JASMINE UW

Overview:

Introduction

Data

Findings

Summary

References

JASMINE - The Joint Air-Sea Monsoon
Interaction Experiment
7 April - 8 June, 1999
University of Washington Participation Report

Introduction

Studies find that the onset of the Asian monsoon varies significantly from year to year at any given location (Webster et al., 1998). In addition, the onset period itself is marked by 'active' and 'break' conditions separated by about 30 days. Such variability greatly affects the agricultural communities within the monsoon region. During active conditions, anticyclonic circulations set up over the Bay of Bengal resulting in strong westerly winds and heavy precipitation. The anticyclonic gyre is associated with a Kelvin-Rossby complex which is initially observed in the west Indian Ocean along the equator. The gyre propagates northeastward around the tip of the Indian continent and into the Bay of Bengal. A weaker gyre is also observed to propagate south of the equator. The goal of JASMINE is to observe the atmosphere and ocean environment during an onset event, including both an active and break period, so that we might begin to understand the conditions responsible for the variability of the monsoon season on short time and space scales.

The JASMINE project is a collaborative effort among scientists from the University of Colorado, NOAA Environmental Technology Laboratory (ETL), the University of Washington, the University of Hawaii and CSIRO in Australia to study the atmospheric and upper ocean structure which characterize the onset period of the south Asian summer monsoon. The University of Washington's primary interests are in observations of the atmospheric thermodynamic and dynamic structure, and the structure of the atmospheric convection in relation to the larger scale environment during the active periods of the monsoon.

Data Description

The JASMINE cruise took place in the east Indian Ocean and the Bay of Bengal between 7 April - 8 June 1999 on the National Oceanic Atmospheric Association's (NOAA's) research vessel the Ronald H. Brown. The cruise was broken up into two segments shown in Figure 1. The first leg (green) was from 7 - 22 April and collected data along 89 ^o E, between 17^o N and the equator. The second leg (red) was from 28 April - 8 June, collecting data during two north-south transits along 89 ^o E, between the equator and 12 ^o N. At the north end of each transit during the second leg two 5-day star patterns were traced, each approximately 125 km in diameter. The length scale and pattern was chosen to be large enough to calculate the upper ocean budget, but small enough to observe the development of atmospheric processes while effectively in one place.

A wide range of instruments were used in addition to the basic oceanic and surface meteorological observations available on the Brown, including; radiosondes, upward and downward looking IR radiometers, a 35 GHz vertically pointing Doppler cloud radar, S-band radar, a 915 MHz profiler, a scanning C-band Doppler radar, an air-sea flux system, optical rain gauges, and CTDs. These on board measurements are placed in a larger scale context through 3-hourly European Meteosat-5 satellite data, NOAA-12, -14 and -15 polar orbiting satellite data and ECMWF reanalyses run specifically for the JASMINE project.

This report introduces the preliminary findings of the C-band Doppler radar and radiosonde data sets. Surface meteorological observations from the Woods Hole Oceanographic Institution's IMET system on a bow-mast are used to describe the conditions at the location of the ship. In addition images from the Meteosat-5 geostationary satellite and the NOAA GOES-12, -14 and -15 polar orbiting satellites are used to provide an overview of the regional-scale convective activity during the project. As UW scientists were not on board for JASMINE leg 1, this preliminary report will focus on data collected during leg 2.

A. C-band Radar

The Brown has a permenantly installed, stabilized scanning C-band Doppler radar. The C-band has a 5 cm wavelength and a 1^o beam width. Three scan sets were used during JASMINE. The surveillance scan set consisted of 2 low elevation tilts and took about 2 minutes to complete. The two volume scan sets consisted of 21 elevation tilts, slightly staggered to provide better vertical resolution of the precipitating region. Each of these sets took about 8 minutes to complete. The radar alternated between the surveillance and volume scan sets, totaling about 20 minutes for a complete cycle. Figure 2 shows the tilt angles for the volume scan sets. The surveillance scans were at 0.4 and 0.8 ^o above the horizon, the same as the lowest 2 tilts of the volume scan sets shown. The tilt at 80^o above the horizon in scan set C was done to provide a comparison with the up-looking cloud radar. Because JASMINE leg 1 was not initially planned, but instead was the result of making use of extra ship time, no C-band data could be organized for this leg. For JASMINE leg 2 the radar data cover the period from 6 - 31 May, providing both reflectivity and radial velocity within precipitation regions.

