Wayman Baker (NOAA/NWS) opened the meeting and welcomed the participants to Frisco. Debra Hallmark reported on the status of the on-line minutes from the Oxnard, CA meeting of February 2001. She requested that electronic versions be provided as soon as possible for the Frisco meeting. Wayman Baker led a review of the action items from the Oxnard meeting:
Steve Mango (Integrated Program Office - IPO) gave an update on the status of NPOESS. The NPOESS Preparatory Program (NPP) mission is set for 2005 and will carry three of the five major NPOESS payload instruments. The CMIS contractor down select is imminent. Risk reduction contracts with the two major candidates for system integrators are in place.
An aerosol/cloud sensing polarimeter may be added to NPOESS based on a late established climate requirement. This leads one to ask whether winds and DWL could also be a late addition. The NPOESS Airborne Sounder Testbed (NAST) could provide a means of demonstrating the capability of and need for DWL. Steve indicated another possibility: the predicted 2013 ìNPOESS Liteî mission that will make up for a schedule conflict with METOP. NPOESS Lite will be under-populated with instruments, so DWL sensor accommodation could be possible.
Reiterating the need for winds, Steve expressed satisfaction that the GTWS data requirements are being developed in via a NASA/NOAA partnership. He noted that the Integrated (IORD) would soon be revised. Global winds remain at top of the list. The Pre-Planned Product Improvement list totals 21 EDRs. Windsat Coriolis is being tested via airborne validation campaign. Windsat data (surface winds over oceans) are to be available to the research community.
Bob Menzies (JPL) presented the status of the NASA New Millennium Program (NMP) for Chris Stevens (who is assuming the job of NMP Program Director.) There is NMP interest in winds. For example, the EO-3 mission will carry GIFTS, permitting derivation of winds from water vapor motion. The Navy provides the vehicle and launch while NASA provided GIFTS and control. First to be applied to CONUS, GIFTS will later be moved over the Indian Ocean. NOAA will derive meteorological products and handle data archival.
Bob reviewed the January 2001 Earth Science Enterprise(ESE) technical workshops, noting several that might relate to lidar, such one on large deployable telescopes. He commented that NMP validation support flights are to be partnered with science missions, and technology will be chosen to amplify the science missions with which they are partnered. NMP has identified science requirements for lidar aerosols, CO2 and other chemistry topics as well as winds. For example, DIAL is the driver for large deployable telescopes, and powerful laser transmitters for DWL.
Ron Ticker (NASA/HQ) and John Pereira (NOAA/NESDIS) reported on programmatic aspects of the joint NASA/NOAA Global Tropospheric Wind Sounder (GTWS) mission studies.
Ron reviewed the objectives of GTWS‚exploring the usefulness and feasibility of a science data buy. He discussed the establishment of the GTWS Executive Steering Team and the formation of the Science Definition Team (SDT), noting their intention to have Web-posted draft requirements in September 2001. A Technology Assessment Team (TAT) has also been created. He summarized three items that need to be in place to proceed with a GTWS:
John reviewed NOAAís role in monitoring GroundWinds DWL development. He listed work done recently, and noted that several presentations on the GroundWinds demonstration campaign are scheduled on the Frisco agenda. He indicated that new data collection campaigns being conducted to investigate aerosols, turbulence, and other phenomena using GroundWinds NH. He briefly discussed the development of the GroundWinds Hawaii instrument.
John commented that NOAAís approach to a wind data buy is to work in concert with NASA. He emphasized the need to complete and analyze OSSEs to assist defining wind data requirements.
Bob Atlas (NASA/GSFC) and Jim Yoe (NESDIS/ORA) summarized the recent activities of the GTWS SDT. Foremost among these was the February, 2001 data requirements workshop.
With the objective of defining a GTWS mission that will be ìuseful, doable, and affordableî, the workshop featured both plenary talks and closed discussions. The state of the current wind observing system and the prospects for enhancing it with DWL and other new systems such as GIFTS were discussed. Researchers reviewed the use of winds in a variety of disciplines, some of them indicating the need for very challenging wind measurements, such as high-resolution data to capture divergent processes, or profiles extending above 20 km for chemistry investigations.
Following the workshop, a sub-committee summarized the input into a preliminary draft, designed additional OSSEs at DAO to complement those being conducted at NCEP, and organized a July, 2001 SDT meeting in Boulder. There the OSSE results were discussed, revisions made to the first draft requirements, and actions assigned to meet the goal of publishing a version for public comment in September.
The requirements themselves were not shown in the interest of fairness. However, a ìblankî table was presented to give an idea of how science and operational requirements were being reconciled at both threshold and objective levels. Finally, Bob explained the need for new OSSEs based on a high resolution Nature run (NR); the ECMWF NR was too smooth and underestimated features such as jet streaks. Also, it was too short for science applications. The new NASA/NCAR NR has .5-degree resolution, is of 3.5 months duration, and includes hurricane landfall.
