ccr index

Climate Change Research Section

The Climate Change Research Section (CCR) is part of the Climate and Global Dynamics (CGD) Division at the National Center for Atmospheric Research (NCAR) in Boulder, Colorado.

CCR uses appropriate climate system models to study the sensitivity and stability of the Earth system to a variety of forcings, including changes of greenhouse gases, aerosols, solar irradiance, volcanic forcing, land characteristics, and land use change.

CCR provides a focal point for NCAR and university paleoclimate research and serves as a resource to the paleoclimatic and climate change research community in the use of the Community Climate System Model (CCSM).

CCR is involved with Department of Energy (DOE) Laboratories in developing and using high-performance coupled climate models.

Paleoclimate

Climate of the Last 150,000 Years

Bette Otto-Bliesner (CCR), Esther Brady (CCR), Christine Shields (CCR), and Sang-Ik Shin and Zhengyu Liu (visitors, University of Wisconsin) used the Paleo Climate System Model Version 1 (PaleoCSM1) to investigate the coupled climate system in the tropical Pacific region over the last glacial-interglacial cycle. CSM reproduces recent estimates, based on alkenones and Mg/Ca ratios, of sea surface temperature (SST) changes and gradients in the tropical Pacific and predicts weaker El Niños/La Niñas compared to present for the Holocene and stronger El Niños/La Niñas for the Last Glacial Maximum (LGM). Changes for the LGM (Holocene) are traced to a weakening (strengthening) of the tropical Pacific zonal SST gradient, wind stresses, and upwelling and a sharpening (weakening) of the tropical thermocline. CCSM results suggest that proxy evidence of weaker precipitation variability in New Guinea and Ecuador over this time period are explained not only by changes in El Niño/La Niña but also changes in the atmospheric circulation and hydrologic cycle.

This figure shows the annual mean results for the equatorial Pacific Ocean from the CSM simulations. Top: sea surface temperature (°C), Middle: upper-ocean temperature (°C) averaged over 2°S-2°N, and Bottom: upwelling (10^-4 cm s^-1) at a depth of 50m.

Carrie Morrill (Advanced Study Program) and Otto-Bliesner used this series of PaleoCSM1 simulations to also study the influence of El Niño/La Niña on extratropical climate since the LGM. These simulations indicate that the effects of tropical Pacific SST anomalies on the extratropics were greater than today at the LGM and less than today during the Holocene. This is due both to changes in the strength of El Niño/La Niña events and changes in the base state of the climate. These results are important for interpreting paleoclimate records and for understanding strong linkages between the tropics and extratropics during the LGM.

Otto-Bliesner and Amy Clement (University of Miami) compared the Hadley cell response to altered external forcing and climatic boundary conditions in two climate models, the NCAR CSM and the Geophysical Fluid Dynamics Laboratory (GFDL) slab ocean model. The experiments used past and future climatic forcings that range from the LGM to six-fold increases in atmospheric CO₂. They found that both models have a generally consistent response to all forcings in the depth and meridional extent of the Hadley cell in both solstitial seasons. The depth of the winter Hadley cells increases as tropical surface temperatures increase from LGM to present to future greenhouse climates. The strength of the December-January-February (DJF) cell varies consistently between the two models, increasing in colder climates and weakening in warmer climates. For June-July-August (JJA), however, the sign of the response is opposite between the models for a given forcing. It is suggested that the origin of this different response in the JJA cell is related to the type of ocean model used. Nevertheless, the results from all seasons show a consistent correlation between the surface temperature gradient between the winter and summer hemisphere tropics and Hadley cell strength: when the temperature contrast is larger (smaller), the Hadley cell is stronger (weaker). These results have implications for paleoclimate studies where surface temperatures are derived from proxy records.

