August 1990

5.0 EXTENDED TIME OBSERVATIONS 75

6.0 COLLABORATIVE EXPERIMENTS 85

7.0 REFERENCES 91

APPENDIX A DESCRIPTION OF AIRCRAFT AND INSTRUMENTATION FOR

CIRRUS IFO-II A-1

APPENDIX B LISTING OF NWS STATIONS AND CODES B-1

APPENDIX C FIRE SCIENTIFIC ORGANIZATION C-1

APPENDIX D STANDARD DATA FORMATS FOR FIRE D-1

APPENDIX E NCDS USERS ACCESS INFORMATION E-1

APPENDIX F ACRONYMS F-1

3.0 SCIENTIFIC OBJECTIVES AND STRATEGIES

and the corresponding cloud physical properties, including,

Numerical modeling is a fundamental component of FIRE research into the physics and dynamics of cirrus clouds and the impact of cirrus on the global climate. Current techniques used to model cirrus clouds in GCMs are extremely crude. Nonetheless, GCM modeling studies have demonstrated the potential importance of changes in the distribution and composition of high clouds to global climate change due to greenhouse gases or external forcing (section 2). Greater reliability and credibility of GCM-based predictions of climate change requires development of physically-based cirrus cloud parameterizations that have been experimentally verified. Because GCMs diagnose the ice and liquid water contents, cloud heights and amounts, and cloud radiative properties of cirrus from the resolved large-scale fields, each of these must be measured as accurately as possible and with as much temporal resolution as possible in constructing suitable verification data sets for testing cirrus cloud parameterizations.

Many processes acting to determine the physical properties and distribution of cirrus clouds operate on scales smaller than those resolved in GCMs (section 2). Extended cirrus systems are typically organized into distinct mesoscale regions of enhanced cloudiness which are often comprised of numerous small scale convective-appearing cellular structures. The small-scale circulations strongly influence the microphysical development of the cloud (Starr and Cox, 19851)). The resultant sizes and habits of the constituent cloud particles determine, to a large extent, the net downward vertical flux of cloud water via gravitational settling. In turn, it is the balance between the rates of ice water generation and depletion, primarily via ice particle fallout, that determines the overall cloud water content and radiative properties as well as the vertical development of the cloud layer and feedback to the larger scale (Starr, 1987a). Clearly, development of a physically-based GCM cirrus cloud parameterization requires knowledge of these processes.

Consequently, FIRE cirrus research includes a strong effort to model cirrus cloud development on smaller scales. Models are used to investigate the basic physics and dynamics governing cirrus cloud formation, evolution, and structure. These include regional mesoscale models nested within larger scale models, very high resolution mesoscale cloud models, and models of fundamental turbulence interactions and microphysical growth. Detailed modeling studies have significantly increased our understanding of cirrus clouds (section 2) and have served to focus the observational objectives for Phase II. In Phase II, these models will serve as surrogates for high resolution data sets to assist in the formulation and testing of cirrus parameterization schemes.

As in the case of GCM parameterizations, the credibility of these models must also be established which requires that suitable initialization and verification data sets be obtained. Accurate measurements of thermodynamic and dynamic conditions at the appropriate time and space scales in conjunction with appropriately detailed and quantitative physical descriptions of the corresponding clouds are essential. In particular, accurate descriptions of vertical motions, the water vapor budget, cloud structure, and cloud composition (especially the populations of small particles) are essential if the uncertainties are to be reduced to an acceptable level. The data sets must adequately resolve the fundamental scales at which the physical processes are organized. Observations of the aerosol particles involved in the nucleation process and of the characteristics of turbulence are also needed.

There is also a need to assess the predictability of cirrus clouds on large and small scales. Do the large-scale fields of temperature, moisture, and winds contain sufficient information to diagnose the locations, heights, water con tents, and radiative proper-ties of cirrus clouds? To what extent is cirrus predictability enhanced by incorporating the temporal evolution of the large-scale fields in a model-based data assimilation? To what extent is cirrus predictability enhanced by incorporating special mesoscale data sets into mesoscale models nested within the large-scale models? Is the absence of routine, quality upper tropospheric observations of moisture a major impediment to improved diagnosis of cirrus cloudiness? To what extent is this compensated by incorporating time-dependent data into model-based data assimilations? These questions can only be answered by obtaining high quality data sets of cirrus cloud structure and evolution along with simultaneous measurements of the large-scale fields.

Forecast experiments using FIRE Phase I data sets (e.g., Westphal and Toon, 1989 and 1990; Nicholls et al., 1990) have served to define key issues and observational requirements for Phase II. In Phase II, forecast experiments with both mesoscale and global-scale models will be used to test current and new cirrus parameterization schemes and to examine the predictability of cirrus. Specific models that will be used include the CSU/RAMS and PSU/NCAR mesoscale models and the UKMO global model. Large Eddy Simulations (LES) experiments using nested grids will be performed with the mesoscale models and will draw heavily on results from the highly detailed models. Large-scale data sets with increased temporal resolution over a wide area are required for improved definition of the development of the dynamic and thermodynamic environment of extended cirrus cloud systems, especially the distribution of upper tropospheric water vapor. This is particularly true for the LES experiments that serve as the bridge between the very detailed models and the global-scale models and provide the framework for investigating scale interactions that are central to the problem of relating cloud response and feedback to large-scale control.

The strategy for obtaining the data sets (Phase II) required for initializing numerical models and verifying models results over a range of scales, developing and testing parameterizations of cirrus clouds, and evaluating the predictability of cirrus is to:

At this latter scale, the descriptions will be derived from in situ (aircraft and sondes) and remote sensing (satellite, airborne and ground-based) observations. The integration of these observations into a coherent description is a formidable task (section 3.2).

3.2 Characterizing Development of Cirrus Clouds

Production of data sets suitable for initialization and verification of cirrus cloud development models requires careful analysis and integration of diverse observations. Even in the case of FDDA analyses using regional mesoscale models, independent analysis of the input data base and comparison to the model-produced fields is essential to establish the integrity of those large-scale fields. As evidenced by the results from FIRE Phase I, there is much to be learned simply from analyzing the data. While the data requirements for characterizing cirrus cloud development are equivalent to those described in section 3. 1 . there are a number of areas where analysis of these data are particularly critical and/or problematic. These areas are briefly highlighted here.

3.2.1 Structure of Cirrus Cloud Fields

Quantitative descriptions of the time-dependent, horizontal, and vertical structure of the observed cirrus cloud fields are required. These descriptions must include accurate distributions of cloud height, areal coverage, and cloud radiative properties over a large region encompassing the domain of regional mesoscale models. Additional parameters such as cloud thickness and the bulk cloud microphysical properties (vertically integrated ice water content and mean particle size) are highly desirable. Characterization of the vertical distribution of cloud parameters is also desirable (e.g., delineating multilayered structure). Satellite-based observations provide the only means to produce such analyses. However, achievement of the level of detail and accuracy needed to meaningfully constrain regional mesoscale models requires significant improvements in current algorithms used to process satellite data. Such improvements will only result from detailed comparisons of satellite observations with quantitative 'ground truth' descriptions derived from surface-based (remote sensing) and airborne (remote sensing and in situ) observations.

To be useful for application to cloud process models, which are appreciably more detailed but consider a much smaller domain than regional mesoscale models (and may even be nested within a larger scale model), the cloud physical descriptions required for meaningful comparison to model simulations must be correspondingly more detailed. The time-dependent, vertical structure of cloud water content, particle size distribution, and particle phase are needed over a relatively small mesoscale region (~20 - 50 km). Both Eulerian and Lagrangian descriptions are highly valuable. In addition, the structure of these fields must be characterized in terms of the magnitude and spatial scales of variability, including that of the cloud radiative properties. Although in situ observations (aircraft) can provide the basic required microphysical and radiative measurements, the sampling limitations of aircraft platforms do result in uncertainties associated with spatial coverage and representativeness. Integration of those observations with satellite, airborne, and surface-based remote sensing observations can potentially provide a significantly more complete characterization.

Extended cirrus cloud systems exhibit a significant degree of persistent mesoscale organization. This may reflect corresponding structure in the ambient dynamic and thermodynamic fields (Starr and Wylie, 1990). Mechanisms potentially responsible for generating such structure include the effects of up-scale energy transfer in two-dimensional turbulence, inertial instability due to strong horizontal wind shear, deep tropospheric gravity waves associated with ageostrophic flow in the jet stream, and longitudinal roll circulations associated with jet flow. At the present time, it is unknown to what extent these or other mechanisms operate in organizing cirrus or to what extent they maintain the observed structure. It is also unknown to what extent the radiative (section 3.2.5) and thermodynamic (latent heating and cooling - section 3.2.4) effects of cirrus clouds act to maintain this structure. In order to understand the origin and maintenance of mesoscale cirrus cloud features, the time-dependent structure of the cloud fields (section 3.2.1) must be known in conjunction with the corresponding dynamical and thermodynamical fields. Given that high wind speeds are typical at cirrus altitude, significant enhancements to the routine synoptic upper air observing schedules will be required to resolve these fast moving and relatively small-scale dynamical features. Fortunately, a new capability exists to actually resolve upper tropospheric wind fields on a near-continuous basis (-10 min values) at a relatively fine scale (-200 km) using the NOAA/NWS wind profiler network that is presently under construction. Acquisition of thermodynamic information at increased frequency (supplemental rawinsondes) will facilitate analysis of these mesoscale features.

3.2.3 Vertical Motions at the Mesoscale and Convective Scale

Observations of vertical motions in cirrus clouds has always been very problematic and have significantly retarded progress in quantitatively testing cloud development models since model simulations are very sensitive to the prescribed large-scale forcing. In particular, accurate knowledge of vertical motions on the scale of individual mesoscale cloud features (20 - 100 km) is essential. Recent technological and methodological advances have provided various ways to observe or estimate vertical motions in cirrus clouds at this scale. However, the quality and representativeness of these data (and analysis methods) have yet to be definitively established. Thus, it is clear that the required approach is to apply every available means for characterizing vertical motions in cirrus clouds. Only in this way can the present uncertainties in independent analysis of data from individual systems be reduced to an acceptable level.

Data from a mesoscale network of wind profilers would enable calculation of the upper tropospheric vertical motion field at the required space and time scales using kinematic techniques (derived from divergence analysis of observed horizontal winds). Similar techniques can also be applied to scanning Doppler lidar (CO2) and Doppler radar (mm) observations and to aircraft-observed horizontal wind profiles (Gultepe and Heymsfield, 1990) to generate profiles of vertical velocity at a scale of about 20 km. Kinematic and adiabatic techniques can be used for deriving vertical motions from rawinsonde observations at a larger scale. Vertical motion fields are also derived from FDDA (section 3. 1).

