Some NCDC systems are temporarily unavailable. This message will be removed when all systems are back online.
Some NCDC systems are temporarily unavailable. This message will be removed when all systems are back online.
|
Joint Workshop of the TOGA COARE Flux and Atmospheric Working Groups
Boulder, Colorado, USA 11-13 July 1995
Frank Bradley and Robert Weller (editors)
TOGA COARE International Project OfficeUniversity Corporation for Atmospheric Research
15 October 1995
Table of ContentsI. Introduction II. Summary III. Workshop Programme IV. Precipitation VI. Atmospheric Forcing of the Ocean VII. Discussion of Mapping, Aircraft Doppler Analysis and Data Assimilation VIII. Modelling and Parameterization IX. Large-scale Space/Time Observations
Appendices B. Electronic Access Points to COARE Flux Data
Introduction
A Joint Workshop between the COARE Air-Sea Interface (flux) and Mesoscale Atmospheric Working groups was held at the NCAR Foothills Laboratory on 11-13 July 1995. The main purposes of the meeting were to have investigators of the flux and mesoscale groups discuss together the current status of COARE research and analysis within these groups, to exchange ideas, to initiate collaborative research efforts, and to work together on problem areas such as the discrepancy between various estimates of precipitation during the IOP. The workshop agenda was developed by the flux group co- chairs, Frank Bradley and Robert Weller, based on suggestions of topics by a cross-section of the principal investigators from the flux and mesoscale groups. The version of the agenda distributed just before the workshop, slightly amended to include the names of all presenters, is given in section III. The five research topics on the agenda were: precipitation, diurnal cycles, atmospheric forcing of the ocean, modelling and parameterization, and large-scale space/time observations. For each topic, there was a series of brief presentations followed by a discussion session. The presentations and discussions are summarized by the respective discussion leaders of each topic in sections IV-IX of this report. At the end of the workshop, data management issues were discussed and a consensus developed on action items, tasks that needed attention within the next year, and dates for case studies to initialize collaborative research. These discussions are summarised in the next section. Appendices provide a list of acronyms, some of the electronic access points for COARE flux data, and the list of participants. We wish to thank Richard Chinman, Director of the TOGA COARE International Project Office, and the staff, Jeanette Walters, Cheryl Jones and Crista Kippes for hosting and assisting with the organisation of the workshop.
II. SummaryBecause of the broad representation of COARE scientists at the workshop, covering a wide spectrum of scales and interests, discussion included a range of expertise and perspectives on each of the topics. In all cases, a group of plans was put in place to resolve outstanding problems and advance science toward COARE goals. This section summarises the main thrust of the discussion and the most crucial of the proposed actions within each topic. Precipitation. Discrepancies between the various methods for determining rainfall during the IOP still exist, but the reasons are becoming more clear. Under close scutiny, differences in analytical technique between similar systems became apparent. The disagreement of around a factor of two between remote sensing and in situ gauges is not resolved, but numerous possibilities were advanced for further study, particularly the vastly different sampling involved. Most importantly, seven specific action items were recommended as a way to make progress, including intercomparison case studies on which investigators should concentrate. Everyone's most recent estimates of IOP-mean precipitation, since the Toulouse workshop, are documented in this report. Diurnal cycles. These were evident in many critical parameters and processes, and discussion focussed on the relative timing of peak values, and what this tells us about diurnal and sub-diurnal interactions. This cuts across all scales and both atmospheric and ocean processes. The peak SST warming found around 1400 local time is almost out of diurnal phase with the large MCSs which occur overnight, and likewise supports stable stratification in the ocean surface layer until after midnight. However, there is much interest in the clear evidence for secondary peaks, signalling an important role for processes on more localised space and time scales. Central to further progress with the analysis of atmospheric soundings data must be the resolution of a serious discrepancy in the lowest level humidities. Atmospheric forcing of the ocean. This session brought into sharp focus the need for close collaboration between COARE scientists studying atmospheric boundary layer, air-sea interface and ocean mixed layer processes. Furthermore, the need for refinement of models in all these areas came through strongly, and for mapping of fluxes, and surface velocity and temperature fields down to 1 km resolution. Many suggestions were made with a view to advancing the science in this area, and 10 specific action items were agreed upon. These included the selection of several case study days, to coincide with aircraft flights, cloud-resolving modelling efforts and proposals to bridge small-scale and satellite-scale observations. Mapping. Discussion centred on the requirements for surface maps of wind, flux, etc. fields, their resolution and the need to select case studies. It was agreed that this was an area for close collaboration between the surface flux group, the mesoscale modellers and the aircraft radar group using their most innovative analysis techniques. Modelling and parameterization. Information that the GEWEX Cloud Systems Study program has selected the modelling of convection during COARE for an international numerical model intercomparison study was welcome news for those on the observational side. Two case studies have been selected. Processes of particular interest to the modelling community are the role of mesoscale processes on surface fluxes, particularly deep convection and the effects of downdrafts. Plans are to use a hierarchy of cloud-resolving and limited area models. For the observationalists, it was encouraging to note the fine resolution which is being aimed for, and the advanced state of prototype work. This has already established the importance of coupling these atmospheric models with the ocean mixed layer to obtain the correct atmospheric budget. The timetable plans for a workshop to evaluate the ensuing model datasets against COARE observations during fall 1996. Large-scale space/time observations. The improvement of satellite retrieval schemes is receiving a great deal of effort, with the COARE surface dataset providing validation in some areas. Surface heat, freshwater and momentum fluxes can now be determined to useful accuracy on time scales of 1-3 hours and spatial scales of 30-100 km. Again, the message was very clear from discussion within this session: progress depends critically on close interaction between those refining surface datasets and those who understand the strengths and limitations of satellite products, with the mesoscale modelling community and radar/soundings analysts to bridge the gap. Recommendations were made with respect to modelling needs, and particular concern was raised about the paucity of data assimilation projects being funded within COARE. Concern was also expressed about the diversity in location and quality of satellite datasets, and it was recommended that every effort be made to establish a central archive of all relevant satellite datasets with the best possible calibration, both raw and derived products. It was also recommended that an intercomparison of surface fluxes on all time/space scales be arranged as a first step towards understanding the process of scaling up from point observations, that this be the topic of the next flux workshop, and that the issue of scale interactions be a focus of the next major COARE-wide workshop. Progress made at the Boulder 1995 workshop in cross-group collaboration is encouraging for the state of COARE science over the coming few years and for the successful achievement of the overall COARE plans and objectives.
