around 16 km altitude particularly during latter half of the observations and changes to eastward above 20 km. In general, meridional winds are southward in the troposphere
4.4. Vertical coupling of lower and middle atmosphere during cyclone The winds obtained from WRF 3DVAR assimilated outputs are further subjected to 1–5 h band pass filter.Fig. 8(a–c) shows the time series of 1–5 h band pass filtered zonal, meridional and vertical wind over Tirupati/Gadanki during Madi cyclone. Clear downward phase propagation is noticed with alternating updrafts and downdrafts with magnitudes 6
H. Hima Bindu et al.
Journal of Atmospheric and Solar-Terrestrial Physics xxx (2017) 1–13
Fig. 6. Time – altitude variation of (a) zonal wind, (b)meridional wind, (c)vertical wind and (d)temperature over Gadanki (13.5 N, 79.2 E) during Madi cyclone obtained from WRF 3DVAR assimilated outputs.
Fig. 7. Wavelet analysis of (a) zonal wind (b) meridional wind and (c) vertical wind at 20 km altitude obtained from the 3DVAR assimilation.
range from 50 to 150 km with dominant horizontal wavelengths of 100 km. Since convective sources produce a broad spectrum of waves, the location of an instrument site (Gadanki/Tirupati) relative to the source can heavily influence the spectrum of waves seen. Fig. 8d-f shows the time-altitude section of 1–5 h band pass filtered variations observed
reaching as high as 0.1 m/s. FFT in the vertical domain revealed the vertical wavelengths ranging from 4 to 16 km with dominant vertical wavelengths of 8 km. The horizontal wavelengths during the cyclone over Tirupati/Gadanki are estimated by the dispersion relation using observed periods and vertical wavelengths. The horizontal wave lengths
7
H. Hima Bindu et al.
Journal of Atmospheric and Solar-Terrestrial Physics xxx (2017) 1–13
Fig. 8. Time-altitude section of 1–5 h band pass filtered variations observed in (a) zonal, (b) meridional and (c) vertical winds over Gadanki (13.5 N, 79.2 E) during Madi cyclone obtained from WRF 3DVAR assimilated outputs. (d), (e) and (f) Same as (a), (b) and (c) but over 13.45 N and 88 E.
in zonal, meridional and vertical wind over 13.45 N and 88 E (same as that Gadanki/Tirupati but at different longitude) during Madi cyclone obtained from WRF 3DVAR assimilated outputs. Dominant downward phase propagation is noticed with alternating updrafts and downdrafts with magnitudes reaching as high as 0.1 m/s. All these characteristics suggest these waves are of high frequency GWs. Thus, it is again clear that WRF assimilated outputs can be used to investigate the cyclone generated GWs. The advanced SVU meteor radar located at Tirupati is specially designed to provide continuous high-resolution wind measurements
round the clock for observing both small and large scale oscillations in the MLT region (70–110 km). It is advanced meteor radar due to its high meteor count rate sensitivity, configured by ATRAD, Australia. The SVU MR provides an opportunity to elucidate the vertical and lateral coupling of the MLT region (Rao et al., 2014). Horizontal wind profiles observed by SVU meteor radar during the cyclone period (06–13 Dec. 2013) are subjected to FFT analysis. Waves with periods 1–7 h with dominant period 2–5 h is noticed in both zonal and meridional winds (figure not shown). Fig. 9a-b shows the time series of 1–5 h band pass filtered zonal and meridional winds over Tirupati during Madi cyclone. Clear 8
H. Hima Bindu et al.
Journal of Atmospheric and Solar-Terrestrial Physics xxx (2017) 1–13
Fig. 9. Time-altitude section of 1–5 h band pass filtered variations observed in (a) zonal and (b) meridional winds over Tirupati during Madi cyclone obtained from SVU metoer radar. (c) and (d) same as (a) and (b) but during normal weather condition (without cyclone days).
