A21D-0191
Using ice core data to constrain the Southern Ocean's climate sensitivity
Models of the current generation differ significantly in the Southern Ocean response to doubling in the greenhouse gas concentration, due to their differing simulations of climate feedbacks. Since these feedbacks may operate similarly in the context of natural low-frequency climate variability, one way to constrain these feedbacks in the context of anthropogenic climate change is to constrain them in the context of natural climate variability. Here, using the "pre-industrial control experiments" with the 15 state-of-the-art climate models, we demonstrate that there is equally large inter-model spread in the magnitude of Southern Ocean low-frequency variability. By applying linear stochastic theory, we attributed most of this inter-model spread to the lag-one autocorrelation coefficients of climate variables, largely representing the effects of climate feedbacks on the Southern Ocean low-frequency variations. The lag-one autocorrelation coefficients of climate variables in the Southern Ocean is closely linked to those at the coastal Antarctica. The latter can be extracted from the past climate records stored in ice cores drilled at the coastal Antarctica. This maps out a clear strategy to constrain climate sensitivity in anthropogenic climate change using ice core data.
A21D-0192
In Current Climate Models, an Unrealistic Negative Feedback in the Arctic?
The aim of this study is to explain the origin of the large spread of the response to anthropogenic forcing simulated by state-of-the-art climate models in the Arctic. After having re-examined the notion of Arctic amplification and analyzed some basic features of Arctic climate change, we develop a specific feedback analysis framework in order to address this issue. The analysis is firstly focused on oceanic temperature change as it is a key feature of Arctic climate change. The feedback analysis shows that a large part of the spread of oceanic temperature change is explained the longwave feedback parameter. Then, we show that the negative longwave feedback parameter is strong because the positive cloud cover and the water vapor longwave feedbacks are small in the Arctic. The large spread of the longwave feedback parameter is mainly due to the temperature feedback. The vertical structure of the atmosphere in the Arctic, characterized by a surface inversion during wintertime, is a key factor of the climate response to anthropogenic forcing. In particular, the climatological strength of the inversion appears to exert a strong control on the temperature feedback and consequently on Arctic climate change. As most of the current climate models are likely to overestimate the climatological strength of the inversion, we conclude that their climate sensitivity could be to weak in the Arctic, and the ensemble mean response to anthropogenic forcing over-optimistic.
A21D-0193
Why is There a Minimum in Projected Warming in the Tropical North Atlantic Ocean?
In IPCC projections for the 21st Century, the sea surface temperature shows several regions of minimum warming in the tropical ocean. These include the South-Eastern tropical Pacific, the South tropical Atlantic, and the North tropical Atlantic. These patterns appear both in fully coupled ocean-atmosphere general circulation models (GCMs) and also in atmospheric GCMs coupled to mixed-layer ocean models, and are robust across a multi-model ensemble. The present study focuses on the minimum warming in the tropical North Atlantic, as it has implications for the influence of greenhouse gas-induced climate change on hurricane development. The surface heat budget is analyzed in order to determine the causes for this minimum warming. It is found that the primary contribution is through the influence of the climatological mean wind speed on the efficiency of latent heat flux. In regions of high wind speed, radiative heating can be balanced by latent heat flux with a smaller change in surface temperature.
A21D-0194
Climate Sensitivity Related to Cloud Processes in the General Circulation Model ECHAM5
Climate change as the response to an external forcing is highly related to radiative feedback factors. Among those feedback factors cloud-related feedbacks are of high relevance, as parameterizations of cloud processes in general circulation models (GCM) account for the biggest uncertainties in recent climate projections. Clouds affect radiation in the solar and terrestrial spectra and are sensitive to climate conditions. Discrepancies between different GCMs are still mainly due to cloud depiction and their related uncertainties. Different feedback factors for cloud-related fields in the general circulation model ECHAM5 are estimated. These estimates are calculated from all radiation-relevant output fields for climate sensitivity experiments with varied well-mixed greenhouse gases. Stratospheric-adjusted radiative forcing calculations as a first approximation for potential climate change is used. A single column radiation model based on the ECHAM5 radiation kernel is used to obtain feedback factors for different cloud parameters and their impact on climate sensitivity. The results will be evaluated with observational data using skill scores to analyze short wave and long-wave radiative effects.
