& Microscale Meteorology Division
National Center for Atmospheric Research
|Science - Science Plan|
Five Years and Beyond
Scientists engaged in mesoscale and microscale meteorology research are poised to make major progress on improving forecasts of precipitation, and accounting more accurately for mesoscale and microscale processes in models of the climate and weather. This progress is made possible by recent advances in observing systems, computers, and theoretical understanding.
Spectacular advances have taken place in our ability to observe the atmosphere on the meso- and microscales. These include the installation of NEXRAD, wind profilers, and observations from commercial aircraft. There are new radar techniques that allow detection of the microphysical properties of clouds; mobile radars on the ground and airborne radars that allow mapping the structure of mesoscale convective systems and even tornadoes at unprecedented resolution; satellite observations that take advantage of the GPS system to sample the atmosphere globally with near-mesoscale resolution; new particle imagers that can clearly define the drop and ice-crystal structure of clouds; and, new airborne instrumentation measuring chemical, physical and dynamical properties of the atmosphere.
Computing power has advanced to the point where it is now possible to perform idealized simulations of atmospheric flow over large domains with high resolution; in particular, small-scale processes, which cannot generally be resolved in models for climate and weather forecasting, can be explicitly computed. These idealized simulations will allow investigations of the effects of the unresolved processes on larger scales, and perhaps even a method to account for them in realistic lower-resolution models. These simulations are necessary steps on the path to the development of modeling systems that accurately forecast weather, and simulate climate change on long time scales.
Other advances have occurred in the theory of cyclones and mesoscale convective systems, in the predictability of mesoscale weather systems, in the interactions of the atmosphere with orography at the meso- and microscales, in the development and structure of turbulence and boundary-layer processes, and in ice processes at very low temperatures. New mathematical techniques involving adjoints of prediction models, and the 3DVAR and 4DVAR data assimilation systems promise to make impressive strides in weather forecasting.
The Mesoscale and Microscale Meteorology Division of NCAR has contributed strongly to all of these areas. MMM is unique because of the breadth of its expertise in mesoscale and microscale meteorology and its ability to integrate its program to focus this diverse expertise on important problems that require an interdisciplinary approach. The division will continue to play a major role in mesoscale and microscale meteorology research and, as it plans for the future, will structure its program to address a few specific problems that have been identified as objectives of the U.S. Weather Research Program (USWRP) and the Global Change Research Program (GCRP). The two main research themes of the division are Prediction of Precipitating Weather Systems and Cloud and Surface Processes Parameterizations.
For prediction of precipitating weather systems, the main goal is to advance the understanding and prediction of significant precipitation events in order to reduce substantially forecast errors toward the limits of predictability. To accomplish this goal, there will need to be research to develop the numerical forecast tools, including high-resolution forecast models and the data assimilation systems needed to initialize those models. In addition, there will need to be efforts to understand the limits of mesoscale predictability and what it is going to take to drive the forecast systems toward those limits. Understanding the dynamics and life cycles of the weather systems responsible for major precipitation events will be necessary to understand the factors that limit predictability and its dependence on the weather systems themselves. Programs within the division will address each of these areas. Accurate prediction of precipitation is one of the major objectives of the USWRP and the division's program is intentionally aligned with this national program.
For assessing the cloud and surface processes parameterizations, the main goal is to quantify the large-scale effects of mesoscale and microscale processes and to develop physically based methods to account for these effects in large-scale models. The emphasis is on understanding how the moist atmosphere, land and ocean surface, and hydrological processes interact and how these processes can be quantified. This effort will include four key research areas: deep convective cloud systems, boundary-layer clouds, surface-atmosphere interactions, and the physical chemistry of clouds. Nonhydrostatic fine-scale modeling using large-eddy simulation and cloud-resolving models will be an important component of our approach to these problems since they allow high-resolution definition of the mesoscale and microscale systems involved, and are therefore a good means for testing methods to quantify the effects of these processes on larger scales. Critical for the success of our program will be the evaluation of these models against detailed observational studies of the underlying physical processes. The goals of this program will contribute to the objectives of the GCRP Global Energy and Water-cycle Experiment (GEWEX) and the Cloud System Study (GCSS) component of GEWEX.
The present document is the "Science Plan" of the Mesoscale and Microscale Meteorology Division of NCAR for the next five years and beyond. The plan describes what the division intends to accomplish during this period and the rationale for the choice of topics. Details on how the Division plans to meet these goals are outlined in the Director's Message of the FY 1999 Annual Scientific Report (ASR) where we discuss our research plans for a two-year period; the ASR gives a detailed account of all the current research of the division. The Science Plan has been developed with the existing expertise of the division in mind and with the expectation that fulfillment of the goals will require extensive collaboration with other groups outside the division, such as other NCAR divisions and the university community. It is recognized that additional expertise may be needed to fully implement our plans, but we are optimistic and excited about the prospects of achieving the research goals outlined within our Science Plan.
I. SCOPE OF THE PROGRAM
The mission of the Mesoscale and Microscale Meteorology Division of NCAR is to advance the understanding of the meso- and microscale aspects of weather and climate, and to apply this knowledge to benefit society.
The division strives to address the most important and fundamental scientific themes in mesoscale and microscale meteorology, with emphases on understanding and forecasting weather and on evaluating the influence of meso- and microscale processes on larger-scale phenomena. These basic scientific themes are reflected by the nature of the division's subgroups:
To maintain this scientific program and respond quickly to new research opportunities, the scientific staff must have diverse expertise. Therefore, a central aspect of our science plan is to attract and retain leading scientists in the field and to develop the careers of young scientists toward leadership roles. To achieve this goal, the division will respect scientific uniqueness and will promote collaboration and teamwork through maintenance of a congenial, supportive, and professional working environment.
The division also devotes resources to community service activities. This service is not only regarded as an important part of our role as a national center, but it enhances our scientific program. Current service functions include the maintenance and support of a community mesoscale numerical prediction model (MM5), including training on its use and making it generally available to the community; the maintenance of access to a suite of data analysis software packages; and the support of a strong and viable visitor program.
