Strategic Priority: Developing community models

Community Climate System Model: Development of Scientific Capabilities

Historical context and rationale

The development and continuous improvement of a comprehensive climate modeling system that is at the forefront of international efforts to understand and predict the behavior of the Earth's climate is a high priority of NCAR research. This includes the Community Climate System Model (CCSM) as well as its component models. The CCSM, run on some of the world's most powerful supercomputers, simulates the many interconnected events that drive Earth's climate. These include changes in the atmosphere and oceans, the ebb and flow of sea ice, and the subtle impacts of forests and rivers.

CCSM is unique among the most comprehensive of global climate models. Primarily supported by the National Science Foundation (NSF) and the Department of Energy (DOE), with additional support from the National Aeronautics and Space Administration (NASA) and the National Oceanic and Atmospheric Administration (NOAA), it belongs to the entire community of climate scientists, rather than to a single institution. Hundreds of specialists from across the United States and overseas collaborate on improvements to CCSM. The model's underlying computer code is freely available on the Web. As a result, scientists throughout the world can use CCSM for their climate experiments.

The CCSM project was started in 1994, although climate modeling at NCAR has a much longer history; stretching back to about 1980. The first version of CCSM was unveiled in 1998, and the most recent version, CCSM-3, was released in 2004. CCSM-3 represents a major advance over earlier versions of the model because it contains far more information about Earth's physical processes. For example, it tracks the flow of major rivers that empty into the oceans and it now resolves five different thickness categories of sea ice within each grid cell, such as the thickness and the melt rate. Moreover, the finer scale and lower viscosity of its ocean allows scientists to capture significantly greater detail about ocean currents and the mixing of salt and fresh water.

CCSM is constantly being updated and improved; CCSM4 is most likely to be released in 2009. In preparation, an interim version, CCSM3.5, was assembled in mid-2007, and is being evaluated from a number of perspectives, with carbon system spin-up a particular focus. In addition to remaining at the forefront of international modeling efforts, the scientific goals of the CCSM project are as follows:

• to use the modeling system to investigate and understand the mechanisms that lead to interdecadal, interannual, and seasonal variability in Earth's climate;
• to explore the history of Earth's climate through the application of versions of the CCSM suitable for paleoclimate simulations; and
• to apply this modeling system to estimate the likely future of Earth's environment in order to provide information required by governments in support of local, state, national, and international policy determination.

ESSL/CGD (in collaboration with scientific and software engineering partners from Universities, DOE, NASA and Industry) has been busy in developing component models, integrating those components as candidates considered for the next generation of the CCSM, and exploring these in a variety of ways for understanding the Earth System and climate change. One of the topics identified in last years report as a "Plan for 2007" was "a concerted effort to address systematic model biases in the tropics on seasonal and longer timescales." We describe one contribution to that effort here.

Accomplishments in FY 2007

A significant reformulation of the parameterization of convection within Community Atmosphere Model (CAM) took place in 2007. The parameterized convection was made much more sensitive to the dilution of air as it ascends in the atmosphere. The convecting parcels were also made sensitive to the change in phase of condensate between liquid and ice, and momentum transports were included in the formulation. The result was a substantial improvement in many aspects of the atmospheric simulation when driven with observed sea surface temperatures, and in the coupled system when used with CCSM.

As an example of this improvement we show the Taylor Diagram of a version of CAM with these modifications. (Figure 1.)

Another major accomplishment of the CCSM project over the first half of 2007 is that an interim version called CCSM 3.5 has been developed. This uses the new finite volume dynamical core in the atmosphere component, the updated POP 2 code for the ocean component, the latest CICE 4 version for the sea ice component, and a much updated version of the land component compared to the CCSM 3. There have also been significant parameterization improvements in all the components. In addition to the atmospheric convection changes discussed earlier, the land component, CLM 3.5, has a very much improved hydrology cycle compared to CLM 3, aspects of which are shown in Fig. 2. Also, the global partitioning of evapotranspiration is greatly improved using the CLM 3.5 compared to CLM 3.

The ocean component has a new near-surface eddy flux scheme, which nicely reduces the eddy-induced meridional overturning near the surface. This gives a much smoother vertical profile of total ocean advective heat flux. The updated sea ice component uses an improved parameterization of sea ice ridging, and an improved treatment of the snow that accumulates on top of the sea ice.

The most significant aspect of CCSM 3.5 is in its simulation of the El Nino - Southern Oscillation (ENSO) in the tropical Pacific Ocean. All previous versions of the CCSM, and most other climate models, had a peak in the ENSO frequency near two years, which is much shorter than in reality. This problem has now been corrected in CCSM 3.5, which shows a frequency peak between 3-6 years. In addition, the correlation between the Nino 3 SST timeseries and SST anomalies across the globe has also been substantially improved. Figure 3 shows this correlation from the HadiSST observational data set, the previous CCSM 3 control, and the new model version with the CAM 3.5. The improvement is quite remarkable, with the new model showing a correlation pattern that is very like the pattern from observations. The correlation is much broader in the eastern tropical Pacific, and the correlation patterns in the Pacific and other oceans are also much improved. We believe this is primarily due to the two modifications of the deep convection parameterization in the atmosphere, which include convective momentum transport and a dilute approximation to calculate CAPE.

During FY 2007, CSEG has ported CCSM to the NCAR IBM Bluegene/L. In addition, significant progress has also been made in the creation of a sequential, single-executable CCSM. The goal is to create a sequential system that contains backwards compatibility with the current concurrent system, provides "plug and play" capability of data and active components and produces the same climate as the current concurrent system. A prototype version of this sequential, single-executable version should be ready by early in FY 2008.