B. Radiosondes

The Brown is equipped with a Vaisala radiosonde system. The launch procedure and processing of the radiosonde data can be found by following the links. A total of 23 soundings were launched from April 10 - 17 for JASMINE leg 1, and 171 soundings were launched from May 6 - 30 for JASMINE leg 2, giving a total of 194 profiles. During the first leg nominally 3 soundings per day were collected, while nominally 6 soundings per day were collected during the second leg, going up to 8 per day for the star patterns. These data include pressure, temperature, humidity and GPS-determined horizontal winds. Table 1 (adopted from the processing report) provides some basic statistics on the soundings. A more complete discussion is given in the radiosonde documentation.

Initial Findings

A. Monsoon Convective Variability

The JASMINE data encompass two active periods and a break period of the south Asian monsoon. Just as climatology predicts, the onset of the monsoon arrived in early April, during leg 1. This active period was followed by a break period in early May, followed by a second active period in late May, encountered during leg 2. A time-latitude plot of the OLR data from the European Meteosat 5 satellite is shown in Figure 3 for leg 2 only (courtesy of David Lawrence), with the ship track shown in white for reference. The second active period is seen clearly, propagating from the equator, around Julian Day 139, and passing through the Bay of Bengal where the activity continues through Julian Day 150. The break period is also evident up to Julian Day 139, during which time the convection is seen to reside on the equator.

The time series of relative humidity profiles from the soundings collected during leg 2 is shown in Figure 4 along with the ship track latitude for reference. The time periods for star 1 and star 2 are bracketed by the thick black lines and are from 10-15 May and 21-26 May, respectively. The break or suppressed period of star 1 is characterized by relatively dry conditions above the low level cloud layer, which extended to approximately 2 km (800 mb). Cirrostratus clouds, recorded frequently in the hourly cloud observation log and detected by the cloud radar and upward looking radiometer, are apparent in the relative humidity profiles between 200-300 mb. These clouds were also seen with the C-band and S-band radars when optically thick. A time series of the C-band radar data is shown in Figure 5. The data shown are the percent of 4 km x 4 km regions within 125 km of the ship with reflectivities from 15-35 dBZ (dotted line) and greater than 35 dBZ (solid line). As in Figure 4, the heavy solid black lines bracket the times of star 1 and star 2. Consistent with the relative humidity profiles, the C-band indicates little convective activity occurred during the first star.

Table 2 shows the integrated water vapor calculated from the soundings for star 1, star 2 and the total for leg 2. The values for star 1 indicate more than half of the total moisture between the surface and 50 mb is found in the lower troposphere (1010-850 mb) during this period. From these data it is also evident that the total water vapor below 850 mb is approximately 25 kg/m² and does not change significantly between the active and break periods. The somewhat lower total value for leg 2 is lower than either of the two stars, indicating the variability in low level moisture is more a function of latitude, being drier towards the equator, than the local convective activity. The sea surface temperature is higher in the Bay of Bengal than at the equator at this time of year producing the surface moisture dependence on latitude observed here.