Our research in mesoscale and microscale meteorology spans a broad spectrum, as reflected by the diverse research interests within the Division subgroups mentioned above. For example, MMM scientists have research projects ranging from the mechanics of the freezing process, wildfire modeling, advanced numerical methods in computational fluid mechanics, to theoretical models of synoptic-scale cyclones and anticyclones. However, as is appropriate at a national center, our scientific program has a few major themes that require extensive collaboration among collections of scientists within MMM, NCAR, and the university community. MMM has chosen the following two major themes:
These themes will be supported through an optimal application of MMM's existing expertise in observing, modeling, and describing the fine-scale structure of the atmosphere and through given national and international initiatives in advancing the ability to predict the weather and climate change. As such, close collaboration, coordination, and shared leadership will be required with researchers within universities and NCAR and with mission agencies, both nationally and internationally. The following two sections give a more- detailed account of how we plan to contribute to the achievement of these goals.
II. Prediction of Precipitating Weather Systems
1. Mesoscale Predictability
Goal: To estimate upper bounds on the time over which forecasts of precipitation within mesoscale weather systems retain useful skill, and to identify the key physical processes that limit forecast skill.
Mesoscale weather systems are typically embedded within synoptic-scale flows, while at the same time contain a variety of smaller-scale motions such as moist convection, symmetric instabilities, gravity waves, and frontal and boundary-layer circulations. Since these motions are unique to the mesoscale, studies of mesoscale predictability with models in which these mesoscale motions are parameterized are of limited utility. Moreover, the fact that these motions are highly intermittent and localized casts doubt on predictability results based on turbulence closures, which assume the turbulence to be homogeneous and isotropic. These facts suggest that further progress in understanding mesoscale predictability hinges on knowing how uncertainty in mesoscale forecasts is influenced by uncertainty in both smaller and larger scales.
These issues are being addressed by examining the growth of small differences (or "errors") in the initial conditions for forecasts of interest, such as the snow storm of 24--25 January 2000 that brought Washington, D.C. to a standstill. These experiments, which focus on the evolution of the scale and amplitude of the initial error, employ local horizontal resolution of a few kilometers in order to minimize spurious effects of parameterized physics and limited resolution. It appears to date that moist processes exert a controlling influence on the error growth.
To gain further understanding of how mesoscale forecast errors evolve in the presence of moisture, idealized numerical simulations are planned of the up-scale organization of moist convection in simple environments, and of synoptic-scale flows with embedded precipitation systems. The simulations of convection initiation and organization, which are intimately related to efforts to understand the life cycle of precipitation systems discussed below, will assess the sensitivity of initiation, cell evolution and up-scale organization to perturbations of initial conditions. The simulations of synoptic-scale flows will begin with simulations of idealized baroclinic waves and fronts to assess the degree to which uncertainties in the forcing at synoptic scales can alter the predictability of meso- and convective-scale flow and, conversely, the way in which variability at sub-100 km scales can influence larger scales.
Through these and other sensitivity studies, guidance will be provided for the development of the model forecast systems by identifying those physical processes most crucial for mesoscale forecast accuracy.
2. Life Cycle of Precipitating Weather Systems
Goal: To understand the life cycles of precipitation events and the implications of their behavior on predictability, data-assimilation requirements, and the treatment of forecast-model physics.
a. Convection initiation
An important element of this effort is to improve the treatment of convection initiation in forecast models, particularly during the spring and summer when large-scale forcing may be weak. Achievement of this goal is impeded by the lack of understanding of the mechanisms responsible for convective triggering. This lack is in large part due to the absence of fine-scale observations which might indicate the ways that convection is initiated (or "triggered''), such as convergence lines, boundary-layer rolls, gravity waves, differential heating, and orographically modified flow. To better document the local initiation processes and linkages with larger-scale forcing, MMM expects to participate with NOAA/NSSL, ATD, RAP, and the University of Oklahoma in the planning and execution of field experiments, such as the International H2O Project (IHOP) which will use new observational tools to probe the fine scale structure of water vapor within the planetary boundary layer.
b. Long-time-scale dynamics of mesoscale convective systems
Until quite recently, much of the effort devoted toward understanding three-dimensional organized convection stemmed from the Pre-Storm Experiment in the Midwest U. S. in 1985. Over the past few years, the rich spectrum of organized convection in midlatitudes and the tropics was realized through the examination of new, large-domain radar and satellite data sets made possible because of the NWS modernization. In addition, newer cloud-resolving models are able to capture the growth stage of a variety of MCSs. Despite these advances, there is relatively little understanding of MCS behavior in the mature and decaying stages. Cloud-resolving simulations are unable to capture the mature-to-decaying phase of systems while observational field studies tend to be too confined geographically to document both the mesoscale environment and convective scale circulations.
A collaborative effort with ATD, RAP, NOAA/NSSL, and university scientists to analyze radar and satellite information over domains of approximately 1000-2000 km is being conducted to document the mature convective system and its interaction with the environment. This effort complements field-project studies and, combined with modeling efforts incorporating ever-more complicated environmental structure, will allow us to address the relative importance of synoptic-scale forcing versus internal dynamics in MCS evolution, the mechanism(s) by which convective systems regenerate, the interaction of mesoscale convective vortices (MCVs) with convection, and the feedback of microphysical processes and radiation onto system-scale dynamics. The Bow Echo and MCV Experiment (BAMEX), planned for 2002 or 2003, will address many of these issues in the context of convective systems developing in both strongly sheared and weakly sheared environments.
c. Tropical cyclones
The single most damaging weather phenomenon is the landfalling tropical cyclone. The need for continued research on formation, intensification, and track of tropical cyclones, the promise for major advances in prediction in the near future, and the large potential societal impacts made landfalling tropical cyclones one of the main foci of the USWRP. Research in MMM on tropical cyclones will emphasize integration of observations (i.e., airborne dual Doppler winds, cloud-track and water vapor winds from GOES, TRMM data, and additional data from field experiments such as CAMEX and EPIC) with mesoscale and cloud-scale numerical simulations to study hurricane formation and the predictability of landfall intensity, location and timing. We will emphasize the interaction of tropical cyclones with extratropical systems and the upscale organization of convection within the tropical cyclone circulation. We expect strong collaboration with researchers at CSU and Albany in these efforts.