A21D-0195
Defining and Quantifying Feedbacks in Earth's Climate System
Feedbacks in Earth's climate system are increasingly being examined to identify processes controlling Earth's climate sensitivity, to quantify the effects of these processes, and to assess the ability of climate models to accurately represent the actual climate system and changes due to increases in greenhouse gases and other forcings. At present differing explicit or implicit choices of the measure of climate change, of definitions of feedbacks, and of the underlying non-feedback climate to which feedbacks must be referred have resulted in differing measures of feedbacks. The single variable that is most commonly taken as a measure of climate response to radiative perturbation is global (and annual) mean (near) surface (air) temperature GMST; climate models indicate that many other changes in Earth's climate scale with change in GMST. The choice of GMST as the index of climate change together with recognition that Earth's energy content H is controlled by shortwave absorption and by longwave emission at the top of the atmosphere as dH/dt = γJS/4 - εσTs4, where Ts is GMST, γ is the planetary coalbedo (complement of the Bond albedo, ~0.70), JS is the solar constant (~1368 W m-2), σ is the Stefan-Boltzmann constant, and ε defines an effective emissivity (~0.62) as the ratio of the longwave flux emitted at the top of the atmosphere to that emitted by a black body radiator at the global mean surface temperature, leads to the choice of reference no- feedback or "open loop" climate sensitivity S0 as the equilibrium change in GMST that would result from a small change the planetary energy budget, forcing ΔF, normalized to that forcing, for γ and ε held constant. This definition yields to first order a climate sensitivity in the absence of feedbacks S0 = (dTs/dΔF)0 = Ts/γ0JS, where the subscript 0 denotes absence of feedback. For Ts = 288 K, S0 = 0.30 K/(W m-2); for forcing from doubled CO2 taken as ΔF2X = 3.7 W m-2, the corresponding CO2 doubling temperature change (interpreted as a derivative quantity) ΔT2X0 = 1.1 K. This no-feedback sensitivity serves as the basis for considerations of feedbacks which would increase or diminish Earth's climate sensitivity from this no- feedback value. Formally this sensitivity is S = fS0 = S0/(1-Φ), where f is the feedback factor and where the total feedback strength is Φ = (d lnγ/d lnTs)0/4 - (d lnε/d lnTs)0/4; a positive value of Φ denotes positive feedback, an increase in the sensitivity over the no-feedback value. The two contributions to feedback strength denote changes in planetary coalbedo and effective emissivity with change in GMST. A decrease in cloudiness (increase in coalbedo) with increasing GMST would result in positive shortwave feedback (positive contribution to Φ); similarly an increase in atmospheric water vapor content and associated longwave absorption with increasing GMST (decrease in emissivity) would result in positive longwave feedback. For climate sensitivity ΔT2X = 3 K (IPCC, 2007) the feedback factor 2.7 corresponds to feedback strength Φ = 0.63. Determining the contributions of individual climate processes to the feedback strength and the sensitivity of feedback to representations of these processes is a major challenge facing the climate modeling community.
A21D-0196
Observing Climate Sensitivity
The reduction of uncertainty in simulated climate sensitivity has proven to be illusive in the past few decades despite major advances in modeling the climate system. Here we explore the hypothesis that relationships between model sensitivities and both the models' mean states and annual cycles can be exploited with the goal of providing an observationally constrained basis for climate sensitivity. We focus on fields involving the energy and water cycles and use simulations from the CMIP3 archive. Relevant model fields are identified based upon: 1) their association with climate sensitivity, 2) their large uncertainty in future climate projections, and 3) their central role in the global energy budget. Strong relationships are identified for several proxies and significantly revised PDFs of climate sensitivity are presented based on their associations with sensitivity among models. Implications for other statistically based methods of deriving climate sensitivity are also discussed.
A21D-0197
Measured Surface Radiative Budget Changes in the Northeastern US and Resultant Regional Climate Change
While numerous numerical climate hind- and forecasts exist, empirical characterizations of changes in the surface radiative budget are exceedingly rare. This situation---which reflects the dearth of long-term, reliable, surface radiation measurements---is most unsatisfying, as it renders a central element of the climate simulation machinery, the radiation code, imperfectly tested and validated against data on an on-going, spatially dense, basis. I report results obtained by analyzing a rare 21-year long dataset comprising hourly measurements of most relevant climate variables, including upwelling and downwelling short- and long-wave radiation, in Millerton, NY, a typical rural southern New England locale. The analysis shows that, while temperatures are indeed rising significantly, consistent with anthropogenic greenhouse forcing, the surface net radiation is in fact decreasing significantly in parts of the diurnal and seasonal cycles. Indirectly, I demonstrate that the changes are due to significantly enhanced probability of southerly flow and resultant advective warming. I also show indirectly that the reduced solar radiation is consistent with the augmented southerly flows which advect water vapor into New England. The increased q enhances relative humidity beyond the dictates of the Clausius-Clapeyron relation and the observed warming, and increase the likelihood of moist ascent. I conclude that while New England is warming, the warming in not directly related to increased atmospheric greenhouse gas concentrations. The possibility that the elevated southerly flow probability is itself a consequence of the thickening atmosphere cannot be unambiguously resolved by the data.
A21D-0198
A Comparison of Climate Feedback Strength between CO2 Doubling and LGM Experiments
Studies of past climate potentially provide a constraint on the uncertainty of climate sensitivity, but previous studies warn against a simple scaling to the future. The climate sensitivity is determined by various feedback processes and they may vary with climate states and forcings. In this study, we investigate similarities and differences of feedbacks for a CO2 doubling, a last glacial maximum (LGM), and LGM greenhouse gas (GHG) forcing experiments, using an atmospheric general circulation model coupled to a slab ocean model. After computing the radiative forcing, the individual feedback strengths: water vapor, lapse rate, albedo, and cloud feedbacks, are evaluated explicitly. For this particular model, the difference in the climate sensitivity among experiments is attributed to the shortwave cloud feedback in which there is a tendency that it becomes weaker or even negative in the cooling experiments. No significant difference is found in the water vapor feedback between warming and cooling experiments by GHGs despite the nonlinear dependence of the Clausius-Clapeyron relation on temperature. The weaker water vapor feedback in the LGM experiment due to a relatively weaker tropical forcing is compensated by the stronger lapse rate feedback due to a relatively stronger extratropical forcing. A hypothesis is proposed which explains the asymmetric cloud response between warming and cooling experiments associated with a displacement of the region of mixed- phase clouds. The difference in the total feedback strength between experiments is, however, relatively small compared to the current intermodel spread, and does not necessarily preclude the use of LGM climate as a future constraint.