Eugenia Kalnay recommended updating vorticity on smaller scales which would improve the analysis more than wind would itself. However, to do so, wind components, or line of sight winds from multiple perspectives would be required.
Dave Emmitt commented that to derive vorticity, resolution and accuracy have to be considered jointly. Otherwise, one might end up trying to extract derivatives from inaccurate winds, and achieve inaccurate derivatives as a consequence.
Rod Frehlich asked whether the GTWS requirements are valid for an NPOESS mission. The answer is no, at least not for a threshold mission, although the SDT chairs would prefer to see a mission that could be scaled to NPOESS orbit.
Pierre Flamant (CNES) presented an update of ESAís Atmospheric Dynamics Mission (ADM) Activities. ADM is designed to use both molecular and aerosol direct detection channels at 355 nm in a 400 km orbit. ESA is aiming for a 25% duty cycle with a single LOS perspective. ADM is to be a demonstration mission. ESA recognizes that the single LOS is not optimal, but contends that it will provide an excellent opportunity to learn. Planned height resolution is 250 m and 1 km, respectively, for the PBL and free troposphere, and perhaps 2 km in the stratosphere. A three-year life is specified.
Pierre discussed science results related to ADM, including the VALID campaign. He noted that the VALID location caused many undesired orographic effects. On the other hand, cirrus was abundant, which was favorable for the direct detection systems. In thick clouds, the 10-micron system excelled. Large discrepancies in the wind observations were found sometimes, which must be understood in terms of representativeness (from jet stream extremes, for example). New signal processing studies devoted to non-stationary conditions to improve frequency estimation have been undertaken, and an error covariance study by KNMI. ESA is planning to look at representativeness issues using the WIND instrument in the Euro-THORPEX campaign over North Atlantic in 2003. (Note: WIND is a CO2 TEA DWL being installed on the Falcon 20 aircraft.)
Technology development for ADM is being pushed on several fronts. Two contractors are working instrument concepts, soon to be down-selected. There are also two contractors working on the laser transmitter; it is acknowledged that the 25% duty cycle in space is challenging for the technologists.
Steve Mango asked how there could be parallel development for the DWL instrument and the laser transmitter. Pierre indicated that each contractor had been given relevant, consistent specs as needed.
Bob Brown inquired whether ESA considered a data buy approach to ADM. Pierre replied that it had not.
John Pereira asked how ADM would achieve operational data delivery. Pierre indicated that with two ground stations (Kiruna, Sweden and Point Barrow Alaska), it should be possible to achieve 3-hour delivery.
Bob Menzies questioned whether aerospace corporate mergers in Europe will pose a risk to ADM. Pierre thought not, as two big contractors may prove more competitive than several small ones.
Upendra Singh (NASA/LaRC) reviewed the NASA lidar systems strategy team report. Because several NASA lidar missions (SPARCLE, PICASSO-CENA, and the VCL) all had laser problems, Ghassem Asrar independent review panel to identify technical challenges and make recommendations. The panel concluded that NASA needs a strategy to develop lidar technology for space missions. In the past it has relied too much on mission PIís to develop their own technology. Bill Heaps (GSFC) and Upendra Singh were assigned to develop the NASA laser strategy, which theyíve sought to make independent of missions (and PIís).
Numerous missions with lidar requirements have been identified, including:
Key technologies for these missions have been identified as well:
Upendra argued that there is a need to develop 2 micron and 1 micron (which can be frequency multipled) technology. Common problems include issues related to: laser diodes, laser damage, frequency control, efficiency, heat removal, ruggedness and lifetime, and contamination tolerance. Upendra recommended sorting out the respective roles of the DoD, DoE, NASA, and industry in addressing these problems. Then NASA could efficiently leverage efforts by the other. Three specific recommendations were made:
Upendra emphasized the wide range of laser applications that could be realized by progress in a few key technologies, and expressed the conviction that this could be done for relatively modest investment. Steve Mango asked about the timeline for accomplishing the three recommendations. Upendra estimated four years.
Farzin Amzajerdian (NASA/LaRC) presented a paper on improving the receiver sensitivity of coherent lidar systems. The objective is to find a way to ease requirements for power or telescope size. Farzin noted that a vast amount of literature exists on this topic. His approach is to express the transimpedance transfer function (TTF) in terms of detector parameters, and he derived an alternative SNR equation in these terms. The optimum LO power is then determined by setting the derivative of the SNR function with respect to the SNR to zero. Farzin argued that the commonly used SNR equation overestimates lidar sensitivity; this is offset, though, by optimizing the LO power. He maintained that he has derived more accurate formulations describing the performance of heterodyne receivers, and that the approach of optimizing the LO power will lead to increased sensitivity.