Plans for FY 2008

One of the higher-priority ongoing activities of the CCSM program is a concerted effort to address systematic model biases in the tropics on seasonal and longer time scales. A particular focus in 2008 will be the seasonal cycle of the near surface ocean and atmosphere in the eastern tropical Pacific. It is expected that this will be strongly linked with long-standing biases such as the appearance of a double Inter-Tropical Convergence Zone and warm sea surface temperatures (SST) under the stratocumulus regions off the west coasts of North America, South America and Africa. Several hypothesis-driven activities will be in collaboration with colleagues outside of NCAR to address seasonal variability. In addition, collaborative efforts will continue within ESSL to examine, in climate simulations with embedded regional models, the importance of explicitly resolving mesoscale and smaller processes that govern weather and local climate but that may also have significant impacts on the large-scale circulation.

The reduction of such biases becomes even more important as the complexity of CCSM increases. Several of the most pressing scientific questions regarding the climate system and its response to natural and anthropogenic forcing require that physical models be extended to include the interactions of climate with biogeochemistry, atmospheric chemistry, ecosystems, glaciers and ice sheets, and anthropogenic environmental change. While the ultimate goal is a comprehensive Earth System Model (ESM), practical considerations suggest that there will be a multitude of versions with different capabilities. The CCSM project will work towards developing a first-generation coupled chemistry-climate model in the next two to three years. A project of this scope will necessarily involve scientific partnerships across ESSL, NCAR, and the external CCSM community. This model could be used to study the complex interactions among biota, chemical processes, and physical climate for paleoclimate studies or scenarios for future climate change. It could also be used to study variations of the chemistry of the present-day atmosphere driven by external forcing from solar variability and major internal natural modes of variability, such as the El Niño-Southern Oscillation.

The current long term plan of the CCSM project is to develop and freeze the next version of the model, CCSM-4, by the end of 2008. In addition to several other improvements, this version will most likely have new components for the carbon cycle and interactive atmospheric chemistry. This will enable a whole new range of scientific questions to be asked of, and answered by, the CCSM. In addition, the CCSM-4 will be the model used to contribute to the next IPCC report.

return to top

Space weather: Model Development and Data Analysis

Space weather research seeks to understand and work towards predictions of the physical conditions in the geospace environment, particularly when disturbed by energetic events occurring on the Sun. This is a multidisciplinary field of research which requires understanding of solar, solar wind, magnetospheric, and ionospheric physics. It covers a broad range of time scales, including solar cycle variations (years), recurrent solar wind streams (months), coronal mass ejection (CME) propagation and geomagnetic storms (days), flares and energetic particles (minutes). Understanding these phenomena is important for human spaceflight, satellite design, communication and navigation systems used by our increasingly technologically dependent society.

HAO scientists have made substantial progress in understanding this broad range of topics during the past year. 3D magnetohydrodynamic simulations, conducted by Sarah Gibson and Yuhong Fan, of the emergence of twisted magnetic flux ropes into the low corona have highlighted the importance of interactions with ambient coronal fields for triggering unstable behavior of the kind that can lead to the eruption of a CME. These eruption mechanisms play an important role in determining the observational signatures of the CME as well establishing the orientation of the magnetic field. Determining the correct magnetic field orientation is an important first step in making reliable space weather predictions; understanding the evolution of the magnetic field and the development of plasma parameters as the structure propagates through the heliosphere is the second step in the process of space weather prediction. A team lead by Mark Miesch has begun development of a heliospheric solar wind model which uses advanced numerical methods, including adaptive mesh refinement, to accurately track CMEs as they interact with the ambient solar wind. In close coordination with the Center for Integrated Space Weather Modeling (CISM), scientists within the AIM section of HAO have completed development of the second generation of the Coupled Magnetosphere Ionosphere Thermosphere (CMIT) model. This model couples the LFM global magnetospheric model with the Thermosphere Ionosphere Electrodynamic General Circulation Model (TIEGCM) to provide a mechanism for studying how CMEs create geomagnetic storms. Early scientific results from CMIT 2.0 have shown the model is capable of reproducing the temporal and spatial variations total electron content (TEC) measured by Global Positioning Satellites (GPS) during magnetospheric storms. Comparison with radar observations made at the Jicmarca indicates that electrodynamic forcing from the magnetosphere at high latitudes improves the agreement with observations. Of course all of these modeling efforts are dependent upon observations to provide the verification that their predictions are physically reasonable. Observations from the Mauna Loa Solar Observatory as well as a variety of space based platforms are used to test these models and guide their continued development.

In the coming year, simulations of flux tube emergence, coronal magnetic evolution, and CME onset will continue, with effort directed toward performing "event-driven" simulations. Specifically, Hinode observations of the evolving photospheric vector magnetic field will be used to drive conditions at the lower boundary of the computational domain; comparison of numerical results with multi-wavelength observations will provide insight into the nature of the coronal magnetic field evolution that produces eruptions. The next steps in the development of a Sun-to-Earth space weather model will be taken by extending the computational domain used for this year's studies of 2D solar wind configurations closer to the Sun. This modification will ultimately permit the introduction of simulated CME-like disturbances into the flow, enabling the evolution of the ejected plasma and fields to be investigated. The plasma and magnetic field parameters at L1 will be extracted from this model and used to drive the CMIT model. We will continue development of the CMIT focusing on three areas, 1) adding mass outflow from the ionosphere into the magnetosphere, 2) including a ring current model in the inner magnetosphere, and 3) extending the lower boundary down into the mesosphere.

return to top

Nested regional climate modeling

Mesoscale processes can have a major impact on large-scale circulations. Yet climate models do not adequately represent this upscale influence of mesoscale processes. This major limitation to improved predictions of the physical climate system has been recognized by the international community through the establishment of The Observing system Research and Predictability Experiment (THORPEX). In a complimentary manner, NCAR is addressing this challenge by

• developing the Nested Regional Climate Model (NRCM), which will embed a high-resolution version of WRF within CCSM, with two-way communication between the models, and
• experimenting with the NRCM to investigate the impact of upscale development from mesoscale organized tropical convection on capturing large-scale modes of tropical variability in global climate models.

• improved downscaling from global climate simulations to allow for accurate regional predictions;
• upscaling from regional processes, including resolved moist convection and the effects of land and ocean processes;
• understanding and simulating the manner in which mesoscale organization of moist convection impacts larger scales; and
• providing a means of distinguishing processes that are, or may be
  adequately parameterized, from those that cannot and must be resolved in climate simulations.