fitting in order to evaluate the amplitude and phase of each component. A 3-day data window, sliding forward by 1day each time, is used for extraction of the tide. The resulting amplitude and phases of the tide are attributed to the middle day. Each component of the wind can be expressed in terms of the prevailing and tidal components, as follows:
downward phase propagation is noticed with magnitudes reaching as high as 20 m/s. FFT in the vertical domain revealed the vertical wavelengths ranging from 4 to 16 km with dominant vertical wavelengths of 16 km. The horizontal wavelengths during the cyclone over Tirupati/Gadanki are estimated by the dispersion relation using observed periods and vertical wavelengths. The horizontal wavelengths range from 5 to 150 km with dominant horizontal wavelength of 50 km. All these characteristics suggest that these waves are of high frequency GWs. Since high frequency waves are thought to be fairly ubiquitous in the MLT region, the SVU MR data without cyclone activity has been taken for the similar analysis to check whether these waves are generated by the cyclone or not. The time-altitude section of 1–5 h band pass filtered variations observed in zonal, meridional winds over Tirupati during normal weather (no cyclone activity) obtained from SVU meteor radar shown in Fig. 9(c–d) reveals no significant wave activity during the non-cyclone days. Observed amplitudes are below 10 m/s during non-cyclonic periods. Thus, the upward propagating waves obtained from the model outputs shown in Fig. 8 are again noticed in the SVU MR horizontal wind observations (from 70 to 110 km altitude) which showed a clear propagation of GWs into the mesosphere. Thus, it is clear that the imprints of cyclone generated GWs reach as high as MLT region. As mentioned earlier, the zonal and meridional winds from SVU MR are obtained with 2 km and 1 h resolutions from 70 to 110 km. In order to investigate the behavior of the background tides during cyclone period, the hourly averaged zonal and meridional winds were subjected to Fourier analysis. The diurnal (24 h), semi-diurnal (12 h) and terr-diurnal (8 h) tidal oscillations are clearly noticed (figure not shown). The hourly mean winds recorded by the SVU MR are further subjected to harmonic
AðtÞ ¼ Ao þ Ak cos
2π k ðt ϕÞ Tp
[2]
where t is the local time, A[t] is the parameter [U and V] to be fitted, Ao is the prevailing component, A and φ are the amplitude and phase of the tidal, and Tp ¼ 24 (e.g. Venkateswara Rao et al., 2011). The tidal parameters are individually estimated for U and V. Similar analysis is followed for WRF assimilated outputs to know the tidal activity in the troposphere and stratospheric altitudes. Tidal amplitudes observed in zonal wind and meridional wind from WRF assimilation and SVU Meteor radar during 3–19 Dec. 2013 is shown in Fig. 10. Enhancement in the tidal activity in the MLT region during the peak cyclone activity is clearly noticed in the zonal component of SVU meteor radar. However, semi-diurnal component is suppressed in the meridional component. In the troposphere and lower stratosphere, enhancement in the diurnal and semi-diurnal tides is seen at little later times in both diurnal and semi-diurnal components. No significant terrdiurnal amplitudes are noticed in both zonal and meridional components in the troposphere and lower stratosphere. This shows the clear evidence of vertical coupling between the lower and middle atmosphere during a VSCS Madi even due to tides. However, note that there is a big gap between 30 and 70 km and we do not have any observations at these
9
H. Hima Bindu et al.
Journal of Atmospheric and Solar-Terrestrial Physics xxx (2017) 1–13
Fig. 10. Time-altitude sections of (a) diurnal, (b) semi-diurnal and (c) ter-diurnal tides observed in the zonal component at MLT region using SVUMR. (d)–(f) same as (a)–(c) but observed in the TLS region using WRF assimlated outputs. (g)–(i) same as (a)–(c) but for the meridional component. (j)–(l) same as (d)–(f) but for the meridional component.