Liu, Brady, and J. Lynch-Stieglitz (Lamont-Doherty Earth Observatory) studied the global upper ocean evolution in the Holocene due to changed orbital forcing. They found that the annual mean SST changes in the early to mid-Holocene are forced primarily by the annual mean insolation forcing with an overall symmetric response of colder equator (<0.5°C)/warmer high latitudes (<0.4°C in the Southern ocean and >1°C in the Arctic). This SST change is consistent with a synthesis of mid-Holocene paleo-SST records. In contrast, the temperature response in the thermocline is dominated by an antisymmetric pattern with cooling (warming) in the Northern (Southern) Hemisphere midlatitudes. The thermocline response is determined predominantly by surface water subduction, and ultimately, the insolation forcing in local late winter.

Otto-Bliesner, Jonathan Overpeck (University of Arizona) and Gifford Miller (University of Colorado) are using simulations with the Community Climate System Model Version 2 (CCSM2) to study arctic climate change during the Last Interglacial (LIG) (130 ky before present (BP)). Proxy data for the Last Interglacial indicates that terrestrial sites warmed rapidly and early (most by 4-6°C), boreal forest reached the Arctic Ocean coastline except for the North Slope of Alaska, arctic glaciers melted above 5 km elevation, and the Greenland ice cap melted except for the summit. The primary forcing of the LIG was the anomalous seasonal solar radiation at the top of the atmosphere due to orbital changes. At 70°N, maximum anomalies occur in May (+70 Wm^-2) and minimum anomalies occur in September (-50 Wm^-2). Forced with these anomalies, CCSM indicates significant decreases in arctic sea ice and warming of Greenland during the boreal summer months. Greenland warms by 1-5°C with temperatures above freezing in southern and coastal areas. Warming is comparable to that at 3 x CO₂.

This figure shows June-July-August surface temperature change, 130 ky BP (pccsm.08b) versus present (pccsm.08) simulated by CCSM2.

Otto-Bliesner and Brady show that the significant improvements in the component models of CCSM2 now result in improved simulation of the "Greening of the Sahara" during the early to mid-Holocene. Mid-Holocene proxies for northern Africa suggest that lake levels were significantly higher and steppe and xerophytic vegetation grew in present-day desert regions. PMIP atmosphere-only general circulation models (GCMs) and previous atmosphere-ocean GCMs (including Community System Model Version 1 (CSM1)) underestimated both the northward shift and the magnitude of the precipitation increase required to maintain steppe vegetation. A CCSM2 simulation for 8.5 ky BP gives a more realistic northward expansion of precipitation into the Saharan-Sahel region with both a longer African summer monsoon season and greater rainfall during the monsoon season.

This figure show the annual precipitation change (cm) over Africa for 8.5 ky BP compared to present simulated by CCSM2. The insets give details of the monthly precipitation changes simulated for 8.5 ky BP.

Brady and Otto-Bliesner investigated the response of the circulation in the Mediterranean Sea and the exchange with the Atlantic Ocean at Gibraltar to orbital and greenhouse gas forcing in CCSM2 simulations for the mid-Holocene (8.5 ky BP) and LIG(130 ky BP). Sediment cores taken from the eastern Mediterranean show periodic depositions of dark, meter-thick, layers of rich organic sediments called sapropels. Compared to the present-day simulation, results from these paleointegrations, which have perihelion in boreal summer, are consistent with the hypothesis that sapropels may have been deposited because an enhanced African monsoon increased the Nile river discharge into the eastern Mediterranean weakening deep water formation. Enhanced Nile river discharge may have increased the nutrient supply enhancing productivity in the Eastern basin, and the weaker deep water formation may have allowed for anoxic benthic conditions that would have led to better preservation of the organic matter. Although this model lacks the biogeochemistry component that would be able to model these conditions definitively, the results of Brady and Otto-Bliesner are consistent with the physical changes hypothesized. A transient increasing CO₂ (by 1%/year) simulation shows a weak increase in the Mediterranean Overflow Water (MOW) salinity and transport at the Strait of Gibraltar due to an increase in evaporation and decrease in precipitation over the basin. This is consistent with the observed trend over the last 50 years.