III. Workshop ProgrammeTuesday 11 July 1995
0815 Welcome, Introduction, Announcements
IV. Precipitation (Dick Johnson and Steve Rutledge - discussion leaders and rapporteurs)The session began with a presentation of calibration issues for rain- sensing platforms including IR, SSM/I, ship radar and gauge techniques. After a short discussion and break, rainfall statistics from the various platforms were presented. These statistics included monthly time series and means, and spatial and temporal statistics. Some of the key goals of this session were as follows:
1.Identify the errors and biases inherent to each sampling platform. Calibration Issues Phil Arkin from NOAA led off with a presentation of the GOES Precipitation Index (GPI) technique. This technique relates cold cloud areal coverage to rainfall. It was noted by Phil that there is an unexplained effect of changing averaging scale on the relationship between cloud cover and rainfall. The relationship between cold cloud pixels and rainfall amounts should be independent of the averaging area. In GATE the 235 deg.K proportional constant used was 3.0, whereas for COARE it was unity. Guosheng Liu from the University of Colorado discussed SSM/I-based techniques (in collaboration with Judy Curry). Energy received at the satellite radiometer is first converted to brightness temperature after which emission and scattering models are applied to deduce hydrometeor profiles. Beam filling is accounted for before a final rainfall rate estimate is made. The multi-wavelength capability of the SSM/I instrument appears to be advantageous since the higher frequencies are most sensitive to scattering by ice and the lower frequencies are most sensitive to emission from water. Hence, the SSM/I technique looks at column depth properties and how they relate to rainfall. The footprint of the SSM/I instrument is a concern at the low frequencies; for example, at 19 GHz the footprint is on the order of 50 km, well below the typical width of a convective cell. The Global Precipitation Climatology Program (GPCP) AIP-III compared different SSM/I precipitation algorithms as they performed during the IOP. Depending upon the assumptions employed in the retrieval algorithm, substantial variations in rainfall totals are obtained. Steve Rutledge presented information on the MIT shipboard radar. He gave an overview of the stabilization method employed on both the MIT and TOGA ship radars. There was rainfall within range of the MIT radar virtually every day this radar was in operation in the IFA, 30/30 days of cruise 1, 29/30 days during cruise 2, and 24/25 days during cruise 3. The Vickers was on station for 30 days of cruise 3 but the radar stabilization failed on day 25. He summarised several meetings regarding the ship radar operations and calibrations that were held prior to the workshop. As a result of these meetings and some post calibration work to the MIT radar system, the reflectivities from this radar were increased by 2.4 dBZ to account for a discrepancy in the antenna gain figure (1.8 dB) and a bandwidth term that was not in the SIGMET version of the radar equation (0.6 db). The stability and repeatability of the MIT calibration data was noted during all of COARE. Rutledge also discussed calibration methods employed during COARE which included near-daily receiver calibrations, sphere calibrations (three), and solar calibrations (several). No further modifications of the MIT reflectivity data are planned at this point. Dave Short gave a discussion of the TOGA radar operation during COARE, and NASA efforts to adjust the reflectivity data from this radar during the post-COARE period. Early calibrations during COARE were complicated by hardware/software incompatibilities, unsteady ship power during cruise 1, and a few hardware failures late in cruise 1. For the second and third cruises of the PRC No. 5/TOGA radar, the reflectivities of the TOGA system were stabilized and generally fell within +/-2 dBZ of the MIT radar, for common echoes. The present version of the TOGA data for cruises 2 and 3 were calculated by adjusting the TOGA reflectivities to match the area-time averaged MIT data. Cruise 1 reflectivities are still uncertain. A satisfactory adjustment of the cruise 1 reflectivity data may not be possible. Based on drop size distributions from Kapingamarangi Atoll, current Z-R relationships are Z=374R^1.43 for stratiform rain and Z=139R^1.43 for convective rain. NASA has produced a beta-level product for AIP-3 which consists of a merged rainfall field for the MIT and TOGA radars for the 20-day overlap periods of cruises 2 and 3. Maps are 2x2 km, 10-minute rain rate data (snapshots, not averages), for a total of 93 days in the IFA. Mike McPhaden (NOAA/PMEL) gave a presentation on Optical Rain Gauges (ORGs). He discussed the calibration history of these instruments. Not all of the ORGs used in COARE had pre-COARE calibrations in natural rain. Since several of the ORGs were returned damaged after COARE, even post-COARE calibrations are not possible on some instruments. A calibration effort for the ORGs being conducted at NASA-Wallops needs to be finished. McPhaden discussed error sources for the ORGs which include electronic noise (0.1-1.0 mm/hr, cosine response due to rocking (~1%), vibration (negligible), sea spray (negligible), variations in drop size distribution (~10%), calibration bias (15-30%), and manufacturer's QC (+/-20-30%). Obviously some of these errors are quite significant. Frank Bradley reported on ORG/siphon gauge studies. At the 1994 Toulouse Workshop, large discrepancies between ORG and conventional (siphon and funnel) gauges were reported. Subsequently, Bradley and Paulson (1995, unpublished manuscript) undertook an evaluation of rainfall measurements from various instruments during COARE. At the Boulder Workshop, Frank Bradley reported on a pre-COARE calibration (September 1992) of the ORG used aboard the Franklin against the Bureau of Meteorology funnel gauge at Canberra and during a 1994 Indian Ocean cruise. The two gauges compared extremely well at the site on land. Most of the ORG discrepancies, however, have been reported aboard ships underway. Indeed, without any corrections applied, the siphon gauge aboard the Franklin reported nearly half the accumulated rainfall as the ORG during COARE. Further study suggested, however, that the siphon gauge catch is underestimated during strong winds due to the Jevons' effect, namely, the deflection of raindrops away from the capture area by flow distortion around the gauge. Using Koschmeider's (1934) correction table for the Jevons' effect, Bradley corrected the siphon catch values, with the final result that the ORG and corrected-siphon accumulated rainfall amounts came into good agreement, provided wind direction relative to the ship was not from astern of the beam. Some questions were raised at the workshop about the applicability of Koschmeider's correction data to the COARE region, considering that drop sizes in COARE are very large and deflection by flow around gauges may be minimal. Additional calibration of the gauges in a rain tower and tests of the sensitivity of the ORG measurements to ship vibration were conducted by Bradley, but no discrepancies or sensitivities were found that could account for anything close to a factor of 2 in the rain rates. However, imperfect cosine response to rainfall angle could produce over-estimates of 15-20%. Chris Fairall pointed out that this could be corrected, given the wind field. (Subsequently, using Bradley's results, Fairall found that the Moana Wave IOP-average rainfall dropped from 11.2 mm/day to 9.4 mm/day. Rainfall Statistics Several speakers presented rainfall statistics and associated methodologies. The following is a brief report of the presentations and discussion, and a table outlining everyone's current best estimate for the COARE IOP-average rainfall, which appears at the end of the section. Arkin reported that satellite IR and Visible data have been combined to generate leg-averaged rain maps. Nine different algorithms have been used to obtain estimates to compare with the radars. For instantaneous rain rates, SSM/I is far superior to the others. ECMWF data is not representative of the observed rain. 0.5x0.5deg. spatial maps of rain rates are possible. Arkin gave IOP rainfall estimates of 6-10 mm/day for GPI and ~7.0 for MSU. Liu pointed out that rain-rate maps from SSM/I for the four-month mean reveal that the highest rain rates were in the IFA. Low rain rates were found to the northwest. Comparision of the SSM/I, mixed and sounding estimates showed good agreement in the IFA on five-day time scales. The Curry/Liu estimate is an IOP mean of 5.6 mm/day and other SSM/I estimates range from 4-11 mm/day for the radar area. McPhaden reported that daily averaged rain rates from the Atlas ORGs had correlation coefficients of 0.2 to 0.4 at 1 deg to 3 deg separations. Correlations are even better at 5 to 10 day averaging. GPI estimates are generally lower by 50% than the ORG averages. IOP averages were 8 mm/day on the equator, 16 mm/day at 2S, 156E and 13 mm/day over the IFA. For the MIT radar out to the 120 km range, Rutledge and Rickenbach gave mean values for the three Vickers legs:
10 Nov -10 Dec4.14 mm/day Dave Short, reporting on the merged TOGA and MIT radars, gave the following estimates, out to the 150 km range:
Leg 12.8 mm/day The NASA group calculates R from Z in polar coordinates, then smooths the R field to Cartesian, using a 2-km x 2-km grid. The CSU group first smooths Z to Cartesian, then calculates R from Z-R relationships after the reflectivity field has been subdivided into convective and stratiform components. CSU also employs a 2x2-km grid. The latter (CSU) procedure would tend to produce higher rainfall rates than the former (NASA) by 10-20%, thus helping to explain some of the differences in rain amount results. Convective rain accounts for 65-80% of the net rain and stratiform rain is 20-35% of the net rain. Different methods of analysis lead to different results in sounding budgets, as discussed in the Recommendations section below. Briefly, Xin Lin and Dick Johnson from CSU obtained an IOP mean of 5.7-6.1 mm/day, while Bill Frank from PSU reported 10.5-11.8 mm/day using the same soundings data. Since there were large discrepancies in budget determinations of surface rainfall rate over the IFA, different objective analysis schemes, including the Barnes objective analysis scheme and the optimal interpolation scheme, were used to generate gridded data. The IOP-mean rainfall rates estimated from these gridded dataset appear to be very similar over the IFA (ranging from 5.7 mm/day to 6.3 mm/day. Surface flux data predicted by the ECMWF model are corrected by buoy data over the IFA and the IOP-mean evaporation over the LSA can be obtained. There was a band with minimum evaporation along the equator over the western Pacific warm pool and the mean evaporation rate was around 3.0 and 4.0 mm/day. It increased polarward north and south of the equator due to increasing trade winds. The IOP-mean rainfall rate over the LSA clearly indicates a double-ITCZ structure at about 5 deg N and 5 deg S west of 165 deg E. The IFA was located within a minimum rain band along the equator. To the east of the IFA was a broad rain area with rainfall rate of 10 mm/day, consistent with satellite OLR data. Clayton Paulson gave the Wecoma rainfall from the siphon gauges (uncorrected for wind effects) as 3.1, 12.5 and 10.1 mm/day on legs 1, 2 and 3 respectively, with an overall mean of 8.6 mm/day. The Wecoma's ORGs read about 70% higher. Chris Fairall reported an IOP-average of 11.2 mm/day by the Moana Wave's ORG (corrected down for cosine over-estimation to 9.4 mm/day as discussed above). The Franklin was in the IFA for a total of 43 days in two cruises before and after Christmas, but not during the December WWB. Frank Bradley gave the ship's ORG average as about 11 mm/day. Chris Williams and Ken Gage discussed the use of the Aeronomy Lab profilers to obtain rainfall by measuring the vertical velocity of droplets. The profiler operates up to 12 km and the maximum vertical velocity of water droplets is 6 m/s, observed at 4 km. Lower velocities are seen above and below. They distinguished between stratiform precipitation, mixed stratiform/convection, deep convection and shallow convection. A comparison with the Japanese radar on Manus Island is planned. Paul Willis reported on the work at NOAA/AOML on Electra PMS drop size distributions (DSDs) which show a shift to larger drops as rain rates increase. They are using this data to verify the DSD response of Eldora, and also propose calibrations against the P3s and radar ships. Flights at 500 m are used. Colleen Leary and Tim Doggett noted the availability of a number of their MIT radar rainfall products recently posted on the TCIPO web site (http://www.coare.ucar.edu): radar reflectivities; freshwater accumulations for periods ranging from 20 mins to 6 hrs using the GATE relationship Z=230R^1.25, animated loops and time series of area-averaged rainfall rates at 10-min resolution. They have also performed analyses of rainfall during the December WWB, and showed a 100-km cross-section of a freshwater puddle on 25 Dec, extending to 20-m depth. During a short discussion of freshwater budgets, Frank Bradley illustrated the large effect of advection on freshwater distribution over the small area surveyed by the Franklin as she worked around a drifting buoy drogued at 20 m. The budget was closed satisfactorily when the ORG rainfall data was used for P-E, or if the buoy's siphon gauge was corrected for the wind effect using Koschmeider's result, but otherwise functioned poorly. In an attempt to reconcile the various differences between the rain measurements, the following recommendations were formulated. It must be kept in mind, though, that the various rain estimates were produced by a very diverse instrument suite, ranging from point measurements by the siphon and ORG gauges, large area samples at crude resolutions compared to the scale of convective precipitation elements (SSM/I, GPI, MSU, sounding budgets) and radar, which provides high resolution measurements (2-km averages) within the field of view, nominally 120-150 km. (The ISS technique will be another point sample when these rainfall statistics become available.) Each platform/technique has its own calibration/bias/error issues as well. Recommendations 1.Work on the GPI estimate. Work should proceed in pinning down the relationship between cold cloud and rain for the COARE area. Is the 3 mm/h -235 deg K coefficient best for COARE? A recommendation was made that IR and ship radar data be used to fine-tune this coefficient. 2.Work should proceed to reconcile the small differences in the rain statistics presented for the MIT and TOGA radars. It appears that these differences can be explained by the fact that the NASA group computed rainfall to a range of 150 km whereas CSU calculated rainfall to only 120 km. The effects of beam broadening are probably a factor here. Also, there are differences in the methods used to compute rainfall by these two groups. Although the two groups use exactly the same Z-R relationships for convective and stratiform rain, there are differences in how the rainfall is partitioned between convective and stratiform components, and when the correction for attenuation is applied. NASA applies the attenuation correction in polar coordinates before interpolating to a Cartesian grid, whereas CSU applies the attenuation correction on the Cartesian grid. Work will continue between these two groups to consider the overall effects of these differences. 3.Efforts should be made to conduct a detailed post-COARE calibration of all ORGs. Some of the ORGs were returned from the field in a damaged state so post-COARE calibration is not possible on all ORGs used in the field phase. The Wallops intercomparison-calibration exercise should be completed and the findings reported to the community. The applicability of the Koschmeider study (1934) for correcting siphon gauges for wind effects to the COARE area should be examined. This study developed a correction based on rather small drop sizes associated with light rain situations at high latitude. Certainly, the larger drop sizes in COARE might be less subject to catch effects due to wind. Also, it was recommended that recent WMO reports on rain gauge corrections be consulted. To examine the spatial variability in time averages from the various gauge measurements, we plan to simulate the catch of a point gauge within the field of view of the radar to test the effects of using a point sample to map rainfall when sharp gradients in rain rates exist. 4.The vertical structure of rain could introduce a small error in the radar estimated rain rates. The radar-based rain rates are computed at a nominal height of 2 km (in order to map rain over as large an area as possible). If reflectivity increased from this level to the surface (most likely in convective regions only), the rain estimate at 2 km would underestimate that at the surface. A statistical study of the vertical reflectivity gradient in COARE convection is presently being undertaken. At this time, it is thought that this correction will have a minimal (~10% or less) effect on radar-based rain estimates. Drop-size data collected by the NCAR Electra (being analyzed by Paul Willis of NOAA/ERL/HRD) will be used to test whether the Z-R relationships over the COARE IFA region are representative. The Z-R relationships presently being used to develop the radar-rain estimates were generated from drop-size data at Kapingamarangi atoll. 5.The radar data from the Keifu Maru, although only two weeks in duration in the IFA, can be used to intercompare with the MIT reflectivity data. 6.Rainfall estimates from atmospheric heat and moisture budgets were presented by Lin/Johnson and Bill Frank. Lin and Johnson used three different techniques to come up with a 5.7-6.1 mm/day average rain rate for the IFA. Frank employed a line-integral technique to obtain about 10 mm/day IOP average for the IFA. The large discrepancy between these two groups may be related to a procedure used by Frank to set to zero negative rain rates larger than a specified value. This procedure has the effect of introducing a positive bias to the average rainfall estimate. Bill Frank plans to recompute the rain rates without using this procedure and will report his revised results. Both the Lin/Johnson and Frank groups obtained similar rain rates for the OSA (8-9 mm/day). Lin and Johnson found values smaller than this for the IFA. This is consistent with independent studies of rainfall over the COARE domain (involving ECMWF, SSM/I data) that show that the IFA is at a local minimum of precipitation for the four-month IOP. 7.Finally, we recommend that detailed intercomparison studies be conducted to focus on certain time/space intercomparisons. The following periods are recommended: Cruise 3 (29 January -19 February), Cruise 2 (21 December -9 January). We also recommend that radar, satellite, point and budget estimates be intercompared in detail for the radar area. References: Koschmeider, H., 1934: Methods and results of definite rain measurements. Mon. Wea. Rev., 62, 5-7. TOGA COARE International Project Office, 1995: Summary Report of the TOGA COARE International Data Workshop, Toulouse, France, 2-11 August 1994. TCIPO/UCAR, Boulder, Colorado, 170 pp.