A21D-0199
Modelling Uncertainties in the Climate of the Last Millennium: the ASTER Project
The LOVECLIM model (Driesschaert et al., 2007; Goosse et al., 2007) is used to simulate the climate of the last millennium with several 'climate' parameter sets yielding different sensitivities of the climate and the carbon cycle model. The purpose of these simulations is twofold. We intend to assess first the role of the carbon cycle on the climate, and second, the ability of the different selected parameter sets to drive the model within the range of the observed climate, and further to assess the uncertainty related to these parameters. The high frequency variability of the forcings is taken into account. For each set of parameters, LOVECLIM is driven by the natural evolution of insolation, solar irradiance and stratospheric aerosol concentrations due to volcanic activity as well as by changes caused by human activities such as deforestation, CO2 emission or concentration changes, changes in concentrations of greenhouse gases other than CO2 (including ozone) and in sulphate aerosol load. Several transient experiments are conducted for each parameter set. A first transient simulation (CONC) is forced with reconstructed atmospheric CO2 concentration. In the next two simulations, the emissions of carbon are taken into account, the model computing the corresponding atmospheric CO2 concentration. First (EMIS), the emissions due both to the land use changes and the fossil fuel burning are provided. Second (EFOR), only the emissions from fossil fuel burning are provided in addition to the vegetation change related to deforestation. The Northern Hemisphere annual mean temperatures simulated by the model according to the different parameter sets and carbon cycle sensitivities do not show striking differences. The general pattern shows slightly warmer conditions in the early part of the simulation and cooler ones during the Little Ice Age. At last, the global warming of the last century is also clearly simulated. The response of the carbon cycle to the evolution of forcings over the last millennium does not differ much among experiments although there is a much larger spread when considering different emission scenarios (e.g. EFOR and EMIS). Sensitivity tests to the amplitude of the variation of the total solar irradiance (TSI) are performed; a very first quick look at the simulations does not show significant differences in the pattern of the simulated climate in response to modification in the TSI amplitude. Further analysis must be conducted. Climate response to different schemes of deforestation will also be presented. Driesschaert E., Fichefet T., Goosse H., Huybrechts P., Janssens I., Mouchet A., Munhoven G., Brovkin V., and Weber S. L., 2007. Modelling the influence of Greenland ice sheet melting on the Atlantic meridional overturning circulation during the next millennia. Geophys. Res. Lett., 34:L1070, 2007. Goosse H., Driesschaert E., Fichefet T., and Loutre M.F., 2007. Information on the early Holocene climate constrains the summer sea ice projections for the 21st century Clim. Past 3, 683-692.
A21D-0200
Analysis of fast feedbacks in coupled climate-carbon experiments in response to 2x and 4xCO2
We performed an ensemble of twelve five-year experiments using a coupled climate-carbon-cycle model with scenarios of prescribed atmospheric carbon dioxide concentration; CO2 was instantaneously doubled or quadrupled at the start of the experiments. Within these five years, climate feedback is not significantly influenced by the effects of climate change on the carbon system. However, rapid changes take place, within much less than a year, due to the physiological effect of CO2 on plant stomatal conductance, leading to adjustment in the shortwave cloud radiative effect over land, and a 10-20% enhancement to the effective radiative forcing due to CO2. The implications for calibration of energy-balance models are discussed.
A21D-0201
Role of water vapor feedback on the amplitude of season cycle in the global mean surface air temperature
We have analyzed the seasonal variations of global mean surface air temperature (SAT) and surface energy
budgets of 17 AR4 models. Considerable differences in the amplitude of seasonal cycle (A) in the global
mean SAT in the pre-industrial control simulations among the models have been traced, to a large degree, to
differences in their simulated clear-sky downward longwave radiation (LW ) and latent heat flux (LH). We
suggest that water vapor feedback process influence the seasonal changes of SAT through its roles on the
seasonal variations of LW and LH. This implies that the simulated seasonal change of global mean SAT
might contain a clue about the sensitivity of water vapor feedback and the A of in SAT thus provides some
constraint on climate sensitivity since both are subject to the same feedback process.