d. Orographic effects
Orography significantly modulates the initiation and evolution of precipitation systems, and can promote substantially higher precipitation amounts and greater spatial variability. The quantitative prediction of the amount as well as the type of precipitation, together with accurate treatment of land surface characteristics (vegetation, snow cover, topography, and soil type), are extremely important for an accurate prediction of run-off and stream flow. In spite of its importance, the accurate prediction of precipitation in mountainous terrain has remained an elusive goal. However, recent studies suggest that prediction of the spatial and temporal distribution of orographic precipitation can be significantly improved with high-resolution mesoscale atmospheric models that adequately capture the orographic influence on the flow. In particular, simulations based on cases observed during the Mesoscale Alpine Experiment (MAP) have shown that the horizontal variation of relative humidity along mountain barriers can produce localized blocking of an impinging air stream and, when juxtaposed with an area of saturated, unblocked flow, can lead to mesoscale convergence and heavy rainfall. Further analysis of MAP datasets will be undertaken. Modeling techniques, including microphysical parameterizations, to predict accurately both the phase and quantity of orographic precipitation, will be developed and evaluated, possibly in collaboration with RAP.
e. Cloud microphysics and precipitation
It is well known that the structure and evolution of precipitating weather systems depend strongly on the microphysics and, in particular, on the conversion of water to ice. Microphysical processes affect the dynamics of systems through their influence on the strength of updrafts, downdrafts, and cold outflows, as well as important forecast parameters such as precipitation type and amount. As new understanding is gained of the factors that regulate the amount and type of precipitation in cold clouds from ice microphysics research, this knowledge will be applied to improve the prediction of ice properties in cloud-resolving models. As part of this effort, the specific sources of error must be identified in current microphysical parameterizations, and physically based improvements to the model physics must be developed, particularly for ice formation, which are responsible for these deficiencies. A comprehensive assessment of microphysical processes within precipitation systems will be conducted, using multi-parameter radar observations and where possible, detailed in-situ microphysical observations. Observations obtained in the Stratosphere-Troposphere Experiments: Radiation, Aerosols, and Ozone (STERAO) and from the Severe Thunderstorm Electrification and Precipitation Study (STEPS) will be utilized together with model simulations to provide the basis for evaluation and improvement of the cloud microphysics in precipitation forecast models. Cooperation with ATD and RAP to develop new capabilities to estimate hydrometeor type and shape from multi-parameter radar will facilitate these efforts.
3. Mesoscale Data Assimilation
Goal: To develop and support state-of-the-art data assimilation systems for application in mesoscale models.
These data assimilation systems can be used for a variety of purposes including the assimilation of data from new observing systems, analysis of the optimal use of observations, and understanding the observational requirements for accurate precipitation forecasts and optimal strategies for obtaining targeted observations.
a. Advanced data assimilation systems for community use
While mesoscale data assimilation is a critical component of USWRP research, relatively few researchers have access to sophisticated data assimilation systems. To facilitate broader research on this important topic, a community mesoscale data assimilation system based on a state-of-the art mesoscale forecast model and its full physics adjoint will be established. A suite of supporting components, including observational operators, minimization software, and necessary background and observational error covariances will be developed, and user support for the data assimilation system will be provided. This system will initially be based on the MM5 Model; as the next generation WRF model is developed, the data assimilation will be migrated to the new modeling system. Through continued interactions with scientists at Florida State University, NOAA/FSL, and CGD, improved techniques for assimilating data on the mesoscale that benefit this assimilation system will be developed and evaluated.
b. Optimal use of existing observations and the potential benefits of new observing systems
Over the past decade, many new observing tools have been developed, including Doppler radars, wind profilers, GOES (water vapor winds), GPS/MET systems (radio occultation technique), GPS (ground-based precipitable water observations), ACARS, and many other in-situ and remote sensing systems. These new observing platforms provide mesoscale observations with greatly enhanced spatial and temporal resolution. A major challenge is to develop techniques that provide optimal benefit in using these observations to improve mesoscale weather prediction, and to assess quantitatively the potential value of new observing systems. Data assimilation experiments using the MM5 4DVAR systems will be conducted to assess the impact of GPS/MET data, GOES water vapor winds, and other observing systems. MMM will also participate actively in the NAOS (North American Observing System) program and will conduct necessary experiments to help determine the observational requirements for future forecast models.
The assimilation of Doppler radar data is a particularly important aspect of this research. Ground-based and airborne Doppler radars provide important wind and precipitation observations of weather systems, including mesoscale convective systems, wintertime banded precipitation, and tropical cyclones. In complementary research with RAP and the University of Oklahoma, MMM will continue its efforts to determine how to best use the observations from Doppler radars for cloud-scale and mesoscale model initialization and prediction. Interactions with scientists at ATD and NOAA Hurricane Research Division on the assimilation of Doppler radar data for tropical cyclones, both in the open ocean and near landfall, are also planned.
c . Data assimilation and ensemble forecasting
Approximations of the statistics of forecast error, typically in the form of a forecast-error covariance matrix, are central to data assimilation. These statistics determine the relative weighting of observations and the first-guess (or background) forecast, as well as how a given observation influences other locations and other variables in the analysis. Although in the past it has been applied primarily at the medium range, the goal of ensemble forecasting is to predict just such statistics of the forecast. It is therefore natural to combine data assimilation with ensemble forecasts at the short range, say 3--6 h.
The ensemble Kalman filter is a promising avenue to a combined ensemble forecast and data-assimilation system. In essence, one begins with an ensemble of analyses and makes an ensemble of short-range forecasts (using the full nonlinear forecast model) to the time of the next available observations. This ensemble of forecasts is used to estimate the forecast covariances required to assimilate the new observations via the standard Gaussian formalism of the Kalman filter. Each ensemble member is then updated given the new observations by assimilating a set of perturbed observations (that is, the actual observations plus noise consistent with the observational uncertainty). This approach shares with four-dimensional variational techniques (4DVAR) the benefit of flow-dependent forecast covariances, but, unlike 4DVAR, it does not require the linearized or adjoint versions of the forecast model. The ensemble Kalman filter also has the attractive feature of providing a short-range ensemble forecast and of ``initializing'' the ensemble members for longer-range ensemble forecasts.