This figure shows the zonal overturning streamfunction calculated for the Mediterranean Sea. Units are 10⁶m³/s and positive streamlines exhibit clockwise flow. The streamlines show surface Atlantic water flowing eastward, sinking in the eastern basin and returning as Mediterranean overflow water. In the paleosimulations, a reversed circulation is found in the Eastern basin.

Climate of the Past 1000 Years

Caspar Ammann (Advanced Study Program) and Fortunat Joos (University of Bern, Switzerland) completed transient simulations with the coupled PaleoCSM1 covering roughly one millennium of natural climate variability prior to significant anthropogenic influence. In collaboration with David Schimel (TSS), Otto-Bliesner, and Robert Tomas (CCR), experiments were designed to test the fidelity of the previously chosen approach of implementing natural external forcing factors in the coupled simulations of the 20th century. A new new multi ice-core-based volcanic forcing series for the period 850-2000 AD and the latest atmospheric composition data from Antarctica were used as forcings. Emphasis of the simulations was focused on the magnitude of potential solar irradiance changes on climate. A best guess scenario was run that included a background trend of solar irradiance back to the Maunder Minimum (when sunspots almost entirely vanished for decades) of roughly double the magnitude of the instrumentally confirmed 11-year cycle range. This was compared with a simulation forced by a solar series scaled by a significantly larger trend (~6x). The base history for generating past irradiance was based on 10-Beryllium isotopes recovered from polar ice cores. The temporal statistics of this series is consistent with the commonly employed sunspot based forcing series. The best guess scenario, for reference, is also consistent with the forcing used in previously reported 20th century simulations. Therefore, the long pre-anthropogenic millennium simulations test the coupled models ability to reproduce naturally forced climate variations.

The two simulations (small solar and large solar forcing) are compared to different proxy-based climate reconstructions. Remarkable agreement of the best guess run (low solar) is found with most climate reconstructions. Although the experiment remains on the cooler side, it stays almost entirely within the uncertainty bounds. Only very large explosive eruptions cause the hemispheric and global surface temperature to drop out of the expected range, a problem commonly encountered in climate simulations. The large solar experiment, on the other hand, is significantly colder and is thus less consistent with the climate data. In the 20th century, the simulation that was quite successful in generating pre-anthropogenic natural variations remains 0.5 to 0.7 degrees Celsius too cool if anthropogenic forcings are held constant at 1870 AD conditions. Yet, if the greenhouse gas changes and tropospheric sulfate aerosol (only direct effect) are included, the simulations reproduce observations very closely.

This figure shows global annual average surface temperature anomalies from 1850-2000 AD simulated by PaleoCSM compared instrumental record (black) and Jones et al. reconstruction (yellow). PaleoCSM1 simulations are small solar (red), large solar (blue), and large solar without anthropogenic forcing after 1870 (green).

Given the good match over the last 1000 years including the instrumental period, Ammann then extended the 'best-guess' simulations in collaboration with Warren Washington (CCR) and Gerald Meehl (CCR and Climate Analysis Section) to the year 2100 using the Intergovernmental Panel on Climate change (IPCC) scenario A2 for future anthropogenic forcing. Natural forcings were held constant at 2000 AD conditions.

This figure shows the comparison of this run with a multiproxy reconstruction and the instrumental record. Based on the good agreement of the low-frequency magnitude in the past confirming, at least to a first order, the climate sensitivity of the model, the projected future warming of +2.5 degrees Celsius to the end of the 21st century can now be seen in a different, more striking perspective.

Under the umbrella of the Weather and Climate Change Assessment Science Initiative, several projects investigated aspects of the millennium simulations from the standpoint of uncertainty. Eugene Wahl (Environmental and Societal Impacts Group) and Ammann tested stationarity of teleconnections to El Niño Southern Oscillation (ENSO) throughout the millennium. Many paleoclimate reconstruction techniques rely on the assumptions that stationarity is preserved. Results based on the model output largely support this notion, but some exceptions carry potentially important implications. Additionally, Ammann, Otto-Bliesner and Linda Mearns (Environmental and Societal Impacts Group) have worked with the UCAR Education and Outreach Office (Roberta Johnson, Linda Carbone, Susan Foster and Doug Haller) to generate the framework for a teachers guide to the new NCAR Climate Discovery exhibit focusing on the Little Ice Age.