http://www.agu.org/pubs/crossref/2008/2008GL033454.shtml
A21D-0202
Dynamical Polar Warming Amplification and a New Climate Feedback Analysis Framework
An idealized coupled general circulation model is used to demonstrate that the surface warming due the doubling of CO2 can still be stronger in high latitudes than in low latitudes even without the negative evaporation feedback in low latitudes and positive ice-albedo feedback in high latitudes, as well as without the poleward latent heat transport. The new climate feedback analysis method formulated in Lu and Cai (2008) is used to isolate contributions to the total temperature change obtained with the full GCM model from both radiative and non-radiative feedback processes. The new feedback analysis method considers not only feedbacks that directly affect the TOA radiative fluxes but also those feedbacks that do not directly affect the TOA radiation (such as vertical convections and poleward energy transport). The feedback analysis reveals that the net effect of the external and water-vapor induced radiative energy flux perturbations gives rise a temperature change pattern showing a stronger warming in low latitudes than in high latitudes at the surface and throughout the troposphere and a cooling in the polar upper troposphere and in the stratosphere. The low latitude warming reduction and high-latitude amplification due to non-radiative energy flux perturbations, or dynamical feedbacks, revert the meridionally decreasing warming pattern due to radiative energy flux perturbations at the surface and in the lower troposphere, but not be able to do so in the upper troposphere. As a result, the final warming pattern shows the co-existence of a reduction of the meridional temperature gradient in the lower troposphere and at the surface with an increase of the meridional temperature gradient in the upper troposphere. In terms of the global mean, the external and water-vapor induced radiative energy flux perturbations cause the strongest warming at the surface and strongest cooling in the stratosphere sandwiched with a vertically decreasing warming profile from the surface to the upper troposphere. The dynamical feedbacks increase the global mean warming in the troposphere at an expense of reducing the surface warming. In our idealized coupled GCM, the change in the vertical redistribution of energy is smaller than the radiative energy flux perturbation. As a result, our simple climate model shows a somewhat stronger global mean warming at the surface than in the troposphere.
A21D-0203
Uncertainty analysis of climate sensitivity using a novel spatial reconstruction of sea- surface temperature over the past 500,000 years
Our research strives to use the Earth's past climate history to help bracket the deep uncertainties in future climate projections. We estimated a probability distribution for equilibrium climate sensitivity using a novel reconstruction of global temperature and two independent strategies for estimating changes in surface albedo. Reconstructions of the Earth's past climate have the potential to provide important bounds on climate sensitivity. However, along with large signals, paleoclimate records have large potential errors. Thus, we rigorously applied statistical methods to include multiple sources of uncertainty. We have made advances on four fronts. First, we created a novel reconstruction of global sea-surface temperature over the past 500,000 years that we calculated from over fifty ocean-core proxy records from across the globes, including multi-proxy comparisons. Second, we compared two independent strategies for estimating surface albedo changes: scaling from multiple sea-level reconstructions and parameterization of a simplified one- dimensional climate model. Third, we applied rigorous statistical methods and the tools of uncertainty analysis to the paleoclimatic records, including numerous sources of uncertainty in the final probability distribution. Lastly, we leveraged the strengths of the paleoclimate records and explored what would have to be different in our understanding of the past to support the low and high tails of current climate sensitivity probability distributions.
A21D-0204
Annular Mode Time Scales and Climate Sensitivity in the Intergovernmental Panel on Climate Change Fourth Assessment Report Models
A significant component of recent and future climate change consists of a poleward shift of the extratropical jets (and correspondingly, the storm tracks) and this shift projects strongly on the annular mode patterns of variability (e.g. Yin 2005, Miller et al. 2006, and Son et al. 2008). Recent work with idealized models has shown a possible fluctuation-dissipation relationship with the annular modes (Ring and Plumb 2008). Hence their temporal structure may be a gauge of climate sensitivity (Gerber et al. 2008). Here the ability of climate models in the Intergovernmental Panel on Climate Change Fourth Assessment Report (AR4) to capture the temporal structure of the annular modes is evaluated. Models capture the key qualitative features of the Northern and Southern Annular Modes: Northern Hemisphere time scales are shorter than those of the Southern Hemisphere and peak in boreal winter, while Southern Hemisphere time scales peak in austral spring and summer. Models, however, systematically overestimate the time scales, particularly in the Southern Hemisphere summer, where the multimodel ensemble average is approximately twice that of reanalyses. Fluctuation-dissipation theory suggests that long time scales in models could be associated with increased sensitivity to anthropogenic forcing. Comparison of model pairs with similar forcings but different annular mode time scales provides a hint of a fluctuation-dissipation relationship in the AR4 models.
A21D-0205
Comparison of Climate Feedbacks in Atmospheric Versus Coupled GCMs
Climate sensitivity is generally studied using two different types of models. Atmosphere-ocean general circulation models (AOGCMs, or coupled GCMs) include interactive ocean dynamics and detailed heat uptake, but they are computationally expensive to run. Atmospheric GCMs (AGCMs) with mixed-layer oceans cannot fully simulate the ocean's response to and influence on atmospheric changes. However, AGCMs require much less computer time and thus are often used to quantify and understand climate feedbacks and climate sensitivity. This work compares physical climate feedbacks (water vapor, lapse rate, surface albedo, and clouds) between coupled GCMs and atmospheric GCMs with mixed layer oceans from the World Climate Research Programme's (WCRP's) Coupled Model Intercomparison Project phase 3 (CMIP3) multi-model dataset. We use the new radiative kernel technique to calculate feedbacks. This technique decomposes each feedback into two parts: 1) the radiative effect of climate variable changes and 2) the response of these feedback variables to an imposed forcing. Since kernels are similar for different models, a single kernel can be used to estimate feedbacks for any model simulation. We find differences in feedbacks between coupled models and their corresponding atmospheric models with mixed layers. Some of these global-average feedback differences appear to be of the same sign for most GCM pairs, such as the water vapor and lapse rate feedbacks, while others, such as the albedo and cloud feedbacks, show differences of either sign. For some models, the choice of IPCC scenario is important. We also examine the zonal-average feedback patterns to better understand the differences. These results indicate that care should be taken when climate feedbacks and sensitivities from AGCMs are used to provide information about AOGCM feedbacks and sensitivities.