Bob Menzies expressed some reservations, as did Gary Spiers. The parties agreed to discuss the matter offline. Rod Frehlich noted that the broadband formulation used in this analysis runs counter to intuition, since coherent DWL sensitivity depends on narrowband SNR.
Frank Peri (NASA Earth Science Enterprise Technology Office‚ESTO) discussed Laser and Lidar Technology roadmaps for NASA. These roadmaps are intended to support outstanding needs, working in parallel with the NMP. The ESTO approach is to work on missions that define ìenablersî, then tackle the enabling technologies independently.
For 2003-2007 development efforts ESTO is focusing on laser and lidar enablers. These are related to science drivers in chemistry, winds, and altimetry, in concurrence with Upendra Singhís presentation. Projected budgets for lidar transmitter development showed same tasks listed by Upendra. Frank noted that the total needs of ~$20 million is probably too much for ESTO to bear alone.
Steve Mango asked how ESTO plans to dovetail science capabilities with technology readiness and mission readiness. Frank indicated that the instrument incubation process (IIP) looks at instruments end-to-end. While this talk has focused on the tall poles, the rest of the tent is not being forgotten.
Bruce Gentry asked whether ESTO is receiving external input. Frank replied that ESTO has a person dedicated to maintaining external linkage, especially with DoD. The ESTO also depends on feedback from grass roots efforts like the Working Group.
Zoltan Toth (NCEP/Environmental Modeling Center) reported on the status of the NCEP Observing System Simulation Experiments (OSSEs), which have been undertaken to provide simulated data impact assessments. These are intended to help select configurations of future sensors, and to enable users to assimilate data early in sensor life cycle.
Zoltan reviewed the role of the Nature Run (NR) as ìtruthî in the OSSE process, and the generation of simulated observations from NR. He emphasized the need for calibration to make sure that real and simulated current observations give similar impacts, as a prerequisite to having confidence in the ability of an OSSE system to evaluate impact of simulated ìfuture observationsî. He itemized pitfalls to which OSSEs are subject - poorly simulated observations, identical twin contamination, and focus instrument concept - to underscore the importance of OSSE calibration and quality control of the NR.
Quality control of the ECMWF NR revealed the need for low-level cloud adjustment. This was essential for proceeding with lidar OSSEs. Zoltan noted that the constant sea surface temperature SST in the Southern Hemisphere is bound to affect the apparent forecast skill (and shows up in the calibration of real data Observing System Experiments (OSES) versus OSSE). Calibration of the NCEP OSSE system indicates that it is not perfect, but suitable for proceeding with the bracketing DWL forecast experiments.
Steve Mango commented that metrics need to be pre-selected and applied to ALL instruments in the OSSE series to avoid bias. Eugenia Kalnay noted that OSSEs are also important for optimizing data use prior launch.
Adrian Marroquin (NOAA/FSL) reported on the status of FSLís Regional OSSEs, concentrating on the NR derived from the NCAR MM5. The domain consists of 10 km x 10-km grid with 43 vertical levels and covering North America. The domain boundaries were initialized with the ECMWF NR. One shortcoming was the lack of soil moisture in the NR fields; this was overcome by initializing with climatological values. Another problem is the huge file size --and the number of files, given 15-minute resolution for the 11 day run. Therefore, a quick NR of all basic variables with no mean or variance was first completed July 13, 2001 and is available for examination. An observation extractor has been set up for the NR. and a bias error simulator. Their model does not stray from the ECMWF NR, The next step is to quantify and analyze MM5 minus ECMWF differences.
Matt McGill (NASA/GSFC) reported on the status of simulation activities that are intended to help assess fringe imaging direct detection DWL for NOAA/ETL. As a first step, Matt is trying to simulate what would be seen in a ìrealî meteorological system (i.e., a northern hemisphere frontal passage) with varying clouds, winds, etc. He expects to have simulated winds for this situation by the next meeting.
Matt also has undertaken simulation of fiber optic recycling for fringe-imaging systems. His first-order results support Paul Haysí estimate of a factor of five improvement as a theoretical upper bound. He will attempt to establish the practical (with losses) upper bound next.
Rod Frehlich noted that coherent DWL is limited by atmospheric variability, and asked whether is there any plan to include that in the direct detection simulations. Matt indicated that he is open to this generally, but would not want to do it with this particular model atmosphere.
Eugenia Kalnay (U of Maryland) presented a paper addressing Low Dimensionality of the Atmosphere and the use of ìBred Vectorsî to determine the so-called Errors of the Day. She noted that the 6 h forecast errors of 1958 were similar for the Eastern and Western USA, but were dominated by day-to-day variability ‚ errors of the day. These got better over 40 years, but are still the dominant feature, as seen in the ìspaghettiî ensemble trajectories.