Philippe Naveau (University of Colorado), Hee-Seok Oh (University of Alberta, Canada), and Ammann have generated and applied new statistical tools to extract external forcing fingerprints from climate time series. Using discrete wavelet decomposition on decadal to century timescales, the detection was successful in isolating potential solar influence in different multiproxy reconstructions and in simplifying the detection in the coupled model simulations. For the very short-lived volcanic effects, a state-space model approach was chosen that allows quantification of the volcanic cooling with associated posterior probability.

Alan Robock (Rutgers University) and Ammann, with support from the global ice-coring community, started to collect the latest high-resolution ice core sulfate data from around the world. A meeting at the Niels Bohr Institute in Denmark launched this effort to generate a new and improved volcanic forcing dataset specifically designed for climate modeling. This record will be made available to the community through the NOAA paleoclimate website.

Brad Adams and Michael Mann (University of Virginia) and Ammann have analyzed the effects of past explosive volcanism on the ENSO system as reconstructed in multiproxy networks. For different volcanic eruption lists, including the largest events during the past 400 years, and evaluating climate response in both Niño3 as well as largely independent Solar Oscillations Investigation (SOI) reconstructions, results point to a statistically significant tendency of the preferred El Niño state for the first few years after a large volcanic eruption in the tropics (probability almost twice than expected by chance). The significance still holds to five years after the eruptions when anomalous cooling (La Niña conditions) is found in the composites. This work contributes to the current focus of research on the response of the tropical climate system to largescale external or anthropogenic forcing.

Future Climate Change

As part of the DOE/UCAR Cooperative Agreement, Washington and Meehl have carried out the largest set of publicly available climate change simulations with the Parallel Climate Model (PCM) that we are aware of anywhere in the world. The data from these simulations are available from the Program for Climate Model Diagnosis and Intercomparison (PCMDI) at Lawrence Livermore National Laboratory (LLNL), National Energy Research Scientific Computing Center (NERSC) located at Lawrence Berkeley Laboratory, and at NCAR. This archive of simulations includes a 1000-year control run, Special Report on Emissions Scenarios (SRES) scenario runs for the 21^st century, business as usual, stabilization of greenhouse gas runs to the end of the 22^nd century, special simulations for the demonstration project of the Advanced Climate Prediction Initiative (ACPI), a large number of single and combined forcing runs for 20th, and other simulations. Information on the simulations can be found at www.cgd.ucar.edu/pcm/. Gary Strand (CCR) has been the coordinator of the large data archive and he has been the lead CGD person involved in the DOE Earth System Grid (ESG) project that will provide web access to PCM and CCSM that are distributed at several supercomputing centers. This archive complements the smaller number of CSM simulations with some of the same external climate forcings. The computer time for these simulations was obtained from the Scientific Computing Division of NCAR, NERSC, Oak Ridge National Laboratory (ORNL), and Los Alamos National Laboratory (LANL).

Higher Resolution Versions of CCSM

Both the Climate Change Working Group (CCWG) and the DOE Climate Change Prediction Program (CCPP) are particularly interested in regional aspects of climate change. We have seen that by increasing the horizontal resolution with the uncoupled version of the Community Atmosphere Model (CAM) at T85 (developed by the Climate Modeling Section (CMS) and the CCSM Atmosphere Model Working Group) that there is substantial improvement in the regional aspects of precipitation especially in mountainous areas. This is particularly important for the western North American regions. Also, there appears to be improvement in the surface winds in the Arctic Ocean region, which in turn will affect the flow of sea ice. The present T42 version does not drive the sea ice correctly in several areas of the Arctic Ocean. During the last year we have performed some limited tests using T85 coupled to the CCSM ocean, sea ice, and land components. The tests indicate several improvements. This is very encouraging since these have been persistent systematic errors in previous lower resolution versions.