This approach has been tested within the simplified context of a quasigeostrophic model using simulated data and performed well with 100 ensemble members. Further testing is planned in a nonhydrostatic model at convective scales using simulated Doppler radar data, where comparison against the 4DVAR scheme mentioned above will also be made.
d. Research in adaptive observations
A benefit of a combined ensemble forecasting and data assimilation system is that an estimate of the uncertainty in the short-range forecast is then available. This estimate is crucial to rigorous algorithms for adding observations to the network (that is, adaptive observations), as it allows one to calculate the expected impact of an additional observation given the location and the observation error. Techniques based only on adjoint sensitivities or singular vectors, such as those employed in previous field experiments (FASTEX, NORPEX, Winter Storms), do not utilize such estimates and thus are likely suboptimal. In practice, the calculation of the expected impact of an observation must be approximated if a number of observation locations are to be considered. Various approximations will be tested in the context of the quasigeostrophic model.
4. High-resolution Weather Research and Forecast Model Development
Goal: To provide a new mesoscale forecast and assimilation system that will advance both the understanding and prediction of important mesoscale weather, and promote closer ties between the research and operational forecasting communities.
The recent effort to develop a new model, called the Weather Research and Forecast (WRF) Model, will be continued as a collaborative effort among NCAR, NCEP, FSL, CAPS, AFWA, and a number of university scientists. With this model, these researchers seek to improve the forecast accuracy of significant weather features across scales ranging from cloud to synoptic, with priority emphasis on horizontal grids of 1-10 kilometers. The model will incorporate advanced numerics and data assimilation techniques, multiple relocatable nesting capability, improved physics and treatment of complex terrain. These advances will help enhance the ability to simulate convection and mesoscale precipitation systems, including precipitation systems in mountainous regions. The model should be well suited for a range of applications, from idealized research to operational forecasting, and will have flexibility to accommodate future enhancements.
The WRF Model has the potential benefits of providing a more direct path for research advancements to feed into operational forecast models, and an easier transition for personnel moving between university research and the operational modeling and forecast centers. A functional version of the WRF Model is imminent, and will be maintained, supported, and freely distributed by MMM as a community model. When the WRF model becomes sufficiently mature to be used operationally, NCEP will consider implementing it initially as a very high-resolution nest within the Eta model (by 2002). At that point, the WRF model would also be a candidate for replacing the regional operational models.
To receive acceptance in the research and NWP communities, specific model features are being designed based on convincing theoretical analyses and evaluation of controlled model testing. With this in mind, the model is being developed in stepwise fashion, beginning with the solver for the basic dynamical equations and progressing to include more complex physical processes and data-assimilation techniques. Comparisons with known analytic solutions and converged idealized simulations are being used to the extent possible in evaluating alternative approaches for specific components of the model system. As model complexity increases, the system will be evaluated for real-data NWP applications.
The focus on high-resolution regional prediction will require new efforts in model verification. Because of the highly intermittent and localized nature of mesoscale weather systems, traditional measures of forecast accuracy developed for synoptic-scale forecasts may be inadequate to provide useful statistics on model performance. In addition, datasets representing details of mesoscale systems over appropriately large domains are generally lacking. Our efforts in improving the quality of model verification will thus proceed along two paths. First, in collaboration with ATD, RAP, and NOAA/NSSL, we will create regional data sets from WSR-88D radar, rain-gauge networks, and other local observations to document adequately the structure of precipitating weather systems. Second, as needed, new verification techniques will be developed that consider the physical character of mesoscale systems as well as whether prediction is inherently deterministic or stochastic.
III. Cloud and Surface Processes Parameterizations
1. Deep Convective Cloud Systems
Goal: To understand the physics of convective cloud systems on time scales up to intraseasonal, how they influence large scales, and how they can be parameterized.
Deep convective cloud systems can now be numerically simulated for long periods over large domains with high resolution. These simulations provide a way of understanding large-scale circulations of which deep convection is an integral part, and they provide a consistent basis for the development of parameterizations of the effects of deep convection cloud systems in larger-scale motions. However, microphysical processes, cloud-radiation interaction, and sub-grid turbulent processes have inherent uncertainties that feed back to produce uncertainties in predicted large-scale motions. The work described in this section is part of the NCAR Clouds in Climate Program (CCP) which is a concerted effort to bring together process studies and parameterization of deep convection relevant to climate modeling and numerical weather prediction.
a. Cloud systems on long time scales
Emphasis will continue to be on cloud systems in the tropics, on time scales from a week or so up to the intraseasonal, and on space scales from about a kilometer to planetary. The focus on the tropics is motivated by the fact that the tropics play a key role in the climate system in terms of the energy and water cycle. In spite of its importance, the coupling between moist convection and large-scale dynamics in the tropics lacks a fundamental basis. The difficulty has been that large-scale processes in the tropics depend directly on the continued and systematic action of small-scale processes. For example, the concerted effect of deep cumulus clouds may influence a slow, large-scale tropical oscillation, but the tropical oscillation equally affects the cumulus convection. Hence the small and large scale motions must be solved for together. The need for long-time-and-space-scale integrations means that processes that can usually be neglected for short-time-scale weather prediction (e.g., cloud-radiation and air-sea interaction) must be accounted for in models of climate.
In simulating convection on a time scale of a few hours, the primary issue is how to approximate the rates of change between the three phases of water. On time scales longer than a day or so, the evaporation rate and fall velocity of the hydrometeors, and the interaction of water in any of its three phases with solar and long-wave radiation becomes progressively more important. When cloud-resolving models are integrated for long times a key issue is the nonlinear coupling between the parameterized microphysics and radiation through explicitly resolved dynamics. We will continue to use the single-column-type approach in which the model (i.e., the cloud-resolving model (CRM), or the single-column model) is driven by large-scale conditions derived from the data gathered in major observational campaigns (e.g., GATE and TOGA COARE). This approach is a valuable testbed to evaluate various parameterizations (cloud microphysics in particular) used in the CRM and to assess the impact of cloud systems on surface processes and on the radiative transfer.
Building on our experience with GATE and TOGA COARE cloud systems, we aim to couple the cloud-resolving model with an ocean model. Multiscale modeling of a coupled system of clouds and the ocean has only very recently become computationally feasible. This kind of multiscale modeling is necessary to achieve realistic Hadley and Walker circulations in domains of order 10,000 km, and will build on ongoing idealized prototype simulations. A new approach, whereby convection is explicitly simulated is also in a prototype stage and will soon be brought to maturity.