In FY2006, with the availability of significant computing resources made available by the NCAR Directorate, MMM and CGD scientists have performed a set of NRCM simulations with the Advanced Research WRF (ARW) using a tropical channel model domain driven by prescribed SST forcing and lateral boundary conditions along the north and south boundaries. Some preliminary analyses were performed to compare the WRF simulation from 1996 – 2000 with observations and global atmospheric simulations by CAM.

In FY2007, we successfully utilized additional computational resources available on NASA's Columbia Supercomputer to extend the tropical channel simulation to cover 2000 - 2005, hence providing 10 years of model climatology to investigate tropical biases in the model simulations. In addition, we repeated the simulation for the hurricane season of 2005 with a high-resolution (12-km) nest spawned over the tropical Atlantic to investigate the causes for this anomalously active and destructive season. Finally, we completed a reprocessing of all tropical channel model outputs to monthly data at standard pressure levels to facilitate further analysis and evaluation of the simulations.

Our main research goal in FY2007 was to perform a more detailed and extensive analysis of the tropical channel simulations to advance our understanding of tropical wave modes, and evaluate the model's skill in simulating various tropical phenomena. Building on strong collaborations between MMM and CGD, and between NCAR and a host of external investigators, significant progress has been made in several areas, with a few key examples highlighted below:

1) Analysis of tropical wave modes

A preliminary space-time spectral analysis of the outgoing longwave radiation in FY2006 indicated that the tropical channel model is able to capture important wave-like modes corresponding to Kelvin waves and Madden-Julian Oscillation (MJO). In FY2007, a more detailed analysis of the Kelvin waves showed that their horizontal propagation speeds and vertical dynamical signals are reasonably well simulated, although the climatological variance of these waves is significantly under-predicted in the deep tropics (between 5S-5N). A similar under-prediction problem was found for the MJO, which also appears less well organized than the observed MJO.

To investigate the causes of this poor MJO representation, an additional analysis was performed of two simulations of an MJO event that occurred during May - June of 1997. Careful processing of the continuous run initialized on January 1, 1996, shows that no MJO events occurred during the May - June period. However, in a simulation initialized on May 1, 1997, the MJO event was successfully reproduced. These findings suggest that model biases built up in the continuous run have a significant negative impact on the simulation of MJO. Additional analyses are being performed to understand the separate influences of initial conditions, lateral boundary conditions, and model physics on the ability of the model to capture various tropical modes.

2) Analysis of the East Asian monsoon

Comparison of the observed and simulated East Asian summer monsoon circulation and precipitation revealed a major weakness in the NRCM simulation. This problem was particularly significant during 1997 when the West Pacific Subtropical High in the simulation was displaced much further east and a low pressure center was generated near the South China Sea that created a northeasterly flow into central and southern China, as opposed to the southwesterly monsoonal flow that brings abundant moisture and precipitation. On the other hand, the simulation compared well with observations during the 1998 summer.

To investigate the reasons for the large difference in model skill from year to year, we performed several sensitivity experiments to isolate the impacts of model initialization, sea surface temperature (SST), lateral boundary conditions, and physics parameterizations. These experiments suggest that SST plays a dominant role in explaining the large difference between the simulations in 1997 and 1998, possibly due to the strong dynamical feedbacks between SST forced convective heating in the western Pacific and the large scale circulation. However, the erroneous circulation in 1997 was insensitive to the convective parameterizations used. We hypothesize that deep convection may be triggered too easily in the model because of the lack of air-sea coupling that should provide a negative feedback to modulate deep convection in areas with warm SST. This highlights an important area for model development, in addition to boundary layer and cumulus convection parameterizations that are important to correctly simulate tropical convection and atmospheric response to warm SST.

3) Analysis of tropical cyclones

We have begun a more quantitative analysis of tropical cyclones (TC) simulated by the models. A TC tracking algorithm has been developed and implemented to compare the TC genesis locations, storm tracks, and TC numbers based on observations and model simulations. Figure 1 shows that the simulation well captured the TC genesis locations in various ocean basins, except the North Atlantic, where the simulated TC genesis locations are shifted much further north as compared to observations. However, analysis of the relative vorticity suggests that the African easterly waves are well simulated by the model. We hypothesize that the shift in TC genesis locations in the Atlantic may be an artifact of the TC tracking algorithm, which starts tracking a TC only after the formation of a low pressure center, despite the fact that other criteria (e.g., enhanced relative vorticity at low levels) may have been met much earlier. Additional analyses are being performed to test this hypothesis and to investigate other aspects of TC, including genesis mechanisms.

NCAR plans that the NRCM will eventually be provided and maintained as a community resource for use by all the academic, government, and private sector communities. Indeed, some of the NRCM components such as the CAM radiation parameterization have already been added to the standard WRF release in November 2006. The NRCM project is funded by NSF and supported by the NCAR Directorate, ESSL Directorate, and MMM and CGD divisions.

return to top

WRF

The Prediction Across Scales initiative is a collaborative effort between CGD and MMM to coordinate research and system development activities across weather and climate scales. Recent major advances in petascale computing coupled with rapid advances in scientific understanding are enabling advances in simulating a wide range of physical and dynamical phenomena with associated physical, biological and chemical feedbacks that collectively cross the traditional weather-climate divide. Such simulations and predictions are essential to a society that is becoming much more sophisticated in its requirements for weather, air quality and climate predictions and that is able to make useful economic and social use of such improvements. Moreover, fundamental barriers to advancing such prediction on time scales from days to years, as well as long-standing systematic errors in weather and climate models, are partly attributable to our limited understanding and capability to simulate the complex, multiscale interactions intrinsic to atmospheric and oceanic fluid motions. The scientific and societal questions and issues to be addressed are many. A limited sample includes better understanding of:

• The water cycle and its predictability, particularly the limitations of available water and the impacts on food production;
• The limits of weather, air quality and climate predictability including the impacts of mega-cities and the stressed Earth’s capacity to sustain quality of life;
• The interaction of hydrological, chemical and biogeochemical cycles and their feedback on weather/climate processes;
• The mechanisms by which solar variations influence the chemistry and dynamics of the upper atmosphere, and how these effects are manifested in the lower atmosphere;
• The interactions between climate change, ENSO and other natural modes of variability, including changes to the behavior of phenomena like hurricanes;
• The mechanisms of abrupt climate change and potential tipping points.