PCM Climate Change Simulations of the 20^th and 21^st Centuries: Sensitivity and Decadal Variability

Meehl, collaborated with Aiguo Dai (CCR), Washington, Tom Wigley (Climate Analysis Section), and Julie Arblaster (CCR) to run the model and analyze the influences of various anthropogenic (greenhouse gases (GHGS), ozone, sulfate aerosols) and natural (volcanic and solar) forcings over the 20th century in the PCM. Four member ensembles of the single forcings and an additional eight experiments with combinations of forcings in four member ensembles produced the largest number of forced 20th century climate experiments ever run with a global coupled climate model. Early century warming was shown to be mainly due to solar forcing, with enhanced tropical convection contributing to amplifying the solar forcing. The global temperature signal from the forcings was shown to be roughly additive.

Meehl, Washington, Thomas Bettge (SCD), Arblaster and Wigley collaborated with Ben Santer (LLNL) and a team of scientists to analyze tropopause height as an indicator of climate change in models and observations, noting a warming troposphere to be associated with increased tropopause height in the model and observations. Through this work they were able to quantify the contributions of the combination of natural and anthropogenic forcings on changes in tropopause height, and confirm that the troposphere actually warmed during the second half of the 20th century due to anthropogenic forcing factors, mainly GHGs and ozone.

In collaborative work with Arblaster, Meehl documented the mechanisms involved with the model result of increased mean south Asian monsoon precipitation and increased interannual variability in a future warmer climate. Sensitivity experiments with the CCM3 showed that increases in mean monsoon precipitation were caused mainly by warmer Indian Ocean SSTs, and increased variability was associated with forcing from the tropical Pacific via the Walker Circulation.

Meehl, in collaboration with Ping Liu (University of Hawaii), analyzed results from the global coupled models in the Coupled Model Intercomparison Project (CMIP) to study possible future changes of the Sahara Desert. Five of the eighteen CMIP models were chosen for analysis based on their ability to simlulate a reasonable present day climatology of the Sahara Desert with similar rainfall distributions and meridional boundaries as in the observational data. When CO₂ concentration was increased at one percent per year for 80 years in these models, the Sahara Desert moved north, became hotter, and became drier. The local enhanced greenhouse effect from increased CO₂ increased the net surface sensible heat flux, which in turn contributed to the warming trend.

Meehl also collaborated with Washington and Arblaster to analyze four recent NCAR global coupled models (PCM, CSM, Parallel Climate Transitional Model (PCTM) and CCSM) to show that the relevant feedbacks (ice/albedo, water vapor, and clouds) for climate system response to increasing CO₂, are managed by the atmospheric model. The ocean, sea ice and land surface play secondary roles for the globally averaged response. Two models with identical atmospheres but different ocean and sea ice components (PCM and PCTM) have the most similar response to increasing CO₂, followed closely by the CSM with comparable atmosphere and different ocean and sea ice. The most dissimilar model, the CCSM, has a different response from the other three. In particular, it differs from PCTM although having very similar ocean and sea ice but different atmospheric model components. Analysis of ocean heat uptake in these models showed that processes and responses were model dependent, and must be studied more thoroughly in each model to understand the global climate sensitivity of the coupled system.

Meehl and Arblaster collaborated with David Karoly (University of Oklahoma) and a team of scientists to document a set of simple indices for climate change. These indices capture aspects of regional climate such as land-ocean temperature contrast for North American climate. The PCM agrees with observations for 20th century climate change for these indices. This suggests that such indices can be used for climate change/detection to provide more detail than globally averaged temperature, while still capturing the essential aspects of regional scale climate change. This can be done without going to pattern-based techniques that may have problems with data coverage in observations, or processes at unresolved spatial scales in the current medium resolution climate models.