The following table summarises the various estimates of COARE rainfall. Each should be read in conjunction with the above text.
Observer Method Cruise and mean (mm/day) IOP mean (mm/day)
Rutledge/ MIT Radar 1 4.1
Rickenbach (120 km) 2 5.5
3 5.2 5.0
Short et al. merged TOGA and 1 2.8
MIT for cruise 2,3 2 4.6
(150 km) 3 3.2
Curry/Liu Satellite SSM/I 5.6
(for radar area) 4-11
Arkin Satellite GPI 6-10
Satellite MSU ~7.0
Lin/Johnson Budgets 5.6-6.1
Frank Budgets 10.5/11.8
McPhaden Buoy ORG 12-14
Paulson Wecoma siphon 1 3.1
2 12.5
3 10.1 8.6*
Fairall Moana Wave ORG 11.2
(cosine corrected) 9.4
Bradley Franklin ORG 1 17.72
2 6.81 11.65
Gage/ ISS (to be determined)
Williams
* Siphon estimates not corrected for wind effect.
V.Diurnal Cycles (Chris Fairall and Peggy LeMone - discussion leaders and rapporteurs)This session concentrated on the diurnal changes that take place in atmospheric and oceanic boundary layer, and upon those processes (namely wind, convection, clouds, and radiation) that influence the changes in the interfacial layers. The talks revealed that convection and its rainfall, and consequently, related atmospheric and oceanic boundary-layer processes, are affected by processes throughout the troposphere. Chris Fairall opened with composites of the diurnal variation of sea surface temperature (maximum 1400), air temperature (maximum 1900), sensible and latent heat fluxes (maximum 1300), and stress (maximum 0300), based on a grand average of data from the Moana Wave. The mean air-sea temperature and specific humidity differences were 1.5deg.K and 6.5 g/kg^1 respectively. The variation of the directly measured fluxes was well captured by the bulk method. A second analysis was done, adding data from the Franklin and Wecoma (bulk variables only). The data were stratified according to wind. In the light-wind case (less than 3 m/s, a surrogate for fair weather) the diurnal trend was the strongest and cloudiness the least, so that the diurnal variation was forced by stronger variation in solar flux and mitigated little by vertical mixing in the sea. On the other hand, the diurnal variation was the least for most variables with the strong wind cases (greater than 6 m/s, a surrogate for disturbed weather). Enhanced mixing of the upper levels of the ocean and greater cloudiness helped to reduce the diurnal changes. In this case, some diurnal variations associated with deep convection were observed. The timing of the air temperature maximum (average amplitude about 0.3 deg C) stimulated considerable discussion. LeMone noted that a sunset maximum would be expected for a constant-depth mixed layer with net heating during the day. However, Chen noted that her analyses had revealed maxima early in the afternoon, and McPhaden noted the frequent occurrence of double maxima during the day: one in the late morning, and a second in the afternoon. Several noted that correction for radiative heating of the instrument and its environment (housing, tower, ship, buoy) was needed, and wondered whether it was being done correctly. Carol Ann Clayson showed results of a numerical model of the evolution of the upper levels of the ocean, with atmospheric variables over COARE as input. The diurnal variation of the sea-surface temperature (at -5 cm to be consistent with Fairall's results) revealed the same patterns: strong diurnal variation in lighter winds, less diurnal variation with stronger winds or when continuous precipitation stabilized the upper levels of the ocean. When impulses of rain fell in strong winds, the model mixed the water down rapidly. The modeled depth of the ocean mixed layer matched observations, with values around 60 m in stronger winds, and of the order of 10 m in light winds. Peak surface warming was about 2.5 deg K at 1 m/s. In a comment concerning this problem, Roger Lukas showed observations of the diurnal variation of the temperature difference between depths of 3 and 5 m below the ocean surface. He showed that the thermal stratification was unstable between midnight and sunrise, and it became stable in the afternoon. He noted that compilation of data on salinity is underway. The diurnal variation of convection and associated precipitation was summarized by Shuyi Chen, based on her work with IR satellite data, and by Tom Rickenbach, based on examination of radar and lightning data. Chen showed the large mesoscale convective systems reached maximum extent during the night, but had a secondary peak in the afternoon. Such systems could originate at any time of the day, but were about twice as likely to be initiated the previous afternoon. Smaller convective systems, being more short-lived, peaked shortly after their afternoon inception. She associated the afternoon inception with afternoon heating and moistening. Rickenbach's analysis of MIT radar data showed a very clear semidurnal cycle for convective rain but a diurnal cycle for stratiform. His measurements of stratiform precipitation and lightning patterns peaked during the early morning which, as expected from Chen's work, were associated with larger convective systems. Convective precipitation had an early morning and afternoon peak. Chen's and Rickenbach's talks stimulated considerable speculation about the peaks in convective occurrence. Short suggested that mergers might account for growing of the larger convective systems in the afternoon. LeMone cited the importance of windshear for growth of larger systems, and recalled Kuettner's idea that stronger winds, which enabled stronger moisture flux, might favor growth of more convective cells closer together and increase the chance of mergers. McPhaden inquired about the possibility of the influence of tides on the semidiurnal variation of convection, and Rickenbach alluded to Lindzen's work (1978) which suggested a positive feedback between deep convection and the semidiurnal tide. Leslie Hartten showed the diurnal variation of the low-level profiler-derived wind vectors in the COARE region, noting that diurnal variation of convection over land (Kavieng) might have influence over the convergence several hundred kilometers from the island, suggesting another factor which might influence diurnal variation of convection at least at stations near significant land masses. Grossman and others pursued the idea of reinforcement of convection by the destabilization of the cloud top by infrared radiation after sunset. It was suggested that this was dynamically more important for the stratiform rain, thus explaining its diurnal (rather than semidiurnal) variability. An additional idea was supplied by Mapes, who discussed the diurnal variation of temperature and humidity. He began by showing the balance of solar and longwave heating profiles in clear sky conditions. He then discussed the role of radiation in the diurnal warming of the boundary layer. He also pointed out that the temperature in the upper troposphere reached a minimum 1.0deg.K lower than the diurnal maximum: if the boundary-layer diurnal amplitude is smaller (and Fairall's data suggests it is), air parcels rising from the boundary layer will have larger buoyancy and nighttime Convective Available Potential Energies (CAPEs) should be larger, favoring convection. Mapes also noted an early morning (0800- 0900 local time) minimum in the boundary layer and lower troposphere humidity, and speculated that the drier air might be due to radiation-induced subsidence. George Young considered that it might be simply due to turbulent transport. Zipser was also skeptical about Mapes' explanation. He followed with his own analyses of soundings which showed maximum CAPEs at night. However, Zipser presented evidence that the humidity data at the lowest levels of the soundings still had problems. The surface observations that accompany the soundings show a surprising uniformity across the network (Ta about 28.2 in the day and about 27.8 at night with Td about 24 either time). However, there are big differences in the actual soundings from different locations. Thus, CAPE was only about 200 J/kg at the north end of the network and near 2000 J/kg at the south end. This led to a lunch working session to discuss problems with the COARE soundings. It was recommended that an effort be made to explain and resolve these inconsistencies in the soundings, possibly using the aircraft profiles. Peggy LeMone and Bob Grossman have a subset of the data available. Ed Zipser will continue to pursue this problem. Reference: Lindzen, R.S., and J.M. Forbes, 1978: Boundary layers associated with thermally forced planetary waves. J. Atmos. Sci., 35, 1441-1449.