Idealized modeling will also be used to study cloud systems and their impact on the large-scale tropical dynamics. We will continue to study convection organization in long-term simulations of convective-radiative equilibrium, applying two- and three-dimensional domains within the equatorial waveguide using both cloud-resolving and parameterized convection. A nonhydrostatic global model will be used in a series of studies of convection organization on a rotating planet, starting with a constant sea surface temperature (SST) aquaplanet, and proceeding to a planet with idealized distribution of the SST and land masses, including topography. Using such an idealized setup we aim to study the role of convection in monsoons.
Finally we plan to study the influence of organized convection (e.g., mesoscale convective systems) on the large-scale momentum budget in the tropics. Convection organization, how it is modulated by the large-scale dynamics, and how it feeds back into the large-scale flow are key issues. MMM has a strong heritage in the dynamics of the organized convection that will be essential in theoretical and observational studies of the large-scale impacts of organized convection.
b. Convectively generated tropical ice clouds
Tropical cirrus covers a significant part of the tropics and has a key effect on the Earth's radiation budget and dynamics. Because anvils occupy an area much greater than the deep convection producing them, studies are needed to understand better the interactions among radiation, dynamics, and microphysics. Modeling studies and satellite data will be used to determine the complete life cycle of tropical anvils and factors responsible for their persistence.
In-situ and remote-sensing measurements of cloud microphysical and radiative properties help derive distributions of cloud microphysical properties as a function of altitude as well as their relationship to cloud radiative properties. Convectively generated cirrus have sufficiently high optical depths near cloud top to produce localized areas of bright or optically thick cirrus, reflecting more than 40 of the incoming solar radiation. However, the upper parts of cirrus cannot alone account for the high albedos. The lower parts sometimes extend down to the melting layers in the so-called stratiform cloud regions that are usually necessary to produce high albedos. These aspects need to be further quantified and we will carry out studies to do so using microphysical, radiometer and conventional, polarimetric, and Doppler-radar data from TRMM.
Using observational data from earlier tropical cirrus measurements and ongoing TRMM field programs, parameterizations will be developed of the tropical cirrus microphysical properties in terms of diagnostic or prognostic variables for GCMs and CRMs. For example, expressions for the ice-particle effective radius, extinction coefficient, absorption coefficient, and mean terminal velocity in terms of cloud ice water content and temperature will be developed. Additionally, a characterization of ice particle shapes will be provided.
Aside from the difficulty of parameterizing convection per se, the calculation of convective cloud-system areal extent, or cloud fraction, is very important for radiative transfer in GCMs. A primary area of research is based upon satellite data and analyses, and progress will be accelerated by using cloud-resolving models. In turn, observational data will aim to develop new microphysical schemes to be used in CRMs and provide important validation for CRM simulations.
c. Impact of tropical cloud systems on radiative transfer
Clouds impact radiative processes in a complicated way. Large-scale weather and climate models apply a plane-parallel approach to deal with the transfer of shortwave (solar) and longwave (thermal) radiation. Such an approach can be argued appropriate for models with grids featuring large aspect ratio (i.e., the ratio between the horizontal and vertical grid spacings). Cloud-resolving models, featuring grids with aspect ratios close to one, usually apply a similar approach because of the lack of affordable alternatives. However, the validity of the plane-parallel approach can be questioned when vertical columns with a few-kilometer horizontal extent are treated independently, and only vertical radiative fluxes are considered. We will evaluate the effects of detailed three-dimensional radiative transfer on the energy budget and evolution of the resolved convection models. The traditional "independent pixel" radiative parameterizations adapted from GCMs will be replaced by multidirectional quasi-exact calculations. The objectives are to (1) determine the changes in the surface and top-of-atmosphere radiative fluxes when radiation is allowed to interact with the 3D structure of the cloud field; (2) determine the relative importance of changes to cloud microphysics and changes to radiative parameterizations for the energy budget; and (3) quantify the effects of three-dimensional radiative transfer on the structure of radiative equilibrium solutions for tropical convection. The principal objective is to determine what aspects of radiative interactions with cloud geometry and inhomogeneity are important for large-scale model parameterizations.
Radiative processes have long been postulated to strongly influence tropical deep convection (e.g., the early morning maximum of convective intensity over tropical oceans). Such an influence was reproduced using cloud-resolving models. It is not clear, however, what influence radiative transfer has on the large-scale tropical dynamics and on the SST through the interaction of radiation with water vapor and clouds. For instance, idealized studies suggest that convective-radiative equilibrium is unstable in the sense that self-maintaining circulations have to develop to balance the differential radiative cooling between dry/cloud-free and moist/cloudy large-scale regions. It remains to be seen if this "moisture-radiation instability" is relevant for the intraseasonal variability in the tropics. A need for a cloud-resolving approach to address the large-scale impacts of radiative transfer is apparent.
Microphysical parameterizations used in cloud-resolving models have been developed to represent phase changes of water substance and precipitation fallout. They are not designed to predict parameters relevant for the radiative transfer (such as effective radius or single scattering albedo of cloud particles). Moreover, impact of some hydrometers (e.g., graupel or rain) is often neglected in radiative transfer models. One can argue, however, that parameterizations of cloud microphysics and of radiation transfer should be closely coupled in order to address the cloud-radiation interaction in a meaningful way. We will attempt to develop such microphysical parameterizations.
d. Parameterization of deep convection
Current parameterizations do not account for the effects of convective organization, which influences the life cycle and spatial coherence of large-scale circulations in the tropics. We have a hierarchy of models to tackle such problems ranging from idealized process models, to cloud-resolving models, to an intermediate model in which convection is parameterized but dynamical interactions are resolved, and finally to a large-scale "cloud-resolving parameterization" model which applies a cloud-resolving model instead of a parameterization scheme. This unique suite of models will allow comprehensive study of not only the parameterization problem, but also the underlying fundamental problem of understanding the interaction between convection and the mean flow.
2. Boundary Layer Clouds
Goal: To understand the physical processes of PBL (shallow) clouds and represent their effects in climate models.
In the following we describe our observational and modeling studies of the different types of boundary-layer clouds. These cloud studies will continue to be coordinated with the GCSS (GEWEX Cloud System Study) program, which seeks to develop physically based parameterizations of cloud-related processes for climate and global numerical weather prediction models. We will also work as the NCAR CSM Atmospheric Modeling Working Group (AMWG) to improve PBL clouds in climate models.