Conventional in situ measurements from aircraft (inertial and gust probe) provide a means to characterize convective scale (~1 km) motions which are needed for initialization of microphysical models and verification of high-resolution mesoscale cloud models. Alternate strategies and methods should be utilized to the extent possible in order to provide the broadest possible observational basis. For example, direct inference based on Doppler velocity spectra obtained using C02-lidar or millimeter radar (Eberhard et al., 1990; Kropfli et al., 1990) or wind profiler measurements taken in a vertically-pointing configuration is possible. Such capabilities are highly desirable. In addition, radar tracking of chaff released in a cloud and in situ sampling and analysis of inert chemical tracers, such as sulfur hexafloride, can be used to assess transports within a cloud. Each of these techniques will be tried.

3.2.4 Water Budget of the Upper Troposphere

A key factor influencing cirrus cloud development and also manifesting the effects of cloud processes is the distribution and redistribution, respectively, of upper tropospheric water vapor. However, useful observations of water vapor content at cirrus altitudes have not been available, even at the synoptic scale. This severely constrains our ability to quantitatively test cirrus cloud parameterizations since current treatments are typically formulated in terms of relative humidity. Although imagery from the GOES 6.7 m m water vapor channel has long been very useful to forecasters, it has proven very difficult to quantify and incorporate into operational or research models since the detected signal is also strongly coupled to atmospheric temperature structure and the vertical location of water vapor (Blackwell, et al., 1988). Similarly, information on tropospheric water vapor derived from the three HIRS water vapor channels on the operational NOAA polar orbiting satellites has also seen little use in operational forecast models. Recent advances in aircraft instrumentation (cryogenic frost-point hygrometer) now provide the means to accurately observe atmospheric water vapor contents at cold temperatures. Raman lidar has been used to remotely sense water vapor contents to altitudes of 6 km at night (Melfi et al., 1989). On-going improvements (a more powerful lidar and more sensitive detectors) promise to extend the useful Raman lidar operating range into the upper troposphere at night and permit daylight observations. Longer integration times (30 minutes) can also be used to enhance the sensitivity at these levels (presently, 2-minute values are derived). These technologies should be utilized to the fullest extent possible to provide the required observations. In addition, the possibility of using these new observations to improve our ability to quantify the satellite observations should be fully explored.

Characterizing the effects of cirrus cloud latent heating and cooling and the rate of water vapor incorporation into cirrus clouds provides an essential constraint on model-based studies and our understanding of cirrus cloud development. Model evaluation in terms of system throughput is most meaningful. Knowledge of the upper tropospheric water vapor budget is a key ingredient in this methodology.

Radiative processes significantly affect the development of cirrus clouds. Net radiative cooling or heating can enhance or suppress the overall rate of production of condensate by altering the relative humidity within the cloud and adjacent layers. Vertical gradients in radiative heating can lead to stabilization or convective overturning of a cirrus cloud. - Horizontal gradients in net radiative heating can regulate the intensity of convective or forced circulations.., The intensity and structure of internal cloud circulations and dynamic entrainment at cloud boundaries has a major influence on the microphysical development of the cloud and its overall composition and properties (Starr and Cox, 1985b). Thus, efforts to model cirrus cloud development using detailed cloud models must necessarily incorporate the effects of radiative processes. The radiative component of these models are uncertain, especially in the treatment of solar absorption. An accurate characterization of the radiative energy budget of cirrus clouds is essential both for development and verification of suitable parameterizations and models.

It is important that broadband radiative properties be resolved in a consistent manner for the infrared, near infrared, and visible spectral regions. The objective and requirements here are somewhat different than in sections 3.3 and 3.4 where a high degree of coincidence is required to resolve fundamental issues regarding the transfer of radiation in cirrus. For the application here, a statistical approach is desired in which the means and variances of the observed radiative fluxes (rather than radiances) and cloud water content can be used to derive representative bulk relationships enabling reproduction of the gross features of the observed cloud radiative budget. Such relationships can then be incorporated into detailed cloud models to simulate the structure of radiative forcing in cirrus clouds.

Cirrus clouds are microphysically complex consisting of a broad range of particle sizes and particle habits (e.g., Heymsfield et al., 1990). Even descriptions of the crystal habits are difficult in that combined polycrystalline forms, including aggregates, are prevalent in many cases. The microphysical composition of the clouds is the dominant factor determining the cloud radiative properties. In addition, the water and energy transports associated with the gravitational settling of larger crystals is a dominant factor regulating overall cloud properties and exchanges with surrounding air. FIRE Phase I has led to significant improvements in our fundamental understanding of how cirrus particle populations evolve. These studies have relied on observations obtained in previous field experiments (FIRE Cirrus IFO-I and pre-FIRE data) and incorporated these data into theoretical microphysical growth models. While much has been learned, these studies have raised a number of basic questions that need to be answered. For example, does air-mass origin and composition have a pronounced effect on the microphysical composition of cirrus clouds through the effects of the contained populations of condensation and ice nuclei? How effective are cirrus at cleansing the upper troposphere of particulates? What is the effect of this cleansing on subsequent cloud development in the same air mass? Furthermore, there are fundamental uncertainties with respect to the nucleation mechanisms actually operating. A well-supported explanation of how significant populations of the very large ice crystals typically observed (order 100s m m) can develop in the presence of large quantities of small crystals, especially if the small crystals are as prevalent as indicated by remote sensing observations (section 2), within geometrically thin cloud generating layers would significantly enhance our basic understanding of how cirrus clouds work.

Observations of the ambient cloud condensation nuclei (CCN) and ice nuclei concentrations and compositions in conjunction with detailed observations of the entire ice crystal populations are essential for resolving these issues. Such observations should attempt to define these parameters on the scale required for microphysical growth models, usually that of an individual convective updraft (-l km). In addition to optical probes and collection devices on aircraft, balloon-borne ice crystal replication sondes provide an additional means to sample ice crystal populations and ice water contents.

3.2.7 Turbulence and Convection Processes in Cirrus Clouds

Turbulent heat and moisture fluxes play an important role in the evolution of cloud layers. Knowledge of the characteristic scales and intensity of these processes would provide another constraint on the performance of cloud development models and cirrus cloud parameterizations. In addition, results from FIRE Phase I (Smith et al., 1990b; Flatau, et al., 1990) and ICE (Quante et al., 1990a and 1990b) indicate that turbulent energy cascades may be an important factor influencing the organization of cirrus cloud fields. Although measurement of turbulence and turbulent fluxes at cirrus (or any) altitude is very difficult, the potential applicability of these data warrants the effort. Even relatively simple information on the characteristic scales of cloud organization and scale dependence of variance of dynamic and thermodynamic parameters would be quite useful.

Analysis of turbulence characteristics typically employ spectral techniques. These analyses require high-frequency sampling (20 Hz) using aircraft platforms. Adequate sample length is also required for meaningful results (flight legs more than six times the length of the analyzed scale). In addition, characterization of turbulence intensities using remote observations (e.g., wind profilers) may be possible and should be exploited.

Characterizing relationships between cirrus cloud radiative properties in various spectral regions and cloud microphysical properties supports each of the other key science objectives. While the determination of cloud microphysical parameters in a GCM necessarily proceeds from the resolved large-scale fields of temperature, moisture, and winds, the radiative effects of cirrus clouds on the resolved fields are computed based on the radiative properties associated with those microphysical properties. These radiative effects represent the most important impact of the clouds on the simulated climate. To first order, it is the cloud radiative properties that are the primary desired output of a GCM cirrus cloud parameterization. At the scale of individual cirrus clouds, radiative processes also directly influence the development of the cloud to a significant degree by modulating the scales and intensities of convective and turbulent circulations and transports as well as the overall production rate of condensed cloud water. Thus, the cloud radiative; properties help determine the cloud physical properties which, in turn, determine the radiative properties. Conversely, monitoring of cirrus cloudiness and its radiative impact on the surface and atmospheric energy budgets, whether from space or from the earth’s surface, requires that the detected radiances be interpreted in terms of more useful quantitative measures of cloud physical properties. This is particularly important given the general lack of opacity and great natural variability of cirrus radiative and physical properties in space and time.

The radiative properties of cirrus clouds critically depend on two poorly known fundamental properties: the effective radiative particle size and scattering phase function of the particle population. Both depend on the sizes, shapes, orientations and, possibly, the composition of the particles. Larger particles affect the longer wavelength infrared radiation more strongly than the smaller particles, while smaller particles, if present in sufficient concentrations, dominate the transfer of solar radiation. Results from FIRE Phase I clearly demonstrate the uncertainties in present knowledge (section 2). Model computations of effective particle size based on observed multispectral radiances (reflected and emitted) were inconsistent with the observed particle size spectra. Similarly, computations of cloud bidirectional reflectance patterns based on the microphysical observations and present best-estimates of the phase function for citrus crystals were inconsistent with the corresponding radiative observations. The differences are very significant and have been attributed, variously, to the presence of substantial numbers of undetected small ice crystals (the called 'small particle radiative anomaly'), the presence of significant numbers of small liquid phase particles, inadequate knowledge of the phase functions of actual cirrus particle populations, and uncertainties in the refractive index of ice.

Relationships between the radiative and physical properties of cirrus clouds can only be determined if adequate observations of the incident and emergent radiative fields are available in conjunction with coincident measurements of cloud composition. The FIRE Phase I data sets are inadequate for resolving these basic uncertainties primarily due to deficiencies in observing capabilities. Motivated by these findings, new instruments have been developed or are in the process of being developed. Improvements in our ability to observe the microphysical composition of cirrus clouds, especially the numbers of small particles, and enhancements in our ability to resolve the associated radiative fields, including the possibility of direct measurements of scattering phase functions, will yield significant improvements in our knowledge of and ability to model fundamental relationships between cirrus cloud optical properties and microphysical properties.

Obtaining these data requires insitu observations from aircraft and balloon-borne ice particle replicators to define the microphysical composition (and possibly the scattering phase function) and remote observations from satellite, aircraft, and ground-based sensors to define the radiative fields and cloud optical depths.

This last requirement necessitates a concentrated effort to calibrate and intercompare all radiometric sensors. To the extent possible, all instruments should be referenced to a common standard.

This diverse set of observations will be incorporated into a hierarchy of radiative transfer models as initial conditions and verification data sets. These models range from present radiance and simplified scattering models to experimental multi-dimensional models and explicit treatments of the effects of crystal morphology on individual scattering events. In this way, self-consistent relationships between the microphysical and radiative properties of cirrus can be derived.