Our plans for 2007 were to complete a series of simulations with the Advanced Research WRF model reconfigured as a Nested Regional Climate Model (NRCM), and to commence the analysis of the outcomes.

We have completed a 10-year simulation of the tropical circulation with the NRCM configured in a channel mode using NCEP/NCAR reanalysis data on the poleward boundaries and specified surface conditions. A set of comparative simulations were also made using the CAM at T170 resolution and configured in a similar channel mode with relaxation towards reanalysis in the polar regions. These simulations were run on the Blue Vista and the NASA Columbia computers. The base simulation was made at 36 km, with one year simulation using a 2-way nest over the Maritime Continent at 4 km resolution. A set of high-resolution simulations were made off the west coast of North and South America for use in determining oceanic simulation issues there. And a full simulation of the 2005 hurricane season was accomplished with a 12 km nest over the North Atlantic. A subset of the data has been analyzed for tropical modes, hurricanes and the East Asian monsoon. This has shown a range of successes, with good simulation of tropical cyclones and tropical modes, but some failures with the East-Asian monsoon. Several university groups have joined the program to conduct research on the data.

Our immediate plans are:

• To diagnose the problems that have arisen and how they can be removed in future simulations;
• To continue the analysis of the previous simulations in collaboration with our university colleagues;
• To nest the NRCM inside CAM and advance development of a full nested configuration including atmospheric and oceanic components;
• To conduct high-resolution simulations with both the CCSM and WRF Global models

return to top

WACCM

The Whole-Atmosphere Community Climate Model (WACCM) is a comprehensive numerical model, spanning the range of altitude from the Earth's surface to the thermosphere. WACCM is built upon the numerical framework of NCAR's Community Climate System Model (CCSM), and is envisaged as a flexible model environment, whose domain and component modules can be configured according to the specific problem under study. WACCM incorporates physical and chemical processes required to investigate the coupling among atmospheric regions from the surface to 140 km. The current version of WACCM (WACCM3) has fully interactive chemistry and dynamics, and can also be coupled to the ocean component of CCSM. Addition of upper thermospheric physics and chemistry is currently underway, and will allow the model to extend upward to about 500 km.

WACCM has contributed to the Chemistry-Climate Model evaluation (CCMval) effort, an international activity under the auspices of the Stratospheric Processes And their Role in Climate (SPARC), and to the most recent Ozone Assessment (2006) of the World Meteorological Organization. In addition, the following topics are currently under investigation using WACCM:

• the effects of solar variability in the middle atmosphere using time-slice simulations during solar minimum and maximum conditions with and without the quasibiennial oscillation in the Tropics (see Figure);
• the role of parameterized gravity waves and their tropospheric sources in simulations of the whole atmosphere;
• the dynamical variability in the middle atmosphere, in collaboration with University colleagues and other modeling groups;
• the response of the Brewer-Dobson circulation to climate change;
• the validation of chemistry and the dynamics of WACCM against observations (satellite and ground-based);
• developed of a version of the model that can be driven by assimilated meteorological data to study the exchange of mass and constituents in the upper troposphere/lower stratosphere;
• atmospheric predictability in the whole atmosphere context.

Recently published papers on these and other subjects may be found in the WACCM website ( http://waccm.acd.ucar.edu/Pubs/index.shtml ). During 2007-2008, a major emphasis of the WACCM project will be the study of the effect of the middle atmosphere on the tropospheric climate. Other areas of interest in 2007 and beyond are polar mesospheric clouds and their connection to climate change; extension of WACCM to 500 km; improvements to the parameterization of mesoscale gravity waves; and internal generation of the quasibiennial oscillation.

return to top

WACCM development and extension

When first proposed, the goal of the Whole Atmosphere Community Climate Model (WACCM) is to develop a model that extend from the Earth surface to the upper thermosphere, and self-consistently resolve the dynamical, chemical, radiative, and electrodynamical processes and the coupling between atmospheric regions. The current standard WACCM version (WACCM3) extends from the surface to the lower thermosphere (~140 km).

We have recently extended WACCM with 1.9x2.5 degrees horizontal resolution to the upper thermosphere at 3.4x10^-7 Pa (the same as the upper boundary of the TIME-GCM, approximately at 500 km). In the process of achieving this upward extension, we have

(1) Implemented modules to resolve the major species diffusion, which becomes increasingly important above 110 km.

(2) Implemented modules and revised the codes to reflect the constituent-dependency of the specific heats, gas constant, and mean molecular weight. They were set to constants in the previous codes.

(3) Revised the treatment of the vertical diffusion equations for minor species and heat conduction equation. The EUV/FUV photochemistry, non-LTE radiative processes, and ion-drag and Joule heating have already been implemented in the WACCM3.

We have made test runs using the extended WACCM, and obtained the wind structure, thermal structure, the vertical profiles of major species (N2, O2, O), oxygen species (O3, O1D, O2_1S, O2_1D), hydrogen species (H, H2O, HO2, OH), nitrogen species (NO, NO2), carbon species (CO, CO2, CH4), ions (O+, O2+, NO+, N+, N2+) and electron density from the ground to the upper thermosphere. Initial comparison shows that these structures in the middle and upper atmosphere are in good agreement with those obtained from TIME-GCM.

We will further validate the WACCM results, including seasonal and annual variation of thermosphere density against TIME-GCM and MSIS. We also plan to implement ambipolar in the WACCM.

return to top

MOZART

Two versions of the Model for Ozone and Related chemical Tracers (MOZART), an offline chemical transport model, are being used currently in ACD at NCAR and by collaborators. Both versions are updates of the MOZART-2, which is publicly available from the NCAR Community Data Portal and is used widely around the world. MOZART-3 is an extension of MOZART-2 into the stratosphere, with the addition of halogen chemistry and heterogeneous processes on polar stratospheric clouds. MOZART-4 has been updated over MOZART-2 to improve tropospheric chemistry simulations, with more detailed representation of hydrocarbons and tropospheric aerosols.