Errors of the day cause the uncertainty in the forecast, and propagate from one day to the next. Eugenia described the use of ìbreedingî to study the unstable regions. The errors of the day are localized in space and time, and breeding allows one to identify them. The BV-dimension can be used to determine if the BVs are orthogonal--low dimensional BV identifies the unstable area. Low dimensionality, high errors of the day have vertical structure and move coherently, so they can be tracked.
Instabilities are intrinsic to background flow and do not occur by chance. The implication for data assimilation is to find a BV surface in best agreement with the model. Only a few BVs ‚ 5 to 10 are critical, and only a few days required to ìtuneî the assimilation system.
Finally, Eugenia stated that knowing the shape of the errors of the day allows observations to be used better. To accomplish this, she stated that she will need to be able to calculate vorticity, which will require multiple-perspective DWL observations.
Steve Mango asked what are the typical time scales of the instabilities seen in the 6-h forecast. Eugenia indicated that five to seven days is the life of these instabilities.
Zoltan Toth (NCEP/EMC) gave a presentation entitled: ìAdaptive Targeting (AT) Past, Present, and Future.î Questions for the future include: whether AT can be used on global scales, if features in the 4-6 day forecasts can be targeted, and whether case selection can be automated. Other issues include those related to deployment of the observations and what instrument design feedback will be accepted from AT.
The usefulness of AT has been based on the ability to produce a set of ensemble forecasts, from which areas of instability could be identified and targeted. The method has been tested in numerous field programs ‚ FASTEX, NORPEX, CALJET, WSR99, WSR00, WSR01, PACJET (winter 2002), and pre-THorPEX, WSR 2002 (Note: WSR = ìWinter Storm Reconnaissanceî.)
WSR is now a fully operational collaboration between NWS field offices and EMC. Each mission is requested by field forecasters to support critical weather forecasts. Among predefined flight tracks, the Senior Duty Meteorologist selects the best. Dropsondes are carried out by AOC and USAF reserve aircraft, and are released 20 minutes or about 20 km apart.
Overall results show ~55% forecasts improved while approximately 15% is degraded. Gains of 12 to 24 h in lead-time are typical.
Michael Kavaya asked if the sensitive area moves w/lead-time. Zoltan affirmed that it does. Wayman Baker noted that there are implications for a low duty cycle DWL such as ADM -- how to choose shots for the greatest benefit. Dave Emmitt expressed the opinion that the WSR data could be valuable for doing data denial experiments. For instance, use a fraction of the targeted observations, or a fraction of each one. Zoltan concurred, and noted that the data are available for this type of investigation. Eugenia Kalnay inquired whether Bred Vectors could be used to automate selection of targeted observations globally. Zoltan agreed they are a viable candidate. Rex Fleming asked whether humidity data from the dropsondes is used. Zoltan indicated; QC of these observations proved them to be suspect.
Bob Atlas (NASA/GSFC) described the new Joint Center for Satellite Data Assimilation. The JCSDA has been created in response to a growing problem: Operational data assimilation fail to use most of the available satellite data. In fact only approximately 1/7 is used. Moreover, it takes two years for a new satellite to become operational in terms of having its data stream used. NASA and NOAA scientists and managers have defined the following JCSDA goals:
The JCSDA was established as a virtual entity with participation of NASA/DAO, NWS/EMC, and NOAA/OAR and NESDIS. The NSF has also been involved. The Acting Director is Ricky Rood (NASA/DAO)with Steve Lord serving as Deputy Director. Thirteen data types have been prioritized and responsibilities assigned among the constituents.
Bob explained how this approach was applied to QuikScat, and showed an example of how scatterometer wind data can make forecast improvements. Michael Kavaya noted the large magnitude of the QuikScat benefit. Bob pointed out that the case show was an extreme one.
Upendra Singh (NASA/LaRC) discussed the development of 2-micron laser transmitter for coherent DWL at LaRC. The issues, challenges, and strategy have changed little since the Oxnard meeting. High among these is development of dual DWL and DIAL capability. Upendra is looking for ways to work with industry, and would like to implement an airborne DWL as a stepping stone to space.
Asked whether the DIAL and DWL could use all of the same components, Upendra replied that two seed lasers would be needed, one for DIAL and one for DWL. Most other components are thought to be shareable.
Dave Emmitt (Simpson Weather Associates) presented an overview of NASAís ìDecadal Visionî for a Geosynchronous Lidar Observatory, a concept he reviewed for ESTO. The concept is aimed 2-3 decades ahead, at notional missions that would revolutionize certain types of science.
With the assumption that 100 m optics would be available and that the measurement will be technically feasible, Dave attempted to predict whether the resulting science would be revolutionary. Potential exists since GEO would give rapid updating as well as the coverage and resolution. However, competing data from both POES and other GEO observations must be presumed, which would erode the margin for revolutionary impact.