Satellite observations provide a means to quantify global cloudiness on a climatological basis. In comparison to surface-based observing systems, satellite platforms provide the advantage of uniform and representative sampling and truly global coverage. Since cloud parameters are derived from observed radiances, consistency is insured with respect to the associated cloud radiative impact on the observed radiative budget of the earth-atmosphere system (top-of-atmosphere budget). Satellite observations permit monitoring of the physical and radiative properties of cirrus clouds over extended time on a global basis enabling resolution of chances in an internal component of the climate system that will potentially play an important role in modulating any global climate change. Moreover, the credibility of models of the global climate system (GCMs) will be significantly enhanced if the realism of crucial internal components, such as the treatment of cirrus clouds, can be verified against suitable observations of the present climate, e.g., the annual cycle in the global distribution of cirrus clouds.

Cirrus present a particularly difficult target for remote detection from space (Rossow et al., 1985) because of their general lack of opacity and great natural variability of radiative and physical properties in space and time. Due to the tendency for low cloud albedo, they may be difficult to distinguish from the surface background at visible wavelengths. Although much more evident in the infrared observations, derivation of cloud height (temperature) and cloud emittance is problematic given the intrinsic coupling of these parameters with respect to the observed radiances in window regions. Inference of cloud physical properties from satellite-observed radiances requires assumptions about cloud structure and composition. In essence, a radiative-physical cirrus cloud model is invoked in any retrieval procedure in order to make the desired inference. An objective of FIRE is to test the underlying assumptions in these models. It is only through a rigorous examination of the validity and representativeness of these assumptions and models that improvements in satellite-based cloud observation capabilities will result.

Explicit quantitative knowledge of the capabilities and limitations of present methodologies will enhance the value of the derived cloud observations by providing a measure of the applicability of and confidence in the cloud data. This will facilitate the optimal use of ISCCP and other satellite-derived cloud data by the modeling community. It will also provide a basis to improve current techniques using existing satellite observing systems. Moreover, dramatic increases in observing capabilities (satellite instrumentation) and data processing capacity are planned (EOS). FIRE seeks to provide a basis for the design and development of future cloud retrieval methodologies that will take full advantage EOS as a global cloud observing system.

Optimally, the cirrus cloud properties that are desired from present and future methodologies and that will be considered in FIRE include:

The data requirements to support the required investigations are nearly identical to those given in section 3.3. The key elements are for high spatial resolution, multispectral observations at multiple viewing angles in conjunction with coincident observations of cloud composition and fundamental radiative properties as well as thermodynamic structure. These data can only be obtained by simultaneous measurements using satellite, aircraft, and surface-based instruments. Accuracy and representativeness are essential. Cross-calibration of all radiometric sensors is critical (satellites, aircraft and surface). In addition, a high resolution, quantitative, satellite-independent characterization of the horizontal and vertical structure of the cloud fields over an area larger than the resolution of operational satellite sensors (-30 km for sounding channels) is needed. Scanning cloud lidar or radar provide this capability.

Some data sets must be acquired over a water background to provide suitable simplest cases for treating cloud reflectances. In addition, data sets corresponding to land surfaces, to nighttime conditions (a cold background), and to snow-covered backgrounds should also be obtained to provide a means to evaluate and contrast model performance in more demanding situations. In order to insure the representativeness of the data and derived models for a wide range of naturally occurring conditions, routine Comparisons of satellite-derived cloud parameters to 'ground truth' cloud observations (lidar and passive radiometric observations) will be conducted at a number of sites over the course of Phase II (section 5). Though these data sets will not be as comprehensive as the intensive observations, they will provide coverage of a variety of climatological situations (solar elevation, surface conditions, and cloud conditions) and will benefit from efforts to improve the utilization of surface-based remote sensing for observing cirrus clouds (section 3.6).

The importance of cirrus clouds in modulating the radiative energy budget of the earth's surface and atmosphere is a prime motivation for FIRE cirrus research. FIRE research directed at quantifying the radiative impact of cirrus clouds on the surface, top-of-atmosphere (TOA), and atmospheric radiation budgets follows two general approaches.

3.5.1 Monitoring the Radiative Impact of Cirrus

Cirrus clouds constitute an important component of global climate data sets such as the cloud climatology being produced by ISCCP and the TOA radiation budget climatology being produced by ERBE. Efforts to develop techniques for monitoring the global surface radiation budget from space, both in the solar and infrared spectral regions, must necessarily account for the effects of cirrus clouds. In conjunction with efforts to quantify the limitations and capabilities of methods to derive cloud properties from satellite observations (section 3.4), FIRE seeks to better understand relationships between cloud parameters and the TOA and surface radiative energy budgets. The strategy is to obtain data sets suitable for investigating relationships between cirrus cloudiness and the surface TOA radiative budgets. The required data sets includes:

It is important that these quantities be broadband and that the infrared, near infrared, and visible spectral regions be resolved in a consistent manner. Moreover, the data must be as simultaneous as possible in time and space. In addition, rawinsonde observations of atmospheric temperature and humidity structure are also required since atmospheric temperature and moisture also impact the surface and TOA radiation budgets.

As in the case of methods to derive cirrus cloud properties from satellite observations (section 3.4), continuing efforts to collect data representative of a wide range of situations over extended time periods will be made (section 5). The acquisition and analysis of these data will supplement the intensive field observations and will provide a broader data base for these studies.

FIRE also encompasses efforts to evaluate models of atmospheric (gaseous) radiative transfer in the infrared spectral region. The need for testing our capabilities in this area have been clearly established by the results of ICRCCM (Intercomparison of Radiation Codes in Climate Models) where an unacceptably high level of variance was found among infrared radiative transfer models. The reasons for the disparities have proven difficult to isolate but given the importance of infrared radiative processes in climate change scenarios, especially exchanges due to water vapor, renewed efforts to verify such models with observations are warranted. SPECTRE (Spectral Radiance Experiment) is a collaborative experiment (section 6. 1) to be conducted in conjunction with FIRE intensive field observations. Although SPECTRE emphasizes clear sky situations, observations under cirrus cloud conditions will be obtained and analyzed. These observations should prove valuable in assessing the radiative impact of cirrus. SPECTRE adds the additional requirement for observations of-.

Remote sensing techniques will be required for water vapor (Raman lidar) and for temperature (RASS; Radio-Acoustic Sounding System) since rawinsonde data will not be adequate.

Large-scale atmospheric models provide an alternate means of quantifying the radiative and climatic impact of cirrus clouds. Although present GCM cirrus cloud parameterizations are extremely crude and uncertain, investigations using these models have provided important insights into the role of cirrus in the climate system (section 2). As such, GCM studies can serve to focus research on specific critical questions. For example, inadequate knowledge of solar absorption in tropical cirrus clouds has been identified as a probable cause of persistent problems in reproducing upper tropospheric temperatures in tropical and subtropical regions with resultant impact on the strength of upper tropospheric general circulation and climate (Ramaswamy and Ramanathan, 1989). GCM studies on the radiative and climatic impacts of cirrus clouds will continue during FIRE Phase II, Efforts will be made to intercompare the simulated TOA and surface radiation budgets and cloud parameters with global climatologies such as ERBE and ISCCP. New cirrus parameterizations will be implemented as they are developed and verified during FIRE Phase II (section 3. 1).

Routine ground-based remote sensing observations offer considerable promise for characterizing the properties of cirrus clouds and their environment over extended time periods. Active sensing technologies not presently available from satellite platforms can be applied including visible and infrared lidars, short-wavelength cloud radars, and long-wavelength wind radars (profilers). These observations provide 'ground-truth' corroboration for properties derived from passive sensing techniques, especially retrievals from global satellite-based systems. A wide variety of passive remote sensing technologies can also be utilized with considerably more spectral coverage and resolution than presently available from satellite platforms. Although, in most instances, the physical interpretation of remote sensing observations is significantly more uncertain than for in situ measurements, the benefits in terms of reduced costs and increased coverage in comparison to aircraft or balloon-borne observations are appreciable. Moreover, these techniques provide the possibility for needed quantitative observations of the climatologically important, high-altitude, tropical cirrus systems that are largely inaccessible to research aircraft (section 6.2).

A primary advantage of active remote sensing of cirrus is the capability to accurately measure, with unparalleled precision and resolution, the temporal evolution of the height-dependent structure of the clouds. The fundamental parameter measured is the vertical profile of backscattered intensity at the probing wavelength. Although this parameter has intrinsic value, inference of more widely applicable cloud parameters is highly desired, including profiles of cloud optical depth, particle phase and habit, and cloud water content. The Doppler velocity spectrum (radar and infrared lidar) and polarization of the returned signal are also measured by some systems. The Doppler velocity spectrum provides a means of estimating fall speeds of the ice crystals which are highly dependent on crystal mass (and habit). These may be used to derive the net vertical flux of ice mass which is a dominant component of the cloud water budget. Polarization provides a measure of the phase and habit of the cloud particles. In addition, high spectral resolution lidar offers a much more direct measurement of the vertical structure of cloud optical depth than achievable with other systems where appreciable uncertainty is associated with the assumptions required to make this inference.

Although current lidar and radar probing techniques offer the promise of reasonably quantitative analysis of cloud composition, the analysis methods are presently based on rather scant preliminary evidence in this developing field. Since the interaction of radiation with cloud particles depends both on the size of the particles and the wavelength of the radiation and since a wide range of particle sizes typically occur in cirrus, systems probing at different wavelengths may respond quite differently to the same population of cloud particles (Intrieri et al., 1990). This indicates that the aggregate information derivable from coincident remote sensing measurements using a variety of these systems may be increased in a highly synergistic manner. This possibility has only begun to be explored (e.g., Sassen et al., 1989; Eberhard et al., 1989 and I 990).

Similar possibilities exist in the combination of lidar and radar observations with passive high spectral resolution measurements in visible, near infrared, and infrared window regions, especially with respect to determination of cloud optical properties (Grund et al., 1990). When combined with surface-based remote sensing observations of ambient winds. (wind profiler), moisture (microwave profiler), and thermodynamic structure (infrared and microwave profilers), the possibility exists for a rather comprehensive surface-based observing system. Given the demanding observational requirements described in the preceding sections, the potential of this approach needs to be fully explored.