VI.Atmospheric Forcing of the Ocean (Clayton Paulson and Roger Lukas - discussion leaders and rapporteurs)
Presentations A theme that ran through the early presentations was the multiplicity in structure and scale of the phenomena which contribute to processes at the ocean surface. Shuyi Chen addressed the question of the wide variability of scales of convective systems and their interaction. She contrasted the 100-200 km horizontal scale of the generic anvil structures to the group of convective systems associated with an ISO which can be 10,000 km in extent, in effect spanning one-third of the globe. She described the detail of these systems during COARE, the active superclustering of eastward-moving convective systems, combined with the shorter-lived IR less than 208 deg K cloud clusters which propogate westward with clear air in between. Peggy LeMone pointed out the intrinsic relationship between convection and tropospheric shear, which provides the supply of air to maintain the system. In the analysis of mesoscale convective systems it is now common to distinguish between those which develop with their orientation predominantly parallel or predominantly perpendicular to the low-level shear. Shear-perpendicular systems tend to be two-dimensional, rather intense systems which propagate rapidly with strong winds at the surface. They are accompanied by a large "cool pool" of descending air, typically 0.5 deg C cooler than the environment, thus enhancing both heat fluxes and surface stress. Shear-parallel systems tend to be associated with shallow, weak, low-level shear and travel slowly, thus producing a different boundary-layer response and smaller flux changes. LeMone stressed that these terms are just convenient working definitions. Ed Zipser illustrated the Jan/Feb 1993 Mesoscale Convective Systems, as determined from SSM/I images. He pointed out that the COARE IFA sampled only five of a vast number of such systems in the general region, 18-19 Jan and 10, 17, 20 and 22 Feb. Potential temperature can be an indicator of the source of air masses, from inflow or outflow regions for example, or air which has simply been left in the wake. In discussing the ocean response, Bob Weller referred to salinity structure on 11 Feb after a storm with 16 m/s winds. A freshwater lens on the surface has its buoyancy enhanced by solar heating so that wind momentum can be trapped in a shallow surface layer. There are two possible responses to a sudden squall under these conditions: the wind may overcome the stable density gradient and mix the freshwater down, or the lens may be swept along the surface like a hockey puck (Grossman's analogy). Good radar rain pictures, to obtain the spatial coverage of such an event, combined with estimates of the near surface currents will help to clarify the processes involved. Steve Anderson discussed the response of the ocean surface over a longer time scale, to the competing influences of salt and temperature stratification, and the processes of wind and convective mixing. He showed that different mixed layer depths could exist simultaneously for salinity, temperature and density using profiles for 27 Nov. The bottom of the mixed layer was defined by changes: Delta S = 0.0133 psu, Delta Theta = 0.03 deg C or Delta rho = 0.01 kg/m^3. The net surface buoyancy flux was predominantly positive during COARE. The effect of freshwater flux was illustrated with results from the Price, Weller and Pinkel surface mixed layer model (1986) with and without rainfall included. With no rain, nighttime convective overturning deepens the mixed layer to 80 metres; including the rainfall amount measured on Moana Wave, deepening was limited to 40 metres, in good agreement with observations. With double this rainfall, the mixed layer would shoal to only 30 metres depth.
Clayton Paulson described the characteristics of freshwater lenses in terms of differences from the surrounding ocean surface. For lenses newly formed by heavy rainfall, Delta T = 1 deg C, Delta S = 7 psu, horizontal scale Delta x = 20 km and vertical scale Delta z = 10 cm. An important goal would be a statistical, dynamical model to describe the movement and mixing of fresh lenses. The scale of raindrop size can also influence mixing.
Ed Walsh showed SST data from the aircraft radiometers, associated with mean square slope measurements made with the scanning radar altimeter during the very light wind period on 28 Nov, the first platform intercomparison day. Under these conditions, the spatial pattern of SST was remarkably persistent. Along a 65 km track, both the overall distribution of SST pattern and fine-scale variations of 0.1-0.3 deg C, on horizontal scale of order 3 km, were almost unchanged when overflown one hour later. Roger Lukas also reported on spatial variation of ocean temperature (0.2 deg C) and density (st=0.1 kg/m^3) with scales of 8x8 km, observed from Moana Wave at night. He questioned whether such anomalies with the potential to inhibit mixing and trap heat the following day, could influence the place where convection started. Recommendations The following items requiring action arose from the presentations and subsequent discussions:
1.Idealized convective cases (e.g., a squall line) could be used to test the sensitivity of ocean models to different kinds of forcing and would be useful for interpretation of forcing by more complicated fields. The paper by Young et al. is a source for idealized flux fields.