The Model for OZone and Related chemical Tracers, version 3 (MOZART-3), which represents the chemical and physical processes from the troposphere through the lower mesosphere, was used to evaluate the representation of long-lived tracers and ozone using three different meteorological fields (Kinnison et al., 2007). The meteorological fields are based on: 1) the NCAR Whole Atmosphere Community Climate Model, Version 1b (WACCM1b); 2) the European Centre for Medium-Range Weather Forecasts (ECMWF) operational analysis; and 3) a new reanalysis for year 2000 from ECMWF called EXP471. In this work, model-derived tracers are compared to data climatologies from satellites. In the Figure, an example of such a comparison is shown for water vapor. Here the tropical time-height propagation of water vapor in the equatorial region (averaged between 12° S and 12° N) is shown. This propagation pattern is known as the H2O "tape recorder signal". For this comparison MOZART-3 driven with WACCM1b meteorological fields best represents the data climatology, followed by the new ECMWF reanalysis meteorological product. This work showed that there was a significant improvement in the new ECMWF reanalysis product over the operational version.

MOZART-3 has been selected to be used for chemical weather forecasting by ECMWF as part of the Global and regional Earth-system (Atmosphere) Monitoring using Satellite and in-situ data (GEMS) project ( http://www.ecmwf.int/research/EU_projects/GEMS/ ).

MOZART-4 is being used to assist in the analysis of the MIRAGE and NASA/INTEX-B field experiments. Several regional models are using MOZART-4 simulations for boundary conditions, including the University of Iowa and the Pacific Northwest National Lab for MIRAGE and INTEX-B simulations, and Colorado State University and NCEP for TexAQS simulations.

FY2008 work will include continued use of MOZART-4 for analysis of the MIRAGE and NASA/INTEX-B field experiments. MOZART-3 and MOZART-4 will be published and released to the public in the coming year. This work is funded by NSF/NCAR and NASA.

return to top

Community Atmosphere Model combined with the MOZART chemical Mechanism (CAM CHEM)

CAM is the latest in a series of global atmosphere models developed at NCAR for the weather and climate research communities. CAM also serves as the atmospheric component of the Community Climate System Model (CCSM). The incorporation of interactive chemistry capability in the Community Atmosphere Model (CAM) has made considerable progress over the last year and now encompasses a variety of options to accommodate the needs of the coupled climate model; in particular, using the implemented MOZART framework, CAM-chem can now be configured to combine prognostic and diagnostic variables. As a result, aerosols can either be prescribed, simulated using simple input oxidant fields, or simulated using the full MOZART-4 aerosol parameterization. In addition, it was modified to enable the use of very detailed aerosol scheme developed by S. Ghan (PNNL). The aerosol indirect effect has been incorporated through collaboration with CGD scientists. All of the features developed under the MOZART framework are now included in CAM-chem: these include interactive photolysis rates, interactive dry deposition and interactive biogenic emissions, where the last two are fully included into the land model CLM and the full capability of running CAM-chem in an offline mode (i.e. driven by meteorological analyses). For extended chemistry-climate studies, a number of different options exist for simulating aerosols and chemistry to facilitate using the model in the optimal configuration. ACD scientists have developed a mechanism in which the number of hydrocarbons is quite reduced from the MOZART-4 mechanism, leading to a speed-up of a factor of 2. This reduction has been analyzed and evaluated against measurements and it reproduces many of the main features of atmospheric chemistry relevant for climate studies, e.g. the ozone distribution and the methane lifetime. In addition, a version of CAM-chem with a representation of stratospheric chemistry was developed as a tool to represent ozone changes in the lower stratosphere; simulations over 1970-2005 indicate (Figure) a very good comparison of ozone trends with respect to observations. FY2008 plans include continued evaluation of the model performance under the different options described above. This work is funded by NSF/NCAR, NSF Biocomplexity, and DOE.

return to top

Upper Atmosphere Community Models

HAO scientists have developed a suite of upper-atmospheric models, in collaboration with scientific visitors and scientists at universities, government labs, and other organizations. These models are made available for use by the community, typically through collaborations between HAO scientists and scientists in the community. A central model is the Thermosphere-Ionosphere-Mesosphere-Electrodynamics General Circulation Model (TIME-GCM) and simplified variants of it, for which Ray Roble has been the leading developer for the past 25 years. The TIME-GCM simulates the three-dimensional, time-dependent global dynamics, chemistry, energetics, and electrodynamics of the mesosphere, thermosphere, and ionosphere, for given inputs representing solar, magnetospheric, and lower-atmospheric effects. Other HAO models with extensive use by the community are the Assimilative Mapping of Ionospheric Electrodynamics (AMIE) procedure for synthesizing high-latitude observations of ionospheric electric fields and currents, the Global-Scale Wave Model (GSWM) for calculating atmospheric tides and planetary waves from the ground through the thermosphere, and the GLOW model for calculating the effects of solar ultraviolet and X-rays as well as energetic particles. These models are used to understand the processes affecting the dynamical, electrodynamical, thermodynamical, and chemical conditions in the Earth's upper atmosphere, its response to the Sun's variable radiative, particulate and magnetic emissions, and its coupling to the lower atmosphere and the magnetosphere.

The models have been upgraded through improvements to photochemical reaction rates, gravity-wave parameterizations, eddy coefficients, and upper-boundary fluxes. A preliminary version of a plasmasphere model has been tested with the TIE-GCM. Different ways of specifying magnetospheric energy and momentum inputs at high latitudes have been tested and analyzed. Driving of the TIME-GCM with non-migrating tides from the GSWM has been implemented, and used to demonstrate how these tides can help explain the observed longitudinal structure of the ionospheric equatorial anomaly. Model results have been provided to collaborators in the community. This work has been sponsored by NSF base support to NCAR, NSF Space Weather special funds, and NSF CEDAR special funds. It has also been supported by NASA and DOD programs.