The strategy for improving utilization of remote sensing observations for quantitative characterization of cirrus clouds and their environment during FIRE Phase II is to intercompare coincident observations from a wide variety of active and passive systems together with in situ observations of cloud physical composition and atmospheric state. Although maximizing the quantitative information from each system is an obvious objective, results from FIRE Phase I indicate that the capabilities of individual systems are limited and that coincident observations are highly desirable. Thus, the emphasis of Phase II research will be on defining the potential synergistic increase in quantitative information from combined systems encompassing:

In addition to plans for intensive field observations (section 4), pilot studies (extended time) have been initiated (sections 5.3 - 5.6) and will continue during Phase H. It should also be noted that utilization of remote sensing observations of cirrus clouds from aircraft and satellites will also benefit from these activities.

4.1 Synopsis

A second intensive midlatitude cirrus field experiment (FIRE Cirrus IFO-II) is required to meet the objectives of FIRE Phase II (section 3). Findings from the first experiment (IFO-I) have led to the development of new instrumentation and observing capabilities for the purpose of resolving some of the most critical unresolved questions (section 2). Advantage will also be taken of other new observing capabilities that will markedly increase the overall utility of the data sets for use in understanding cirrus cloud development. FIRE has grown to include active participation by mesoscale (regional) and large-scale (GCM) modelers in the area of cirrus cloud parameterization and climatic effects. The plans for FIRE Cirrus IFO-II have been strongly influenced by the data requirements of these models. Experiences from IFO-I have also played a major role in defining the measurements that will be needed and in redesigning the sampling strategies that will be used. This experiment will address the problem of cirrus cloud development and impacts over a wide range of scales through the integration of models and observations. FIRE Cirrus IFO-II will provide a unique observational basis for improving the quantitative utilization of remote sensing observations from satellites and from the surface for investigating and monitoring cirrus clouds. The observational elements are summarized below and described in more detail in succeeding sections. It must be emphasized that the true value of the observations will derive from their use in evaluating and improving models and observing systems (data analysis algorithms) which can be used to extend the knowledge gained into the context of the global climate system.

Table 2 - Active remote sensing systems participating in FIRE Cirrus IFO-II. Planned location

Category Instrument Location/Organization

Hub R2 R3

Kernel Polarization Lidar U. Utah LaRC

Group High Spectral Resolution Lidar U. Wise.

Scanning Doppler Radar WPL (9 mm) PSU (3 mm)*

Highly Water Vapor Raman Lidar** GSFC

Desirable Radio-acoustic sounder (RASS)** WPL

Wind Profiler PSU CSU NWS

Scanning C02 Doppler Lidar WPL CSU*

Scanning Cloud Lidar (VIL) U. Wisc.***

content, optical thickness, and particle vertical motions. These basic measurements are integral to the success of the FIRE Cirrus IFO-II. The combined data allow for a unique characterization of cirrus cloud development and structure. For example, particle phase, habit, and orientation can be derived from polarization lidar observations. A similar capability has recently been added to the WPL cloud-sensing radar (Kropfli et al., 1990). These data can be combined with the Doppler-observed particle vertical motion to estimate particle mass and terminal velocity and, thus, updraft and downdraft wind speeds. These data can then be combined with the radar derived ice mass content to characterize the vertical ice mass flux and the mechanisms and scales involved in cirrus cloud generation. Simultaneous lidar and radar observations at multiple wavelengths will serve to refine the ice content estimates by providing additional information on particle size distribution in comparison to single-wavelength probing. Simultaneous measurements will also provide information on the relationship between ice content and optical cross-section of the cirrus clouds. Optical depth measurements provided by high spectral resolution lidar provide a near-absolute calibration for optical depth retrievals using polarization (or non-polarized) lidar observations enabling a characterization of cloud optical properties in the visible at very high temporal and vertical resolution.

The group of highly desirable remote sensors takes advantage of state-of-the-art and developmental type systems with additional research capabilities. Continuous lower tropospheric water vapor profiling (Raman) and temperature profiling (RASS) are essential for SPECTRE (sections 6.1 and 4.3.2.2). Provided that data integration times at cirrus cloud altitudes are not excessive and that daytime capability can be achieved (section 3.2.4), Raman lidar water vapor measurements preceding local cirrus cloud development, and in the vicinity of cirrus, may help answer important questions about humidity fields and cirrus formation. Obtaining an accurate characterization of upper tropospheric water vapor distributions is a very high priority (section 3.2).

Infrared lidar (CO2) observations are expected to provide a measure of the vertical structure of cirrus optical properties (emittance) in the important 10.6 m m window region. Polarization data at this wavelength can be used to address important questions about the infrared absorption efficiencies of the particles. The Doppler velocity spectra from the C02 lidar has some potential for characterizing particle size distributions (the distribution of particle fall speeds). Combining these data with coincident radar-derived ice mass concentrations (radar response is weighted according to the particle size distribution) and in situ microphysical observations obtained by aircraft should lead to significant improvements in the interpretation of the data from each of these sources (e.g., Sassen, 1987; Eberhard et al., 1989; Eberhard et al., 1990). Problems with signal attenuation that affect each system to differing degrees can be addressed in comparative studies (Sassen et al., 1989; Uttal et al., 1990).

A rapidly scanning lidar has considerable value for capturing essentially Eulerian views of cirrus structure along and across the mean wind direction within a 50 km radius. These very high resolution data will provide a unique means to characterize cirrus cloud structure over scales corresponding to satellite observations and the field of view (fov) of radiative flux observations. Under conditions involving strong vertical wind shear, the distortions present in the cloud structure obtained from zenith operations can be assessed. Moreover, scanning lidar data would be a very useful tool for pre-flight aircraft mission planning, particularly if the images can be made available in near real-time to the mission scientist and aircrews (section 4.4). Potentially, these data could also be used to direct aircraft during mission execution.

The planned location of the VIL at a distance of 10 to 20 km to the southwest of the Hub site attempts to maximize collection of data in an optimal sampling pattern for the science objectives while minimizing potential safety problems (below) and consequent interference with aircraft operations (section 4.3.3). Given that a southwesterly wind direction dominates during large-scale cirrus events at this location at this time of year , this location will permit continuous cross-wind scanning just upwind of the Hub site and in situ aircraft patterns. This crosswind VIL scanning pattern will map the time-dependent cirrus cloud structure in a plane parallel to and of the same scale (100 km) as that being sampled by the remote sensing aircraft which will operate just downwind of the VIL sampling volume. The VIL site should be as close to the Hub as possible but at a sufficient distance to permit continuous aircraft operations over the Hub without their intersecting the VIL scan volume. In addition, this location permits VIL scanning in the upwind direction (away from the Hub and aircraft operations). Situations may require further restrictions on the VIL scan volume.

Observations of vertical motions in cirrus clouds has always been very problematic and have significantly retarded progress in testing cloud development models (section 3.2). The array of wind prowlers at the remote sensing sites (section 4.3.4) provides a means of deriving area-averaged vertical motions on the scale of mesoscale cloud features (-50 km, see fig. 4). One of the wind profilers (five-beam system at R2) will have the capability to independently observe vertical motion on a scale of about 20 km. Scanning Doppler systems provide an alternate means to derive vertical motions at a comparable scale (-20 km) also using kinematic techniques (VAD scans). Similar techniques can be applied to aircraft-observed winds as in Gultepe and Heymsfield (1989). Doppler spectra derived from the zenith-pointing C02-lidar observations can potentially, if sufficient resolution is achieved, provide an independent direct measure of vertical air motion and particle vertical motions within individual convective cells for comparison to values retrieved from Doppler radar observations at similar scale. These observations can be compared to in situ aircraft observations at a similar scale. The approach here is to apply every possible means to definitively characterize vertical motions in cirrus clouds. Only in this way can the present uncertainties in independent analysis of data from individual systems be reduced to an acceptable level.

Since maximum information content is derived from the multiple remote sensor approach when the various systems are examining the same cloud volume, emphasis will be given to zenith-pointing operations (except for the scanning cloud lidar--VIL). However, the scanning Doppler systems designed to collect Velocity-Azimuth-Display (VAD) scans will coordinate their activities at predetermined time intervals to periodically characterize three-dimensional cloud structure and vertical motion fields which are of very high priority.

An important concern is that lidar operations are conducted in such a way as to be eyesafe to ground and aircraft personnel. (The infrared C02-lidar is eye-safe.) This is generally not a problem for vertically pointing systems. However, the scanning lidar (VIL) can be a hazard. Spotters can be used to monitor intruding commercial, private, and military air traffic and scan patterns can be coordinated with research aircraft activities. Operations may have to be suspended during aircraft missions if the level of ground-to-air communications required to ensure crew safety proves burdensome to the pilots or if situations become too complex to insure safe operations. A plan to adequately deal with eye-safety issues will be formulated and submitted to the appropriate government agencies.

Another concern is the potential for electronic interference among various active and passive sensing systems, including those borne on aircraft or balloons, and communication equipment. In particular, potential interference problems between wind profilers and rawinsonde systems operating near 400 MHz and between wind profilers and aircraft operations need to be understood and resolved, if possible. Efforts will be initiated to investigate this issue and to recommend possible solutions which may include altering the location of specific problem systems. In this vein, the location and operation of the RASS system will minimize the potential negative impacts of acoustic pulses (noise) on project personnel.

4.3.2.2 Radiance and Radiative Flux Observations

There are four aspects to the measurements of downwelling (and upwelling) radiances and radiative fluxes at the surface using passive radiometric instrumentation. Although separately presented here, exchange and cross-integration of data from each effort will be of mutual benefit as will the availability of coincident active remote sensing data and aircraft and satellite data. Efforts to measure the surface radiation budget and to determine the effects of cirrus clouds on the surface radiation budget will be conducted in collaboration with the NASA Surface Radiation Budget (SRB) program and the Spectral Radiation Experiment (SPECTRE, section 6. 1). A summary listing of these instruments is given in table 3.

Data from three instruments are particularly useful for characterizing cirrus cloud optical properties, especially in combination with lidar and other active measurements. They are:

These measurements will serve to define cirrus cloud optical depth in the visible and infrared, and cirrus cloud infrared effective radiating temperature and emittance.

The effects of cirrus clouds on the downwelling solar radiation at the surface will be assessed using narrowband and broadband solar flux radiometers at the Hub and R2 sites where cloud imaging systems will also be deployed. Besides the satellite observations (section 4.3. 1), direct measurements of cloud and aerosol optical depths using a high spectral resolution lidar and sun photometers at the Hub site as well as cloud cover observations from a scanning lidar system are needed. Low level aircraft observations of the spatial distribution of surface albedo and solar radiative fluxes are also desired (section 4.3.3).