A21D-0206
Isotopic Fingerprint of Polar Sea Ice Extent in High Latitude Precipitation
One of the oldest theories of ice ages posits the crucial role of sea ice in controlling the precipitation necessary to support the growth of continental ice sheets. This mechanism would also affect albedo, thus providing a negative feedback to global climate change. However, data has been lacking to estimate the magnitude of these effects. Our analysis of two decades of monthly isotopic data provides the first such estimation. It is widely accepted that deuterium excess contains information about moisture source conditions, i.e., sea surface temperature and relative humidity. Based on this notion, we developed a method to use the observed deuterium excess from Vernadsky, Antarctica and Eureka, Canada to determine the amount of meteoric water that originated as vapor from the Bellingshausen and Amundsen Seas and the Baffin Sea, respectively. We examine the amount of evaporated polar sea water appearing in precipitation at both observatories, as determined by this method, and correlate these variables with sea ice extent in the respective seas. Our findings suggest that sea ice, by controlling the contribution of evaporation from polar seas to polar coastal precipitation, plays a significant role in global climate. Our results can be used to evaluate the role of sea ice in the looping chain of events that characterizes the Quaternary ice ages. They also provide a basis to evaluate the moderating effect of the temperature-sea ice-precipitation-albedo-temperature feedback on global climate change.
A21D-0207
Investigating Tropical Cyclone-Induced Feedbacks Using an Ocean General Circulation Model
It has been hypothesized that tropical cyclones actively contribute to the dynamics of the climate system by redistributing ocean heat through enhanced vertical mixing [Emanuel, 2001]. A recent observational study adds support to this idea, by providing evidence for a tropical cyclone-induced ocean heat pump [Sriver and Huber, 2007]. This heat pump appears to behave as a tropical thermostat, regulating tropical temperatures by pumping surface heat down into the thermocline, where it can be advected poleward by the Meridional Overturning Circulation. Since tropical cyclone variability is linked to changes in the mean climate state, this heat pump hypothesis provides for the possibility of feedbacks that are only beginning to be explored using climate models. Here we present new results from a sensitivity study using an ocean general circulation model, in which we analyze the model's response to varying levels of tropical cyclone surface wind forcing. Model results are compared against observational records for multiple spatial and temporal scales, and we examine potential feedbacks within the context of explaining past greenhouse climates, exhibiting small equator-to-pole temperature gradients.
A21D-0208
Carbon Cycle and Climate Sensitivity in an Earth System Model
The sensitivity of the potential feedbacks between climate and biogeochemical cycles (BGC) is adressed with the help of LOVECLIM, a global three-dimensional Earth system model of intermediate complexity. Key physical or biogeochemical parameters of LOVECLIM are varied within their range of uncertainty in order to provide an ensemble of parameter sets resulting in contrasted climate and global carbon cycle sensitivities. The selected climate parameter sets lead to a climate sensitivity ranging from 2 to 4°C and a reduction of the Atlantic meridional overturning circulation (MOC) ranging from 20 to 60% after 1 kyr in response to identical external forcings. The key parameters for the carbon cycle were chosen among those with the largest impact on the marine biogeochemical cycle and on the response of atmospheric CO2 to emission scenario. We then analyze the results of freshwater hosing experiments in which both the climate parameters and the BGC parameters are modified. These experiments allow to examine the impact of changes in climate sensitivity and of MOC reduction over the biogeochemical cycles as well as to assess the potential feedback from the carbon cycle onto the climate. A decreasing MOC directly impacts the ocean biogeochemistry. Most of the model setups show a decline in export production although some parameter sets yield reorganisation of the large scale ocean circulation, which leads to different behaviour of the ocean biogeochemistry. The atmospheric carbon is also affected by a decrease of the MOC. While most parameter sets cause a modest increase in atmospheric CO2, consecutive to the decrease of the continental vegetation, some model versions exhibit an amplification of the atmospheric CO2 response to the forcing. The mechanisms leading to the different responses for the different parameter sets are examined and discussed.
A21D-0209
The Strength of Snow-Albedo Feedback in GCMs Explains a Significant Fraction of the Spread in the Simulated Circulation Response to Climate Change
Climate change is expected to cause reductions in the amount and spatial extent of snow cover on land. Recent work suggests that this will exert a local influence on the atmosphere and the hydrology of snow- margin areas through the snow-albedo feedback (SAF) mechanism. A significant fraction of variability among IPCC AR4 general circulation model (GCM) predictions for future summertime climate change over these areas is related to the models' representation of springtime SAF. In this study, we present new work that demonstrates a nonlocal influence of SAF on the summertime large-scale circulation in the Northern Hemisphere. Our results show that increased land surface warming in models with stronger SAF is associated with large-scale sea-level pressure anomalies over the oceans and a poleward intensified subtropical jet. We demonstrate that 20-30% of the spread in GCM projections of changes in, for example, near-surface circulation patterns, is linearly related to SAF.