For FY08, we plan to continue model upgrades, testing, and scientific analysis in collaboration with the community. Model developments will be documented, and results of scientific studies will be published. Particular foci will be testing and implementing the plasmasphere model with the TIME-GCM, continuing to transfer and document process modules from the TIME-GCM to WACCM, and continuing to test and implement modules coupling the magnetosphere with the ionosphere and thermosphere.

return to top

Coupled Carbon-Nitrogen Cycle Modeling

Overview, historical context, and significance

An important trend in the development of numerical models of the global climate system is the incorporation of biological mechanisms into the main prognostic model equations. Early development included the introduction of vegetation canopies with representations of biophysical and biochemical controls on the fluxes of sensible and latent heat between the land surface and the atmosphere. Later development, beginning around 2000, introduced coupling between fluxes of carbon, water, and energy. This first generation of coupled climate-biogeochemistry models eventually developed to include not only carbon fluxes, but also carbon state variables such as biomass in leaf, wood, and root tissues, and organic matter in soils.

Another line of model development has been taking place in the ecological research community, and while there are many similarities between the state-of-the-art terrestrial ecosystem models and the first generation of coupled climate-carbon cycle models, an important distinction has been in the treatment of nutrient limitations. Ecologists have long appreciated that nutrients and nutrient limitation play a critical role in regulating carbon, water, and energy fluxes on a wide range of time scales, from days and seasons, to years and centuries. Nitrogen is a dominant limiting nutrient controlling carbon uptake and growth in many ecosystems, and accordingly most ecosystem models have incorporated some treatment of the mechanisms coupling the carbon and nitrogen cycles.

ESSL scientists have recently completed a model development project to bring together these parallel efforts in the climate modeling and ecological modeling communities, integrating a detailed treatment of carbon-nitrogen cycle interactions (from the Biome-BGC model) with a state-of-the-art land surface model (the NCAR Community Land Model - CLM) . The resulting model, CLM-CN, has now been tested and its performance documented both in offline and fully-coupled configurations. A critical application of the model has been to study the influence of carbon-nitrogen cycle coupling on present-day and potential future climate-carbon cycle feedbacks.

In the early stages of development for CLM-CN, ESSL scientists identified the need to introduce a more detailed treatment of radiation interception and carbon and water fluxes within the vegetation canopy, compared to what existed already in CLM. Based on a combination of new theory and new observations, scientists developed and implemented a canopy integration scheme that deals explicitly with the observed vertical gradients in leaf morphological characteristics and the related variation in photosynthetic behavior (Thornton and Zimmerman, J. Climate, 2006). This work showed that the new model solved a problem with low productivity under the influence of nitrogen limitation and prognostic leaf area calculation experienced with the original formulation in CLM, and the new treatment has now been adopted for all CLM configurations (Dickinson et al., J. Climate, 2006, and Lawrence et al, J. Hydrometeorology, 2007).

A preliminary test of the newly completed CLM-CN was to drive the model with a reanalysis of surface weather, to document its present-day predictions for carbon fluxes and stocks and to evaluate the influence of changes in atmospheric CO2 concentration and changes in the rate of mineral nitrogen deposition on carbon, nitrogen, water, and energy cycles. This study (Thornton et al., Global Biogeochemical Cycles, in press) found that the introduction of carbon-nitrogen coupling significantly altered the model response to increasing CO2, and the sensitivities of net land carbon flux to variation in temperature and precipitation. That study suggested that C-N coupling would have an important impact on the magnitude and possibly the sign of climate-carbon cycle feedbacks when exercised in the fully coupled Community Climate System Model (CCSM).

ESSL scientists, incollaboration with scientists from several other institutions, have recently completed a series of fully-coupled simulations, using CLM-CN as well as a new ocean ecosystem model as components of CCSM, testing the sensitivity of global climate-carbon cycle feedbacks to carbon-nitrogen cycle coupling in the land model.

ESSL scientists are also participating in a project to evaluate CLM-CN and an earlier carbon-only biogeochemistry module (CLM-CASA'), known as C-LAMP (Carbon Land Model Intercomparison Project). That ongoing effort involves the evaluation of the land model predictions against a broad database of observations (remotely sensed and in situ) of many different ecosystem states and fluxes.

Plans for FY08

Over the coming year ESSL will continue to document the behavior of the fully-coupled model, with several additional manuscripts in preparation or in planning stages at present. Scientists will also be exploring several new avenues of model development, including coupling CLM-CN and the existing capability in CLM for prognostic biogeography (Dynamic Global Vegetation Model - DGVM), and the introduction of coupling between the carbon, nitrogen, and phosphorus cycles. Members of the scientific staff are participating in an NSF-sponsored project focused on fire at the intersection of carbon and water cycles, and in that context will be evaluating and hopefully improving the existing simple treatment of prognostic fire fluxes in CLM-CN. Staff also recently participated in a workshop aimed at coordinating the development of new scenarios within the integrated assessment community and the needs and capabilities of the new generation of coupled climate-biogeochemistry models. Work also continues on exploring the interactions of carbon and nitrogen cycling with landuse and landcover change, where scientists have already performed some preliminary experiments in the CCSM framework. As with most activities in this area, our work is highly collaborative and ESSL continues to explore opportunities for collaboration, such as with another NSF-sponsored project on hydrologic synthesis, led from the University of Illinois, Champaign-Urbana.

return to top

Climate and tropical cyclones

NCAR scientists working with university colleagues at Georgia Tech have shown that the extremely active 2005 hurricane season can be attributed to oceanic and atmospheric anomalies that substantially enhanced the potential for African Easterly Waves (AEW) to develop into tropical cyclones. The oceanic anomalies were very much concentrated in the eastern equatorial Atlantic, a continuation of a long trend for the North Atlantic warm pool to extend eastward. These anomalies appear to have driven a secondary zonal circulation in the atmosphere, which shifted the maximum in the low level easterlies from the Caribbean to near Africa. Over much of the hurricane development region, these conditions of easterly winds increasing towards the east, leads to a strong tendency for AEWs to shrink in size, accumulate energy, and increase their core vorticity

The resulting transfer of large-scale wave vorticity and energy to smaller scales substantially enhanced the potential for tropical cyclone formation. And this is reflected in the statistics, for example: 30% of AEWs in 2005 produced tropical cyclones, compared to the normal 10%; and, an unprecedented 5 storms formed in July from AEWs and 2 became major hurricanes, also a record. The importance of this mechanism can be seen in the figure, which shows that all equatorial tropical cyclones in 2005 formed in regions of wave accumulation.