2.Bob Grossman suggested case studies for which aircraft coverage was extensive. These include: 27-28 Nov, 16 Dec, 9 Jan, 16-18 Jan and 9 and 10 Feb. Conditions for these days range from "severe blue" up to class 3 convection. Additional possibilities for case studies include 17 Feb and 22 Feb, the two days which have been simulated by Jean-Luc Redelsperger. Priorities for the above might depend on the availability of ship radar and other data. It may be worth noting that cases selected by the modelers for intercomparison lie outside the IFA. 3.Use the ship rain radars to bridge small-scale and satellite-scale observations. The combination of the radar rain data and models could provide realistic ocean forcing fields down to scales of a km. 4.Explore the sensitivity of the upper ocean to time/space variations of fluxes. This overlaps item 1. What processes produce the observed space-time variations in the upper ocean? One concern is the extent to which variations in turbidity may affect solar heating and thereby SST. There may already be papers on this topic. 5.Compute space/time statistics from buoys, ships and aircraft. Auto-and cross-correlation and/or structure functions and other statistics should be computed. In addition to insight into processes, the results would bear on the issue of temporal vs. spatial resolution. 6.The flux group should influence and cooperate with the cloud-resolving and mesoscale modeling efforts. 7.An attempt should be made to close the ocean salinity budget to constrain the rainfall rate estimates. Modelling by Steve Anderson suggests that such an effort might yield useful results. 8.Models which predict the diurnal variation of SST should be refined. These include PWP used in the bulk flux code, the model presented by Clayson and models embedded in GCMs. This item overlaps item 4 above. Additional matters discussed include: 9.Compare SST measured by satellite with measurements of near-surface temperature by ships, buoys and aircraft. A comparison of individual nighttime passes over the IFA with underway measurements from Wecoma is planned as part of this effort. Comparison of aircraft and underway Wecoma measurements might also be fruitful. 10.The need for surface fields of velocity for estimates of bulk fluxes was discussed. Surface velocity from a tidal model is anticipated and this could be melded with longer term velocities. Ship drift data should be taken into account in determining surface velocity fields. As partial motivation for the efforts listed above, we note the following: 11.GCM fluxes need improvement. 12.Satellite flux estimates need validation. The promising start by Clayson and Curry of estimating fluxes down to 30-km separation and 3-hr intervals from satellite observations needs validation and development. 13.Fluxes down to 1-km scales are needed for ocean modelling. Reference: Price, J.F., R.A. Weller, and R. Pinkel, 1986: Diurnal cycling: Observations and models of the upper ocean response to diurnal heating, cooling, and wind mixing. J. Geophys. Res., 91, 8411-8427.
VII.Discussion of Mapping, Aircraft Doppler Analysis and Data Assimilation (Bob Weller - discussion leader)Some of the questions which had been posed (see section III) to generate discussion in this section had been addressed during the course of the previous two topics (e.g., the existence of spatial structure in the ocean, flux enhancement due to squalls). But questions of cause and effect and of scale interactions remain open. Discussion centred on the ways in which our understanding of ocean-atmosphere coupling might be advanced, and the requirements for surface mapping in this respect. Could we get away with fairly broad-scale maps? Weller pointed out that during a period of successive squalls during late January and early February, the ECMWF forecast model had the ocean gaining heat, counter to intuition and the IMET observations. So our mapping must resolve squalls. This regenerated discussion on the importance of understanding just why the ECMWF surface analysis performs poorly, both in the sign of the net flux and timing of events (a subject to be revisited in topic IX). In response to questions about the tools available to obtain high-resolution wind maps, and how to select days for case studies, Dave Jorgensen described results obtained from co-ordinated flights by the two NOAA P3s using their radars in "quad-Doppler" mode. This technique enabled much better estimates of radar reflectivity and the 3-D wind field, particularly the vertical velocities. Jorgensen illustrated case studies of mesoscale convective systems (MCSs) of 6, 9, 20 and 22 Feb which were different in nearly all their characteristics, particularly momentum transport. Wind fields were obtained down to 500 m above the surface, the lowest height possible. It is clear that with this and other new techniques of operation and analysis being developed as a result of COARE, and the unique concentration of airborne and ground-based radars deployed during the experiment, the possibilities for high-resolution wind and rain mapping are promising. The above cases are being studied intensively by Jorgensen, LeMone, Smull, Zipser and others. The aircraft group will work with the mesoscale modellers to obtain surface values, and with the flux group to establish time series. It is understood that Frank Marks (AOML) has already compared samples of the aircraft rainfall data with the ship radars. The mesoscale modellers and surface flux investigators agreed to work together to produce a set of case studies which would provide the flux fields from the models for a westerly windburst, a small-scale deep convective event, and other events typical of the IOP. These case studies would assist in the development of maps of the surface flux fields and provide the oceanographic modelling community a means to begin investigating, in high resolution, the response of the upper ocean to the diverse forcing seen during the IOP. This topic closed with discussion about our goal in this work. We must identify suitable cases to work on and extend our single point observations over space to force the ocean, then couple the system with the aim of improving the climate model. Finally, George Young reported on Dave Stauffer's work on data assimilation, using the NCAR/Penn State MM4 model with soundings data as input. Not much is known about the status of this approach for COARE; it may be useful for synoptic scale flux mapping. Latent heat flux is reasonably predicted between squall lines. Because flux enhancement is not resolved the regional flux averages may be somewhat low. This again generated some discussion on the scales needed for flux mapping. A typical space average for the mean observables in the bulk flux algorithm would be a few km. For our work on fluxes, and the need to capture cloud activity, the higher resolution of the MM5 model would seem to be more appropriate.
VIII.Modelling and Parameterization (Mitch Moncrieff, Jean-Luc Redelsperger, and George Young - discussion leaders and rapporteurs)The enhancement of evaporation and sensible heat flux by atmospheric mesoscale systems has been estimated in long term observations from the TOGA TAO buoy array over the equatorial Pacific Ocean, including moorings in the COARE domain. Mesoscale enhancement is due primarily to the lack of wind steadiness on synoptic time scales and is associated with periods of significant precipitation.
The magnitude of the mesocale enhancement of monthly-averaged sea surface evaporation is ~10% or less of the total evaporation. The mesoscale enhancement of monthly averaged evaporation fields can reach 30% of the total during occasional periods with weak and variable winds over the western Pacific warm pool and the other major precipitation zones in the equatorial Pacific. It is necessary to more completely quantify the role of mesoscale processes on surface fluxes, devise ways to represent these effects in general circulation models, and test the resulting parameterizations in these models. The effect of deep convection on surface fluxes is particularly relevant to TOGA COARE because of its decisive role in controlling the sign of the surface energy budget. These are the most comprehensive datasets presently available over the tropical oceans. As regards the role of numerical modelling in understanding this interaction, it is essential to consider both short-time scales (less than a day) and long-time scales (up to a month). To fully investigate these aspects across the range of scales involved it is necessary to use a hierarchy of cloud-resolving models and limited-area models. On short time scales, three-dimensional models can be used to determine convective structure and transports and evaluate these data in detail against observations so that the coupling of deep convection to the enhancement of surface fluxes can be understood and parameterized. This type of process study is best performed as a set of case studies. As far as surface fluxes are concerned, the role of downdrafts is an important convective process to be evaluated (e.g., to improve parameterization of both convective fluxes and surface fluxes). Note that downdrafts affect both the thermodynamic properties and mesoscale wind fields now being incorporated in coupled convective/surface process parameterizations. To understand the effects of these processes on the coupling of the ocean/atmosphere system on climate scales, convection must be handled more explicitly than it is in GCMs. On these longer time scales it is not presently practical to run three-dimensional models. This is because it is essential to have reasonably sophisticated microphysics and radiation schemes in these models as well as the long-duration integration period. The two-dimensional models are particularly valuable for producing synthetic datasets of the effect of convection on the scale of the entire TOGA COARE IFA. These data can, for example, be compared with observationally-derived Q1 and Q2 quantities, precipitation and cloud amount derived from radar and satellites. Morever, the datasets can be used to determine the effect of cirrus anvils on the surface fluxes (radiative interaction) that is significant on climate time scales. Prototype calculations have already been performed. It has been established that the coupling of these models with the ocean mixed layer is almost certainly necessary to obtain the correct atmospheric budget. Nested regional models (with parameterized rather than with resolved convection) are useful to investigate the problems that will be encountered as GCMs evolve toward finer resolution. An advantage of these limited-area models for testing the consequences of improved resolution is that they are computationally less demanding than full GCMs. The GEWEX Cloud System Study (GCSS) Working Group IV: Precipitating Convective Cloud Systems (chair, Mitch Moncrieff) has selected the modelling of convection in TOGA COARE for an international numerical model intercomparison study. This can be viewed as a vehicle for performing the above modelling experiments in a way that helps evaluate individual model strengths and weaknesses. The initiative will include evaluation of the model datasets against observations. There are two distinct modelling projects proposed: 1.Three-dimensional modelling of the COARE squall line cloud system on 22 Feb 1993. The integrations will be from several hours to one day (Project leader, Jean- Luc Redelsperger, Meteo France). 2.Two-dimensional modelling convection over the TOGA COARE IFA using IFA-averaged forcing provided by Lin and Johnson (CSU) (Project leader, Steve Krueger, University of Utah). The details of the model intercomparison study will be ready for circulation in late summer 1995. The first intercomparison workshop will be in the fall of 1996. Modelling has advanced sufficiently to serve as a primary tool in resolving the fundamental issues concerning the coupling between ocean and atmosphere that is a central objective of TOGA COARE.