Dai, with others in CCR, analyzed trends and variability in the surface air temperature, precipitation, ENSO and Arctic Oscilliation variability, and the Atlantic Ocean circulations in the 1200-year control run of the PCM. The PCM control run is among the first few millennial simulations by a coupled GCM without flux adjustments. The PCM shows small drifts in surface temperature and other fields that are comparable to those of flux-corrected GCMs. The Atlantic Thermohaline Circulation (THC) in the PCM control run shows many interesting features, including a multidecadal (~23 year) oscillation, a reversal of the THC trends, the influence of the Antarctic Bottom Water on the THC, and a horizontal current changes in the North Atlantic associated with the THC changes.

Dai also analyzed a PCM climate change simulation in which CO₂ and other greenhouse gases and sulfate aerosols are projected to increase in the 21st century and stablize in the 22nd century (CO₂ at 550ppmv). Thereafter, the greenhouse gases decrease so their concentrations are symmetric around year 2200 (i.e., CO₂ concentration in 2450 equals that in 1950). The results show that the response of the THC is fairly linear, i.e., the THC strength decreases with the CO₂ increases and increases with the CO2 decline with very small phase lag.

Dai (with others in CCR) analyzed the PCM simulations for the DOE-sponsored ACPI Program. In these runs, the oceans were initialized to 1995 conditions by a group from the Scripps Institution of Oceanography and other institutions. An ensemble of model runs then was carried out to the year 2099 using the projected forcing. It is shown that the initialization to1995 conditions removes a large part of the unforced oceanic temperature and salinity drifts that occurred in the standard 20th century integration.

Using the PCM, regional responses of the THC in the North Atlantic to increased CO₂ and the underlying physical processes were studied by Aixue Hu (CCR). The Atlantic THC shows a 20-year cycle in the control run, qualitatively consistent with observations. Compared with the control run, the simulated maximum of the Atlantic THC weakens by about 5 Sverdup (Sv) or 14% in an ensemble of transient experiments with a 1%/year CO₂ increase at the time of CO₂ doubling. Changes in atmospheric forcings, oceanic current patterns, and sea ice conditions produce overall weakening of the THC in the Labrador Sea and South Denmark Strait Region (SDSR), and more vigorous ocean overturning in the Greenland, Iceland and Norwegian (GIN) Seas. The northward heat transport south of 60^oN is reduced with increased CO₂, but is increased north of 60^oN due to the increased north Atlantic water across this latitude.

Hu also preformed a set of CMIP coordinated experiments to investigate the sensitivity of the THC to the variability of the surface freshwater flux using NCAR’s CCSM2. The resulting THC variations show that when the northern Atlantic receives a 0.1 Sv additional freshwater (hosing), the THC slows down by 4.4 Sv, and then recovers to its full strength 50 years after the end of the hosing. When atmospheric CO₂ is doubled (quadrupled) the present day value, THC weakens by 1.5 (2.9) Sv. This weakening of THC is related to the CO₂ induced warming and the transport of melt ice water from the Arctic into the Labrador Sea. When the surface freshwater forcing from the control run is used in the 1%/year CO₂ run, the THC is further weakened due to a higher freshwater input in the Labrador Sea. On the other hand, the use of the 1%/year CO₂ run’s freshwater flux in the control run leads to a stronger THC. This is due to overall less freshwater flux input in the northern North Atlantic.

Land Cover Change

Johannes Feddema and Jerome Dobson (both University of Kansas), and S. Freire (Oak Ridge National Laboratory), have made a great deal of progress on developing data sets and changes to the models specifying anthropogenic land surface changes from years 1750 to 2050 that can be used in PCM and CCSM. This will provide a new forcing for anthropogenic changes in the climate system.

This figure shows the difference in land cover between 1870 and the present (1992). The largest change is the increase in agriculture use as indicated in the bright red.

More information may be found on the Parallel Climate Model and CCSM Climate Change Working Group (CCWG) web pages.