A21D-0210
Generalized Scalar Prediction: Predicting Trends in Arbitrary Climate Variables with Arbitrary Data Timeseries
Climate models persist in their uncertainty in transient sensitivity of the climate. Models' predictions of regional scale trends remain even more divergent. A variety of approaches have been presented intended to utilize ensembles of models to reduce uncertainties and to utilize global data sets to do the same. It has proven challenging to derive metrics that qualify models for their usefulness in prediction, and likewise it has proven challenging to extract information from global data sets appropriate to prediction of future trends in arbitrary climate variables. Using standard evidentiary inference together with an assumption of self-similarity in models' predictions of future trends---models show much better agreement in the patterns of trends than in transient sensitivity---we have derived the equations of generalized scalar prediction. It permits timeseries of arbitrary data types to be used in predicting trends in arbitrary variables. Some data types have little information content in trend prediction for some variables, so good judgment is necessary in choosing data types. We have found that regional surface air temperature change can be predicted 40 years into the future with a precision of 0.1 K given historical maps of surface air temperature. Testing the methodology with independent runs of climate models shows the methodology to be accurate. We will present the methodology of generalized scalar prediction and discuss its potential implementations and caveats.
A21D-0211
How well can we monitor cloud properties over polar regions in winter?
Understanding the impact of clouds on the Earth's radiation balance and detecting changes in the amount and distribution of global cloud cover requires accurate global cloud climatologies with well-characterized uncertainties. To meet this challenge, significant effort has been given to generating climate quality long-term cloud data sets using over 30 years of polar-orbiting satellite measurements [Rossow and Schiffer, 1999; Jacobowitz et al, 2003; Wylie and Menzel, 1999] with plans to continue the cloud record using the next generation of polar orbiting sensors [e.g. Ackerman, et al., 1998]. A "Climate Quality" climatology requires that both the uncertainties and the physical sensitivities are quantified and are smaller than the expected climate signature. Clouds play a critical role in the Arctic climate system, through interacting with other important climate processes, including snow/ice albedo feedback. Clouds modulate the surface radiative fluxes (Wang and Key, 2003) that influence the growth and melting of sea ice. Increasing cloud cover, which keeps the shortwave irradiances at the top-of-atmosphere unchanged, possibly compensates the reduced sea ice extent (Kato et al., 2006). However, assessing changes in polar conditions during winter has been a challenge. Holz et al (2008) presented a global two-month comparison between the Cloud-Aerosol Lidar with Orthogonal Polarization (CALIOP) and the Moderate Resolution Imaging Spectroradiometer (MODIS) cloud properties. Both CALIOP and MODIS are part of the NASA A-Train constellation of satellites and provide continuous near-coincident measurements that result in over 28 million cloud detection comparisons in a month. Globally (includes polar regions), it was found that the MODIS 1-km cloud mask and the CALIOP 1-km averaged layer product agreement is 88% for cloudy conditions in both August 2006 and February 2007. For clear-sky conditions the agreement is 84 (85) % for August (February). The best agreement is found for non-polar daytime conditions while the poorest agreement is in polar regions. The poor agreement during the polar wintertime leaves the following questions: "How capable are passive observing systems in retrieving cloud properties over the winter poles for climate studies?" Recent advances in active remote sensing technology have provided satellite-based lidar and radar measurements that allow us to fully answer this question. A combined data set of CALIOP and MODIS measurements provides a new opportunity to evaluate passive sensor cloud property retrievals over the entire polar region. To facilitate the comparison, a computationally efficient and accurate collocation methodology has been developed by Nagel and Holz (2008). MODIS has several infrared channels, three of which closely match the observations of the historical AVHRR instrument, thus the MODIS can be used to explore the capability of the AVHRR data record. With the collocated MODIS and CALIOP retrievals, nearly instantaneous comparisons are compiled regionally over both poles and are used to quantify our uncertainty in monitoring cloud conditions over the winter poles.