run with 2-way nested two domains with horizontal resolutions of 18 km and 6 km. But all the results are analyzed from the high resolution 6 km domain with the temporal resolutions of 10 min. The assimilated model outputs are verified first with the IMD observed track to check the model consistency (Fig. 2). The WRF model outputs are further compared with the independent Gadanki GPS radiosonde temperature and horizontal winds profiles (Fig. 3), which are available during the cyclone period and also with COSMIC GPSRO profiles (Fig. 4). A good comparison is found between the WRF assimilated outputs and the independent observations. Further, the snapshots of vertical wind component at different times in the lower stratosphere from the WRF model showed a clear evidence of GWs over Gadanki region during TC Madi (Fig. 5). The dominant periods of these waves are found to be 1–5 h (Fig. 7) and have the dominant vertical wavelengths of 8 km with the horizontal wavelengths of 100 km. All the three components of winds (zonal, meridional and vertical) showed clear downward phase propagation (upward propagating waves) in the upper troposphere and lower stratosphere during the cyclone (Fig. 8). All these characteristics reveal that the high frequency GWs is generated during the VSCS Madi. These upward propagating waves obtained from the model are again noticed in the SVU MR (Tirupati) horizontal wind observations in the MLT region (70–110 km) (Fig. 9).Thus, evidence of coupling between lower and middle atmosphere exists through waves generated during the VSCS Madi is shown. It can be further confirmed with ray tracing of the gravity waves, however, getting background information which is very much essential for accurate implementation of ray tracing is almost impossible during disturbed events like cyclones. Enhancement in the tidal activity in both zonal and meridional components during the peak stage of cyclone is also noticed both in the MLT region using SVU meteor radar and troposphere and lower stratosphere using WRF assimilated outputs (Fig. 10). Interestingly, suppression of semi-diurnal tide in the meridional component is noticed. No
altitudes. Thus, above mentioned link between lower and middle atmosphere through waves and tides remains as qualitative study. To further look into this aspect, we make use of recently released ERA 5 reanalysis data sets. As ECMWF ERA5 data is quite recent, we have validated the ERA5 data products with the independent GPS radiosonde observations to check the consistency of reanalysis data. Fig. 11 shows the profiles of zonal (top panel), meridional winds (middle panel) and temperature (bottom panel) during Cyclone period (07–13 Dec 2013). A good agreement is observed between ERA5 and radiosonde profiles. With this confidence on ERA5 products, to cover the gap between 30 and 70 km the similar analysis (Tidal extraction) procedure is applied for this (ERA5) data over Gadanki/Tirupati and obtained all the three tidal components for both zonal and meridional winds. Fig. 12 shows the time-altitude sections of diurnal, semi-diurnal and terr-diurnal tidal components in zonal and meridional components during entire December 2013 month covering both cyclone days and non-cyclonic days. More or less similar features observed in the SVU meteor radar i.e, enhancement in diurnal tide and suppression in semi-diurnal tide and no significant amplitudes in the terr-diurnal tide during cyclone period are again noticed even in ERA5 data sets between 40 and 60 km particularly in meridional component. It is clear that significant changes in the tidal activity during cyclone period exist in lower and middle atmosphere when compared to non-cyclonic periods. 5. Summary and conclusions This paper presents the vertical coupling of lower and middle atmosphere by gravity waves (GWs) and tides over Gadanki (13.5 N, 79.2 E)/ Tirupati(13.6 N, 79.4 E) during the evolution of an intense tropical cyclone Madi (2013) using Weather Research and Forecast (WRF) version 3.8.1 model simulations using three dimensional variational assimilation technique and SVU Meteor Radar observations. The model is 10
H. Hima Bindu et al.
Journal of Atmospheric and Solar-Terrestrial Physics xxx (2017) 1–13
Fig. 11. Profiles of Zonal wind (top panel), Meridional wind (middle panel) and Temperature (bottom panel) from ERA5 reanalysis data products and GPS Radiosonde observations over Gadanki during the period of Cyclone ‘Madi’ (07–13 Dec 2013).
Fig. 12. Time-altitude sections of diurnal (left panle), semi-diurnal (middle panel) and ter-diurnal (right panel) tides observed in the ERA-5 reanalysi data sets of zonal (left panels) and meridonal winds (right panles) during December 2013.
Enhancement in the diurnal tide, suppression in the semi-diurnal tide and no significant amplitudes in the terr-diurnal tide during cyclone activity is observed particularly in the meridional wind. There are few interesting observations being obtained from the present study. Strong updrafts and downdrafts surrounding the cyclone are observed that are mainly due to the cyclone generated cloud bands which
significant amplitudes in the terr-diurnal component in the lower atmosphere are noticed though significant amplitudes do exists in the MLT region. Thus, coupling between lower and middle atmosphere through tides is also possible during the cyclone activity. Similar features are also observed (Fig. 12) in the recently released ERA 5 reanalysis data sets covering the gap between WRF outputs and SVU meteor radar.