IX.Large-Scale Space/Time Observations (Judy Curry, Bill Rossow and Dick Johnson - discussion leaders and rapporteurs)Extension to larger scales A critical issue in achieving the TOGA COARE science objectives is to document and understand the temporal and spatial variability of surface fluxes in the tropical oceans and its role in air/sea interactions. Although much information must come from in situ observations, obtaining an integrated perspective of multi-scale variations that influence the surface fluxes over the entire domain requires numerical models and analysis of satellite measurements. Large- scale fields of surface fluxes will have application to the following:
a)a spatial context for observations obtained from a given platform in the IFA; Three-dimensional numerical models and satellite remote sensing techniques that have been validated by the TOGA COARE IOP data can then be used to determine surface fluxes over extended periods and over the global tropical oceans. Comparison of the ECMWF model with the buoy and ship observations revealed several shortcomings of the model flux fields. The ECMWF wind stresses are significantly lower than those from the other platforms. The ECMWF shortwave fluxes and precipitation are uncorrelated with the surface observations. The ECMWF latent heat fluxes are on average higher than the surface observations. There is a five-day lag in the SST from the ECMWF surface analysis that has not been accounted for. The most striking result is that, during the third leg when there were many squalls in the IFA, the ECMWF model recorded the sign of the net heat flux incorrectly, as it does not resolve the small scale convective systems that contribute significantly to the surface heat flux. Mesoscale models have the potential to provide substantially improved analyses of flux components on scales ranging from 1 km to the scale of the OSA, incorporating such diverse datasets as radar, satellite, and aircraft into a 4-D data assimulation scheme (4DDA). Before mesoscale models can provide accurate analyses of surface fluxes, the model parameterizations must simulate accurately the relevant physical processes. Cloud-resolving models that are coupled to an ocean mixed layer will provide the basis for improved parameterizations of convection, cloud microphysics, radiation transfer, and the atmospheric boundary layer that can be incorporated into the mesosale models. Improved mesoscale models can then provide improved analyses of surface fluxes. It is recommended that the three related modelling activities--cloud-resolving model simulations, simulations using mesoscale models, and 4-D data assimilation--proceed in parallel (rather than sequentially), because of the high priority of 4-D data assimilation products. A concern was raised that there is an insufficient number of data assimilation projects that are being funded for COARE. Determination of surface fluxes from satellite have seen major advances in the past few years as a result of COARE. Previously, only monthly-averaged values of surface flux components averaged over large areas were made, with unacceptably large errors. Recent results show that satellite retrieval of all components of the surface heat, fresh water and momentum fluxes can be determined to a useful accuracy on time scales of 1-3 hours and spatial scales of 30-100 km. These new advances were made possible by use of the following:
a)improved physical models to determine the flux components from input variables determined from satellite, A summary of the individual surface flux (and input) parameters is as follows: 1.Sea surface temperature. Values of skin SST are being directly retrieved from satellite using AVHRR data. Direct retrievals of skin SST are still plagued by cloud discrimination problems, but are believed to be accurate to within 0.5 deg.C. Use of the ATSR dataset from the ERS-1 satellite will improve the retrievals. Since the direct retrievals can only be conducted under clear sky conditions, a new algorithm has been proposed, tested, and validated to supplement the clear sky retrievals using the results from an ocean mixed layer model to impose a diurnal cycle on the SST time series. The diurnal cycle parameterization is a function of daily-averaged wind speed and rainfall rate, and peak solar insolation (all of which can be derived from satellite data). 2.Surface radiation fluxes. Using the most recent version of ISCCP-derived cloud properties (from AVHRR satellite data), surface radiation fluxes are determined using a sophisticated radiative transfer model. Ongoing improvements include:
a)use of temperature and humidity profiles obtained from soundings and improved determination of aerosol optical depth, Comparisons of satellite-derived fluxes with ship-measured values show biases in the downwelling shortwave flux of about 18 W/m^2 (the accuracy of the ship measurement is 15 W/m^2) and in the downwelling longwave flux of 5 W/m^2 (the accuracy of the ship measurement is 5-10 W/m^2. The discrepancy in the downwelling shortwave flux appears to be associated with patchiness in aerosol optical depth related to biomass burning in New Guinea. 3.Precipitation. GPCP recently conducted an intercomparison of satellite precipitation algorithms for retrievals obtained during the TOGA COARE IOP. Because of remaining uncertainties in the radar dataset, the absolute accuracy of the satellite products remains uncertain. It is anticipated that forthcoming improvements to the radar rainfall rate dataset will clarify this issue. The most promising method for obtaining rainfall rates on the time/space scales required for COARE is to combine SSM/I microwave data with the visible and infrared radiances measured by AVHRR. It is also anticipated that the MSU data can provide useful measurements of precipitation, particularly when merged with the other datasets. 4.Surface turbulent fluxes (sensible and latent heat, momentum). The required input variables for bulk models are surface wind speed, air temperature, water vapor mixing ratio, and SST. Surface wind speeds can be determined using SSM/I and ERS-1 scatterometer data. Since both instruments are on polar orbiting satellites, combination of the datasets is required to provide winds over the diurnal cycle. New techniques have been developed to determine surface air temperature and water vapor mixing ratio, by assuming that the atmospheric surface layer structure is reflected by the cloud characteristics (including precipitation) determined using AVHRR and SSM/I data. Concerns were raised at the workshop over the diversity of satellite datasets, the lack of a central source for the satellite data, and calibration differences among different versions of the same datasets. The workshop participants concurred that every effort should be made to archive all relevant satellite datasets with the best possible calibration in a central archive. Derived satellite products should also be archived, with the datasets containing all input variables used for determining the satellite parameter. This will allow future investigators to use the products to determine the parameter using different methods and make use of the input parameters for different applications. The following satellite datasets are deemed useful for TOGA COARE: AVHRR (GMS, polar orbiters), SSM/I, HIRS2/MSU, ATSR, scatterometer. Concerns were raised about calibration of the AVHRR radiances and cross-calibration between different satellites. This issue is being addressed by the ISCCP project, and the calibration information needs to be disseminated to the appropriate COARE PIs. Several different versions of the SSM/I dataset are circulating. The most accurate dataset is processed by Remote Sensing Systems, Inc. The ATSR is a new dataset on the ERS-1 satellite that has limited availability. Subsequently, Dave Legler pointed out that some satellite data, including ERS- 1 scatterometer data, AVHRR data, and SSM/I data can be obtained through the JPL Physical Oceanography Data Center (JPL PO-DAAC). However, the status of this site as a formal NASA data acquisition centre, and whether this dataset has the "best" calibrations, navigations, etc., in terms of the preceding paragraph, are unknown at this stage. To improve the intercomparison of surface fluxes made from different platforms and on different space/time scales, a systematic intercomparison is proposed, whereby a spectrum of time and space scales is specified, and each platform provides estimates to the extent possible for this scale. This intercomparison will be the first step towards understanding the scaling-up of the surface fluxes from point measurements to larger scales. Space scales
1.IFA (budget studies, satellite, objective analysis of all surface-based measurements, ECMWF analyses), Time scales
1.Entire four-month period (budget studies, satellite, surface buoys, ECMWF analyses), It is recommended that this intercomparison be the topic of the next flux workshop, and that the issue of scale interactions be the focus of the next major TOGA COARE workshop.