Table 3 -Summary of planned surface instrumentation and responsible organizations for the FIRE Cirrus

IFO-II Campaign at the Hub, R2, and R3 surface remote sensing sites (see figure 4).

Location/Organization

Sensors Hub R2 R3

Active Remote Sensing Systems

Polarization Lidar U, Utah LaRC

High Spectral Resolution Lidar U. Wisc.

Scanning Doppler Radar WPL (9 mm) PSU (3 mm)*

Scanning C02 Doppler Lidar WPL CSU *

Scanning Lidar (VIL) U. Wisc.**

Laser Ceilometer LaRC CSU

Raman Water Vapor Lidar GSFC

Radio-Acoustic Temperature Sounder WPL

Dual-channel Microwave Water Vapor Sounder WPL

Microwave Wind Profiler PSU, WPL CSU NWS***

Infrated Spectrometer GSFC, U. Wisc., CSU

Sun Photometer ERL

Infrared-Window Radiometer U. Utah, WPL CSU

Direct Spectral Solar Radiometer U. Denver CSU

Solar Flux Radiometers PSU CSU

Broadband Infrared Flux Radiometer many

Calibration Facility SPECTRE CSU

Sky-Imaging Systems (VCR and 35 mm) LaRC CSU

Rawinsonde Station (CLASS) ## NCAR, LaRC NCAR

Cloud Ice Particle Replicator Sondes NCAR

Ozone Sonde ERL

Dobson Total Ozone Instrument ERL

Tethered Balloon Sounder GSFC/WPL

Instrumented Meteorological Tower TBD

Surface Observations LaRC CSU

Trace Gas Flask Samples ERL

* - Scanning capabilities may not be available

** - VIL will be located within 10-'!O km of Hub

An SRB objective is to perform an intercomparison of various infrared flux radiometers under all-sky conditions (continuation of an international intercomparison project under the

WMO).

The SRB measurements will be obtained only at the Hub site. Instrumentation will include:

In addition to the VCR all-sky cloud imaging system, a dedicated CLASS-type rawinsonde station will also be deployed at the Hub site (section 4.3.4). Soundings will be made during each overpass (ascending and descending) of the two operational NOAA polar orbiting satellites. TOVS temperature soundings will also be collected for comparison.

SPECTRE's main objective is to accurately measure the zenith infrared radiance at high spectral resolution while simultaneously profiling the radiatively important atmospheric characteristics. These data can then be used to meaningfully test detailed infrared radiative transfer models of the gaseous atmosphere (section 6.1). The key features of SPECTRE are continuous, instantaneous profiles of atmospheric temperature, humidity, aerosol and cloud; continuous high resolution spectra of downwelling infrared radiation; radiance rather than flux measurements; careful and frequent radiometric calibration in the field; and redundant measurements. Daytime and nighttime observations will be taken in clear sky situations and when cirrus are present. Measurements will be taken during FIRE missions.

Additional instrumentation and measurements will include Raman lidar (water vapor profiling to at least 6 km), radio-acoustic sounder (RASS, virtual temperature sounding to 6 km), ozone sondes (ozone sounding to 30 km), Dobson total ozone instrument, tethersonde (temperature, water vapor, and ozone sounding to I km), meteorological tower (temperature, water vapor, and ozone sounding to 10 m), surface meteorological observations (pressure, temperature, humidity, and aerosols), and flask samples of trace gases (carbon dioxide, ozone, methane, nitrous oxide, and freons).

The use of multiple spectrometers provides redundancy and together with a strong effort at absolute calibration will insure that the highest possible radiometric accuracy is achieved. The tethersonde and tower observations will fill the near-field 'blind spot' typical of remote sensing prowlers while FIRE rawinsonde observations will serve to define the atmospheric state above the 6-km level. Data from the FIRE lidars will be used to determine atmospheric aerosol optical depths and to characterize the properties of cirrus clouds.

4.3.3 Aircraft Observations

Airborne observations are an essential component of the FIRE Cirrus IFO-II strategy. As remote sensing platforms, research aircraft provide a means to collect very high resolution, multispectral radiometric measurements at wavelengths used by present and future satellite observing systems (satellite simulators) and at viewing geometries (scanning systems) specifically designed to facilitate use of the data in resolving fundamental uncertainties with respect to cirrus radiative properties (science objective 3, section 3). The airborne remote sensing observations are a key element in efforts to quantify the capabilities and limitations of satellite observing systems (science objective 4). The value of the passive sensor data is greatly enhanced by the acquisition of simultaneous observations of cloud physical properties using active systems such as lidar and radar. The total information derived from coincident lidar and passive radiometric observations is significantly greater than the sum of the independent data sets as shown, for example, by Spinhirne and Hart (1990). Direct measurements of broadband radiative fluxes enable the effects of cirrus on the radiative budget of the atmosphere and surface (science objective 5) to be quantified without the deconvolution required of satellite observations.

As in situ observing platforms, aircraft provide a means to directly observe cloud physical properties, especially the microphysical and radiative properties, and to characterize the dynamic, thermodynamic, and radiative structure of the cloud and its environment. Acquisition and analysis of these observations is crucial to the success of efforts to characterize and model cirrus cloud development (science objectives I and 2), particularly with respect to understanding some of the most problematic but critical aspects (section 3.2). Major improvements in aircraft instrumentation have been made in response to the findings from IFO-I (section 2,) specifically, a capability to observe the numbers of ice crystals down to about 10 m m in size and to accurately observe atmospheric water vapor contents. Significant improvements have also been made in the capability for resolving the spectral and angular distribution of radiance, for characterizing the turbulent fluctuations of dynamic and thermodynamic quantities, and for observing the habits of ice crystals. Efforts to develop an airborne polar nephelometer to make direct measurements of cirrus particle phase functions are in progress and, hopefully, the nephelometer will be installed on the NCAR King Air in time for the Cirrus IFO-II. When combined with satellite, airborne, and surface-based remote sensing observations, the in situ observations provide 'ground truth' for analysis of these data and improving remote sensing capabilities (science objectives 4 and 6) as shown, for example, by Sassen et al., (1999), Wielicki et al., (1990), Spinhirne and Hart (1990), Kinne et al., (1990), Ackerman et al., (1990), and Hammer et al., (1990). The microphysical data provide basic information needed to characterize fundamental relationships between cirrus cloud optical and physical properties (science objective 3).

4.3.3.1 Aircraft Platforms and Instrumentation

The ER-2 and the C-131 are remote sensing platforms and will operate at constant altitudes well above (ER-2 at 65,000') and below (C-131 at 10,000') the cirrus clouds. The Sabreliner, King Air, and the other aircraft are primarily in situ platforms. The Sabreliner will be utilized to sample high cirrus layers and the upper portion of cirrus above the operating ceiling of the King Air. The King Air will sample lower cirrus (and altostratus) layers and the lower portions of high cirrus to its operating ceiling. In some instances, the Sabreliner and King Air have conflicting mission objectives with respect to the required sampling strategies. For example, the flight patterns needed to properly sample radiative, dynamical, and moisture fields can preclude adequate sampling of the microphysical composition of the subject clouds (section 4.3.3.2). Similarly, patterns designed to observe the water budget of cloud layers may not permit adequate sampling of cloud radiative characteristics. This is particularly problematic given the prevalence of cirrus occurring in multiple layers over a deep region (5 km or more) of the atmosphere (Sassen et al., 1989; Starr and Wylie, 1990) with a high degree of spatial variability in both the vertical and horizontal distribution of ice water (Heymsfield et al., 1990). Accurate knowledge of the vertical distribution of cirrus ice water content is essential for achieving many of the FIRE Phase H science objectives (section 3). Thus, a third in situ platform is needed. This aircraft will be used solely for obtaining vertical profiles of microphysical parameters. The platform should be a jet aircraft capable of operating at higher levels (and greater range) than the Sabreliner since extended systems of subtropical jet stream cirrus are anticipated where cloud top heights may exceed the operating ceiling of the Sabreliner on some occasions. The cloud top region is of great concern for satellite-based remote sensing studies. The University of North Dakota Citation would satisfy these requirements.

The scanning radiometers will duplicate channels found on present and future satellites and other channels potentially important for the remote sensing of cirrus and other clouds, specifically in the near infrared (MCR). These sensors have very high spatial resolution. The CLS will provide 'cloud-truth' as in IFO-I. The HIS instrument is nearly identical to one of the interferometer spectrometers at the Hub site. The high spectral resolution and accurate calibration of the HIS provides a capability to simulate most any present or future satellite instrument operating in the infrared. The broadband instruments measure both the upwelling and downwelling fluxes.

As discussed in section 4.3.2.1, the potential scientific return from combining polarization lidar observations with coincident millimeter wavelength radar observations for characterizing cirrus is very high as these systems are quite complementary. The solar flux measurements at a relatively low altitude should prove very useful in efforts to characterize the radiative effects of cirrus on the surface. The meteorological and cloud physics instrumentation are not really needed for the remote sensing objectives of this platform but are in-place and may provide useful measurements of low and middle level cloud properties if deemed desirable for a particular mission (This turboprop aircraft generally cannot operate at sufficiently high altitude to serve as an effective in situ cirrus observing platform). The aerosol observations will provide a 'ground truth' for remote sensing observations of the vertical profile of aerosol concentrations and types which is needed by SPECTRE (sections 6.1 and 4.3.2.2).

The broadband flux radiometers are being upgraded to the same instruments flown on the ER-2 and will monitor upwelling and downwelling fluxes. The TDDR and SPERAD instruments also represent major improvements in radiometric instrumentation on this platform. Both have high spectral resolution and high accuracy and are specifically designed for application in sensing cloud radiative properties.

Of particular note are the new small ice particle probes for measurement of the particle number density size distribution down to 10 m m particle sizes. The 2D-C probes are unable to respond to particles smaller than about 50 m m at Sabreliner airspeeds and 25 m m at King Air airspeeds where significant and uncertain 'corrections' are required for particles smaller than about 100 m m. This basic limitation is associated with much of the remaining uncertainty in analysis and modeling of IFO-I data (section 2). For IFO-II, the Sabreliner will also be equipped with ice particle collection devices. The new cryogenic frost point hygrometer (Sabreliner) will permit accurate observations of upper tropospheric water vapor content which have not been available. The meteorological data will be recorded at 20 Hz (analysis at 10 Hz) to better resolve turbulence characteristics and turbulent fluxes in comparison to IFO-I (1 Hz). In addition, measurements of aerosol and cloud condensation nuclei (CCN) concentrations and composition will be obtained from the Sabreliner which will contribute to resolving basic uncertainties with respect to cirrus generation and microphysical development (e.g., Sassen and Dodd, 1989; Sassen, 1989; and Heymsfield and Sabin, 1989). A cold room and laboratory facilities for analysis of ice crystal and aerosol samples will be needed at (or near to) the airfield. A suitable laboratory exists at the University of Missouri-Rolla.