11
H. Hima Bindu et al.
Journal of Atmospheric and Solar-Terrestrial Physics xxx (2017) 1–13
is the source for the observed GW activity. These GWs propagate in all the directions and those reached over Gadanki/Tirupati are being observed closely. Thus, strong GW activity in the troposphere and lower stratosphere is not throughout the cyclone period but observed as episodic over the Gadanki/Tirupati. Since these high frequency GWs prefer to propagate vertically than in the horizontal direction, enhanced GW activity is observed throughout the cyclone period in the MLT region over Tirupati. Further, stronger amplitudes in the meridional component are observed as the zonal component may be filtered due to strong shears present at 30 km altitude (Fig. 6). Thus, one to one relation between the wave activity in the lower and middle atmosphere may not present mainly due to wave-mean flow interaction. Vertically propagating GWs generated by cyclone (convection) may provide an important contribution to the momentum budget of the middle atmosphere as suggested by Song et al., 2003. Another interesting feature noticed is the enhancement in the tidal activity during the cyclone period which also has source in the water vapour distribution. Tidal amplitudes in the meridional component is almost double than that observed in the zonal component. In continuation to our earlier work (HimaBindu, 2016) and this work, a set of experiments are planned in NARL and SVU covering the atmosphere from surface up to 110 km to investigate more on the coupling between the lower and middle atmosphere during cyclone passage. Future studies will also include detailed description of source mechanisms for convectively generated GWs through severe cyclones.
Dewan, E., Coauthors, 1998. MSX satellite observations of thunderstorm-generated gravity waves in mid-wave infrared images of the upper stratosphere. Geophys. Res. Lett. 25, 939–942. Dutta, G., Kumar, M.C.A., Kumar, P.V., Ratnam, M.V., Chandrashekar, M., Shibagaki, Y., Salauddin, M., Basha, H.A., 2009. Characteristics of high-frequency gravity waves generated by tropical deep convection: case studies. J. Geophys. Res. 114, D18109 https://doi.org/10.1029/2008JD011332. Evan, S., Alexander, M.J., Dudhia, J., 2012. WRF simulations of convectively generated gravity waves in opposite QBO phases. J. Geophys. Res. 117, D12117 https:// doi.org/10.1029/2011JD017302. Fovell, R., Durran, D., Holton, J.R., 1992. Numerical simulations of convectively generated stratospheric gravity waves. J. Atmos. Sci.49 1427–1442. Fritts, D.C., Alexander, M.J., 2003. Gravity wave dynamics and effects in the middle atmosphere. Rev. Geophys. 41 (1), 1003. https://doi.org/10.1029/2001RG000106. Gemmill, W., Katz, B., Xu, L., 2007. Daily Real-time, Global Sea Surface Temperature High-resolution analysis.NCEP/EMC off. Note 260, 39 pp., Environ. Modell. Cent. Camp Springs, Md. Guo, Y.R., Kusaka, H., Barker, D.M., Kuo, Y.H., Crook, A., 2005. Impact of ground based GPS PW and MM5 3DVar background error statistics on forecast of a convective case. Sci. Online Lett. Atmos 1, 73–76. https://doi.org/10.2151/sola.2005-020. HimaBindu, H., Venkat Ratnam, M., Yesubabu, V., Narayana Rao, T., Amit Kesarkar, Naidu, C.V., 2016. Characteristics of Cyclone generated Gravity waves observed using assimilated WRF model simulations over Bay of Bengal. Atmos. Res. 180, 178–188. Holton, J.R., 1972. Waves in the equatorial stratosphere generated by tropospheric heat sources. J. Atmos. Sci. 29,368–375. Hong, S.Y., Noh, Y., Dudhia, J., 2006. A new vertical diffusion package with explicit treatment of entrainment processes. Mon. Weather Rev. 134, 2318–2341. Ibrahim, C., Chane Ming, F., Barthe, C., Kuleshov, Y., 2010. Diagnosis of tropical cyclone activity through gravity wave energy density in the Southwest Indian Ocean. Geophys. Res. Lett. 37, L09807 https://doi.org/10.1029/2010GL042938. Ide, K., Courtier, P., Ghil, M., Lorenc, A.C., 1997. Unified notation for data assimilation: operational, sequential and variational. J. Math. Soc. Jpn. 75, 181–189. Kain, J.S., 2004. The Kain-Fritsch convective parameterization: an update. J. Appl. Meteor 43, 170–181. Kim, Y.