NCAR has now completed a comprehensive simulation of the entire tropical belt using the Nested Regional Climate Model at 36 km resolution from 1995-2005. In addition a 12 km simulation has been made for the entire North Atlantic 2005 hurricane season. These simulations show realistic tropical cyclone frequencies and distributions, including the correct number of 2005 storms. The data are now being analyzed to further our understanding of the underlying mechanisms, including AEW development. This work is expected to help ascertain the types of conditions that high-resolution climate models will need to simulate to enable realistic projections of hurricane activity. This work is being carried out with the support of NSF.

return to top

Comparative Solar System Studies

Dr. Feng Tian has been working on a 1-D multi-component hydrodynamic thermosphere/ionosphere model which can be used to investigate the evolution of planetary atmospheres, including that of early Earth. The first paper from this project, through collaboration with James F. Kasting at PSU, Ray Roble and Hanli Liu at HAO, exposes the Earth's current atmosphere to 1x -- 20x present EUV fluxes and discusses the transition of the thermosphere from a hydrostatic regime into a hdyrodynamic regime, in which the adiabatic cooling associated with the hydrodynamic flow must be taken into the energy consideration. This paper is currently under review in JGR-planets. Collaborating with Stan Solomon and Liying Qian at HAO, Feng has been able to couple the GLOW model with the 1-D hydrodynamic model successfully. The coupled model can self-consistently calculate the contributions of photoelectrons to ionization, excitation, and ambient electron heating in the upper atmosphere. They are currently working on the expansion of the GLOW model so that the coupled model can be applied to a broader range of planetary atmospheres.

return to top

Intraseasonal/tropical climate variability

The important problem of quantifying the interactions between organized convection and representing them in parameterizations for global models has received relatively little attention. Cloud-system resolving models (CRM) represent convective organization and its scale interaction properties explicitly. When used in conjunction with dynamical models and observational analysis, this explicit approach provides a comprehensive way to evaluate and improve the representation of convection, an ever more important aspect of global prediction. The following summarizes the results of three activities that have the organization of precipitating convection as a common factor.

a) Hybrid parameterization of convective organization.

A new hybrid parameterization was designed primarily for next-generation global models (grid-spacing ~10 km). At this intermediate resolution, parameterized convection and explicit convection usually coexist. When applied in MM5 at 60 km grid spacing, Fig. 1 shows that the hybrid parameterization improves the propagation speed and precipitation intensity of propagating convection in summertime conditions over the continental United States. A simplified form of this approach represents stratiform heating and mesoscale downdrafts suitable for contemporary climate models. Plans are underway to implement the latter in the Community Atmospheric Model (CAM) in order to test the large-scale effects impact of convective organization. This study is collaborative with NCAR’s TIIMES Water Cycle Program.

b) Stratospheric gravity waves generated by multiscale tropical convection.

Previous studies of gravity waves in the stratosphere generated by deep convection in the troposphere exclusively examined single convective systems. Herein, gravity waves generated by multiscale systems and fields of convection are examined using a cloud-system resolving model. Results for an initially motionless horizontally uniform atmosphere reveal significant nonlinear effects. Long gravity waves and the convective organization provide layered inflow/outflow to cloud systems which, in turn, affects the convective organization (dynamic feedback). Slower-moving short gravity waves (about 5 m/s) are the main source of the vertical flux of horizontal momentum and cause a peak in the momentum flux spectrum corresponding to the lifetime and horizontal scale typical of cloud systems (Fig. 2). This selective layering (shear) filters the short-wave spectrum. Convection penetrating the stratosphere is not the primary wave generation mechanism, since the simulated convection does not reach the tropopause. As a continuation, the effects of tropospheric shear on the stratospheric gravity waves are being examined. This research is collaborative with Todd Lane, The University of Melbourne, Melbourne, Australia.

c) Madden-Julian Oscillation (MJO).

The first part of an idealized investigation of convective momentum transport and super-rotation properties of idealized MJOs using the Intraseasonal Planetary Equatorial-scale Dynamics (IPESD) multiscale model was completed in 2007.(collaborative with Andrew Majda, Courant Institute, NYU and Joseph Biello, University of California/Davis). Because of the balanced dynamics on the synoptic scales, the synoptic-scale component of the meridional momentum flux convergence must vanish at the equator. MJO analogs are driven by synoptic-scale heating fluctuations that have vertical and meridional tilts. Irrespective of the sign of the direction of the tilts in each of four MJO examples, the zonal and vertical mean meridional momentum flux convergence from the planetary scales always drives westerly winds near the equator: super-rotation. Equatorial super-rotation occurs when the planetary flow due to the vertical upscale momentum flux from synoptic scales reinforces the horizontally convergent flow due to planetary-scale mean heating.

Weakly constrained simulations of tropical convection usually result in a biased mean state with accompanying detrimental effects on large-scale convective organization, such as the MJO. Case studies of natural MJOs began this year to address this issue. The first objective is to address the mean state issue by adopting a numerical experiment setup consistent with the WRF-based tropical channel model, where such bias occurred. Multi-week simulations of nested WRF with the outer domain of Indian Ocean/Pacific basin scale and a grid spacing of 36 km, specified SST, and lateral boundary conditions provided by NCEP and/or ECMWF global analysis. Newtonian relaxation (nudging) applied to simulations produced an environment for the MJO which agrees better with satellite observations of precipitation than simulations without relaxation. This success sets the scene for the second objective: cloud-system resolving modeling to simulate multi-scale convective organization and its large-scale effects. This work is collaborative with the MJO Focus Theme of the Multiscale Modeling Framework NSF Science & Technology Center based at CSU.