The large-scale wind patterns over the Indian Ocean and Bay of Bengal during the monsoon indicate that active and break periods are defined in the flow field by the location of eastward propagating Kelvin-Rossby wave-like gyre. During a break period the gyre is furthest west, producing a relaxation in the anticyclonic flow over the Bay of Bengal. As in Figure 4, Figure 6 shows profiles of the sounding data from leg 2 except for zonal wind. Weak easterlies prevail for all of star 1 throughout most of the troposphere consistent with a weakening of the monsoon circulation. At this time the unorganized convection observed at the ship originated primarily off the coast of Burma to the northeast. Westerlies descend down to 700 mb by the end of this star, but the shear remains fairly weak. The meridional winds are shown in Figure 7 and are seen to be overall weaker than the zonal winds in this region during this season.

The second 5-day star was during an active period of the monsoon. Convective systems during this time either passed over the ship or were well within the 125 km radius of the C-band radar 3-D volume scans. The activity over the ship is evidenced in both the relative humidity profiles (Figure 4) and the radar convective index (Figure 5). The ship encountered convective activity about a day prior to the start of star 2 and continued to be in the convection until reaching about 5 ^o S, on the return transit. Satellite images indicate the activity continued as the ship headed south to Darwin. Table 2 shows that while the low level moisture remains equivalent to that observed during the break period, mid- and upper level moisture values increase by 8 and 3 kg/m², respectively, with an increase of 11 kg/m² overall during star 2. This value is double the estimates of total water vapor increase for the COARE and Tropical Eastern Pacific Process Study soundings for convective verses suppressed conditions. Average moisture fluxes are higher for JASMINE as well (C. Fairall, personal communication), indicating a larger local moisture source for the convection in this region.

The zonal winds shift during the second star to strong westerlies as the Kelvin-Rossby wave gyre moves north-eastward into the Bay of Bengal (Figure 6). Wind speeds are greater than 10 m/s below 500 mb, transitioning to easterlies of the same magnitude above 300 mb. This vertical structure can be contrasted to that observed during star 1. The Kelvin-Rossby wave structure is thought to be at least in part responsible for the variability of the monsoon during the onset phase. Climatology indicates that as the wave complex propagates eastward, one gyre heads northward into the Bay of Bengal, while another weaker gyre heads southward towards the coast of Malasia. These gyres enhance the circulation favorable for convection, bringing on the active phase of the monsoon. These dynamics are consistant with what was observed during JASMINE, supporting a strong connection between the Kelvin-Rossby wave dynamics and variability of the monsoon precipitation patterns.

An example of the horizontal and vertical extent of an organized convective system observed on 22 May during the second star is seen in the sequence of radar scans in Figure 8. The left column shows the surveillance scans for 09:58, 16:58 and 20:58 UTC, corresponding to local times of 14:58, 21:58 and 02:58, respectively. The right column shows the vertical cross sections along the red lines indicated on the images for 16:58 and 20:58 UTC. Colors indicate reflectivity, with the lowest values being purples and blues and the highest values being yellows and reds. The green range rings are 20 km apart out to 160 km. The ship is marked by the red "X" at the center of the range rings. The three surveillance scans show the approach and development of a line of convection, which originated northwest of the ship and moved south, and eventually east, by sunrise of the following day. This movement is typical of the systems observed during the second star. During its development stage this system is seen to consist of mainly convective precipitation areas evidenced by the narrow line of high reflectivity in the surveillance scans of 0958 and 1658. The vertical cross section at 1658 indicates the typical echo tops of greater than 16 km and typical maximum echos of about 50 dBZ (red) at the convective core for these systems. By 2058 the system has developed a large stratiform precipitation region seen in Figure 8(c) as the yellow and tan region at the center of the image. The cross section for this time indicates a bright band at about 4 km, where reflectivities of around 40-42 dBZ (pink-rose) are typical. Development of new cells is on the southwest edge of the convective complex, creating a line of cells perpendicular to the surface winds arriving from the west-southwest. It is interesting to note that lightning was regularly observed during hourly weather observations in association with both isolated cells and organized systems during JASMINE.