The following are scientists and programming staff involved in the PCM effort or its components in alphabetical order: J. Arblaster (NCAR), T. Bettge (NCAR), L. Buja (NCAR), A. Craig (NCAR), A. Dai (NCAR), J. Dennis (NCAR), J. Dukowicz (LANL), J. Hack (NCAR), A. Hu, (NCAR), E. Hunke (LANL), R. James (NCAR), P. Jones (LANL), R. Loft (NCAR), R. Malone (LANL), M. Maltrud (LANL), W. Maslowski (Naval Postgraduate School (NPS)), G. Meehl (NCAR), A. Middleton (NCAR), A. Semtner (NPS), R. Smith (LANL), Ilana Stern (NCAR), G. Strand (NCAR), Robert Tomas (NCAR), W. Washington (NCAR), V. Wayland (NCAR), D. Williamson (NCAR), and Y. Zhang (NPS).

Climate Change Sensitivity Studies

Jeffrey Kiehl (CCR) in collaboration with Shields have carried out an analysis of the equilibrium slab ocean simulations of the Community Atmosphere Model Version 2 (CAM2) and Community Atmosphere Model Version 3 (CAM3) atmospheric versions of the CCSM. These simulations show the climate sensitivity of the new model is 2.6 °K for a doubling of CO₂, while the sensitivity of the CAM2 model is 2.2 °K. Analysis of these two equilibrium simulations and simulations of the fully coupled models forced with a 1%/year increase in CO₂, indicate that changes in the low cloud response to warming is the major reason for the increase in sensitivity. In CAM2 low clouds were diagnosed as a function of relative humidity for all cloud types including convective clouds. In the new model convective cloud amount is diagnosed as a function of cloud mass flux. Studies from high resolution cloud resolving models have shown that cloud mass flux is a better predictor of convective cloud amount than relative humidity. This change in the formulation of low cloud amount leads to a smaller increase in low cloud amount for a warmer climate state.

This figure shows the time series of global annual change in shortwave cloud forcing (Wm^-2) due to a 1%/year increase in atmospheric carbon dioxide. The blue curve is for the CCSM2 and the red curve is for the CCSM3.

The more negative the cloud forcing the greater the negative feedback on the climate system. The CCSM2 has a much stronger negative cloud feedback than the CCSM3 model.

In order to better understand the physical processes contributing to these feedbacks, Kiehl and Shields have diagnosed the correlation of various radiative and hydrologic variables from the CAMX against observations. Given the importance in cloud properties to the overall climate sensitivity, the initial focus has been to consider the relationship of cloud amount, cloud liquid water and cloud radiative forcing to large scale meteorological fields. In general there is very good agreement between the model correlations and the observed correlations.

This figure shows the correlation between longwave cloud forcing and 500 hectopascals (hPa) vertical velocity from the new version of the CAM (red) and observations (black).

These analyses have been carried out for tropical, subtropical and extratropical synoptic regimes. The largest differences between model and observations are in the simulation of high cloud amount and high latitude predicted cloud water.

A comparison of the CAM3 with the latest version of the NOAA Geophysical Fluid Dynamics Laboratory (GFDL) atmospheric model has also been carried out. This project falls under the new memorandum of understanding between NCAR and NOAA GFDL. The goal of this collaboration is to understand model sensitivities and the sources of agreement or difference in climate sensitivity. It is hoped that the regional sensitivity analysis will provide a useful means to understand inter-model differences.

Another method of analysis is to force the atmosphere model with sea surface temperature perturbations from coupled CO₂ simulations. Two sea surface temperature perturbation patterns have been used to force the atmospheric model. The first pattern comes from the slab ocean 2XCO₂ of the CAM3, the second pattern was created by B. Soden (NOAA GFDL) from a composite of CMIP2 1%/year simulations. Both the CAM3 and the GFDL Atmosphere Model Version 2 (AM2) model have been forced with the CMIP sea surface temperature perturbation data. These simulations are being used to study the response of regional climates to a prescribed identical forcing.