For a direct wind measure from GEO lidar to be useful, Dave believes it might need to provide tropospheric and stratospheric winds @ 1 m/s accuracy, and water vapor (via DIAL.) A GEO DWL could target ageostrophic winds, and do off-nadir DIAL. From a GEO platform, it would not be necessary to account for satellite motion and Earth rotation. However, multiple platforms would be required for global coverage. Also, there is very little space or even airborne heritage for this.
Gary Spiers (JPL) elaborated on some of the technical challenges for the GEO Lidar Observatory. Scaling indicates that current laser technology coupled to 100 m telescopes could work in principle. In practice, though, the optics will likely be partial or segmented, and more powerful laser transmitters than are now feasible would be needed to compensate for this.
Variation of the nadir angle to give coverage over the entire field of view is a significant issue. The range to the ground is a function of nadir angle, and that is sensitive to the velocity, i.e., therein lies potential to introduce velocity errors.
A big question is how to steer such a system. The 100 m optics proponents are thinking about imaging, not sounding. The angular momentum associated with such a telescope is so large as to preclude a step-stare approach. Gary predicts that it would be necessary to copy the Arecibo radio-telescope design and use a moving secondary, so that at any one time the full aperture of primary optics will not be used. Pointing stability and knowledge would also pose big challenges. Finally, Gary noted that as nadir angle increases, so does atmospheric extinction, so that at 8 degrees off nadir, three orders of magnitude of SNR will have been lost; it will be challenging to maintain coverage toward the poles.
Dave Emmitt proceeded to discuss some DIAL issues for the GEO lidar concept. He foresees the benefit of GEO DIAL might be realized through targeted observations complementing other measurements, with GEO DIAL operated as a ìslave to a global modelî. Dave recommended OSSEs be performed to confirm the potential of this approach.
Eugenia Kalnay questioned whether this vision would allow horizontal gradients to be measured? Dave responded affirmatively, with the qualification that overlapping multiple satellites would be necessary.
Asked how the GEO DWL product would compare to current GOES Cloud Track winds, Dave replied the current observations are accurate to 3 m/s at best, thus 1 m/s accuracy proposed for GEO DWL should represent a significant improvement.
Bob Brown (U of Washington) spoke on the wind data user perspective. He cautioned that users simply want winds and care little about technology.
Bobís opinion is that winds have remained low priority for NASA, because it is presumed that wind fields can be provided by General Circulation Models (GCMs.) However, he pointed out that GCMs often miss storms that were detected by observing systems. GCMs often underestimate wind strength, too. Finally, GCM scales canít handle turbulence or subgrid-scale organization. Directly sensed winds are need for studies of the PBL, fluxes, etc.
He reminded the group that there be competition for winds from space: the SSMI, scatterometers, and GIFTS as well as DWL.
Some wind observations are already provided by scatterometers and by SARS. Passive radiometers such as WindSat will also come into play, and ESA will launch a DWL. Bob remarked that empirical curve fitting often brings wonderful results for scatterometers even when theory fails, and he predicted that the same thing would happen with DWL. However, Bob thinks microwave instruments (active and passive) at low heights and improved GCMs in the free troposphere will give competition to a DWL.
Bob Atlas remarked that the theory for inferring the winds above the PBL would not work for the ageostrophic wind component. Bob Brown downplayed the problem, arguing that those cases can be readily identified.
Dave Emmitt showed results from a numerical investigation of the effect of single versus dual perspective DWL measurements. The preference for assimilating LOS has led some to assume that a single LOS perspective is sufficient for NWP. The experiment Dave described, dubbed a ìrapid responseî OSSE (RROSSE), was undertaken at the NASA/DAO to examine the validity of the assumption.
The LOS velocity is of little use to end-users, but it is convenient and valuable for use in data assimilation. It is desirable to look carefully at various ways of putting lidar LOS's in a space--how many perspectives, how orthogonal, and how close they need to be.
Results were presented for a perturbed background state in a single level analysis. A NASA/DAO background error covariance model was used, with Monte Carlo simulations for 2000-km x 300-km domain. The experiment was set to favor the single LOS perspective, in that the flow was non-divergent. However, it was found that for a fixed total number of LOS in the domain, dual perspective led to smaller RMS errors (by a factor of approximately 2) than single perspective. When the background errors were made more realistic, bi-perspective LOS did even better (i.e., using mixed error covariance), reducing the rms by a factor of six.
Dave showed that the paired perspectives did not need to be very close, and that the need for orthogonal perspectives is quite relaxed ‚ angles between 25 and 155 degrees do OK. These results are in conflict with the OSSE done for ADM but in line with NCEP preliminary DWL OSSE work, and will be used to support the GTWS SDT data requirement determination.
Rod Frehlich asked about the magnitude of random errors assumed in the study. Dave indicated that they were 0.5 m/s, which Rod deemed to be on the small side.