The stability of the lower atmosphere can be estimated from calculations of convective available potential energy (CAPE), convective inhibition (CIN) and integrated CAPE (ICAPE). In addition comparison of the lifting condensation level (LCL) to the level of free convection (LFC) for boundary layer air parcels can also provide information on the degree to which the boundary layer structure is favorable to convection. Table 3 shows these quantities for star 1, star 2 and all of leg 2. While the LFC and CIN indicate that convection is more inhibited during star 2, the low LCL during this time may be important to favoring new cell growth. The CAPE and ICAPE values do not appear to be significantly different between the two stars. Comparing the thermodynamic structure of these very different convective regimes implies the active phase of the monsoon is likely controlled by non-local conditions. The high surface moisture, sea surface temperatures and large CAPE values indicate the region is capable of supporting deep convection once triggered during either phase of the monsoon.

B. Diurnal Variability

Figure 9 shows the pattern of the strong convective systems (>35 dBZ) during JASMINE as a function of local time. Within the Bay of Bengal these systems tend to peak in activity between sundown and sunrise. There is however variability in this diurnal pattern depending on the stage of the monsoon. During the break period of the monsoon the diurnal cycle is bimodal in character, with peaks in convection occurring between 0200 and 0600 LT and between 1500 and 1900 LT, as is seen in Figure 10 (A). The diurnal variability changes for the active period of the monsoon (Figure 10 (B)). As the scale and intensity of the convection increases, the peaks in convection shift to small maxima between 2100 and 2300 LT and between 0200 and 0700 LT, within an overall peak in convection after sundown, between 1700 and 0800 LT. These results are similar the those from TOGA COARE for the western equatorial Pacific. In studies of the warm pool region (Chen, S. S. and R. A. Houze, 1997; Sui, C.H. et al., 1997), the diurnal cycle in convection peaks between 0000 and 0600 LT during active periods. The afternoon maxima in convection observed during suppressed periods of organized convection in this region are attributed to the diurnal cycle in sea surface temperature, which is reduced if not lost during more active periods.

Summary

Overall the sounding and C-band radar data collected during JASMINE provide observations describing the thermodynamic and dynamic structure of the atmosphere and precipitation regions during the active and break periods of the south Asian monsoon. These preliminary results indicate the importance of the basin-scale Kelvin-Rossby wave structures for determining the time scale of variability of the monsoon precipitation. These data provide detailed qualitative and quantitative measurements of the mesoscale structure in the Bay of Bengal during the monsoon onset for use by model and data-based studies to better understand interactions between the convection and the large-scale dynamics.

In addition to characterizing the convection during the monsoon onset period in the Bay of Bengal, JASMINE data will supplement the growing volume of high quality ground-based observations being collected throughout the tropical oceans. Beginning with TOGA COARE in November-February 1992-1993 over the western Pacific warm pool, studies such as the Tropical Eastern Pacific Process Study (TEPPS) in August of 1997, JASMINE in the Bay of Bengal and the Kwajalein Experiment (KWAJEX) in August-September 1999 are providing detailed information on the characteristics of precipitation in the tropics. Precipitation structure is intimately related to the latent heating in this region which drives global circulation. Such data sets are therefore invaluable to understanding the processes required in numerical models used for climate prediction and in climate studies.

References

Chen, S. S., and R. A. Houze, Jr., 1997: Diurnal Variation and Lifecycle of Deep Convective Systems over the Tropical Pacific Warm Pool, Quarterly Journal of the Royal Meteorological Society, 123, 357-388.

Sui, C. H., K. M. Lau, Y. N. Takayabu and D. A. Short, 1997: Diurnal variations in tropical oceanic cumulus convection during TOGA COARE, Journal of the Atmospheric Sciences, 54, 639-655.

Webster, P. J., V. O. Magana, T. N. Palmer, J. Shukla, R. A. Tomas, M. Yanai, and T. Yasunari, 1998: Monsoons: Processes, predictability, and the prospects for prediction. Journal of Geophysical Research, 103, 14,451-14,510.