In the future, when computer resources permit, we intend to combine all of our cloud modeling efforts in the areas of marine stratocumulus, trade cumulus and deep tropical convection to simulate whole cloud systems within the Hadley Circulation over the oceans and to study their role in the hydrological cycle.
a. Marine stratocumulus regime
One of the most climatologically important PBL cloud types is marine stratocumulus. Small changes in its fractional cloud cover or microphysical properties can drastically alter the amount of solar radiation input to the ocean surface. Hence, an accurate representation of this cloud regime in a coupled climate model is required to simulate accurately the energy budget of the Earth's surface. Current climate models treat clouds, turbulence and radiation separately using independently developed parameterization schemes, but these physical processes can interact strongly on a temporal (or spatial) scale that is smaller than the time step (or grid resolution) commonly used in current climate models. Our goal is to develop parameterizations that represent the net effect of all these processes.
One of the key issues in incorporating marine stratocumulus into climate models is the rate of entrainment of warm dry air from above the PBL into the stratocumulus-topped boundary layer (STBL). This rate determines the thermodynamic structure of the STBL, and hence the cloud amount. Our numerical and observational studies of marine stratocumlus have been focused on this particular issue. Based on large eddy simulations, we have recently developed an entrainment-rate formula which differs from those developed elsewhere and requires testing with field observations. Planning is underway for a field experiment (DYCOMS-II) to focus explicitly on entrainment processes and test different entrainment-rate formulae currently used for STBL parameterizations. This proposed study is planned to use new observational techniques on the NCAR C-130 aircraft. In addition, we will continue to analyze data obtained from several previous airborne observational studies in this regime: DYCOMS-I, FIRE-I, ASTEX, and ACE.
Another key issue in incorporating marine stratocumulus into climate models is the effect of mesoscale variations. Mesoscale variations, such as mesoscale cellular convection or cloud streets, are often observed in the marine stratocumulus region. These variations are likely to modify the grid-averaged cloud amount within a GCM mesh, but their effect has never been included in any GCM. Within the next few years, increasing computer power will allow us to simulate explicitly the mesoscale variations, along with the dominant turbulent motions (i.e., large turbulent eddies). Such simulated flow fields can be used to examine the effect of mesoscale variations on the cloud properties of marine stratocumulus.
b. Transition from marine stratocumulus to trade cumulus regime
As air moves downstream towards the equator over the eastern part of large oceanic basins, marine stratocumulus breaks up and gives way to cumulus. During this transition, along the air trajectory the cloud cover is drastically reduced and hence solar radiation input to the ocean is drastically increased. This transition between the two cloud regimes is another focus of our PBL cloud research within MMM.
In the incipient stages of this transition, the stratocumulus layer typically becomes "decoupled" from the well-mixed layer near the surface; here stratocumulus becomes only weakly linked to the surface process, and cumulus often develops under the stratocumulus deck. Important processes for this decoupling and development of cumulus under stratocumulus include evaporation of drizzle, short wave radiative warming of the stratocumulus, and surface heat flux. We will continue to investigate the roles of these processes.
Another mechanism that may also play a role in the transition is cloud-top entrainment instability. When evaporation of cloud due to entrained dry inversion air is significant, the mixture may become colder than its cloudy environment (that is, negatively buoyant), a process known as buoyancy reversal. Whether this buoyancy reversal process can lead to the transition from stratocumulus to cumulus regime is still debatable. Our recent large eddy simulations showed that buoyancy reversal did not lead to a total breakup of stratocumulus cloud deck but that it plays a dominant role in determining the simulated cloud fraction and liquid water path. We will continue looking for other important factors that determine the cloud amount and eventually develop a cloud scheme of the marine stratocumulus regime and its transition to the cumulus regime.
c. Fair weather cumulus
Fair weather cumulus over subtropical oceans is known to play a major role in the hydrological cycle of the Hadley Circulation. Trade cumulus transports moisture from the PBL to the low- to mid-troposphere, pre-conditioning the atmosphere for deep convection further downstream. MMM scientists have a long history of observational and modeling studies of this cloud regime, which will serve as a basis for further study. Examples of field studies in this regime that will continue to be used for comparisons with modeling studies include: BOMEX, GATE, and ATEX. Key issues include how to represent the cloud amount, which affects the global radiation budget, and moisture transport by cumulus, which affects the global moisture distribution.
Fair weather cumulus over land is also important because it modifies the land surface through its effect on incoming radiation. We intend to include fair-weather cumulus in our coupled PBL-land process model, which is described in the following section. We plan to also use observational data from ARM for comparison with modeling results.
The role of fair weather cumulus (over both land and ocean) on transport of biogenic hydrocarbons and other trace gas species, and their chemical reactions is also being investigated. Work is now underway to incorporate these processes in this cloud regime into our large eddy simulation code that is coupled with a chemistry transport model (see below).
d. Observing the boundary layer
Improvements in remote sensing capabilities being conducted jointly with ATD, as well as with NOAA and NASA, will provide new ways to observe the three-dimensional structure of both the clear and cloudy PBL. Water vapor differential absorption lidar (DIAL) aircraft data from SGP and other programs will be used to study the fine-scale structure of scalars in the PBL, as well as mesoscale variability of humidity and PBL structure. Fine-scale measurements of both radial velocity and scalars, for example from the lidars in flat terrain (LIFT) experiment, will be used to document PBL structure in the entrainment layer and provide data for comparison with numerical simulations of entrainment to develop improved parameterizations.
A fundamental limitation in LES modeling is the fidelity of the parameterizations used to represent sub-grid scale motions. This problem is especially acute near fluid interfaces--i.e., near the surface and near the PBL top. To address this problem, we are planning to conduct a series of observational studies that will measure the sub-grid scale motions and allow comparison with parameterizations of these motions used in large-eddy numerical models. The objective is to develop parameterizations that more accurately incorporate the sub-grid scale field of motion into the resolved field. The first of these sub-grid scale experiments, SGS-2000, has been carried out in the Central Valley of California in September 2000 using a two-dimensional array containing 14 three-dimensional sonic anemometers. We are starting to develop plans for a similar experiment to investigate sub-grid scale parameterizations in the entrainment region at the top of the PBL.