Eugenia Kalnay questioned the effect of specifying a perfect error covariance mean, which would be like having prior knowledge of the errors of the day. Why doesnít this give better results? Bob replies that this was likely due to the non-divergence. Several comments cautioned that the results should be verified with more full-fledged OSSEs.
Rod Frehlich (U of Colorado) addressed the verification of DWL velocity error. He reviewed definitions of error and accuracy for benefit of those engaged in preparing and validating data requirements. Error is based on application; therefore, if anticipating a step-stare DWL, the instrument will stare across a random wind field. The determination of accuracy requires definition of truth: the linear average over the measurement plane at that instant.
Rod pointed out that data users have in mind TOTAL error, while engineers think in terms of instrument or measurement error. He argues that his definition of truth permits a straightforward separation of these terms.
He presented analysis of representative errors for troposphere horizontal winds using GASP data. He showed that there are cross-track and along track differences, and recommended that data requirement specifications be written to reflect that. Finally, he argued that there is no need of in situ data to determine random velocity error, due to the domination of random error by the atmosphere, which varies with application. Rod sees a need to refine requirements of bias and random error, and to determine climatology of critical regions such as the middle troposphere.
Dave Emmitt commented that pumping up the uncertainty in one and not the other does not matter to LOS data assimilators, but does to people who in essence want COMPONENTS, including some of NASAís potential science data users. Gary Spiers indicated that some of them want u and v components, which leads to varied weighting of along and cross track. Michael Kavaya pointed out that a customer might want both to have low error; to satisfy them, one must either reduce the LOS accuracy requirement to less that 1 m/s; or limit to viewing at 45 degrees to ensure such a result.
Dave Emmitt discussed cloud statistics and implications to shot accumulation for space-based DWL. This work is funded by IPO to aid with realistic treatment of clouds in OSSEs, especially the new NASA/DAO nature run.
He reviewed cloud climatology work, focusing on LITE data, because LITE was less biased toward or away from clouds than most airborne missions. He noted that HIRS is giving a lot of insight globally, and has ability to look at thinner and thinner cirrus-- just like lidar, in some sense. HIRS sees clouds about 80% of the time, mostly high and tropical ‚ over 50%. LITE adds insight by penetrating many of the high clouds that HIRS sees. Cloud free to surface only 20% of the time; to 2 km 40% of time (if PRF = 10 Hz).
Daveís objective is to determine how to set a specification for what cloud situations are reasonable for a DWL to meet a given wind specification. Acknowledging that it is hard to maintain DWL ìtechnical neutralityî in doing so, he recommends that both high and low cloud specifications should be made.
Bruce Gentry noted LITE extended to 57 degrees latitude, and asked how global significance could be extracted. Dave replied that it canít be, strictly speaking, but since a DWL probably will have the biggest impact in tropics these are relevant statistics.
Dave was also asked whether he had considered cloud porosity as function of nadir angle. He replied that aircraft show a very weak dependence.
Milt Huffaker (Coherent Technologies Inc.--CTI) reported on their recent progress with wind lidar technology and systems.
Airport atmospheric survey is their current commercial market, but CTI is still interested in global wind monitoring. They are also interested in airport capacity improvement, as well as turbulence detection and aircraft safety, and environmental (pollution) monitoring.
CTI continues doing 2-micron development, emphasizing LuAG for space-based DWL. Working with conduction cooled laser rods, theyíve built a 250-mJ demo for NASA (pre-SPARCLE). Milt hope to propose and fabricate a multi-J unit soon.
CTI has also been supported for investigating 10 Hz TMHOYLF DWL for the International Space Station/Japanese Experiment Module (ISS/JEM). Initial risk reduction tests were planned for Summer, 2001. The most challenging factor for this effort is the 1 kW input power limit on the ISS.
Milt described CTIís next generation lidar for clear air turbulence detection a 10 mJ, 10 Hz, 2 micron instrument. It features a very stable injection seeded operation, is virtually chirpless and highly efficient. They expect to have automated alignment eventually, and manufacturing techniques that will be much cheaper for commercial marketing.
Ground-based CTI systems have been used to track aircraft wakes at a number of airports. A movie loop of Aspen airport winds was shown. Occurrences of turbulence were apparent. The system was also used at Dallas Fort Worth Airport, and has tracked wakes at Edwards AFB. CTI lidars also have been used to monitor winds for forest fire-fighting management near Denver in 2000. The data helped forecast select time/place to drop retardants.
Airborne deployments now or soon will be used to support a number of programs, most notably to support JEM Airborne Flight tests (2001).
Eugenia Kalnay asked how CTIís DWL products compare with radar Doppler winds. Milt pointed out that DWL has been used to calibrate the radar wind profilers. Given this answer, Eugenia wished to know whether there are plans to use CTIís instruments for targeted observations. The answer was affirmative.