Appendix A: List of Acronyms
4DDA 4 Dimensional Data Assimilation AIP American Institute of Physics AOML Atlantic Oceanographic and Meteorological Laboratory ATSR Along-Track Scanning Radiometer AVHRR Advanced Very High Resolution Radiometer CAPE Convection and Precipitation/Electrification Experiment COARE Coupled Ocean-Atmosphere Response Experiment CSU Colorado State University DSD Drop Size Distrubutions ECMWF European Centre for Medium-range Weather Forecasting ERL Environmental Research Laboratories ERS Earth Remote-sensing Satellite GATE GARP Atlantic Tropical Experiment GCSS GEWEX Cloud System Study GCM Global Change Model GEWEX Global Energy and Water Cycle Experiment GOES Geostationary Operational Environmental Satellite GPCP Global Precipation Climate Program GPI GOES Precipitation Index HIRS High-resolution Infrared Radiometers Sounder HRD Hurricane Research Division IFA Intensive Flux Array IMET Improved METeorological Instrumentation (WHOI, US) IOP Intensive Observation Period IR Infrared Radiation ISCCP International Satellite and Cloud Climatology Program ISO IntraSeasonal Oscillations JPL PO-DAAC Jet Propulsion Laboratory Physical Oceanography Data Center LSA Large Scale Sounding Array MCS Mesoscale Convective System MIT Massachusetts Institute of Technology (US) MSU Microwave Sounding Unit NASA National Aeronautics and Space Administration (US) NCAR National Center for Atmospheric Research (US) NMC National Meteorological Center (NOAA, US) NOAA National Oceanic and Atmospheric Administration (US) OLR Outgoing Long-wave Radiation ORG Optical Rain Gauge OSA Outer Sounding Array PMS Particle Measuring System PRC People's Republic of China PSU Pennsylvania State University (US) PWP Pacific Warm Pool SIGMET Significant Meteorological Information SSM/I Special Sensor Microwave/Imager SST Sea Surface Temperature TAO Tropical Atmosphere Ocean Array TCIPO TOGA COARE International Project Office TOGA Tropical Ocean Global Atmosphere WMO World Meteorological Organization WWB Westerly Wind Burst
Appendix B. Electronic Access Points to COARE Flux Data
Florida State University surface meteorology archive (Dave Legler) web http://www.coaps.fsu.edu/coare/ ftp coaremet.fsu.edu Pennsylvania State University radiation archive (Tom Ackerman) web http://wwwarc.essc.psu.edu/data/togacoare/TOGACOAREdataindex.html/ ftp ftparc.essc.psu.edu Texas Tech University high-resolution freshwater flux (Colleen Leary) ftp tcdm.coare.ucar.edu directory /pub/COARE_DATA/air_sea_fluxes/Freshwater_Fluxes Soundings data - OFPS/CODIAC (Jim Moore) web http://www.ofps.ucar.edu/codiac/codiac-www.html Colorado University SST archive (Bill Emery) ftp bilbo.colorado.edu/pub/coare JPL Physical Oceanography Data Center (JPL PO-DAAC) email: podaac@podaac.jpl.nasa.gov web http://podaac.jpl.nasa.gov NMC/NCAR Reanalysis products - late Sept 1995 (Muthuvel Chelliah) email: wd52mc@sgi39.wwb.noaa.gov phone 301-763-8227 TOGA COARE Data Management (Richard Chinman) web http://www.coare.ucar.edu ftp tcdm.coare.ucar.edu
Appendix C. Participants and Addresses
Steven Anderson
Phillip Arkin
Steve Bolen
Frank Bradley
Dan Cecil
Shuyi Chen
Richard Chinman
Paul Ciesielski
Robert Cifelli
Carol Ann Clayson
Judith Curry
Charlotte DeMott
Lynn DeWitt
Jennifer Dickey
Arthur (Tim) Doggett
Beth Ebert
Warner Ecklund
William Emery
Steven Esbensen
Christopher Fairall
William Frank
Kenneth Gage
Wojciech Grabowski
Chris Greb
Bob Grossman
Pat Haertel
Leslie Hartten
Hiroshi Ishida
Richard Johnson
David Jorgensen
Lakshmi Kantha
David Kingsmill
Colleen Leary
Margaret (Peggy) LeMone
Xin Lin
Guosheng Liu
Christopher Lucas
Roger Lukas
Larry Mahrt
Brian Mapes
Michael McPhaden
Mitchell Moncrieff
Masato Murakami
Tetsuo Nakazawa
Atusi Numaguti
Noriko Oki
David Parsons
Clayton Paulson
Charles Pavloski
Walt Petersen
Jean Luc Redelsperger
Thomas Rickenbach
William Rossow
Steven Rutledge
Tom Saxen
Wayne Schubert
Rong- Shyang Sheu
David Short
Shawn R. Smith
Bradley Smull
Gilles Sommeria
Moira Steyn- Ross
Jielun Sun
Otto Thiele
Stan Trier
Osamu Tsukamoto
Hiroshi Uyeda
Edward Walsh
Thomas Warner
Peter Webster
Robert Weller
Gary Wick
Christopher Williams
Paul Willis
Xiaoqing Wu
Pingping Xie
George Young
Yunyue Yu
Yuanchong Zhang
Edward Zipser 
  
http://www.ncdc.noaa.gov/oa/coare/meetings/bldr_rep.html Downloaded Thursday, 08-Jan-2009 19:00:18 EST Last Updated Tuesday, 31-May-2005 12:25:39 EDT by David.P.Smith@noaa.gov Please see the NCDC Contact Page if you have questions or comments. |