3. Surface-Atmosphere Interactions
Goal: To understand the interactions between the atmospheric boundarylayer and the underlying surface, and improve the parameterization of air-surface interactions in synoptic-, meso- and large-eddy-simulation models.
a. Land-atmosphere interaction
Land surfaces are typically heterogeneous. This leads to significant horizontal variations in the contributors to the surface energy budget, and thus PBL structure. This, in turn, can result in errors in numerical climate and weather forecast models that do not incorporate these effects. In order to deal with this problem, variables that describe the surface and variations in surface properties need to be properly formulated to satisfactorily represent the effects of the surface on the atmosphere. High quality comprehensive datasets are critically needed for comparison with models and development of parameterization schemes. Techniques need to be developed for comparing observations of fluxes and other statistical properties of the boundary layer over horizontally heterogeneous land surfaces with model results. We address two complementary questions. First, how can the heterogeneity be accounted for in models with large-scale resolution, such as regional and global models? And second, how do we incorporate the effects of surface heterogeneity on the diurnal variation of PBL structure in regional and global models?
Surface heterogeneity includes spatial variations of soil types, soil moisture, vegetation, and topography. The regional variability of soil moisture is important in storm initiation and evolution, and flash floods, making surface-atmosphere interaction important to both the weather (USWRP) and climate (GCIP, ROCEW) communities. Observational data analysis will focus on surface, aircraft, and remotely sensed data collected from various field campaigns, such as the Boreal Ecosystem-Atmosphere Study (BOREAS), the Cooperative Atmosphere-Surface Exchange Study (CASES-97), and the Southern Great Plains (SGP-97) Experiment. Using remotely sensed soil moisture, the effect of spatial variations in soil moisture on the development of the atmospheric boundary layer will be investigated. In addition, the effect of the surface heterogeneity on stable boundary layers will be examined using the observational data collected during CASES-99. High spatial resolution models with land parameterization schemes will be used to examine model sensitivities and compare their performance with observations. Surface land parameterization schemes will be applied to LESs to study the influence of the subgrid surface heterogeneity on the development of the PBL.
In order to evaluate the performance of numerical models, area-averaged turbulent fluxes over heterogeneous surfaces will be estimated using observations from surface-based sites, aircraft, and satellites. Remotely sensed variables include soil moisture from airborne and satellite microwave sensors, long- and short-wave radiation (including radiative surface temperature), and biomass and land-surface types retrieved from satellite and aircraft imagery. In addition, sensitivity of this scale-up process will be investigated using the BOREAS and SGP-97 data sets. Improved formulations of the bulk formulae for estimate of subgrid turbulent fluxes will also be developed.
Surface heterogeneity also plays an important role in the exchange of carbon dioxide between the atmosphere and terrestrial biosphere. This is very important from a global climate perspective. We will investigate the spatial variation of horizontal and vertical transport of carbon dioxide and the role that they play in carbon dioxide budgets, especially in nocturnal stably stratified boundary layers, which are a particular problem for carbon dioxide budget estimates.
Both climate and weather forecast models have particular difficulty predicting the air temperature close to the surface, and surface energy fluxes during the morning and evening transitions, and at night. As a result, the statistically steady-state and homogeneous turbulence assumptions for Monin-Obukhov (M-O) similarity are violated during these periods. Since M-O similarity theory is the current basis for parameterizing land-atmosphere interactions in numerical models, we will be working on new schemes that parameterize temperature and fluxes during the transition and nocturnal periods emphasizing the effects of vegetation, soil moisture, land use, and topography on the evolution of the PBL. One approach will be to examine the temporal and spatial variability of air temperature, wind, and water vapor, and their vertical transport during the diurnal cycle using data from two field programs conducted during the spring (CASES-97) and summer (SGP-97) over the Great Plains. We will use these data to test surface-process parameterization schemes, linked surface-PBL schemes, and mesoscale models. We will also focus on understanding the nocturnal stable PBL, which is especially difficult to parameterize because of intermittent turbulence, by analyzing the field data collected from CASES-99. We are exploring participation in IHOP to further studies of PBL water-vapor evolution.
b. Ocean-atmosphere interaction
In order to treat the ocean and the atmosphere as one system, we need to understand the turbulent processes on both sides of the interface. Since the surface fluxes are the link between these two media, accurate representation of the surface fluxes is our primary goal in ocean-atmosphere interaction studies. Recent field measurements suggest that ocean waves can dynamically alter the turbulent kinetic energy budget in the atmospheric surface layer. Depending on the relative magnitude of the wave phase speed and the local wind, surface waves can either be a source or sink of momentum. As a result, the traditional relationship between surface fluxes and mean atmospheric profiles is altered. The effects of surface gravity waves on turbulence in the atmospheric and oceanic PBLs, and in particular on M-O scaling, will be investigated using LES with a nested-grid, high-resolution surface layer and a moving surface fitted grid. Our goal is to gain understanding of wave effects and then develop a parameterization that links the ocean and atmospheric PBLs as a system that includes the wavy (interface) effects.
Over the coastal zone, variations in oceanic bottom topography lead to shoaling waves. Existing numerical models for surface stress in the shoaling zone fail because of their inability to properly account for wave age, shoaling, and internal boundary layer development. The fetch-dependent wave field in the shoaling zone cannot be adequately studied without information on the spatial variation of the wind and stress fields. Two goals to be pursued are: to study the relationship between the spatial varying mean wind, stress, turbulence structure, and surface wave fields by analyzing the field data collected from the Shoaling Wave Experiment (SHOWEX); and, to model effects of wave age, shoaling, and internal boundary layer development on the drag coefficient and momentum transfer between the waves and the atmosphere.
c. Chemical transports and transformations
The Earth's surface is the source and sink of many trace atmospheric constituents. The PBL acts as a conduit between the surface and the overlying free atmosphere, and as reactor for many of these constituents, which have both natural and anthropogenic sources. These transport and transformation processes occur in both clear and cloudy PBLs, and the budgets of many of these constituents are determined by physical processes in the PBL such as turbulent diffusion, entrainment, PBL growth rate, and cloud cover and transport. These processes will be studied using data from several field programs such as the Aerosol Characterization Experiment (ACE-1) and the Pacific Exploratory Mission (PEM-Tropics).