-J., Eckermann, S.D., Chun, H.-Y., 2003. An overview of the past, present and future of gravity-wave drag parameterization for numerical climate and weather prediction models. Survey Article. Atmos.-Ocean 41 (1), 65–98. https://doi.org/ 10.3137/ao.410105. Kim, S.-Y., Chun, H.-Y., Wu, D.L., 2009. A study on stratospheric gravity waves generated by Typhoon Ewiniar: numerical simulations and satellite observations. J. Geophys. Res. 114, D22104 https://doi.org/10.1029/2009JD011971. Kim, S.Y., Chun, H.Y., 2010a. Momentum flux of stratospheric gravity waves generated by Typhoon Ewiniar 2006. Asia-Pac. J. Atmos. Sci. 46, 199–208. Kim, S.-Y., Chun, H.-Y., 2010b. Stratospheric gravity waves generated by Typhoon Saomai (2006): numerical modeling in a moving frame following the typhoon. J. Atmos. Sci. 67, 3617–3636. https://doi.org/10.1175/2010JAS3374.1. Kim, S.-Y., Chun, H.-Y., 2011. Impact of typhoon-generated gravity waves in the typhoon development. Geophys. Res. Lett. 38, L01806 https://doi.org/10.1029/ 2010GL045719. Kim, S.H., Chun, H.-Y., Jang, W., 2014. Horizontal divergence of typhoon-generated gravity waves in the upper troposphere and lower stratosphere (UTLS) and its influence on typhoon evolution. Atmos. Chem. Phys. 14, 3175–3182. https://doi. org/10.5194/acp-14-3175-2014. Kumar, K.K., 2006. VHF radar observations of convectively generated gravity waves: some new insights. Geophys. Res. Lett. 33, L01815 https://doi.org/10.1029/ 2005GL024109. Larsen, M., Swartz, W., Woodman, R., 1982. Gravity-wave generation by thunderstorms observed with a vertically pointing 430 MHz radar. Geophys. Res. Lett. 9, 571–574. Leena, P.P., Ratnam, M.V., Krishna Murthy, B.V., Vijaya Bhaskara Rao, S., 2012. Detection of high frequency gravity waves using high resolution radiosonde observations. J. Atmos. Sol. Terr. Phys 254–259. Lieberman, R.S., Hays, P.B., 1994. An estimate of the momentum deposition in the lower thermosphere by the observed diurnal tide. J. Atmos. Sci. 51, 3094–3105. Lieberman, R.S., Oberheide, J., Hagan, M.E., Remsberg, E.E., Gordley, L.L., 2004. Variability of diurnal tides and planetary waves during November 1978–May 1979. J. Atmos. Sol. Terr. Phys. 66 (6–9), 517–528. https://doi.org/10.1016/ j.jastp.2004.01.006. Lindzen, R.S., 1974. Wave-CISK in the tropics. J. Atmos. Sci. 31, 156–179. Liou, Y.A., Pavelyev, A.G., Huang, C.Y., Igarashi, K., Hocke, K., Yan, S.K., 2003. Analytic method for observation of the gravity waves using radio occultation data. Geophys. Res. Lett. 30, 2021. https://doi.org/10.1029/2003GL017818. Liou, Y.A., Pavelyev, A.G., Wicker, J., Liu, S.F., Pavelyev, A.A., Schmidt, T., Igarashi, K., 2006. Application of GPS radio occultation method for observation of the internal waves in the atmosphere. J. Geophys. Res. 111, D06104 https://doi.org/10.1029/ 2005JD005823. Malardel, S., Wedi, N., Deconinck, W., Diamantakis, M., Kuhnlein, C., Mozdzynski, G., Hamrud, M., Smolarkiewicz, P., 2015. A New Grid for the IFS. ECMWF, p. 6. Newsletter No. 146-Winter 2015/16. Matsuno, T., 1966. Quasi-geostrophic motions in the equatorial area. J. Meteorol. Soc. Jpn. 44, 25–43. McLandress, C., Alexander, M.J., Wu, D.L., 2000. Microwave Limb Sounder observations of gravity waves in the stratosphere: a climatology and interpretation. J. Geophys. Res. 105, 11,947–11,967. https://doi.org/10.1029/2000JD900097.