A21D-0212
An Assessment of Uncertainties in the NASA GISS ModelE GCM due to Variations in the Representation of Aerosol/Cloud Interactions
Aerosol indirect effects are known to have a significant impact on the evolution of the climate system. However, their representation via cloud/aerosol microphysics remains a major source of uncertainty in climate models. This study assesses uncertainties in the NASA Goddard Institute for Space Studies (GISS) ModelE global climate model produced by different representations of the cloud/aerosol interaction scheme. By varying the complexity of the cloud microphysics scheme included in the model and analyzing the range of results against cloud properties obtained from satellite retrievals, we evaluate the effect of the different schemes on climate. We examine four sets of simulations with the GISS ModelE: (1) using a new aerosol/cloud microphysics package implemented in ModelE (based on the two-moment cloud microphysics scheme recently implemented in CCSM), (2) using a version of the microphysics scheme previously included in ModelE, (3) using prescribed aerosol concentrations and fixed cloud droplet number (the main link between aerosols and the cloud microphysics scheme), and (4) varying the environment conditions with which the new aerosol/cloud microphysics package is run. The global mean cloud properties are analyzed and compared to global mean ranges as obtained from satellite retrievals. Results show that important climate parameters, such as total cloud cover, can be underestimated by 8-15% using the new aerosol/cloud microphysics scheme. Liquid water path (LWP) is particularly affected by variations to the aerosol/cloud microphysics representation, exhibiting both global mean variations of ~20% and strong regional differences. Significant variability in LWP between the various simulations may be attributed to differences in the autoconversion scheme used in the differing representations of aerosol/cloud interactions. These LWP differences significantly affect radiative parameters, such as cloud optical depth and net cloud forcing (used to evaluate the aerosol indirect effect), demonstrating the importance of fine-tuning autoconversion schemes to better represent aerosol effects on cloud properties and climate. Variations to environment conditions, meanwhile, had little effect on the climate simulated by the model.
A21D-0213
Investigating the Sensitivity of Model Intraseasonal Variability to Minimum Entrainment
Previous studies have shown that using a Relaxed Arakawa-Schubert (RAS) convective parameterization with appropriate convective triggers and assumptions about rain re-evaporation produces realistic intraseasonal variability. RAS represents convection with an ensemble of clouds detraining at different heights, each with different entrainment rate, the highest clouds having the lowest entrainment rates. If tropospheric temperature gradients are weak and boundary layer moist static energy is relatively constant, then by limiting the minimum entrainment rate deep convection is suppressed in the presence of dry tropospheric air. This allows moist static energy to accumulate and be discharged during strong intraseasonal convective events, which is consistent with the discharge/recharge paradigm. This study will examine the sensitivity of intra-seasonal variability to changes in minimum entrainment rate in the NCAR-CAM3 with the RAS scheme. Simulations using several minimum entrainment rate thresholds will be investigated. A frequency-wavenumber analysis will show the improvement of the MJO signal as minimum entrainment rate is increased. The spatial and vertical structure of MJO-like disturbances will be examined, including an analysis of the time evolution of vertical humidity distribution for each simulation. Simulated results will be compared to observed MJO events in NCEP-1 reanalysis and CMAP precipitation.
A21D-0214
Explorin the Ocean-Atmosphere Interaction in Observed Surface Heat Flux Datasets and Coupled Model Forecasts for Recent Decades
In regions of the Tropical Oceans where upwelling is not important, the dominant forcing of interannual variability is due to the surface heat flux. In order to simulate realistic interannual variability coupled models need to accurately represent the surface heat flux and its interaction with the evolving sea-surface temperature (SST) field. Based on order of magnitude the two dominant surface heat fluxes in the tropics are the solar and latent heat flux. The latter together with the sensible heat flux are due to turbulent processes and hence are known as turbulent fluxes. Here, we compare the turbulent heat fluxes from several observed products with those simulated by two seasonal forecasting systems. The local relationship between SST and turbulent heat fluxes is examined using the feedback parameter as defined by Frankignoul et al (1998). The observed datasets including NCEP/NCAR Reanalysis2 (Kanamitsu et al., 2002), GSSTF2 (Chou et al., 2003), ERA40 (Uppala et al., 2005), OAflux (Yu and Weller, 2007), and HOAPS3 (Andersson et al., 2007) are investigated. This analysis covers the period of 1988-2000. Some notable disagreements are found among these different observational datasets. In the South and the North Subtropical Atlantic, the satellite datasets, GSSTF2 and HOAPS3 indicate that the negative feedback in February-March-April season may be two times as strong as what the reanalysis datasets indicate. In the Subtropical Western North Pacific in the northern summer and fall, relatively weak negative or positive feedback is indicated by both satellite datasets, while the strong negative feedback is indicated by the other datasets. Strong negative feedback in the South Indian Ocean in May-June-July season is indicated by the two satellite datasets, while the other datasets indicate a negative feedback that is less than half as strong. The feedback parameter from two coupled forecast systems, the NCEP CFS (Saha et al., 2006) CGCM and a model based on the ECHAM4.5 AGCM (Roeker et al., 1996) coupled to a mixed layer ocean with Q-flux climatology are examined and compared with the observed datasets. The negative feedback parameters found in the CFS retrospective forecasts tend to be overestimated in the various regions and seasons. This tendency to overestimate the strength of the feedback parameter is found to be most severe around Southeastern marginal seas of Asia in May-June-July and August-September-October seasons and east of Australia in January-February-March. The feedback parameter estimated from the retrospective forecasts by ECHAM4.5 coupled to a mixed layer ocean agrees with the observed estimates, except that the negative feedback parameters are underestimated in regions where strong thermal interaction is expected during winter months.
A21D-0215
Tropical intraseasonal variability in global climate models
This study examines the tropical intraseasonal variability in global climate models participating the IPCC Fourth Assessment Report (AR4). The modes studied include the Madden-Julian Oscillation, convectively coupled equatorial waves, and subseasonal modes associated with Asian monsoon, North American monsoon, Souther American monsoon, and western United States winter precipitation. We will also present some GCM experiments on improving the simulated tropical intraseasonal variability.