The effect of boundary-layer processes on the mixing and chemistry of biogenic hydrocarbons and their reaction by-products, particularly over forest canopies, is also being investigated. Hydrocarbons emitted from vegetation are relevant for climate because of their role as sources of ozone and aerosols in the troposphere. By coupling a PBL LES with biogenic hydrocarbon chemistry, an understanding of the influence of boundary-layer mixing in chemical constituents can be attained. By combining a forest canopy LES with simple decay chemistry, the role of the forest canopy and homogeneous and heterogeneous sources of hydrocarbons may be assessed. The role of small cumulus upon the fate of ozone and its precursors via boundary layer venting, aqueous chemistry, or scattering of solar radiation will be determined by including cloud microphysics with the PBL LES and biogenic hydrocarbon chemistry model. This coupled LES cloud and chemistry model will then be used as a tool to understand interactions between clouds, chemistry, and aerosols in the marine boundary layer as well as the convective boundary layer over land.
4. Chemistry, Aerosols, and Dynamics Interactions Research
Goal: To develop an understanding of the interactions between atmospheric dynamics, aerosols and chemistry at the meso- and cloud- scales, particularly with respect to the coupling between transport, cloud physics, and chemistry.
Atmospheric chemistry can be greatly influenced by the dynamics governing air motion and meteorology at meso- and cloud-scales. For example, deep convection can rapidly transport species and aerosols, such as anthropogenically produced nitrogen oxides, into the upper troposphere where they have a longer lifetime and are more effective at modifying ozone concentration, which plays an important role in oxidizing trace gases in the troposphere. Similarly, mixing across the top of the planetary boundary layer can redistribute constituents into the free troposphere. In addition, liquid and solid particles in clouds provide locations for chemical reactions to occur. Gas-phase chemistry (in particular sulfur chemistry) influences the quantity and size of aerosols, which can become cloud condensation nuclei, providing surfaces for cloud drop formation. In the clouds, chemical reactions and microphysical processes alter aerosols and their properties as nuclei for cloud and ice formation. Thus, interactions of chemistry, aerosols, and the dynamics of clouds are important to several aspects of atmospheric research.
MMM will continue to take a lead role in research on the coupling of dynamics, chemistry and aerosols at small scales. We are developing several state-of-the-art cloud-chemistry models for cloud-topped convective boundary layers and for deep convective storms.
To understand dynamical interactions with chemical reactions, chemistry has been incorporated into MMM's LES model and into the COMMAS convective-cloud model. The LES focuses on eddy transport and entrainment at the top of the convective PBL. Ongoing development of this coupled LES includes incorporating cloud and aerosol physics and chemistry, so that it can be applied to the cloud-topped marine boundary layer. Transport, turbulent mixing and chemistry within deep convection are being investigated using COMMAS coupled with chemistry. This model is currently being applied to thunderstorms observed during the STERAO-Deep Convection Experiment to determine the contributions of transport from the boundary layer and from lightning to the nitrogen oxides. This model is also being used to assess the relative importance of chemical species transport versus chemical reactions for the high-plains thunderstorms observed during STERAO. On a larger scale, ACD's Regional Chemistry and Transport Model (HANK) and MM5 simulations of STERAO cases are being used to examine the regional/synoptic scale transport and chemistry for the STERAO convective events. Information learned from the convective cloud model coupled with chemistry simulations can be incorporated into HANK to improve descriptions of convective clouds and chemistry.
Sulfur chemistry directly affects the number and mass of sulfate aerosols, which are the predominant cloud condensation nuclei. This in turn affects the development of precipitation through the condensation-coalescence or the ice process when aerosols act as ice nuclei. These changes can have large effects on radiative balance, precipitation rates, and even dynamics. MMM scientists are playing significant roles in the Indian Ocean Experiment (INDOEX) and subsequent analysis, examining how aerosols affected cloud microphysical properties. Studies that parameterize these effects for numerical models and that investigate interactions between radiation, microphysics, and aerosols are also underway. Future studies will further examine effects of aerosols and entrainment on drizzle suppression and cloud radiative properties, possibly through participation in ACE-Asia. MMM is collaborating with the development of a NCAR-wide box model that describes size segregated aerosol physics and chemistry in detail. This model will be used to guide the model development of prediction of mass, number concentration, and composition of aerosols and cloud hydrometeors in the 3-D cloud and meso-scale models mentioned above.
To provide better representations of chemical transport and the interactions between chemical species, aerosols, and dynamics, chemistry is being incorporated in the Weather and Research Forecast (WRF) model. This coupled chemistry-meteorological model will be used for the cloud scale and the regional scale, will replace the COMMAS coupled with chemistry convective model, and will be merged with ACD's HANK model. Work on this project began with a workshop that assessed the approaches and methodologies of chemistry modeling in cloud and mesoscale models. Next a version of WRF coupled with chemistry will be created so that subgrid parameterizations and depiction of chemistry processing can be developed, working towards a complete description of chemical processing on the mesoscale. The completed model can then be used for many studies including chemical redistribution by convection, regional-scale transport of chemical species, and aerosol-cloud interactions.
Coupling of the modeling and theoretical studies with field programs is crucial; many specific questions being addressed do not have sufficient observational datasets to guide the theory and modeling. Efforts in organizing and leading the STERAO experiment will be continued with similar participation in upcoming field programs emphasizing deep convection, chemistry, and aerosols in the mid-latitudes and the tropics. The Dynamics and Chemistry of Marine Stratocumulus (DYCOMS-II) has been proposed for summertime 2001 off the California coast and will use several trace species as tracers of entrainment and mixing within and across the top of marine stratocumulus. The ongoing work represents cooperative and interdisciplinary investigations coupling small-scale dynamics with chemistry, aerosols, and cloud physics. Ultimately, issues to be addressed are either small-scale in nature, such as air quality (local, regional), or larger scale in nature, such as climate/chemistry issues related to ozone production-loss and the role of sulfur species on cloud microphysics and dynamics. The investigations into clouds and chemistry will also lead to improvements in, or development of, parameterizations for large-scale models. Thus, the work directly supports the goals of the GTCP (Global Tropospheric Chemistry Program) and climate research.