Acknowledgements The authors wish to thank the Sri Venkateswara University Meteor Radar [SVU MR] for providing necessary data for the present study. We are grateful to the National Atmospheric Research Laboratory [NARL] for providing GPS radiosonde data used in the present study. We also thank the India Meteorological Department [IMD] for the best track data through their website. ERA-Interim data from ECMWF are used to initialize the model and boundary conditions. The prepbufr global observational dataset are obtained from http://rda.ucara.edu. The data used in the present study can be obtained on request. Appendix A. Supplementary data Supplementary data related to this article can be found at https://doi. org/10.1016/j.jastp.2018.01.029. References Alexander, M.J., Beres, J.H., Pfister, L., 2000. Tropical stratospheric gravity wave activity and relationship to clouds. J. Geophys. Res. 105, 22,299–22,309. https://doi.org/ 10.1029/2000JD900326. Anthes, R.A., 2011. Exploring Earth's atmosphere with radio occultation: contributions to weather, climate and space weather. Atmos. Meas. Tech 4, 1077–1103. https:// doi.org/10.5194/amt-4-1077-2011. Barker, D.M., Huang, W., Guo, Y.R., Xiao, Q.N., 2004. A three dimensional (3-D-Var) data assimilation system for use with MM5: implementation and initial results. Mon. Weather Rev. 132, 897–914. Bergman, J.W., Salby, M.L., 1994. Equatorial wave activity derived from fluctuations in observed convection. J. Atmos.Sci 51, 3791–3806. Chane-Ming, F., Roff, G., Robert, L., Leveau, J., 2002. Gravity wave characteristics overTromelin Island during the passage of cyclone Hudah. Geophys. Res. Lett. 29 (6), 1094. https://doi.org/10.1029/2001GL013286. Chane-Ming, F., Chen, Z., Roux, F., 2010. Analysis of gravity-waves produced by intense tropical cyclones. Ann. Geophys. 28, 531–547. https://doi.org/10.5194/angeo-28531-2010. Chen, F., Dudhia, J., 2001. Coupling an advanced land surface/hydrology model with thePenn State/NCAR MM5 modeling system, Part I: model description and implementation. Mon. Weather Rev. 129, 569–585. Choi, Y.G., Lee, S.C., McDonald, A.J., Hooper, D.A., 2006. Wind-profilerobservations of gravity waves produced by convection at mid-latitudes. Atmos. Chem. Phys. 6, 2825–2836. https://doi.org/10.5194/acp-6-2825-2006. Chun, H.-Y., Kim, Y.-H., 2008. Secondary waves generated by breaking of convective gravity waves in the mesosphere and their influence in the wave momentum flux. J. Geophys. Res. 113, D23107 https://doi.org/10.1029/2008JD009792.
12
H. Hima Bindu et al.
Journal of Atmospheric and Solar-Terrestrial Physics xxx (2017) 1–13 Sato, K., Hashiguchi, H., Fukao, S., 1995. Gravity waves and turbulence associated with the cumulus convection observed with the UHF/VHF clear-air Doppler radars. J. Geophys. Res. 100, 7111–7119. Skamarock, W.C., Klemp, J.B., Dudhia, J., Gill, D.O., Barker, D.M., Dudha, M.G., Huang, X., Wang, W., Powers, Y., 2008. A Description of the Advanced Research WRF Ver.30.NCARTechnicalNote, NCAR/TN-475CSTR, Mesoscale and Microscale Meteorology Davison. National Centre for Atmospheric Research, Boulder, CO, USA, p. 113. Song, I.S., Chun, H.Y., Lane, T.P., 2003. Generation mechanisms of convectively forced internal gravity waves and their propagation to the stratosphere. J. Atmos. Sci. 60, 1960–1980. Srinivas, C.V., Yesubabu, V., Venkatesan, R., Ramarkrishna, S.S.V.S., 2010. Impact of assimilation of conventional and satellite meteorological observations on the numerical simulation of a Bay of Bengal Tropical Cyclone of Nov 2008 near Tamilnadu using WRF model. Meteorol. Atmos. Phys. 110, 19–44. Srinivas, C.V., Yesubabu, Hari Prasad, K.B.R.R.V., Venkatraman, B., Ramakrishna, S.S.V.S., 2012. Numerical simulation of Cyclonic Storms FANOOS, NARGIS with assimilation of conventional and satellite observations using 3-D-VAR. Nat. Hazards 63 (2), 867–889. Thompson, G., Field, P.R., Rasmussen, R.M., Hall, W.D., 2008. Explicit forecasts of winter precipitation using an improved bulk microphysics scheme. Part II: implementation of a new snow parameterization. Mon. Weather Rev. 163, 5095–5114. Tsuda, T., Nishida, M., Rocken, C., Ware, R.H., 2000. A global morphology of gravity wave activity in the stratosphere revealed by the GPS occultation data (GPS/MET). J. Geophys. Res. 105, 7257–7273. Venkat Ratnam, M., Pravallika, N., Ravindra babu, S., Ghouse Basha,Pramitha, M., Krishna Murthy, B.V., 2014. Assessment of GPS radiosonde descent data. Atmospheric Measurement Techniques 7, 1011–1025. Venkateswara Rao, N., Tsuda, T., Gurubaran, S., Miyoshi, Y., Fujiwara, H., 2011. On the occurrence and variability of the terdiurnal tide in the equatorial mesosphere and lower thermosphere and a comparison with the Kyushu-GCMJ. Geophys. Rep. 116, D02117 https://doi.org/10.1029/2010JD014529. Weger, I., 2015. In: “Preliminary Comparison of ECMWF ERA-interim and ERA-5 Reanalyses” on Core-Climax Workshop, Brussels, January 15-16, 2015, and ECMWF – Computing and Forecast System, ICAS 2015, Annecy, France, September 2015. Wheeler, M., Kiladis, G.N., 1999. Convectively coupled equatorial waves: analysis of clouds and temperature in the wavenumber– frequency domain. J. Atmos. Sci. 56, 374–399.