A21D-0216
Dependency of effective climate sensitivity on the parametrization of land evaporation.
Equilibrium and transient responses of climate models to changes in external forcing are defined by equilibrium and effective sensitivities respectively. Published results show that climate models with similar equilibrium sensitivities may have rather different effective sensitivities and vice versa. Reasons for such differences are not well understood. Here we explore relationship between effective and equilibrium sensitivities for two modifications of a climate model of intermediate complexity (MIT IGSM) employing different land models. One version uses two layer bucket model adopted from the GISS AOGSM (IGSM1a), the other uses Community Land Model version 2 (IGSM2.2). Versions of the IGSM1a and IGSM2.2 with similar equilibrium climate sensitivities exhibit different transient response to an external forcing. In the simulations with an instantaneous doubling of CO2 concentration IGSM2.2 requires much longer to reach equilibrium especially for high climate sensitivity. Analysis of changes in surface fluxes reviled much smaller rate of an increase in the land evaporation per degree of surface warming in the IGSM2.2 compare to the IGSM1a, as well as difference in the sing of changes in sensible heat flux. These differences lead to differences in low level relative humidity and cloud and, as a result, in the strength of cloud feedback. Differences in the response to an external forcing between two versions of the IGSM result in different relationships between equilibrium and effective sensitivities. In the version using two layer surface scheme effective and equilibrium sensitivities are almost identical except for the very high values of sensitivity. In contrast, effective sensitivity of the IGSM2.2 is noticeably smaller than equilibrium sensitivity for values of Seq greater than 5K. Strength of feedbacks at the time of CO2 doubling in simulations with 1% per year increase in CO2 concentration is about 80% and 95% percent of their strength at the equilibrium with doubled CO2 concentration in simulations with the IGSM2 and IGSM1a, respectively.
A21D-0217
Climate Feedback Analysis of IPCC AR4 Global Warming Simulations
We here present climate feedback analysis of IPCC AR4 global warming simulations using the new feedback analysis method developed in Lu and Cai (2008). The goal is to quantify the contributions to the 3-D global warming pattern from the external forcing alone, and from subsequent feedback processes including water vapor and snow/albedo feedbacks, surface sensible and latent heat flux feedbacks, and large-scale scale dynamical feedbacks. We first exam the 3-D profiles of the radiative energy flux perturbations due to the external forcing and due to water vapor feedbacks, and the non-radiative energy flux perturbations due to latent heat flux feedbacks, and large-scale scale dynamical feedbacks. We then calculate the partial temperature changes associated with each of these individual energy flux perturbations. These partial temperature changes are additive and their sum is compared with the temperature change produced by the original IPCC AR4 global warming simulations.
A21D-0218
Polar amplification dynamics: results from an idealized model, reanalysis products and observations
An aquaplanet atmospheric GCM coupled to a mixed layer ocean is analyzed in terms of its polar amplified response to a 2xCO2-like forcing and in terms of phase space trajectory of the relaxation of a free perturbation to equilibrium. In earlier studies concerned with linear stability and fluctuation-dissipation analysis of the same system we have shown that the least stable mode of the linearized operaor of the system has a polar amplified shape. We demonstrate that this shape of the least stable mode is responsible for the polar amplified shape of the response to a uniform forcing and for the manner in which the system relaxes back to equilibrium. We use reanalysis products (ERA-40, NCEP) in order to quantify the signal in the Arctic atmosphere due to the influence of lateral heat transport from the lower latitudes. Robustness of polar amplification obtained from the reanalysis products is tested against radiosonde data from dataset IGRA.
A21D-0219 TI: The observed forcing and response to the 11-year solar cycle provide a critical constraint on climate model performance. Previous GCM simulations mostly involved prescribed sea-surface temperatures, which prevented feedback processes, such as water-vapor feedback, ice-albedo feedback etc. from taking place. We have performed a multi-century simulation using a periodic solar-cycle forcing at 11-year period, which allows the identification of the surface temperature anomaly at the same period as the response. Since this model couples the atmosphere and the ocean, there is a heat flux into the ocean at solar max and vice versa at solar min. This mechanism is responsible for causing a lag of the surface temperature response relative to forcing in both the model and the observation. By comparing the lag in the model with that in the observation, we conclude that the current generation of AOGCMs may have somewhat larger heat fluxes into the ocean during periods of transient heating, as compared to reality . This result has implications on the model predicted transient warming due to the increasing greenhouse gases.
A21D-0220
Capacity of Desert Soils As Organic Carbon Sink Under Elevated Carbon Dioxide
The Nevada Desert FACE Facility project was completed after 10 years exposure to elevated carbon dioxide. Soils exposed to elevated carbon dioxide had 32.5% more organic C (12800 kg/ha vs. 9600 kg/ha) but only 4.0% more N (1400 kg/ha vs. 1300 kg/ha) than ambient controls. The magnitude of increase exhibited large spatial variability and ranged from 120% under Lycium andersonii, a deciduous C3 shrub, to none under Pleuraphis rigida, a C4 grass in surface soils. All carbon accumulation occurred in the upper 0.2m. We conclude that desert soils are likely to be strong organic carbon sinks under elevated carbon dioxide.