Addition of the new small ice particle probe and the new cryogenic frost point hygrometer is highly desirable. Instruments for detecting aerosol and CCN concentration are also desirable.

More detailed descriptions of aircraft instrument characteristics are given in appendix A.

4.3.3-2 Aircraft Sampling Strategies

Three general principles guided development of sampling strategies for the aircraft platforms. First, analysis of results from the Phase I field deployment indicate that greater coordination of aircraft operations is highly desirable (e.g., Kinne et al., 1989; Wielicki et al., 1990) to facilitate the utility of the data obtained in achieving the science objectives. What is required is that all aircraft operate in close proximity to the surface sites (and each other), especially over the Hub site, and that operations be coordinated with overpasses of the NOAA polar orbiting satellites (or Landsat) to the extent permitted by operational considerations and cloud conditions. Second, the patterns flown should be based on the simplest possible configuration suited to the mission objectives and platform and analysis capabilities in order that mission planning be -streamlined and that execution of the mission proceed in a well-coordinated fashion. In essence, too many options create confusion which can negatively impact the success of a mission. Third, calibration and intercomparison of radiometric instrumentation on the aircraft platforms, on satellites, and at the surface sites will have a high priority and will seek to reference all measurements to a single standard. In particular, as much commonalty as possible is desired for radiative flux sensors with respect to instrument design and bandpass. Besides an increased effort at absolute calibration (blackbody and integrating sphere calibrations on the ground), it is important that intercomparisons be performed in the field over the actual range of operating conditions In this way, the effects temperature and density (airspeed and altitude) variations on sensor performance can be firmly established.

For most missions, the remote sensing aircraft will each fly the same basic racetrack pattern (fig. 5). To the extent permitted by navigational accuracy, the patterns will be flown over exactly the same geographical coordinates with the ER-2 at an altitude of 65,000' and the C131 at about 10,000'. The starting time of each leg will be coordinated to ensure the greatest coincidence of me data in the vicinity of the surface remote sensing sites. Adjustments in the length of the racetrack legs or turn patterns will be used to compensate for differences in operating airspeeds. In general, the racetrack legs (level straight lines during which data is obtained) will be 100 to 150 km in length. Longer legs (up to 500 km in length) may be flown for a few selected missions. Except for a few special missions, the approximate midpoint of the racetrack legs will be located over or just downwind of the Hub site (fig. 6) subject, of course, to the continued presence of clouds over the target area. One of two possible orientations (aircraft heading) will be selected for a given mission based upon the science objectives for that mission. In the first, the racetrack will be oriented perpendicular to the solar plane. Thus, the orientation of successive racetracks will change in time as the solar position changes. This will permit the cross-track scanning radiometers on the ER-2 (Daedalus and MCR) to resolve the principal forward and backward scattering peaks of the cloud bidirectional reflectance pattern. Knowledge of the bidirectional reflectance of cirrus is critically needed for satellite-based remote sensing applications. Knowledge of the cloud bidirectional reflectance also places significant constraints on possible ice particle scattering phase functions which is a major source of uncertainty in OUT understanding of radiative transfer in cirrus and the relationship between cirrus cloud optical and physical properties. In the second orientation, the flight legs will be flown perpendicular to the wind direction at cirrus altitude. This provides the most effective mapping of cirrus cloud structure in time and space, as illustrated in figure 7. Movement of the clouds in time is used to provide a second horizontal dimension in this Eulerian cloud mapping mode. Data taken in this manner will be highly useful for characterizing the cloud scene in comparison to satellite observations, for use in cirrus cloud development studies, and for characterizing the effects of cirrus clouds on the surface and atmospheric radiative budgets. It is expected that the cross-wind tracks will often be approximately perpendicular to the solar plane during some part of a mission since extended cirrus systems often occur in southwesterly flow at this time of year (November 13 - December 7) and missions will usually overlap the time of the afternoon NOAA polar orbiter overpass (-1430 LST).

Figure 5 - Basic aircraft flight patterns for FIRE Cirrus IFO-II.

There are three basic flight patterns that will be used for the in situ aircraft (fig. 5). They are:

The racetrack pattern is primarily used for observing the radiative budgets of cirrus clouds and the radiative impact of the clouds on the radiative budgets of the atmosphere and surface. This pattern is also well-suited for characterizing the statistical properties of bulk cloud radiative properties and horizontal cloud structure and for observing turbulence characteristics and turbulent fluxes. it can be used to generate vertical profiles but the number of levels is limited by the time consumed in making the reverse heading legs at each altitude (the reverse heading legs provide a means of normalizing for the cosine response of the flux radiometers).

The step-down/step-up pattern (vertically stacked straight-line legs, fig. 5) is used primarily for observing vertical profiles of cloud microphysical properties and atmospheric state parameters such as temperature and humidity. Useful radiative flux measurements are obtained if the legs are sufficiently long. The spiral descent at a rate of about 1 in s-1 is primarily used for making microphysical measurements (Heymsfield et al., 1990) and is also well-suited for deriving vertical motions from the observed horizontal winds (Gultepe and Heymsfield, 1990). However, the accuracy of the radiative flux measurements may be compromised during such a descent. Similarly, radiative measurements taken during descent or ascent from one level leg to the next in the step-down/ step-up pattern will not be useful. The advantage of the spiral descent versus the step-down/step-up pattern is that the observed profiles of cloud microphysical properties are more representative of the conditions present in an atmospheric column at a particular time because the profiles are obtained in about 15 minutes rather than the time period of an hour or more required for step-down/step-up profiling. Each of these pattern may be flown in both a Eulerian and a Lagrangian manner (e.g., fig. 5). In the Eulerian mode, the orientation of the racetrack and step-down/step-up legs will be along the wind direction at citrus altitude. The upwind end of each leg will generally coincide with the a location just upwind of the Hub (or R2) remote sensing site. Thus, the flight tracks of the in situ aircraft will cross the flight tracks of the remote sensing aircraft, often at a right angle. The Eulerian spiral descent will be flown over the Hub site. The diameter of the spiral is usually about 10 km. Data obtained from Eulerian sampling patterns are best suited for combination with satellite and surface-based remote sensing data.

In the Lagrangian mode, the flight legs are also oriented along the wind. The patterns may be initiated in the vicinity (or upwind) of the Hub site but the upwind terminus of each

In the Lagrangian mode, the flight legs are also oriented along the wind. The patterns may be initiated in the vicinity (or upwind) of the Hub site but the upwind terminus of each successive leg is moved downwind to account for horizontal advection of the cloud by the wind (wind drift). Thus, the legs get progressively farther from where the pattern was initiated. Although the flight legs may remain within the view of the cross-track scanning instruments on the remote sensing aircraft, all the flight tracks may not cross the remote sensing flight tracks. Similarly, coordination with the surface-based observing sites is necessarily less. The Lagrangian spiral also drifts with the wind but can generally be confined to a small area close to the Hub site since each descent takes only about 15 minutes, i.e., the entire pattern can then be repeated beginning at the initial location. An entire step-down pattern may take more than an hour and consequently drift some distance during that time (50 m s-1 wind speeds yield a 180 km displacement in I hour).

The lengths of flight legs for the racetrack and step-down/step-up patterns vary depending on the mission objectives, as described below, and the operating characteristics of the aircraft. While the turboprop King Air can fly relatively short legs (20 km), it is not practical to operate jet aircraft in this manner (50 km legs are a minimum). Moreover, only the King Air is suited to performing the spiral descent pattern. Based on experiences from IFO-I, the King Air will often perform two or more of the above patterns in a given mission (below). This takes advantage of the operational flexibility of this platform and maximizes the overall applicability of the data obtained. Although this may at first seem to violate the maxim of simplicity, experience has shown that a set multi-mode pattern is operationally practical and yields significant scientific advantage.

4.3.3.3 Aircraft Missions

Approximately fourteen (14) coordinated multi-aircraft missions will be flown during the 25 days of FIRE Cirrus IFO-II. [A mission may involve more than one flight of a given aircraft.] Other single-aircraft missions may be conducted on a non-interference basis at the discretion of the Aircraft Principal Investigator. Non-interference means not only that such a mission will not be conducted during a coordinated mission but that it will not impact the operational capabilities for conducting a coordinated mission, e.g., on the same or following day. One of three basic multi-aircraft mission plans will be executed on most aircraft experiment days. These coordinated missions are designed to satisfy the sampling and data requirements for achieving most of the scientific objectives of FIRE Phase II (section 3). Special missions focused on specific objectives requiring unique sampling strategies or conditions will also be performed.

The Type A mission is essentially a Eulerian mission (figs. 6 and 7). It is the mission that will be executed for the three large-scale case studies (intensive rawinsonde launches, section 3.3.4). The primary objectives of this mission are to characterize the time-dependent cloud and radiative fields, cirrus cloud radiative and physical properties, and upper tropospheric water vapor and cloud water budgets within a fixed mesoscale region. The coordination of aircraft flight patterns and surface-based remote sensing observations are maximized in this mission. Obtaining coincident satellite observations, especially by NOAA polar orbiting satellites, is a high priority in planning and executing this mission. Thus, besides supporting studies of cirrus cloud development (science objectives I and 2), this mission will also provide data well-suited for satellite and surface-based remote sensing studies (4, 6) and for quantifying the radiative impact of the clouds on the TOA, atmosphere and surface radiation budgets (5). The data will be useful, but not optimal, for characterizing relationships between cirrus cloud optical and physical properties (3).

In a Type A Mission, the remote sensing aircraft will fly coordinated cross-wind racetracks (100 to 150 km in length) with one leg passing over the Hub site (midpoint). The Sabreliner and King Air will fly combination along-the-wind flight patterns over the Hub site while the third in situ aircraft will operate in coordination with the remote sensing aircraft and R2 remote sensing site. The Sabreliner will begin with a 50-km racetrack just above cloud top followed by a step-down pattern (50 km legs at 2000' intervals) to below the operating ceiling of the King Air (or cloud base). This pattern is then repeated. The King Air will begin with a Lagrangian spiral (initiated over or upwind of the Hub site) from its operating ceiling (or cloud top) to cloud base followed by a 30-km racetrack just below cloud base. A step-up pattern (20 km legs at 2000' intervals) to operating ceiling (or cloud top) will then executed. This pattern is repeated one or more times. At the conclusion of the mission, an Eulerian spiral is performed over the Hub site. This King Air flight pattern was successfully executed in IFO-I and is illustrated in figure 8. The third in situ aircraft will perform 50-km along-the-wind step-up/stepdown profiles between cloud base and cloud top on a parallel heading. The flight track will intersect the flight track of the remote sensing aircraft and extend toward or over the R2 remote sensing site.