Geary Schwemmer (NASA/GSFC) spoke on HARLIE and HOE scanning telescopes. He described recently developed non-Doppler volume scanning aerosol motion drift technique, and presented a movie of successive 36 s scans about a 45-degree cone. Comparisons of these winds to radiosondes have been done in Utah, NH, and OK. Gary recommends this as a complement to Doppler technique, to provide ground-based validation and help provide local turbulence scale.
A wide-angle video camera was used to give local cloud track winds, too, with cloud heights determined by lidar, offering another source of winds to intercompare. VHS tapes are converted to mpegs, then clouds are traced out using GUI software. A Sine fit is made, adjusting the amplitude and phase in the least-squares sense. Jim Yoe question asked whether this was automated, to which Geary replied that to date each guess fit is interactive and labor intensive. But he has developed automated approach to go beyond the wave fitting technique for future trials.
Results show fairly good agreement with sonde speed. The camera and HARLIE direction agree better with each other than they do with the sonde direction, at least in some cases. Some outliers that are difficult to explain or confirm. For example, he showed a bifurcated retrieval, and speculated that it resulted from looking at two sides of a vortex.
Jim Ryan (University of New Hampshire--UNH) presented an update on the GroundWinds (GW) program. He provided an overview of a variety of topics, noting that subsequent speakers would go into deeper detail.
Since the Oxnard meeting, there has been some new science observations, additional analysis of September LidarFest data, and lots of repair and retrofits at the Bartlett Observatory. GW operations have been handed off, from MAC to UNH and Mount Washington Observatory.
A new radiosonde system (the old one was on loan) has been installed at Bartlett. A new chief scientist for MWO is coming on board, plus Ivan Dors as GWNH manager, and additional operators have been trained to facilitate larger duty cycle.
Maintenance includes: new laser table support; GW is now immune to casual misalignment. Telescope-laser alignment is now secured. Realignment no longer necessary with rotation. Field-testing in August is planned, at which time UNH will implement automatic alignment (this should also increase duty cycle). There is new cooling system, not plagued by leaks. All of these improvements (and others) carry into the GW Hawaii design. There are facility upgrades coming at Bartlett: a new dome, paving of the access road, and power conditioning (using smart UPS).
Finally Jim discussed plans for GW Hawaii. The big differences are the transition from ML 532 to 355 nm, more compact optics, onsite and remote operations. Numerous other improvements are based on the NH experience. Site preparation work at Mauna Loa will be complete in December 2001.
Wayman Baker asked whether UNH could give a schedule for moving beyond GroundWinds to a space-based DWL. Jim replied that this is hard to do, since the laser remains as a significant open issue; lasers as reliable and fault tolerant as needed arenít readily available for ground-based applications, let alone in light, space-qualified versions. Moreover, Jim indicated that it would be virtually impossible to present a realistic schedule until the issue of a data-buy is resolved.
Geary Schwemmer inquired as to the source of the GroundWinds New Hampshire laser. Jim replied that the instrument was provided by Continuum, noting that their support had been good, even though the instrument was not reliable as needed for space. A brief discussion of potential users of the GroundWinds Hawaii data rounded out the presentation.
Carl Nardell (Michigan Aerospace Corporation--MAC) reviewed the key technologies of the GW DWL: fiber optic photon recyling to improve system throughput and efficiency. This destroys spatial but preserves spectral coherence of the signal. He also discussed the role of the CLIO optic and the streaking CCD camera detectors. There are several known issues, including CCD read noise, etalon finesse, and system throughput that is not up to what was predicted. The last is attributed to poor CLIO surface quality, low QE of the CCD, lower fiber optic injection efficiency than thought, and large laser power variation. Thus, the primary objective for GW NH is to go do efficiency test and analysis.
Carl noted that all results posted so far exploit only the first fringe, so recycling is not being used to full advantage. There are recycling program issues, too. For example, if lots of aerosols are present, itís better to recycle aerosol than molecular.
GW alignment was sensitive tuned to first fringe, and this caused bias in other fringes. Version two will be tuned to all fringes, will use all. But the system must be more stable than it was in September 2000 time-frame. Craftsmanship of CLIO, Fiber optics must be very high, and need the stability. This was not appreciated in the first instrument, in order to realize the recycling advantage.
Wayman Baker urged Carl to estimate how quickly GroundWinds might be made space-ready in the absence of funding constraints. Carl ventured as early as 2005 - if all the stops were pulled. He indicated entail no intermediate aircraft mission, which he deemed more challenging than a mission in many respects, especially for a direct detection system.
Mike Hardesty (NOAA/ETL) presented an update on activities related to analysis of the ìLidarFestî. Using re-processed (as discussed by Carl Nardell) GW molecular channel data from 9/28/01, Mike illustrated how GW velocity bias has been diminished and variance improved. Re-processed 9/27/01 data look very good to high altitude, with much lower variances and improved correlation between various DWLís than preliminary analysis had indicated.
The LidarFest data set also illustrates certain cloud effects for DWL. For example, the aerosol channels (of GW and the ETL mini-MOPA worked well, while the variance of molecular channel velocity estimates increased dramatically on other (top) side of cloud layers. The data set also provides an opportunity to compare real and modeled relationships of the LOS velocity standard deviation versus the number of photons used to make velocity estimates for the direct detection systems.
To summarize the lessons learned (so far) from the campaign, and issues for subsequent analysis, Mike listed:
Mike discussed new work using the mini-MOPA, including NH3 Plume detection. He is preparing to use mini-MOPA dual DIAL/Doppler to estimate fluxes in the future.
Dave Emmitt asked for a description of the LidarFest backscatter profile. Mike replied that the aerosol was generally quite low and the free troposphere very clean.
Michael Kavaya asked whether velocity error vs photocount agreed with the literature. Mike replied that this part of the analysis has just been undertaken and it is too early to state conclusions.
Rod Frehlich asked whether one always finds a flat floor in spectra at high frequency. Mike replied that this was so, regardless of which spectral estimator was applied.
Bruce Gentry (NASA/GSFC) discussed continued analysis of LidarFest data from the Goddard Lidar Observatory for Winds (GLOW). GLOW LOS velocity data first posted to the intercomparison database were derived with a ìquick lookî retrieval algorithm that used a fixed sensitivity that did not account for temperature sensitivity of the wind estimate properly. A new algorithm was developed to include temperature correctly and the LidarFest data were reprocessed using the U.S. standard atmosphere for the temperature. Subsequently, he determined that the real temperature (determined from rawinsondes) often differed significantly from the standard atmosphere value, resulting in some residual velocity errors, but these should be less than 1 m/s. Bruce mentioned several other refinements to be made to the GLOW data, including:
Steve Mango asked about GLOW errors versus altitude especially above the tropopause. Bruce noted that, in general, GLOWís error bars which are determined by taking the standard deviation of a number of consecutive LOS wind profiles do appear to increase with altitude. However, thre are frequently cases where in fact they decrease. This is interpreted to be an indication of the atmospheric variability which often appears to be higer in the region of the tropopause. Steve asked Bruce to say more about the temperature dependence of wind estimates. It was noted that the bad news is that GLOW needs T to within 3-5 K to reduce velocity errors to less than 1 m/s, although independent temperatures of that quality should be available. The good news is that the sensitivity to temperature of the GLOW data indicate that temperature may be obtained directly from the measured data.
Mike Hardesty asked whether Bruce could sort out aerosol and molecular processing. Bruce thought it might be possible, since the aerosol has no temperature dependence; it might be possible to optimize them independently.
Gary Wilson noted that error bars included all data in the observing period despite temporal variability. Has the temperature variability with time been investigated? Has not been done yet, but Mike and Rama will show data where some of the intercomparison work has been done.
Rama Mundakkara (CIRA and NESDIS/ORA) summarized intercomparisons of the most recently processed LidarFest Data (for all instruments) for 9/25/00-9/2/00. Fundamentally, all of the DWLs were able to provide reasonable wind estimates (with reasonable agreement) when sufficient signal was available. GLOW and GW tend to agree well from about 6 km (the lowest GLOW height gate) to ~11 km, where the molecular signal for GW became unreliably weak.
Rama has also plotted wind error versus photon counts for the direct detection lidars, finding that errors increase sharply (for GLOW and GW both) when the photon count falls below about 103. He cautioned that these results are subject to revision as the instrument teams post subsequent iterations of re-processed data.
Michael Kavaya asked whether wind differences between pairs of systems approach small numbers near ground for small averaging times. Rama indicated that some did for GW and the mini-MOPA. However, the view is incomplete, since the mini-MOPAís coverage did not extend far beyond the aerosol rich boundary layer, and GLOW was gated off below 6 km to avoid receiver saturation.
Carl Nardell commented that the sharp increase in GW error at the apparent critical level of - 103 is due to noise in the detector. Lower noise detectors in the future are expected to move lower the critical level for making good wind estimates.
I. Dors (UNH) discussed recent activities with the NH GW facility and instrument. GroundWinds is being used for a number of science investigations now, including an Air Quality campaign (AIRMAP). On April 7, 2001, a dust plume that originated from China was detected using GW. A turbulence detection campaign featuring intercomparison of GW with Air Force Research Lab balloons found good agreement in locating turbulent layer locations up to 10 km. GW velocity time series are being examined for evidence of of gravity waves. Ivan also hopes to use the observatory to investigate Brillouin scattering and to retrieve temperature profiles.
Steve Mango asked whether the dust observations were supplemented with radiosondes. Ivan affirmed that they were, with balloons launched from the GW Observatory.
These minutes were recorded by Jim Yoe of the NOAA/NESDIS Office of Research and Applications.