Mlawer, E.J., Taubman, S.J., Brown, P.D., Iacono, M.J., Clough, S.A., 1997. Radiative transfer for inhomogeneous atmosphere: RRTM, a validated correlated-k model for the long wave. J. Geophys. Res. 102, 16663–16682. Osuri, K.K., Mohanty, U.C., Routray, A., Mohapatra, M., 2013. Real-time track prediction of tropical cyclones over the north Indian Ocean using the ARW model. J. Appl. Meteorol. Climatol 52, 2476–2492. https://doi.org/10.1175/JAMC-D-12-0313.1. Parrish, D.F., Derber, J.C., 1992. The national meteorological Centre's spectral statistical interpolation analysis system. Mon. Weather Rev. 120, 1747–1763. Piani, C., Durran, D., Alexander, M.J., Holton, J.R., 2000. A numerical study of three dimensional GWs waves triggered by deep tropical convection and their role in the dynamics of the QBO. J. Atmos. Sci. 57, 3689–3702. Pirscher, B., Foelsche, U., Borsche, M., Kirchengast, G., Kuo, Y.H., 2010. Analysis of migrating diurnal tides detected in FORMOSAT-3/COSMIC temperature data. J. Geophys. Res. 115, D14108 https://doi.org/10.1029/2009JD013008. Pramitha, M., Venkat Ratnam, M., Leena, P.P., Krishna Murthy, B.V., Vijaya Bhaskar Rao, S., 2015. Identification of Inertia Gravity Wave sources observed in the troposphere and the lower stratosphere over a tropical station Gadanki. Atmos. Res. 176–177 [2016] 202–211. Rajasree, V.P.M., Kesarkar, A.P., Bhate, J.N., Umakanth, U., Singh, V., Varma, T.H., 2013. 2016. Appraisal of recent theories to understand cyclogenesis pathways of tropical cyclone Madi. J. Geophys. Res. Atmos. 121, 8949–8982. https://doi.org/10.1002/ 2016JD025188. Rao, S.V.B., Eswaraiah, S., VenkatRatnam, M., Kosalendra, E., Kishore Kumar, K., Sathish Kumar, S., Patil, P.T., Gurubaran, S., 2014. Advanced meteor radar installed at Tirupati: system details and comparison with different radars. J. Geophys. Res. Atmos. 119, 11,893–11,904. https://doi.org/10.1002/2014JD021781. Ratnam, M.V., Kishore Kumar, G., Krishna Murthy, B.V., Patra, A.K., Jagannadha Rao, V.V.M., Vijaya Bhaskar Rao, S., Kishore Kumar, K., Ramkumar, G., 2008. Longterm variability of the low latitude mesospheric SAO and QBO and their relation with stratospheric QBO. Geophys. Res. Lett. 35, L21809 https://doi.org/10.1029/ 2008GL035390. Ratnam, M.V., Tetzlaff, G., Jacobi, C., 2004. Global and seasonal variations of stratospheric gravity wave activity deduced from the CHAMP/GPS satellite. J. Atmos. Sci. 61 (13), 1610–1620. Salby, M.L., Garcia, R.R., 1987. Transient response to localized episodic heating in thetropics. Part I: excitation and short-time near-field behaviour. J. Atmos. Sci. 44,458–498.
13