This work is supported by the NSF.

return to top

Analysis, Integration and Modeling of the Earth System (AIMES)

ESSL is the home of the International Project Office (IPO) for the International Geosphere-Biosphere Programme's (IGBP) Earth System modeling project, Analysis, Integration and Modeling of the Earth System (AIMES). The AIMES project endeavors to extend Earth System modeling approaches to include dynamics of human activities alongside biogeochemical and biophysical processes of the coupled climate system. Modelling activities in AIMES include improving biophysical and biogeochemical components of global models and testing the sensitivity of tradeoffs in vulnerability and resilience in terms of economic and ecosystem consequences.

Relevant activities in AIMES include the Coupled Carbon Cycle-Climate Model Intercomparison Project (C4MIP), where the magnitudes of terrestrial carbon uncertainties are still largely uncertain. C4MIP investigates model benchmark and evaluation exercises to explore mechanisms that influence the response of the terrestrial carbon cycle: (1) soil moisture and net primary production, particularly in the tropics, (2) effects of CO₂ fertilisation, and (3) disturbance and land cover. A joint strategy of AIMES and the WCRP Working Group on Coupled Modeling (WGCM) is to collaborate with the IPCC Working Groups to develop climate change stabilization experiments with coupled atmosphere/ocean general circulation models, Earth System and Integrated Assessment models. AIMES, on behalf of the IGBP is leading an applied Earth System Science initiative to foster collaboration and exchange of information between the scientific community and resource managers, policy and and assessment communities and development agencies. The International Nitrogen Initiative (INI), addresses end-to-end problem solving across scales (e.g., spatial, temporal, management) implementing process-based research through mitigation or management. In addition, AIMES sponsors the Global Emissions Inventory Activity (GEIA). AIMES is also working within the international community to develop an integrated synthesis of activities in the northern high latitudes to promote model development.

To date, AIMES has initiated a Young Scientist's Network (YSN) with topics including integrating indirect human activities (e.g., land use) with the Community Climate System Model (CCSM), Urbanization, biogeochemistry and the climate system, and land-use decision making for Earth system models. The Integrated History of People and Earth (IHOPE) activity published Collapse or Sustainability: an Integrated History and Future of People on Earth published by MIT Press, January 2007 and a summary paper in press with the same title to Ambio. An overall conclusion from the Dahlem-IHOPE conference is that societies respond in various manners to environmental (e.g., climate) stress. Extreme drought, for instance, has triggered both social collapse and ingenious management of water through irrigation. Sponsoring entities for workshops, symposia and colloquia include: NSF, NASA, EC-ACCENT (EU EC_FP6), IGBP, MPI (Germany), QUEST (UK), Arizona State University, University of Vermont, IHDP and WCRP.

return to top

Master Mechanism

The NCAR Master Mechanism is an explicit and detailed gas phase chemical mechanism combined with a box model solver. User inputs include species of interest, emissions, temperature, dilution, and boundary layer height. Any input parameter may be constrained with respect to time. Photolysis rates are calculated using the TUV model included in the code package. The model is written in a mixture of F77 and Fortran90, and is managed using C-shell scripts.

ACD scientists continued the development of the Self-Generating Master Mechanism (SGMM), the only fully explicit mechanism for the gas-phase atmospheric oxidation of hydrocarbons. In collaboration with Bernard Aumont and Marie Camredon (U. Paris), ACD scientists incorporated the calculation of vapor pressures of all intermediate compounds. For conditions relevant to Mexico City, the gas phase SGMM consisted of ca. 1.5x106 reactions among ca. 3x105 different chemical species. By allowing for gas-particle partitioning of these condensable compounds, they simulated the formation of secondary organic aerosols (SOA) in Mexico City for the conditions of the MCMA-2003 campaign. Results show significant SOA production for several days, although at rates slower than observed.

In another study, the SGMM was used to simulate the production of SOA in smog chambers, using 1-octene as a case study. SOA formation was found to proceed through several oxidation steps, and is highest at intermediate NOx levels.

ACD scientists, together with researchers from Pacific Northwest National Labs and the U. of Washington used smog chamber measurements to show that hydrophobic aerosols play no role in SOA formation, contrary to current theory.

FY08 work will continue evaluation and use of the SGMM.

This work is funded by NSF/NCAR.

return to top

Tropospheric Ultraviolet and Visible (TUV) - Radiation Model

The Tropospheric Ultraviolet-Visible (TUV) radiation model calculates spectral irradiances and actinic fluxes, biologically active UV radiation at the surface, and photolysis coefficients (J values) for atmospheric chemistry use. TUV version 4.5, was made available to the community through the NCAR/SCD Community Data Portal by ACD scientists. The new version includes updated spectroscopic data for a number of photolysis reactions and biological effects. ACD scientists updated the global UV climatology to use TOMS v.8 ozone and produced maps of UV-A, UV-B, and several biologically weighted exposures (erythema, non-methane skin cancer, and vitamin D production). [Lee-Taylor, J. and S. Madronich, Climatology of UV-A, UV-B, and Erythemal Radiation at the Earth's Surface, 1979-2000, NCAR Technical Note TN-474-STR, August 2007.] An analytic formula for the UV Index in terms of the ozone column and sun angle was developed and allows efficient parameterization of UV effects of ozone changes.

ACD scientists and university colleagues began examining the UV radiation field during the MIRAGE-Mex field campaign. Mexico City's pollutants (ozone, sulfur dioxide, nitrogen dioxide, and aerosols) reduce the UV radiation field in the boundary layer by 10-20% and at the surface by 20-40%. Preliminary analyses suggest that aerosols are more absorbing at UV wavelengths than at visible wavelengths. The UV reductions slow photochemistry significantly, allowing for greater export of yet-unreacted pollutants.

This work is funded by NSF/NCAR.

return to top