There is a potential conflict in that the Sabreliner and King Air are operating in close proximity. It is anticipated that such conflicts can be readily resolved in a safe manner. In a situation of two or more distinct cloud layers, the operating altitude ranges will not overlap, i.e., the King Air will work the lower cirrus layer and the Sabreliner will work the upper cloud layer(s). In this situation, which will likely occur during some missions, an additional racetrack pattern must be flown at cloud top (King Air) and cloud base (Sabreliner) of the respective cloud layers. The third in situ aircraft will work both cloud layers. Furthermore, a capability for some

adjustment of aircraft flight levels based on concurrent surface-based observations (lidar and radar) may be desirable.

The Type B Mission is primarily a Lagrangian mission. The primary objectives for this mission are to characterize the radiative budgets and broadband radiative properties of individual cirrus clouds in relationship to the horizontal and vertical structure of the cloud (science objective 2), to characterize the turbulent structure and associated vertical transports within the cloud, especially in the cloud boundary layers where turbulent interactions have greatest consequence (2), and to characterize the microphysical development of the cloud (2). Thus, this mission is focused on some of the more problematic aspects of understanding cirrus cloud development in support of cloud-scale modeling studies and parameterization development (1). The data will be directly applicable for satellite-based studies (4) and for characterizing the radiative impacts of cirrus (5). Coordination with the surface sites is highly desirable but will be necessarily somewhat limited in the case of the in situ platforms. Nonetheless, the data will provide useful information for improving the utilization of surface-based remote sensing observations for quantitative studies of cirrus clouds (6).

In a Type B Mission, the remote sensing aircraft will operate in the same manner as in a Type A mission, i.e., cross-wind racetracks over the Hub site. The in situ aircraft will conduct along-the-wind Lagrangian flight patterns (drifting with the wind) on specific advecting cloud targets.

The Sabreliner will fly 70 to 100-km racetracks just above cloud top, and at 2000' and 4000' below cloud top. The King Air will fly a Lagrangian spiral from its operating ceiling (or cloud top) to cloud base followed by 70 to 100-km racetracks just below cloud base and at 2000' and 4000' above cloud base. These patterns should cross the flight track of the remote sensing aircraft and overfly one of the remote sensing sites during portions of the mission. Ideally, the Lagrangian spiral would be performed over the Hub site. In this mission, the Sabreliner and King Air may work on the same cloud target or in an independent fashion on separate targets. The third in situ aircraft will perform along-the-wind Lagrangian step-down/step-up microphysical profiling below the Sabreliner (to cloud base) or above the King Air (to cloud top) when they are working separate targets or over the Hub or R2 remote sensing sites (cloud top to cloud base) when they are working the same target.

5.3 The University of Utah ETO Project

Cirrus cloud research based at the University of Utah Facility for Atmospheric Remote Sensing (FARS) is designed to support satellite validation objectives of FIRE Phase II (section 5.1) by providing routine extended time observations (ETO) of cirrus clouds on a 10 'primary' days per month basis at and around the times of NOAA polar orbiting satellite overpasses and GOES satellite observations. In this way, a much greater number of cirrus cloud/satellite observation coincidences are obtained than possible during intensive field observation campaigns (IFO-I or IFO-II). Routine lidar observations are used to provide a direct measure of cirrus cloud base height and often cloud top height, as well as derived estimates of visible cloud optical thickness, These observations of cirrus cloud properties are compared to satellite-derived cloud properties and provide a data base encompassing a broad range of meteorological and surface conditions.

The cirrus cloud observations are also used to enhance basic knowledge of the physical and microphysical properties of cirrus clouds and the characteristic cloud structures and generating mechanisms (Sassen, 1989). Observational periods are typically two to four hours in duration but may be extended to six hours or more on some occasions. In addition to regularly scheduled observations (~10 per month), observations bracketing satellite overpass times are often collected on other days when cirrus ,ire present. Special attention is given to the identification of supercooled liquid water regions (Sassen et al., 1989) within or associated with cirrus, and to the occurrence of oriented planar ice crystals in cirrus which creates a highly anisotropic scattering media. These data enable statistical characterization of cirrus cloud properties, such as areal coverage, thickness, and structure/morphology as functions of season, synoptic conditions, and Cirrus Cloud type. The findings from this basic research component of FIRE Phase II will be applied in generating comprehensive cirrus cloud climatologies and will aid in cirrus cloud modeling studies ranging front the microscale to the regional scale (mesoscale).

In addition to initiating collection of the satellite validation data base observations, ETO activities at FARS during FIRE Phase I focused mainly on the visible light scattering properties of cirrus using dual polarization lidar (ruby), net flux radiometers, and all-sky photography. For Phase II, a broader range of passive remote sensing measurements will be obtained (table 7). A narrow-beam, infrared window radiometer will be co-aligned with the lidar system. These observations together with ancillary rawinsonde observations will be used to derive the infrared emittance of the cirrus clouds (the LIRAD method, Platt, 1979; Platt and Dilley, 1981). Observations using a greatly expanded complement of solar radiometers and the infrared observations will serve to quantify the radiative effects of cirrus clouds on the surface energy budget. Coincident observations using a Ka-band (0.86 mm) radar will also be obtained on

selected occasions. These observations will be used to advance our basic knowledge of cirrus clouds and to explore the potential benefits of multiple wavelength observations for investigating cirrus (e.g., Sassen et al., 1999) as discussed in section 4.2.

The University of Wisconsin at Madison began an extended time study of the optical and physical properties of cirrus clouds in November 1989. The experiment lasted approximately one month (Grund et al., 1990). An additional experiment is planned for the spring of 1990. The objectives of this pilot experiment are:

Instrumentation for the pilot experiments are listed in table S. The VIL scanning lidar produces near-continuous 100-km wide cross-sections of cirrus cloud backscatter at a 100 m resolution. Time sequences of these data are used to reconstruct a three-dimensional images of cirrus cloud fields. The data are used to determine cloud height and structure over the region. The vertically pointing HSRL provides calibrated measurements of the vertical profile of cirrus backscatter cross section, optical depth, and backscatter phase function at a wavelength of 0.532 m m within a 100-m field of view. Two vertically pointing high spectral resolution (better than 1 cm-1) infrared interferometer spectrometers are used to observe downwelling spectra over the spectral region from 3.5 to 20 pm. One is the same instrument that will be flown on the ER-2 in Cirrus IFO-II (section 4.3.3) and was also flown in IFO-I (down-looking), The second is a compact model that is being developed for deployment at the surface sites (Hub and R2, section 4.3.2) in IFO-II. These observations provide estimates of the variability of the infrared spectral emittance of cirrus clouds (Ackerman et al., 1990). The field of view is about 100 meters at cirrus altitude. Satellite observations are collected from all channels on the operational NOAA polar orbiting and geosynchronous satellites, including the multispectral infrared observations used for temperature and moisture sounding (HIRS and VAS). The satellite observations are contained within the sampling domain of the scanning lidar system. A CLASS rawinsonde system is used to observe the vertical structure of pressure, temperature, humidity, and winds for use in processing the data (3-D lidar displays), and analyzing the infrared observations (HIS and satellite). Routine meteorological observations are also collected.

Besides addressing the experiment objectives (above), analysis and integration of this diverse suite of simultaneous remote sensing observation will provide needed experience in use of these data and facilitate similar analysis of Cirrus IFO-II observations. The potential gain in total information content of the combined observations (synergism) will be explored.

One of the key questions to be addressed in the context of FIRE Phase U is me relationship between synoptic and mesoscale circulations and cirrus formation and maintenance. Recognizing the importance of defining the regional circulation, the FST chose to locate the planned experiment in the immediate vicinity of the proposed NWS wind profiler network (figure 10). This network will provide hourly--average horizontal wind speeds and directions. In addition to the NWS network, two or three additional profilers will be deployed for the experiment (sections 4.3.2.1 and 4,3.4).

A crucial variable to be obtained from the wind profilers is the vertical velocity (section 3.2.3). While it is possible to measure updraft vertical velocities on the convective scale, direct measurement of the vertical velocity on the mesoscale is extremely difficult due to the typically small magnitudes on the order of a few cm per second. Alternatively, vertical velocities can be deduced from a network of profilers by computing the divergence of the horizontal wind, and then computing the vertical velocity from continuity. This latter approach has been demonstrated successfully for a few case studies in Colorado and Pennsylvania using arrays of three profilers, but has not been thoroughly examined, particularly for cirrus conditions.

From the perspective of cloud studies, measurements of the vertical velocity alone are inadequate. Cloud occurrence and cloud radiative properties such as optical depth must also be measured. These cloud parameters can then be correlated with mesoscale vertical velocity and average atmospheric thermodynamic structure, leading to quantitative understanding of the cirrus formation and maintenance problem.

Based on these concepts, the Pennsylvania State University has begun a 2-year pilot program designed to address the following objectives:

Preliminary data was acquired during the 1989-90 winter using Penn State's three 50 Hz wind profilers. The prowlers were deployed in central Pennsylvania at the vertices of an approximately equilateral triangle 140 km on a side. Hourly-average wind profiles were recorded continuously for a 3-month period from December to February. Analysis of these data suggests that the vertical velocity can be deduced using the divergence approach, but that missing or incorrect data pose a considerable problem. Fairly sophisticated data smoothing and filtering techniques are being examined to solve these problems.

A second and more extensive phase of the pilot program will be carried out during the 1990-91 winter. The instrumentation to be deployed for this phase is listed in table 9. The three wind profilers will be used again in a triangular array to obtain an average vertical velocity. Temperature and humidity structure will be obtained locally it low levels using the 400 MHz profiler and RASS system. Upper level structure will be found from a combination of both routine NWS radiosondes and special sondes released locally in central Pennsylvania. The 94 GHz (3 mm) Doppler radar will be used primarily to determine cirrus cloud presence and height. In addition, some analysis of cloud structure may be carried out. Cirrus infrared emissivity will be determined by measuring the downward radiance between 9.5 and 11.5 microns with a PRT-5. Broadband solar and infrared downwelling fluxes will be measured with standard Epply instruments. With the exception of the three 50 MHz profilers, all the equipment will be collocated near State College, Pennsylvania.

The Penn State mesoscale model will be run in a real-time forecasting mode for Pennsylvania for some part of the experimental period. Periods of special interest with regard to cirrus presence will be selected. Temperature, humidity, and vertical velocity fields will be extracted from the model runs for comparison with the data.

Preliminary data analysis should be completed by late summer, 1991. The focus of this analysis will be on determining how well deduced vertical velocities agree with model vertical velocities and correlate with cloud Occurrence. Assessments of cirrus predictability based on model simulations and/or measured wind fields will also be made. This information will be made available to participants in the Cirrus IFO-II field program. It is anticipated that this information will be useful in designing operational strategies for the field campaign.

Table 9 - Instrumentation for the Pennsylvania State University Cirrus Remote Sensing Pilot

Experiment

Pennsylvania State University

Active Remote Sensors: 50 MHz wind profilers (3)

Passive Remote Sensors: 9.5-11.5 m m radiometer (PRT-5)

5.6 NOAA/WPL Cloud Lidar and Radar Exploratory Test (CLARET)

NOAA's Wave Propagation Laboratory conducted the field phase of the Cloud Lidar and Radar Exploratory Test for one month beginning September 6, 1989 (Eberhard et al., 1990). The overall objectives were to:

Good data were acquired for 22 episodes (9 cirrus, 5 water cloud, and 6 clear air), where each episode was 1-3 hour duration. Data for many additional hours are available from unattended instruments (including radar) and sometimes one lidar.

CLARET involves both analytical and experimental investigations. Some specific research tasks are:

A follow-on field experiment is planned (pending adequate funding) for spring 1991 that will include the upgraded 8 mm wavelength, dual-polarization, Doppler radar (Kropfli et a]., 1990) and in situ measurements by aircraft for validation of concepts and algorithms. Participation of CLARET components in FIRE Cirrus IFO-IL will be important for validation and application of new remote sensing techniques and achievement of the FIRE Phase II cirrus .science objectives.

APPENDIX A

DESCRIPTION OF AIRCRAFT AND INSTRUMENTATION FOR CIRRUS IFO-II

TBD - Aircraft PI's

APPENDIX B

LISTING OF NWS STATIONS AND CODES

NWS Stations - Type A 15 stations

State City Type Region Call

Arkansas: Little Rock WSFO S LIT

Colorado: Denver WSFO C DEN

Kansas: Dodge City WSO C DDC

Topeka WSFO C TOP

Missouri: Monett WSMO C UNM

Nebraska: North Platte WSO C LBF

Omaha WSFO C OMA

New Mexico: Albuquerque WSMO S ABQ

Oklahoma: Oklahoma City WSFO C OKC

Texas: Amarillo WSO S AMA

Del Rio WSO S DRT

El Paso WSO S ELP

Longview WSMO S GGG

Midland WSO S MAF

Stephenville WSMO S SEP

Type Code: WSFO - Weather Service Forecast Office (main office)

WSO - Weather Service Office

WSMO - Weather Service Meteorological Observatory

WSCMO - Weather Service Contract Meteorological Observatory

Note that observatories are not routinely staffed on a 24-hour basis, although Observers are present to take routine surface and upper air observations.

Region Code:

E- Eastern

C- Central

S- Southern

W- Western

NWS Stations - Type B 36 stations

State City Type Region Call

Alabama: Centerville-Brent WSMO S CKL

Arizona: Tucson WSO W TUS

Winslow WSO W INW

California: Oakland WSCMO/WSO W OAK

San Diego WSCMO/WSO W MYF

Colorado: Grand Junction WSO C GJT

Idaho: Boise WSFO W BOI

Illinois: Peoria WSO C PIA

Salem WSCMO C SLO

Louisiana: Boothville WSCMO S BVE

Lake Charles WSO S LCH

Michigan: Flint WSO C FNT

** Sault Ste. Marie WSO C SSM

Minnesota: ** International Falls WSO C INL

St. Cloud WSO C STC

Mississippi: Jackson WSFO S JAN

Montana: ** Glasgow WSO W GGW

** Great Falls WSCMO/WSFO W GTF

Nevada: Ely WSO W ELY

Mercury WSMO W UCC

Winnemucca WSO W WMC

North Dakota: ** Bismark WSFO C BIS

Ohio: Dayton WSCMO E DAY

Oregon: Medford WSO W MFR

** Salem WSO W SLE

South Dakota: Huron WSO C HON

Rapid City WSO C RAP

Tennessee: Nashville WSMO S BNA

Texas: Brownsville WSO S BRO

Victoria WSO S VCT

Utah: Salt Lake City WSFO W SLC

Washington: ** Quillayute WSCMO W UIL

** Spokane WSO W GEG

West Virginia: Huntington WSO E HTS

Wisconsin: Green Bay WSO C GRB

Wyoming: Lander WSO C LND

** Additional Type B stations

APPENDIX C

FIRE SCIENTIFIC ORGANIZATION

TABLE OF CONTENTS

C.1.0 LEAD MANAGEMENT C- 1

C.2.0 PROJECT MANAGEMENT C- 2

C.3.0 FIRE II SCIENCE TEAM C- 3

C.3.1 FIRE Working Groups C- 4

C.4.0 DATA MANAGEMENT C- 7

C.4.1 Responsibilities C- 8

C.4.1.1 Principal Investigators C- 9

C.4.1.2 Working Groups C- 9

C.4.1.3 FIRE Central Archive C-10

C.4.1.4 Project Office C-12

C.4.2 Data Products C-12

C.4.3 Data Management Plan C-13

C.4.4 Data Management Implementation C-14

C.4.5 Data Protocol and Publications Plan C-15

C.4.5.1 Data Protocol C-15

C.4.5.2 Data Publication C-17

C.5.0 INTENSIVE FIELD PROGRAM MANAGEMENT C-18

C.6.0 NATIONAL AND INTERNATIONAL COLLABORATIONS C-22

C.7.0 FIRE MISSION/MEETING MILESTONES C-23

APPENDIX C

FIRE SCIENTIFIC ORGANIZATION

C.1.0 LEAD MANAGEMENT

NASA will serve as the lead agency for FIRE. The Radiation, Dynamics, and Hydrology Branch, NASA Headquarters, will provide overall cognizance of the FIRE project of those activities associated with processes, especially those with regional emphasis; the Physical Climate and Hydrologic System Branch, NASA Headquarters, will provide cognizance of those activities associated with models, especially those with global emphasis, and will represent FIRE to the Working Group on Data Management for WCRP Radiation Projects. The Office of Naval Research will serve as the lead agency for the ASTEX component of FIRE, under the direction of the Meteorological Research Program Office. Additional support for both FIRE and ASTEX will be provided by NSF, DOE, DOD, and NOAA. Contact points are located at the following offices:

C.2.0 PROJECT MANAGEMENT

The FIRE project activities will be managed by the FIRE Project Office at NASA's Langley Research Center. The FIRE project manager will be responsible for the overall management, coordination, and reporting of the project activities. These responsibilities will include interacting with the Radiation, Dynamics, and Hydrology Branch and Physical Climate and Hydrology Branch management at NASA Headquarters, the Meteorology Research Program management at ONR, and the other agency representatives; overall cognizance of project planning, schedules, and field operations; and allocation of approved funding to support the scientific objectives. The project manager will be assisted by a project scientist and a project staff, which will include: an Operations Manager, a Data Manager, a Satellite Coordinator, and a Project Coordinator.

Project Scientist - The FIRE Project Scientist will be responsible for all scientific aspects of FIRE. The project scientist will (a) act as a scientific advisor to the project manager; (b) represent the members of the science team in their relationship with the project manager; (c) maintain an oversight of the scientific integrity of the project; (d) review and make recommendations of proposed modifications to selected proposals as warranted; and (e) participate in negotiations regarding allocations of specialized project resources among approved investigations.

Operations Manager - The Operations Manager will have overall responsibility for organizing, scheduling, and conducting field operations, preparation of mission plans, establishing mission objectives with the MST Chairman, determining special support requirements, conducting planning and debriefing sessions, and operational procedures. The Operations Manager is also responsible for designing the specific and facilitating operationally feasible flight profiles to meet the scientific objectives and for coordinating the experimental requirements of measurement platforms for resources, testing, and integration for all field measurements. Lastly, the Operations Manager is responsible for the setup, testing, and operation of the mission operations site.

Data Manager - The Data Manager will be responsible for coordinating the types, scope, and quantity of data collected during the intensive field and extended-time activities and its archiving. The Data Manager will serve as the Chairperson for the Data Management Working Group. The Data Manager will ensure that the data is submitted as required to the FIRE Central Archive and eventual release to the scientific community.

Satellite Coordinator - The Satellite Coordinator will be responsible for providing information on the periods and locations of satellite observation within the IFO areas and ETO locations. The Satellite Coordinator is responsible for determining that the appropriate satellite observations are being collected, archived, and made available to the FST.

Project Coordinator - The Project Coordinator will provide administrative support to the FIRE/ASTEX Project Office, including logistical and reimbursement support in the planning and conducting of FST meetings and field missions as required.

C.3.0 FIRE II SCIENCE TEAM

The FIRE II Science Team (FST) will be comprised of the principal investigators of those agency-approved proposals directly related to FIRE research. Appointment to the FST will be based on the recommendation of the supporting agency. The term of participation on the team will continue as long as the approved research continues. The team shall determine its own structure and method for interaction. Ex officio members may be appointed to the team, as needed, by the project manager.

The FST will be responsible for implementing the broad scientific objectives of the project in accordance with the selected investigators. Additional responsibilities will be to:

Table C. I lists the FST members, their institutions, and working groups in which they will be participating. Figure C.1 shows a schematic of the relationships between the agencies, the agency program managers, the FIRE/ASTEX Project, and the FST.

C.3.1 FIRE Working Groups

FIRE has been designed to address a number of complicated scientific problems. The FIRE strategy is to assemble those members of the scientific community with interests in these problems and to create an environment which encourages them to collectively pursue the solutions to these problems. FIRE working groups representing the themes of cirrus cloud systems, marine stratocumulus cloud systems (ASTEX), and data management will be constituted from the FST membership as follows:

It is clear that modeling of physical processes, which includes all processes important to the maintenance and to the characterization of cloud systems, and large scale GCM and climate modeling are critical to the successful attainment of FIRE's goals; it is extremely important that these modeling components be integrated with the observational components of FIRE. For this reason, FIRE investigators with modeling interests will play important roles in the cirrus, ASTEX, and extended time working groups.

The first two data gathering working groups will accept the following responsibilities: