National Aeronautics and Space Administration

Date

Speaker

Title

Abstract:

W. G. Read, T. Flury, H. Su, M. L. Santee, and N. J. Livesey: Since May 2008, the measurement tracks of the CALIPSO cloud profiling LIDAR and MLS constituent measurements of atmospheric trace grases (e.g. H2O) and temperature are nearly aligned. The cloudiness of all the MLS H2O measurements is known which provides insights into the processes by which dehydration is occurring. CALIPSO also determines whether the clouds are of convective origin or isolated thin cirrus. We find, particulary during the Boreal winter, that a high percentage of the driest and coldest air occurs in convective and thick layered cirrus clouds situated above the nominal clear sky level of zero radiative heating (~120~hPa, LZH). The Boreal summer shows fewer such events and hence the height-of-convection shows a strong annual cycle. Another interesting discovery is that over the tropical western Pacific, total H2O (IWC and humidity) at the tropopause has nearly no annual cycle. The mean zonal IWC at altitudes below the LZH shows a weak annual cycle. We will also show some initial results from CH3Cl, a new MLS v3 product. CH3Cl, like CO, is enhanced in biomass burning activities, but has a much longer chemical lifetime in the TTL than CO. Because of this, CH3Cl will be a better tracer of convective transport and supply into the TTL than CO. Preliminary results from measurements and a comparison with a simple 2 D model will be shown.
« Hide Abstract

Abstract:

Recent land surface developments at ECMWF have led to an improved representation of some of the physical processes occurring at the land-atmosphere interface, verified against a variety of independent observational sources. In particular, a satellite-based leaf-area-index climatology, which describes the seasonal evolution of vegetation, has replaced a fixed-in-time vegetation, and a revised bare-soil evaporation has introduced a larger extraction of superficial water in non-vegetated areas. These two schemes revisions are shown to improve near-surface temperatures and soil moisture simulations. In an attempt of moving towards interactive ecosystems, a photosynthesis-based module has also been introduced in order to simulate natural CO2 emissions over land. However, the land-carbon parameterization does not yet interact with the evapotranspiration formulation, and similarly, the vegetation seasonality representation does not interact with the momentum budget. These "ad-hoc" separations of processes have a practical advantage of modularizing the model development (particularly useful in community models) but may present some caveats of realism when representing naturally inter-dependent processes occurring in the Earth system. Examples from recent simulations will be used to illustrate this paradigm and the problems associated to full coupling between processes.
« Hide Abstract

Abstract:

This study aims to explore the dynamical seasonal predictability of tropical storms using high-resolution global climate models. Using a series of 5-member ensemble predictions using the 50-km resolution Goddard Earth Observing System version 5 (GEOS-5) atmospheric general circulation model, which was driven by the observed sea surface temperature (SST) for the period of 1997-2007, we compared the long-term climatology and the interannual variation of the tropical storm activity with the observation database from the International Best Track Archive for Climate Stewardship (IBTrACS). The analysis was specifically focused on the domains in the Northwest Pacific and the Atlantic Oceans. The model in general tends to reproduce realistic statistics of the annnual and interannual variation of the observed tropical storm activity, although it is not able to reproduce the tropical cyclones of category 4 and higher, presumably due to the limitations from the model resolution. Climatological-mean distribution of the tropical storm genesis location and tracks are also realistic, although the model tends to simulate less recurving storms evolving toward the extratropics. We attribute the model bias to the abnormally strong subtropical highs simulated by the model, which results emphasize the important role of the model's simulated basic climate state. The model shows a marginal prediction skill for the yearly tropical storm counts, with a correlation below 0.5 in the North Atlantic basin with the observation, implying that there should be much room to improve the model simulation by improving moist physics parameterization and the initialization of land surface.
« Hide Abstract

Abstract:

An idealized experiment examined the impacts of entrainment in a coupled ecosystem-atmosphere model by implementing an enhanced entrainment parameterization based on the assumption that the heat flux at the top of the PBL is negatively proportional to the heat flux at the surface. This experiment found that entrainment produced a warmer, drier, and deeper PBL and that the surface fluxes of heat and moisture were modified by the vegetative response to the altered atmospheric conditions. A realistic simulation for the summer of 1999 found that enhanced entrainment produced stronger early morning growth of the PBL and a deeper midday depth. This better captured the monthly mean diurnal cycle of PBL depth from observations by a radar-sounding system in northern Wisconsin. Additionally, the complex land-atmosphere interactions produced a time-mean spatial CO2 gradient of 7 ppm over 1000 km. In order to evaluate and improve model simulations, PBL depth has been estimated using the backscatter from the LIDAR onboard the Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observation (CALIPSO) satellite. Using an automated method, millions of estimates have been derived to which model results can be compared. This method evaluates the maximum vertical variance of the backscatter in order to identify backscatter features associated with the top of the PBL and helps to identify the vertical extent of turbulent mixing. This analysis sheds some light on the spatial heterogeneity of boundary layer processes. The derived depths are shallower over water than over land and show a local minimum along the Mississippi River valley. Deeper features are found over the desert Southwest and deeper than expected values are retrieved over the Boreal forests.
« Hide Abstract

Abstract:

Previous studies have shown that there is stronger isentropic mixing between stratosphere and troposphere in summer than in winter, leading to a strong seasonal variation of isentropic stratosphere-troposphere exchange (STE). However, at the extratropical tropopause there are large dynamical and chemical discontinuities, demonstrated by strong quasi-horizontal gradients in potential vorticity (PV) and chemical species on isentropic surfaces. In this study, we introduce a new meridional coordinate �tropopause latitude� (TpLat), defined as the latitudinal distance from the nearest tropopause on an isentropic surface, to investigate the discontinuity feature of the extratropical tropopause. Our analysis shows that the TpLat coordinate is much better than traditional geographic or equivalent latitude coordinates for emphasizing the sharp distinction in tracer abundances between the upper troposphere and lower stratosphere in large global datasets, such as space-borne measurements. Our study also shows that meridional gradients in GEOS-5 PV and Aura/MLS tracer abundances in summer are as large as those in winter. A definition of the tropopause layer boundaries based on PV distributions in TpLat coordinates is proposed and the statistical distributions of features inside the extratropical tropopause layer will also be shown. To summarize, although there is stronger STE inside the tropopause layer in summer than in winter, the distinction between the upper troposphere and lower stratosphere is nearly equally pronounced in winter and summer.
« Hide Abstract

Abstract:

This presentation introduces an approach to estimate system (model) error using observation residuals. Based on the sequential fixed-lag smoother, a diagnostic procedure is developed to allow estimating system error over a dense observing system. Optimality considerations are examined in light of the sequential results. The procedure is re-interpreted in the language of variational data assimilation, such as 4d-Var. Illustrations of the approach are given by studying both identical-twin and fraternal-twin experimental settings for simple linear and nonlinear dynamics. Preliminary results are shown by examining observation residual statistics.
« Hide Abstract

Abstract:

In this presentation a new approach to the estimation of key unknown parameters of water and energy balance equation (moisture and heat diffusion equation) and the closure function which links these two equations is introduced. Parameters of the system are estimated by developing objective functions that link atmospheric forcing (precipitation and incident radiation), surface state and unknown parameters. This approach is based on conditional averaging of heat and moisture diffusion equations on land surface temperature and moisture states respectively. Based on conditional averaging, a single objective function is posed that measures the moisture and temperature dependent errors solely in terms of observed forcings (e.g. precipitation, radiation) and surface states (moisture and temperature). This objective function can be minimized with respect to parameters to identify evaporation and drainage models and estimate water and energy balance flux components. The uncertainty of the estimated parameters (and associated statistical confidence limits) is obtained through the inverse of Hessian of the objective function which is an approximation of the covariance matrix. The accuracy of this method at point scale is examined through the use of synthetic data and various field data obtained from Ameriflux network. The proposed methodology is applied to the arid sahara-sahelian climate of the mesoscale site of Gourma in West Africa. Multi-platform and frequency remote sensing data is used to obtain Evaporative Fraction as a function of soil moisture (EF(s)), neutral turbulent heat Coefficient (CHN) as a function of vegetation phenology and Drainage as a function of soil moisture using the proposed methodology. The estimation results are verified against Agoufa field site data set located in this region and the hydrological characteristics of the sahara-sahelian climate of Gourma region in West Africa. The results of this research project, demonstrate the feasibility of this new scale free technique in providing a good estimate for water and energy balance flux components and their closure relationship, at point scale and at regional scale using remote sensing data.
« Hide Abstract

Abstract:

Speaker: Masahiro Watanabe (AORI, Univ of Tokyo) *** Title: Development, verification, and natural variability in MIROC5 *** Abstract: One of the Japanese community climate model, called Model for Interdisciplinary Research on Climate (MIROC), has been developed for use of various aspects of climate research. We are currently under operation of MIROC with three types of configuration: a high-resolution climate model for the decadal climate prediction, a low-resolution Earth system model, and a low-resolution new version of the climate model (MIROC5). In this talk, I would like to introduce outcomes from MIROC5, in which many of the atmospheric parameterization schemes have been updated. Improvements in the climate simulation with MIROC5 and a high sensitivity of the ENSO simulation to a perturbation in the cumulus convection scheme are remarked. ***************** Speaker: Yoshimitsu Chikamoto (AORI, Univ of Tokyo) *** Title: Overview of decadal climate prediction using a coupled climate model MIROC *** Abstract: Toward coming IPCC AR5, we have performed decadal climate prediction using three versions of the coupled atmosphere-ocean model MIROC. In these hindcast experiments, initial conditions were obtained with an anomaly assimilation procedure using the observed oceanic temperature and salinity while prescribing natural and anthropogenic forcing based on the IPCC emission scenarios. Our hindcast experiments show that initialization contributes to enhance the predictive skills of AMO and PDO indices for several years in advance. In addition to these major climate phenomena, the hindcast experiments tend to simulate a pattern of stepwise sea surface temperature increase in the Pacific during the late 1990s. Although further studies are need to enhance predictive skills of decadal climate variability, our results suggest that the decadal climate prediction has a potential to provide useful information in order to solve socioeconomic problems arising from climate change.
« Hide Abstract

Abstract:

The interannual and the intraseasonal modes of the South Asian summer monsoon are identified in a long free coupled simulation of the National Centers for the Environmental Prediction�s Climate Forecast System version-1 (CFS) by applying multi-channel singular spectrum analysis on the daily anomalies of rainfall. The CFS has two intraseasonal oscillations with periods around 106 and 30 days. The 106-day mode has spatial structure and propagation features similar to the northeastward propagating 45-day mode in the observations. The 30-day mode is northwestward propagating and is similar to its observational counterpart. The dominant interannual signal in the rainfall is related to the El Niño�Southern Oscillation (ENSO), which has maximum variance in the eastern equatorial Indian Ocean. Although an Indian Ocean Dipole mode is not obtained as a separate mode, a dipole structure in the Indian Ocean is almost always accompanied with an ENSO. Regionally coupled experiments are performed to examine the sensitivity of these modes to the air-sea coupling in the Indian and Pacific basins. Regional coupling is employed by prescribing daily mean or climatological sea surface temperature either in the Indian or the Pacific basin while allowing full coupling elsewhere. Results of these sensitivity experiments will be discussed in detail.
« Hide Abstract

Abstract:

Climate feedbacks constitute the largest part of the Earth System response to climate forcings, including the forcing from anthropogenic greenhouse gases. At the same time, feedback mechanisms are the hardest ones to evaluate in climate simulations, as they involve complex interactions between numerous components of the climate system. As a result, feedbacks like those from clouds and water vapor constitute the primary source of uncertainty in climate model projections. Over the past three decades, a wealth of global observations has been collected that, even if they do not contain a sufficiently long climate warming signal, they allow us to resolve the operation of feedback mechanisms for a large subset of time and space scales. In this talk, different data analysis methods are presented that aim to document and study feedback processes in observational data, and to use the results of the analysis in order to evaluate the ability of models to simulate those processes. The examples presented here focus on cloud, radiation, and precipitation processes, but the analysis methods define dynamic regimes that can be used to examine processes in a number of other climate system components.
« Hide Abstract

Abstract:

A new component of the next IPCC assessment report will be decadal time-scale predictions in which models are initialized with states based on the observed state of the climate system. Since the climate system is chaotic, there is an inherent limit on the range at which the information in these initial states can have an impact on the skill of the forecasts. In this presentation we measure this limit for six CGCMs from various research centers. We concentrate on the predictability of upper ocean heat content in the two northern basins, and we use two methods that only require long control runs for estimating predictability limits. One method makes use of analogs and the other uses multivariate linear regression. In contrast to the conventional ensemble technique both methods are able to estimate the average predictability characteristics of very many initial states. When we use relative entropy as a measure of predictability and consider entire basins, we find that on average the effect of initializing a forecast from a specific initial condition can be detected for about a decade, but this limit can vary by as much as a factor of three from one model to another. Furthermore, for a given model, there are variations in predictability of a factor of four at different locations within a basin. The model-to-model variations can be traced to variations in the properties of horizontally propagating disturbances in each model, including prominent modes. Given the large variations in predictability that exist from one model to another, we conclude that a) the predictability characteristics of each model used for decadal predictions must be carefully determined for proper design and interpretation of forecasts, and b) the scientific community currently does not have a reliable estimate of the decadal predictability characteristics of nature.
« Hide Abstract

Abstract:

The Modern-Era Retrospective Analysis for Research and Applications (MERRA) is used to examine the physical mechanisms associated with the development of summer-time climate extremes including heat waves and dry spells that typically occur on weekly to monthly time scales. The results indicate an important role for stationary Rossby waves that can at times span much of the summer hemisphere. Examples are presented that show such waves have played an important role in past major climate extremes such as the 2003 European and 2010 Russian heat waves, and the 1988 US drought. A stationary wave model is used to show that sub-monthly atmospheric transients play a key role in forcing the waves. Some early results from GEOS-5 AGCM experiments will be presented that suggest important roles for SST and land forcing in the development/maintenance of the Russian heat wave as well as the recent (2011) heat wave over the United States.
« Hide Abstract

Abstract:

This presentation will give an overview of the global and mesoscale data assimilation systems developed at the Naval Research Laboratory and run operationally for the U.S. Navy. These systems include the mesoscale three-dimensional variational system NAVDAS (NRL Atmospheric Variational Data Assimilation System) and its global 4D-Var counterpart NAVDAS-AR (for Accelerated Representer). A summary of recent improvements for each system (e.g., new capabilities and newly assimilated observing platforms) will be presented first, followed by a review of ongoing data assimilation research and development activities. Topics to be discussed include preparations for NPP, improving the assimilation of existing sensor platforms, using adjoint methods to estimate mesoscale observation impact, development of mesoscale 4D-Var, and incorporating ensembles DA components into NAVDAS-AR. The talk will conclude the talk with a few words about the Navy
« Hide Abstract

Abstract:

The NASA Moderate Resolution Imaging Spectroradiometer (MODIS), launched onboard the Terra and Aqua spacecrafts, began Earth observations on February 24, 2000 and June 24, 2002, respectively. Among the algorithms developed and applied to this sensor, a suite of cloud products includes cloud masking/detection, cloud-top properties (temperature, pressure), and optical properties (optical thickness, effective particle radius, water path, and thermodynamic phase). All cloud algorithms underwent numerous changes and enhancements for the latest Collection 5 production version; this process continues with the current Collection 6 development. We will show example MODIS Collection 5 cloud climatologies derived from global spatial and temporal aggregations provided in the archived gridded Level-3 MODIS atmosphere team product (product names MOD08 and MYD08 for MODIS Terra and Aqua, respectively). Data sets in this Level-3 product include scalar statistics as well as 1- and 2-D histograms of many cloud properties, allowing for higher order information and correlation studies. To assist in climate model evaluation, we have developed a MODIS cloud simulator with an accompanying netCDF file containing subsetted monthly Level-3 statistical data sets that correspond to the simulator output. We will also show analyses of trends and statistical significance in monthly mean anomalies for a variety of the MODIS cloud and other atmospheric properties, as well as the time required for detection given assumed trends. Correlations of cloud properties with ENSO offer the potential to evaluate model cloud sensitivities as well as provide guidance in interpreting trending results.
« Hide Abstract

Abstract:

The Earth's most powerful year-to-year climate fluctuation -- the El Niño / Southern Oscillation (ENSO) -- affects weather, ecosystems, and economies worldwide. Yet despite great progress towards understanding and predicting ENSO, its future remains uncertain. Will coming decades bring a barrage of strong El Niño and La Niña events, or could there be no events at all? Historical reconstructions, paleo records, and model simulations all display prolonged epochs of active or quiet ENSO, which challenge ENSO theories, hamper detection of anthropogenic impacts, and complicate model evaluation and intercomparison. I shall discuss these challenges, and describe recent efforts to probe the limits of ENSO predictability and its sensitivity to climate change.
« Hide Abstract

Abstract:

Small changes in the surface layer parameterization over oceans leads to a widespread decrease in model biases in the GEOS-5 AGCM. Surface winds, stationary waves, momentum fluxes, heat fluxes, and cloud distribution, are all improved relative to a control run. Fidelity to observations is improved in a 2x2.5 degree run without stratospheric chemistry, in a 2x2.5 degree run with stratospheric chemistry, in a 1x1.25 degree run, and in a series of 0.25 degree forecasts. It appears that other GCMs (both for operational weather forecasting and climate) use a similar class of parameterization; we therefore expect that results from GEOS-5 are relevant to other models as well.
« Hide Abstract

Abstract:

The difficulty of evaluating large scale soil moisture fields presents a major obstacle to the development and application of remotely sensed near-surface soil moisture data sets. Traditionally, remotely sensed soil moisture observations have been assessed against in situ soil moisture observations, however these in situ observations are available at only a few locations, and their utility is greatly limited by the substantial representivity errors between the soil moisture quantities observed by remote and in situ sensors. Consequently, this study pursues an alternative approach to evaluating near-surface soil moisture observations derived from ASCAT, by examining the impact of assimilating those data into the SAFRAN-ISBA-MODCOU (SIM) hydrological modelling suite over France. Specifically, the root-zone soil moisture in the ISBA land surface model has been analysed over a three and a half year period, by assimilating the 12.5 km ASCAT near-surface soil moisture observations disseminated by EUMETSAT (the SM OBS1 product), using an Extended Kalman Filter. For these assimilation experiments ISBA has been forced with the near-real time SAFRAN analysis, which analyses the variables required to force ISBA from all relevant observations available within the real time data cut-off period. The soil moisture profile and surface water budget terms generated from the ASCAT assimilation have then been compared to the corresponding forecasts from the ISBA model forced with the climatological SAFRAN analysis, which ingests additional observations that become available after the near real-time data cut-off (including data from 3000 climatological observing stations). This comparison tests whether assimilating the ASCAT soil moisture observations can correct the ISBA soil moisture profile in response to errors in the near-real time SAFRAN forcing, since this provides strong evidence that the ASCAT data set can accurately detect temporal variations in near-surface soil moisture. Likewise, the impact of assimilating the ASCAT near-surface soil moisture observations on the river discharge simulated by the MODCOU river routing scheme has been evaluated, by comparison to discharge observations from an extensive network of gauging stations throughout France.
« Hide Abstract

Abstract:

As part of the Interim Synthesis effort for the North American Carbon Program, an observational dataset of surface carbon, water, and energy fluxes and associated ecological, biological, and physical data has been developed, drawing on measurements made at over thirty eddy covariance flux tower sites in the U.S and Canada. Extensive data processing and quality control measures have been taken to define observational uncertainty and to provide gap-filled surface weather forcing at sub-daily time steps, suitable for input to ecosystem and land surface models. Based on these observational resources, a simulation protocol has been developed that defines a series of progressively sophisticated modeling experiments, designed for the purpose of drawing quantitative comparisons between observed and predicted fluxes and states across many sites. Following this protocol, we used the Community Land Model, version 4 (CLM4), which includes prognostic carbon and nitrogen cycles, to predict fluxes and state variables at 15 forested sites in diverse climates. We focus here on a comparison of predicted vs. observed diurnal cycle of carbon, water, and energy fluxes at different sites and in different seasons. We found a serious and pervasoive underprediction bias in the CLM4 mid-day carbon fluxes, while water and energy fluxes were more reasonable. We identified the treatment of time scale for growth limitations from nutrient availability as a likely source of this bias, and developed a simple theory that helps to explain and correct the discrepancy with observations. This theory was developed and implemented at one site, then tested across all 15 forested sites. We found a quantitative improvement in model performance across sites, expressed as a reduction in prediction bias and RMSE. This work demonstrates the utility of the site synthesis approach, and also illustrates a major principle of model design in action: the mechanistic representation of the model should be as simple as possible, with additional complexity added as dictated by observational constraints and consistency with understanding of the natural system.
« Hide Abstract

Abstract:

Global reanalysis is an invaluable asset to study weather and climate, and it has been extensively used to understand the pronounced atmospheric variability and improve its prediction using global climate models. As a part of coordinated efforts in examining the realism and limitations of the recently-released, Modern Era Retrospective-analysis for Research and Applications (MERRA) atmospheric reanalysis by NASA Global Modeling and Assimilation Office, this study validates two phenomena: the tropical Madden-Julian oscillation (MJO) and the tropical storm. Representation characteristics of these two phenomena seem to be closely related with the parameterization of moist convection, which tends to produce a large spread in the simulation across various reanalysis products. Results presented are based on the multi-year data comparison between MERRA and observations, as well as with various sources of reanalysis products. It is found that MERRA represents the MJO in a realistic manner in terms of its amplitude and the dominant spatial and temporal scale. Simulations of convection and accompanied precipitation are compared reasonably well with those of CFSR and ERAinterim, and the simulation is relatively better than the old reanalyses from NCEP/NCAR and NCEP R2. We discuss the possible causes for the improvement based on a better representation of the observed relationship between moisture variability and deep convection associated with the MJO. We also present the validation results from the tropical storm simulations over the North Atlantic and the Northwestern Pacific for recent 12 years (1998-2009). Climatological-mean features of the genesis location, tracks and their maximum intensity will be presented with the observed tropical storm characteristics obtained from the International Best Track Archive for Climate Stewardship (IBTrACS).
« Hide Abstract

Abstract:

The speaker will present his recent work on the diagnosis and correction of the excessive summertime precipitation over steep mountains such as the Andes and the Himalayas in the GEOS-5 GCM. This systematic error is common to all global and regional atmospheric models and it impairs data assimilation products. Possible causes investigated include a lack of upward heat flux out of the boundary layer due to sub-grid scale vertical circulations in the affected regions; poor design of horizontal moisture flux; the criterion for cumulus convection being too easily met over high terrain; conditional instability of the computational kind; and insufficient friction in the boundary layer on steep mountain slopes. Of these possible causes the first two are the significant ones. Correction methods and some preliminary results will be presented.
« Hide Abstract

Abstract:

The climatological spatial structure and variability of the African Easterly Jet and Waves (AEJ/AEWs) originating from West Africa are analyzed in the recently completed Modern Era Retrospective-Analysis for Research and Applications (MERRA). The results are compared with other existing reanalyses such as the ERA-40, NCEP-R2, and JRA-25. The basic mechanisms (e.g., barotropic/baroclinic instability) responsible for the generation of the AEWs and how they vary with time scale, as well as the factors controlling the variability of the AEJ and AEWs are examined using a compositing approach that highlights the role of ENSO and the MJO.
« Hide Abstract

Abstract:

A global OSSE has been developed at the Earth Systems Research Laboratory (ESRL) in support of the National Oceanic and Atmospheric Administration (NOAA) UAS Program. The project is part of a multi-institute collaborative Joint OSSE, including a Nature Run provided by the European Centre for Medium Range Weather Forecasting (ECMWF) and synthetic observations generated by the National Aeronautics and Space Administration (NASA) Global Modeling and Assimilation Office (GMAO) and National Centers for Environmental Prediction (NCEP). Using the NCEP Global Forecast System (GFS) as the forecast model, the OSSE has been calibrated through extensive data denial experiments. The impact of observations on the analysis field was used as a metric for calibration of the synthetic observation error for conventional data types. Preliminary case studies of the potential impact of high-altitude UAS-deployed dropsondes on hurricane track forecasts in the Atlantic basin have been performed. An overview of the calibration method will be given, as well as discussion of the OSSE performance and case study results.
« Hide Abstract

Abstract:

In this talk, I will briefly share my personal journey in interdisciplinary research from atmosphere, to land, to ocean, and back to atmosphere. Examples include: using atmospheric hydrostatic subtraction idea to improve the numerical solution of soil moisture Richards equation (implemented in NCAR CESM); using soil temperature diurnal cycle idea to develop the ocean surface skin temperature prognostic scheme (implemented in ECMWF model); and using the ocean-atmosphere turbulence consistency idea to improve turbulence over land (implemented in NCEP model). Furthermore, I will briefly discuss our most recent work on the evaluation of ocean surface turbulent fluxes from 11 flux products (including MERRA and GSSTF2/2b from GSFC) using cruise observations and on the comparison of land surface turbulent fluxes and mean quantities from various reanalyses (including MERRA and GLDAS from GSFC) using 33 flux tower measurements.
« Hide Abstract

Abstract:

Ocean circulation off the west coast of Australia is dominated by the anomalous, poleward-flowing eastern boundary current, the Leeuwin Current. This current flows adjacent to Ningaloo Reef, Australia�s longest fringing coral reef. Using chlorophyll a from satellite-derived ocean colour and in situ field observations, we identified the existence of an autumn bloom in this region and evaluated the physical mechanisms that influence the seasonal variability in chlorophyll a concentration. This involved combining historical field data sets, satellite-derived ocean colour observations and output from a data-assimilating numerical ocean model. The results show that the mixed layer depth deepens to ~100 m in winter and is driven by a combination of surface cooling and the strengthening of the Leeuwin Current. This coincides with an increase in nutrient concentrations and the observed autumn bloom. In the second part of this talk, I will give an overview of the project that I will undertake next year with Watson Gregg. This project will use the NASA Ocean Biogeochemical Model to analyse the spatial and temporal variation of cyanobacteria in the Pacific Ocean. Our preliminary analysis has shown a decrease of cyanobacteria concentration between 1998 and 2007 in the North Central Pacific and to a lesser extent in the Equatorial Pacific. The largest decrease in cyanobacteria concentration was observed in the North Central Pacific and coincided with increasing nitrate concentration. I will present the potential implications of these findings and the way forward for this project.
« Hide Abstract

Abstract:

We present an overview of the Microwave Limb Sounder (MLS) experiments flown on NASA's Upper Atmosphere Research Satellite (UARS) and, more recently, NASA's Aura mission. The MLS instruments measure vertical profiles of atmospheric composition, temperature and cloud ice from the upper troposphere to the mesosphere. MLS observations have been used to characterize and quantify important processes affecting climate, ozone layer stability and global air quality. The fundamentals of the MLS technique will be reviewed, and selected findings from both UARS and Aura MLS will be highlighted. Particular emphasis will be given to Aura MLS observations in the upper troposphere, where water vapor and ozone are strong greenhouse gases. Six years of Aura MLS observations show strong geographical and temporal variability in the behavior of upper tropospheric ozone and carbon monoxide, owing to the complex interplay of horizontal transport, deep convective lofting and chemistry. Finally, concepts for future microwave instruments will be discussed.
« Hide Abstract

Abstract:

TBA
« Hide Abstract

Abstract:

Uncertainty in future climate change presents a key challenge for adaptation and mitigation planning. In this talk, I will discuss an overlooked source of uncertainty, internal climate variability, using a new 40-member ensemble of greenhouse-gas forced simulations with the National Center for Atmospheric Research Community Climate System Model Version 3 (CCSM3) over the period 2010-2060. The dominant source of uncertainty in the simulated extra-tropical climate response is associated with the annular modes of intrinsic atmospheric circulation variability. Coupled ocean-atmosphere variability plays a dominant role in the tropics, with secondary effects over higher latitudes. Uncertainties in the forced response are generally larger for atmospheric circulation than precipitation and temperature. Implications for detection and attribution of observed climate change and for multi-model climate assessments will also be discussed.
« Hide Abstract

Abstract:

Hydrologic uncertainty arises from inherent variability across space and time in the hydrologic cycle. This uncertainty is further exacerbated by our inability to accurately measure all relevant hydrologic states and fluxes at all relevant spatial and temporal scales. The advent of satellite-based remote sensing in conjunction with parallel computing has, however, created new opportunities to better characterize, and ultimately reduce, hydrologic uncertainty via fusion with physically-based models. In this talk, a satellite-derived, ensemble-based framework used to generate downwelling radiation (energy) at the Earth�s surface is presented. A precipitation (mass) model using many of the same satellite inputs is currently under development, but is not presented here. Much of the motivation for this work is the creation of an ensemble of forcing fields (energy + mass) for use in distributed hydrologic model applications. Improved representation of forcing uncertainty within the hydrologic system is expected to improve our understanding of system dynamics thereby improving our ability to better manage freshwater resources. Application of this framework in the Southern Great Plains of the United States shows the resulting prior (unconditioned) ensemble implicitly contains the complex spatiotemporal uncertainty structure overlooked by more commonly used techniques. A data assimilation (DA) scheme is then employed that involves Bayesian conditioning of the prior estimate by relatively coarse-scale, readily-available measurements. Comparison against an independent, ground-based observational network shows the posterior (conditioned) ensemble is more accurate and contains less uncertainty relative to the prior. Furthermore, it is demonstrated that prior uncertainty structure dictates much of the conditioning capability of the DA scheme. Careful consideration of prior uncertainty can lead to an improved posterior whereas commonly used techniques can actually lead to posterior degradation.
« Hide Abstract

Abstract:

In the near future the NASA Global Precipitation Measurement (GPM) Mission will provide new sources of precipitation observations with unprecedented accuracy and coverage of the globe. For hydrological applications, the satellite observations need to be downscaled to the required finer spatial and temporal resolution precipitation fields. A WRF ensemble data assimilation system is developed to explore the potential of using ensemble data assimilation techniques and cloud-resolving models to dynamically downscale satellite observations. A high-resolution regional WRF model with multiple nesting grids is used to provide the first guess and ensemble forecasts. An ensemble assimilation algorithm based on Maximum Likelihood Ensemble Filter (MLEF) is used to perform analysis. Precipitation-affected radiances are assimilated along with the observations from the NCEP regional data stream. Prognostic hydrometeors and dynamical variables are simultaneously updated by the analysis every 3 hours. The experiments using the current available satellite precipitation data (AMSR-E and TRMM-TMI radiances) will be presented to discuss some of the challenging issues, such as model-predicted hydrometeor control variables and associated background error covariance, quality control and bias estimation in radiance space over land.
« Hide Abstract

Abstract:

Accurate coupling between the resolved scale dynamics and sub-grid scale physics is essential for accurate modelling of the atmosphere. Previous emphasis has been towards the temporal aspects of this so called physics-dynamics coupling problem, with little attention towards the spatial aspects. When designing a model for numerical weather prediction there is a choice for how to vertically arrange the required variables, namely the Lorenz and Charney-Phillips grids, and there is ongoing debate as to which is the optimal. The Charney-Phillips grid is considered good for capturing the large scale dynamics and wave propagation whereas the Lorenz grid is more suitable for conservation. However the Lorenz grid supports a computational mode. Further it is argued that the Lorenz grid is preferred for modelling the stably stratified boundary layer parametrisation. This presents the question: which grid will produce most accurate results when coupling the large scale dynamics to the stably stratified planetary boundary layer? The normal mode analysis approach, as used in previous work of a similar nature, is employed to address the question. This is an attractive methodology since it allows one to pin down exactly why a particular configuration performs well. Applying this method encounters issues when the problem is non normal, as it will be when including boundary layer terms. It is shown that when addressing the coupled problem the lack of orthogonality between eigenvectors causes interpretation of mode analysis to become very difficult. It is shown that the Lorenz grid is favoured when the boundary layer is considered on its own; it captures the structures of the steady states and examinable transient modes more accurately than the Charney-Phillips grid. For the coupled boundary layer and dynamics the Charney-Phillips grid is found to be most accurate in terms of capturing the steady state. It is also shown that certain Charney-Phillips configurations for the boundary layer parametrisation should be avoided. Results from examining transient modes are limited due to the problems with the methodology, however important understanding is gained in the application of the overall technique. The Lorenz grid computational mode is examined and found to be suppressed by the boundary layer, but only in the boundary layer region.
« Hide Abstract

Abstract:

We investigate downward wave coupling between the stratosphere and troposphere in the southern hemisphere in the Goddard Earth Observing System chemistry-climate model (GEOS CCM) Version I. Downward wave coupling occurs when planetary waves reflected in the stratosphere impact the troposphere and is distinct from zonal-mean coupling, which results from wave dissipation and its subsequent impact on the zonal-mean flow. A suite of different GEOS CCM simulations of the past and future are used to evaluate the modeled downward wave coupling and understand the impact of ozone changes. The GEOS CCM captures the main features of the observed seasonal cycle of downward wave coupling in the southern hemisphere. Temporal changes in ozone related to past depletion and future recovery significantly impact downward wave coupling in the model during November and December. The results reveal a new mechanism wherein stratospheric ozone changes can affect the tropospheric circulation.
« Hide Abstract

Abstract:

The Navy Operational Global Atmospheric Prediction System (NOGAPS) is the Department of Defense's global numerical weather prediction (NWP) system. In 2000, the Naval Research Laboratory (NRL) inaugurated a collaborative project among scientists in its Space Science, Marine Meteorology, and Remote Sensing Divisions to extend NOGAPS from its current upper boundary (then at ~35 km altitude) to altitudes nearer 100 km. This has led to an Advanced-Level Physics High-Altitude (ALPHA) NOGAPS prototype, known as NOGAPS-ALPHA, that runs at NRL as a fully functional research NWP system extending to ~100 km altitude. In this talk, I'll summarize this NOGAPS-ALPHA development history, focusing on advances in both the forecast model and data assimilation components that were required to operate it at altitudes that extend to the edge of space. Some of the research highlights made possible through forecasting and data assimilation experiments with this new system will then be described. I'll conclude with some future research and development avenues that are planned for this system and its successors, including recent and imminent transitions of NOGAPS-ALPHA technology to the NOGAPS run operationally at the Fleet Numerical Meteorology and Oceanography Center.
« Hide Abstract

Abstract:

Anthropogenic atmospheric aerosols can modify the radiative balance (and climate) of the Earth by altering the properties and global distribution of clouds. Current climate models however cannot adequately account for many important aspects of these aerosol-cloud interactions, ultimately leading to a large uncertainty in the estimation of the magnitude of the effect of aerosols on climate. The seminar will focus on the development of physically-based parameterizations of aerosol-cloud processes for climate models that help to address some of such predictive uncertainty. The parameterizations are analytical solutions to the cloud ice and water particle nucleation problem, developed within a framework that considers the mass and energy balances associated with the freezing and droplet activation of aerosol particles. They explicitly account for the impact of cloud formation dynamics, the aerosol size and composition, mixing, and the dominant freezing mechanism (homogeneous vs. heterogeneous) on the ice crystal and droplet concentration and size distribution. Application of the new parameterizations is demonstrated in the NASA Global Modeling Initiative atmospheric and chemical and transport model to study the effect of aerosol emissions on the global distribution of ice crystal concentration, and, the effect of entrainment during cloud droplet activation on the global cloud radiative properties. The ice cloud formation framework is also used within a parcel ensemble model to understand the microphysical structure of cirrus clouds at very low temperature. The new analytical parameterizations provide an efficient, yet rigorous, microphysical link between aerosol emissions and cloud formation, contributing to the improvement of the predictive ability of atmospheric models and the understanding of the impact of human activities on climate.
« Hide Abstract

Abstract:

We evaluate North American carbon fluxes using a monthly global Bayesian synthesis inversion that includes well-calibrated carbon dioxide concentrations measured at continental flux towers. We employ the NASA Parameterized Chemistry Tracer Model (PCTM) for atmospheric transport and a TransCom-style inversion with sub-continental resolution. We sub-sample carbon dioxide time series at four North American flux tower sites for mid-day hours to ensure sampling of a deep, well-mixed atmospheric boundary layer. The addition of these flux tower sites to a global network reduces North American mean annual flux uncertainty for 2001-2003 by 20% to 0.4 PgC/yr compared to a network without the tower sites. Uncertainty reduction is found to be local to the regions within North America where the flux towers are located, and including the towers reduces covariances between regions within North America. We will also show sensitivity of this inversion to different background terrestrial fluxes and fossil fuel emissions maps, and the impact of increasing the density of the North American surface network.
« Hide Abstract

Abstract:

With the launch of the European Space Agency's Atmospheric Dynamics Mission (ADM-Aeolus) in 2011 and the call for the 3D-Winds mission in National Research Council's decadal survey, direct spaceborne measurements of vertical wind profiles are imminent via Doppler wind lidar technology. Part of the preparedness for such missions is the development of the proper data assimilation methodology for handling such observations. As active measurements, the platforms will have largely predictable lifetimes. With ADM, the lifespan of the instrument is expected to be three years. To maximize the utility of the instrument, an Observing System Simulation Experiment (OSSE) framework is being utilized to generate a realistic proxy dataset for development of the Gridpoint Statistical Interpolation (GSI) data assimilation system utilized at a number of centers through the United States including locally at the Global Modeling and Assimilation Office (GMAO). This effort will be presented, including the methodology and status of proxy data generation, validation of necessary fields in the Joint OSSE Nature Run, and the assimilation of such measurements within the GSI.
« Hide Abstract

Abstract:

The Irminger sea is located between southern Greenland and the Reykjanes ridge, which is south of Iceland. It is an important area of the North Atlantic ocean due to the complex and important consequences of the physical processes that occur in this region. For example, here the Arctic freshwater first meets the North Atlantic saline waters, an important contribution to the meridional overturning circulation (MOC). Detailed study of the various processes and their relationship to another is the main goal of our work. Using an eddy-resolving (~4KM horizontal resolution), regional configuration of the MITgcm ocean general circulation model, we synthesized real, observed data, over a period chosen to coincide with the cruise period of Oceanus-395, during Aug 5-- 13, 2003. Gridded, blended sea-surface-height data was taken from the AVISO mission and the sea-surface-temperature was from GHRSST. In addition, various in-situ temperature and salinity data from the cruise was also assimilated. Our 4d-VAR procedure yielded model initial conditions while minimizing model-data misfit. In this talk, we will discuss these results from our data assimilation experiments and how they could be used for hindcasting for the Irminger Sea.
« Hide Abstract

Abstract:

The ECMWF forecast system was not able to maintain the amplitude of an MJO event for more than a few days until the implementation of a recent version (referred to as CY32r3) which is able to maintain the amplitude of an MJO for at least 30 days. The various changes in the ECMWF model physics that led to this improvement will be discussed. A series of 46-day ensemble integrations starting on the 15th of each month from 1989 to 2008 has been completed to assess the skill of the ECMWF forecast system to predict the MJO. Results indicate that the dynamical model displays skill for up to about 20 days in predicting the evolution of the MJO. However, the MJO simulated by the model has a too slow eastward propagation and has difficulties crossing the Maritime Continent. The impact of the MJO in the Tropics and Extratropics has been evaluated in this set of hindcasts. In particular, the impact of the MJO on tropical storms is consistent with observations which explain the sig nificant skill of the ECMWF forecast system to predict weekly probabilities of tropical storms up to week 4. The representation of the MJO and tropical storms in the ECMWF seasonal forecasting system will also be discussed. At this time range, the ECMWF model displays significant skill to predict the interannual variability of tropical storms over the Atlantic.
« Hide Abstract

Abstract:

We assess the impact of satellite-rainfall error structure on soil moisture simulations with the NASA Catchment Land Surface Model. Specifically, the study contrasts a multi-dimensional satellite rainfall error model (SREM2D) to the standard rainfall error model used to generate an ensemble of rainfall fields as part of the Land Data Assimilation System developed at the NASA Global Modeling and Assimilation Office (NASA-LDAS). We used high-resolution (25-km / 3-hourly) satellite rainfall fields over Oklahoma (derived from the NOAA CMORPH global satellite product) and corresponding rain gauge-calibrated radar rainfall fields (considered as the reference or true rainfall). Land model simulations are evaluated in terms of rainfall and soil moisture errors. Comparisons of rainfall ensembles generated by SREM2D and NASA-LDAS against reference radar rainfall show that both rainfall error models preserve the rainfall error characteristics across a range of spatial scales. The multi-dimensional error-structure in SREM2D, however, generates rainfall replicates with higher variability that better envelope the reference rainfall than those generated by the NASA-LDAS error model. Because rain-to-soil moisture error propagation is a non-linear process, soil moisture simulations are less sensitive to the complexity of the precipitation error modeling approach. Nonetheless, perturbing satellite rainfall fields with a complex error model leads to improved spatial variability in the simulated soil moisture ensemble, which is expected to benefit soil moisture data assimilation.
« Hide Abstract

Abstract:

Some validation results from what we are calling version 1 of the GMAO OSSE will be presented. This uses a 13-month ECMWF model simulation as a nature run and the NCEP/GMAO GSI as the data assimilation system. The simulated observations include most types used operational during the winter of 2005-2006. Metrics examined are O-F and A-B statistics, including horizontal correlations. We obtain close agreements between corresponding results for OSSE and real observation assimilations for most observation types. Although some deficiencies remain apparent, the version-1 simulated observations can serve as a baseline for some less-demanding OSSE studies as well as a step for continuing development.
« Hide Abstract

Abstract:

Construction of the multi-model high-resolution seasonal prediction system is introduced in this study. A set of dynamical and statistical regional models is considered for the construction of this system. In this study, coarsely resolved atmospheric fields from the global climate model (GCM) (or reanalysis with 2.5 degree resolution) are downscaled to a fine spatial scale of ~20 km for an assessment of model fidelity for this fine-scale seasonal prediction. The model test is conducted for summer season for the southeast United States (Florida, Georgia, Alabama, South and North Carolina), where accurate prediction is challenging due to a heavy influence of small-scale subtropical (or tropical) convection. Dynamical climate models considered are the Florida State University/Center for Ocean-Atmospheric Prediction Studies Nested Regional Spectral Model (FSU/COAPS NRSM), Experimental Climate Prediction Center Regional Spectral Model (ECPC RSM), Regional Climate Model 3 (RegCM3), and Weather Research and Forecasting Model (WRF). The conceptual bases for the statistical models in this study are advanced eigentechnique with multiple regression, nonlinear canonical correlation analysis, and geo-spatial weather generator. The present talk discusses the improvement of predictability on seasonal anomaly and frequency of subseasonal extremes by this forecasting system. Specifically, capability of the ECPC RSM for adding smaller scale information that the large-scale forcing fields do not resolve is investigated to suggest the promise for the combination of GCM-RCM for seasonal prediction of summer precipitation at 20km resolution.
« Hide Abstract

Abstract:

The Atlantic Ocean contains a strong signal of decadal to multidecadal variability. Recent work has shown that the associated Atlantic sea surface temperature fluctuations have a significant impact on climate throughout the hemisphere, including droughts from Africa to North America, hemispheric mean temperature, and Atlantic hurricane activity. It has been hypothesized that both radiative forcing changes and internal variability contribute significantly to the observed SST fluctuations. Here we assess the degree to which we understand the causes of Atlantic decadal variability. In addition, we assess the prospects of predicting such decadal scale fluctuations and their impacts.
« Hide Abstract

Abstract:

Environmental monitoring using satellite and in-situ data has become a central focus of several international efforts. In particular, the EU-funded Global Monitoring for Environment and Security (GMES) initiative has funded the GEMS (Global and regional Earth-system Monitoring using Satellite and in-situ data, concluded in May 2009) and MACC (Monitoring Atmospheric Composition and Climate) projects to develop and implement its "Atmosphere Component Service". Both are coordinated by ECMWF. The main goal of these projects is to develop state-of-the-art systems for the retrospective monitoring and real-time forecasting of atmospheric composition (greenhouse gases, reactive gases and aerosol). This talk will present results from the multi-year reanalysis (2003-2009) which was completed under GEMS and the near-real time forecasts that are currently produced daily using these systems. Particular attention will be given to the aerosol component which is based on an aerosol model which is fully integrated in the ECMWF NWP model and relies on assimilation of aerosol optical depth data from MODIS. Examples such as Saharan dust transport over the Atlantic, California fires of July 2008 and the recent Sydney dust storm of September 2009 will be presented. A technical description of the aerosol assimilation system will also be provided along with recent developments and future plans.
« Hide Abstract

Abstract:

Reanalysis has long been considered unsuitable for estimating climate trends, because of the effects of changes in past observing systems on systematic errors in the reanalysis data. This prevailing view is based mostly on experiences with first- and second-generation reanalysis products, but a great deal of progress has been made recently in improving the 'climate quality' of reanalysis data. Ultimately the problems with estimating trends from observations are fundamental and not specific to reanalysis. Inherent uncertainties in the existing observational record can only be exposed by combining multiple sources of information, and this is exactly what reanalysis is designed to do.
« Hide Abstract

Abstract:

In 4D-Var, the forecast model is used to propagate the analysis increment within the assimilation window to the time of observations. Usually, this model is assumed perfect, or, at least, it is assumed that the errors due to the model can be neglected compared to other errors in the system. As most aspects of the data assimilation system have improved over the years, and as the assimilation window might become longer in the future, this assumption becomes less realistic. Weak constraints 4D-Var provides the general framework for accounting for model error. We will present two of formulations of weak constraints 4D-Var. The first formulation, which became operational at ECMWF in 2009, aims at capturing the systematic model error in the stratosphere. This is achieved through the use of a model error forcing term which is kept constant over the length of the assimilation window. Experimentation showed that the error captured by the model error term was not always model error. For example, a feedback effect between the model error term and the Jb balance term of the 4D-Var cost function in the stratosphere was noticed. Results obtained with a model error forcing term in the ECMWF operational 4D-Var system will be presented. The other formulation is more general and comprises a four dimensional state control variable. In principle, this allows for parallelism in the 4D-Var algorithm in the time dimension of the assimilation window which would be essential for computational efficiency with long assimilation windows. However, the nature of the optimization problem in this implementation of weak constraints 4D-Var is more challenging and new preconditioning techniques will have to be implemented. An overview of the expected advantages, challenges and possible approaches towards a fully weak constraints 4D-Var system will be presented.
« Hide Abstract

Abstract:


« Hide Abstract

Abstract:

This presentation is concerned with the potential to improve the root-zone soil moisture in Numerical Weather Prediction (NWP) models by assimilating remotely sensed near-surface soil moisture observations. In many NWP models, including those at Meteo-France and the Australian Bureau of Meteorology, soil moisture is initialised based on errors in screen-level temperature and relative humidity forecasts. While this approach effectively improves low-level atmospheric forecasts, it often degrades the model root-zone soil moisture. An Extended Kalman Filter (EKF) capable of assimilating both near-surface soil moisture and screen-level atmospheric observations has been developed, in the hope that including the near-surface soil moisture observations will improve the model soil moisture while maintaining reasonable low-level atmospheric forecasts. The EKF is tested with the French and Australian NWP land surface schemes, using near-surface soil moisture retrieved from Advanced Microwave Scanning Radiometer (AMSR-E) C-band observations by the Free University of Amsterdam and NASA. In experiments over one year with the Australian NWP land surface scheme (MOSES) the near-surface soil moisture generated by the EKF assimilation of AMSR-E generated a better fit to in situ observations from the Murrumbidgee Monitoring Network than either of the AMSR-E data or the model alone. The assimilation also improved the fit between the model soil moisture and the in situ observations in the root-zone. Experiments assimilating the AMSR-E data into the French NWP land surface model (ISBA) over one month confirmed that while the screen-level and AMSR-E soil moisture data generate conflicting signals regarding the root-zone soil moisture, the EKF can effectively reduce the observation departures for both observations types, although these reductions are modest. Despite these generally positive findings several challenges to implementing the assimilation of near-surface soil moisture in an operational NWP setting have been identified.
« Hide Abstract

Abstract:

The Weather Research and Forecasting (WRF-ARW) model and its associated data assimilation (DA) system WRFDA are developed by the National Center for Atmospheric Research (NCAR) and distributed to the global scientific community. In this talk, we will briefly summarize the recent expertise acquired with advanced variational (4DVar) and variational/ensemble (hybrid) DA systems. The latest developments on the representation of background error statistics will be presented. Specific issues related to the assimilation of satellite radiances will be addressed, with a focus on observation bias correction, cloud detection and steps toward the use of cloudy radiances. We will also discuss the implementation of a diagnostic tool using an adjoint technique to calculate the impact of observations on forecast error. NCAR is leading the developments for the Air Force Weather Agency (AFWA) next generation cloud analysis system, which aims at updating hydrometeors jointly with the dynamical variables. The focus on cloud resolving DA and the decision to transition to the Gridpoint Statistical Interpolation (GSI) system open new avenues for external collaborations that will be explored.
« Hide Abstract

Abstract:

From a large-scale perspective the West African monsoon (WAM) can be described in terms of the annual march of the ITCZ and its associated regional circulations. On the synoptic and mesoscale, the WAM is comprised of a complex collection of wave patterns, organized weather systems and deep convection. These include synoptic systems such as African easterly waves (AEWs) and mesoscale convective systems (MCSs), the main rain-producers in the region. AEWs also initiate many of the Atlantic tropical cyclones (TCs) downstream and thus are an important part of the interactions that take place between West Africa and the tropical Atlantic. The first part of this talk will provide an analysis of the climatological annual cycle of the West African monsoon, emphasizing the coastal rainfall in Spring in addition to the more commonly emphasized summer rainfall in the summer. Particular focus is given to the potential roles played by the Atlantic cold tongue and Saharan heat low in this regard. The second part will emphasize more the nature of the AEWs. It will include discussion of three phases of the observed AEW life-cycle: (i) genesis, (ii) baroclinic development and (iii) West coast developments including tropical cyclogenesis. Some perspectives for future avenues of research will be provided.
« Hide Abstract

Abstract:

A hybrid variational and ensemble transform Kalman filter (ETKF) analysis method has been implemented recently for regional NWP using the WRF model. In the hybrid method, the variational framework is used to calculate the analysis increment using ensemble-based flowdependent background-error covariances. To accommodate flow-dependent ensemble-based covariances within existing variational systems, an "extended control variable method" that includes covariance localization to account for sampling error was implemented. The ensemble transform Kalman filter was used to generate ensemble forecasts. Recent studies have suggested that the hybrid systems may yield the "best of both worlds" by combining the best aspects of variational and ensemble Kalman filter (EnKF) systems. The advantages of the hybrid method will be discussed. A scope of applications of the hybrid VARETKF using WRF for various scales and data types will be discussed in the seminar. These include synoptic scale month-long experiments over North America domain, hurricane forecasts and radar assimilation for continental precipitation.
« Hide Abstract

Abstract:

The seminar introduces the modeling works and the achievement, which have been done at Climate Modeling Group of Seoul National University for last one year. Therefore it updates the progress for the development of new moist physics scheme, which is suitable to the high-resolution climate model of a spatial scale of the order of 10km. The convective parameterization is developed based on the bulk formula and the turbulence ensemble concept is introduced to generate the shallow convection more effectively. In the seminar, also discussed are the problems related to extending the cloud resolving model to the global domain with a grid spatial scale of 10km and the way to include the microphysics in the convectional GCMs.
« Hide Abstract

Abstract:

Marine boundary layer (MBL) clouds can significantly regulate the sensitivity of climate models, yet they are currently poorly simulated in GCMs. In this presentation, I will first describe the seasonal variations of physical properties of these clouds off the California coast by using measurements from several satellite programs. I will then show how simulations of the seasonal cloud properties in the NCAR Community Atmosphere Model (CAM) compare with the observations. I will finally discuss the relevance of seasonal variation of clouds as a test of cloud-climate feedback in GCMs.
« Hide Abstract

Abstract:

As well known, satellite observations constitute a majority of atmospheric observations today, providing an invaluable, nearly continuous flow of information about the state and evolution of the atmosphere. In operations, these data are used routinely for nowcasting and in data assimilation systems to produce initial conditions for numerical forecasts. In the former application, the satellite data are used under all weather conditions, often focusing on locations with significant weather that is characterized by presence of clouds and precipitation. In the data assimilation systems, on the other hand, majority of the satellite data are not utilized in areas that are affected by clouds and precipitation. The data assimilation systems exclude the cloud- or precipitation-affected satellite radiance observations primarily to avoid negative influences of the numerical model forecast background which traditionally has low skill in representing the cloud and precipitation properties, including their spatial and temporal distribution. Besides, simulation of cloud and precipitation affected satellite observations in the data assimilation requires application of more complex radiative transfer models and is more computationally demanding than simulation in clear sky conditions. Recent data assimilation research has demonstrated that operational global forecast models are approaching a level of skill that would allow benefits from using cloud- and precipitation- affected radiance observations (Wilkinson et al., 2008; Bendedetti and Janiskova, 2007). Similarly, data assimilation studies with a cloud-resolving research model show that the assimilation of the cloud-affected IR observations improves model simulations, suggesting potential for extending the benefits of the satellite data assimilation into high-resolution forecast models (Vukicevic et al., 2004, 2006). In this talk a summary of recent past and current research results on the assimilation of GOES Imager (Geostationary Operational Environmental Satellites) observations into a cloud resolving model using 4DVAR data assimilation approach is presented. The past studies include research on the properties of radiative transfer modeling in the cloudy atmosphere, information content of the imager observations with respect to bulk cloud properties, nonlinear 4DVAR, dynamic coupling between the clouds and environment by dynamic data assimilation approach, and benefits of high temporal and spatial resolution in the observations. In the current research several approaches for improving conditioning of the data assimilation for mixed cloud scenes are tested including use of dynamical cloud mask, limit on maximum innovation amplitude in brightness temperature , observation-based estimate of background error decorrelation length for water variables, smooting of adjoint solution and correcting of surface temperature bias in the model. The results are verified against a set of independent ground-based data which include data from AERI (Atmospheric Emitted Radiance Interferometer), MWR (Microwave Radiometer), cloud radar, and radiosonde. The numerical experiments are performed on cases of cloudy atmosphere that were associated with synoptic scale systems in south-central Great Planes in U.S.A. Overall, the results indicate that the data assimilation of the cloud affected geostationary observations is feasible. The results show improvement in the model representation of the cloudy atmosphere and consistent change in the dynamical cloud environment. The latter is achieved by 4DVAR in the current study, but similar dynamical consistency between the cloud and other analysis variables could be achieved with alternative dynamic data assimilation approaches such as Ensemble Kalman Filter. Using the dynamic cloud mask and maximizing the number of observation points by allowing large innovations in the observation space produce the most beneficial control on the conditioning of the cloudy radiance assimilation.
« Hide Abstract

Abstract:

In June 2009 the European Commission awarded funding to the MACC (Monitoring Atmospheric Composition and Climate) project within their seventh Framework programme as a follow-on of the GEMS (Global and regional Earth-system Monitoring using Satellite and in-situ data) project. MACC is a collaboration between the European Centre for Medium-Range Weather Forecasts (ECMWF) and many European research institutes and aims at creating a new European operational system for global monitoring of atmospheric chemistry and dynamics, and an operational system to produce improved medium-range and short-range air-chemistry forecasts through much improved exploitation of satellite data. The main task of ECMWF within the MACC project is to build a monitoring system based on the operational four-dimensional variational (4D-Var) data assimilation system. This allows the monitoring of greenhouse gases, reactive gases, and aerosols in the atmosphere using all available relevant satellite and in-situ data. These four-dimensional fields are then used by the partner institutes to study long-range transport and surface flux inventories. The system is also able to produce accurate boundary conditions for regional air-quality models. I will present an overview of the environmental monitoring system and results representing the various components of the system. Some important issues, such as the background constraint, the observation operators and the interaction with chemistry models, will be discussed as well.
« Hide Abstract

Abstract:

The water vapor feedback is the strongest positive feedback operating in the climate system, with the power to about double the direct warming from long-lived greenhouse gases. In our analysis, we estimate the strength of the water vapor feedback by analyzing the changes in tropospheric specific humidity during El Nino-Southern Oscillation (ENSO) cycles in climate models and in the MERRA and ERA40 reanalyses. We find a wide range in predicted feedback strengths among the models and between the models and the reanalyses, although they all predict a water vapor feedback that is strong and positive. The models and the reanalyses show a consistent relationship between the variations in the tropical surface temperature over an ENSO cycle and the radiative response to the associated changes in specific humidity. However, the feedback is defined as the ratio of the radiative response to the change in the global average temperature. Differences in extratropical temperature variations will therefore lead to different inferred feedbacks, and this is the root cause of large spread in feedbacks among the models and between the models and the reanalyses.
« Hide Abstract

Abstract:

Clouds and convection play an important role in climate and weather related issues over various spatial and time scales. Despite the efforts have been made, representing clouds, convections and their radiative and precipitating processes in numerical weather and regional/global climate models remains a challenge. To help resolve these issues, a CloudSat-centric, multi-parameter A-Train (e.g., CloudSat, Calipso, AIRS, AMSR, MODIS, CERES) and high-resolution ECMWF analyses (INTRIM and YOTC) data set is being developed to characterize dynamical, macro-/microphysical, precipitating and radiative processes associated with clouds and convection to evaluate and constraint the key relevant model physical parameters/processes in numerical models. In this presentation, results from the comparisons between cloud, convection, and precipitation statistics derived from using the above data set will be presented. The errors in the estimated radiative fluxes and heating associated with ignoring convective/precipitating ice (something commonly done in GCMs) relative to the standard CloudSat radiation product as well as results from a stand alone Fu-Liou radiative transfer model will also be discussed.
« Hide Abstract

Abstract:

Characteristics of water vapor transport were analyzed in terms of the flux-gradient (or flux-profile) function for water vapor (Phi_q) and eddy diffusivity (K_q) in the atmospheric surface layer. Roughness length of water vapor (z_0q) over land surfaces is also analyzed. Phi_q and K_q were estimated using the CASES-99 (Cooperative Atmosphere-Surface Exchange Study-1999) filed data and z_0q was analyzed using the FLOSS-II (Fluxes Over Snow Surface-II) data. It is found that, contrary to the previous assumption of scalar similarity, water vapor is transported less efficiently than heat under unstable atmosphere while being transported comparatively or even more efficiently under stable conditions in the atmospheric surface layer. Therefore, it is revealed that the water vapor is dissimilar to the heat. The degree of discrepancy between the fluxes of water vapor and heat is dependent on the atmospheric stability. Optimally determined functions for Phi_q were provided according to atmospheric stability. It is also found that z_0q is significantly smaller than those of the previous studies over bare soil and grass surfaces. Over snow surface, z_0q agrees quite well to the one of reference functions. Modified parameterizations for kB_q^(-1)[=ln(z_0m/z_0q)] for different land surface types are found to estimate the latent heat flux more accurately. Present results suggest dissimilarity between scalars (e.g., water vapor and heat) in the atmospheric surface layer. The cause of the dissimilarity is attributed to the different magnitude of buoyancy effect of the temperature and humidity fluctuation. The differences in the scalar fluxes suggest that different parameterizations for the water vapor and the heat are required for the flux estimations (e.g., Bowen ratio energy balance method, flux-profile method) and numerical models.
« Hide Abstract

Abstract:

Anthropogenic aerosols affect the radiative balance of the Earth-atmosphere system by scattering or absorbing sunlight, by acting as cloud condensation nuclei (CCN) or ice nuclei (IN) and thus modifying the optical properties and lifetimes of clouds, and by altering the local atmospheric thermodynamic status and thus cloud formation and dissipation. To understand the detailed mechanisms and strengths of the climate impacts of anthropogenic aerosols, an interactive aerosol-climate model has been developed based on the Community Climate System Model (CCSM) of NCAR. Its aerosol module describes the size, chemical composition, as well as the mixing states of various sulfate and carbonaceous particulate matters. Modeled aerosol properties including the strength and distribution of solar particulate absorption have been compared with satellite and surface network data. Several sets of long-term integrations have also been conducted to study the climate responses to the direct radiative forcing of various types of anthropogenic aerosols. One interesting finding of the recent study suggests that the radiative forcing of anthropogenic aerosols can cause a significant change in both quantity and distribution of tropical precipitation, ranging from Pacific and Indian to Atlantic Ocean. Because this change occurs mostly in places away from any major source regions of aerosols, it must have been implemented primarily through forced changes in the large-scale circulation. These results along with some detailed features of the model will be discussed.
« Hide Abstract

Abstract:

The population of northern Africa, especially in the Sahel, is vulnerable to climate change because regional economies are highly dependent on climate for agriculture and water resources. Precipitation records of the 20th and 21st c. show that the region exhibits strong decadal variations in climate, with a propensity for long-period droughts, and we know that, abrupt climate change has occurred over northern Africa on millennial time scales. These factors combine to suggest that we should pay special attention to climate change projection for this region. Regional climate change projections for northern Africa at the end of the 21st c. are presented, and the possibility for abrupt climate change in the region is discussed.
« Hide Abstract

Abstract:

Optimal initialization technique to obtain accurate and well-balanced initial condition and generation of fast-growing perturbation for ensemble prediction are developed and examined in coupled model framework. Ensemble Kalman Filter (EnKF) data assimilation system with a SNU coupled GCM has been successfully developed. Parallelism for the data assimilation step is achieved by regionalization of the background-error covariances that are calculated from the phase-space distribution of the ensemble. The hindcast experiments from 1980 to 1999 show that forecast skill of ensemble forecasts with EnKF is superior to that with nudging, especially during longer forecast lead months. Among several benefits of state-dependent error covariance matrix in EnKF, it helps to successfully resolve the small-scale Tropical Instability Waves (TIWs) whose activity is highly dependent on ENSO phase. However, with simple initialization methods (e.g. 3DVAR or OI methods), TIW variability in oceanic initial conditions is excessively suppressed which ruins the interaction between TIWs and climate states. Through 20-yr hindcast experiments with and without TIWs, it is shown that nonlinear relationship between TIWs and ENSO is realistically simulated with state-dependent TIWs, which leads realistic simulation of the El Nino-La Nina asymmetry. It is shown that correlation improvement of simulated NINO3 index is over 0.1 at 4-month lead time. In addition, new initialization method to reduce the forecast errors by assimilating prospective observations are developed to supply the deficiency of conventional EnKF that only observations at analysis time can be used,. The new EnKF scheme is examined with CZ-SPEEDY hybrid coupled model within perfect model context. The analysis and forecast experiments shows that new EnKF scheme successfully cancels out the analysis errors at the target time, therefore, seasonal prediction skill using new EnKF scheme is superior to that using conventional EnKF scheme. To reduce the extra computation and initial shock by adding initial perturbations to initial states, a method for selecting optimal initial perturbations is developed within the initialization with EnKF. Among the initial conditions generated by EnKF, ensemble members with fast growing perturbations are selected to optimize the ENSO seasonal forecast skills. Seasonal forecast experiments show that the forecast skills with selected ensemble members are significantly improved than that with other ensemble members up to 1-year lead forecasts. It is also shown that the forecast skills are significantly improved during ENSO onset and decaying phase, which are most unpredictable ENSO periods.
« Hide Abstract

Abstract:

Composites based on observations and model outputs from the CLIVAR drought experiments were used to examine the decadal modulation of the impact of El Nino Southern Oscillation (ENSO) on the hydro-climate over the United States. The impact of ENSO on drought is not stationary. It is modulated both by the trends and the Atlantic multi -decadal Oscillations (AMO). For warm (cold) ENSO winter and spring, composites show warming (cooling) over the Northwest and cooling (warming) over the Southeast and the Gulf States. However, the patterns and magnitudes of composites change over time. In the recent periods, cooling over the South has been weakening. For precipitation, the largest ENSO impact occurs in winter. While the ENSO impact strengthened over the Southwest, California and the Gulf states for the recent periods, the impact over the Ohio Valley reached a peak in 1930-1970 and has been weakening in recent years. The direct influence of the AMO on drought is small. The major influence of the AMO is to modulate the impact of ENSO on drought. The influence is large when the SSTAs in the tropical Pacific and in the North Atlantic are opposite in phase. A cold (warm) event in a positive (negative) AMO phase amplifies the impact of the cold (warm) ENSO on drought. The ENSO influence on drought is much weaker when the sea surface temperature anomalies in the tropical Pacific and in the North Atlantic are in phase.
« Hide Abstract

Abstract:

The currently available theoretical and observational evidence for a latitudinal structure of thermocline vertical diffusivity is synthesized and included in a state of the art coupled climate model. Compared to the standard background value of 0.1 cm2/s, the simulations with the latitudinal structure show only little change in the meridional overturning circulation or northward heat transport. However, two regions are identified which are sensitive to the value of vertical diffusivity: the equatorial band, where only small changes in sea surface temperature lead to precipitation responses with basinwide teleconnections; and the North Atlantic where diffusivity affects the spiciness of Labrador Sea water and subsequently the Gulf Stream path.
« Hide Abstract

Abstract:

We briefly discuss the Local Ensemble Transform Kalman Filter (LETKF, Hunt et al., 2007), and several extensions recently developed at UMD that extend the usefulness of EnKF, and in particular the 4D-LETKF. Among them: A "no-cost smoother" based on applying the ensemble background weights derived at analysis time--but valid throughout the assimilation window--to the start of the window. "Outer loop" and "Running in place" algorithms that take advantage of the no-cost smoother to handle nonlinearities, non-Gaussianities and spin-up. A coarse resolution LETKF that interpolates weights without degrading the analysis. A simple method to estimate and correct model errors in model space and its application to data assimilation. A method to estimate simultaneously the background error inflation and observation errors. Forecast "adjoint sensitivity" without adjoint in EnKF, and its extension to longer forecasts. Analysis sensitivity to observations and exact cross-validations. The seminar will conclude with a discussion of the potential of hybrid systems.
« Hide Abstract

Abstract:

Tropospheric ozone is a major atmospheric pollutant and a greenhouse gas. Its distribution represents a complex interplay between transport and chemistry. The wealth of recent satellite observations of trace gases in the troposphere offer a unique opportunity to better understand the chemical and transport processes regulating tropospheric ozone and its precursors. Chemical data assimilation provides a powerful means of exploiting these data. We have integrated data from the Tropospheric Emission Spectrometer (TES), the Ozone Monitoring Instrument (OMI), and the SCanning Imaging Absorption SpectroMeter for Atmospheric CHartographY (SCIAMACHY) satellite instruments into the GEOS-Chem global chemical transport model to assess the constraints that these data provide on the processes influencing the distribution of tropospheric ozone, with a particular focus on the summertime ozone distribution in the extratropical northern hemisphere.
« Hide Abstract

Abstract:

Model physics (e.g., cloud microphysics, convection, and radiation) schemes represent an important source of uncertainty in numerical models that range in scale from large eddy simulation to general circulation models. Much of this uncertainty is associated with specification of parameters that control the rates and/or characteristics of physical processes. In contrast to errors in forecast initial conditions, the characteristics of model physics uncertainty are not well understood, hence model physics error is not included in most operational data assimilation systems. As ensembles of simulations are increasingly used in data assimilation and probabilistic forecasting, it is desirable to perturb both initial conditions and model physics parameters. To do so properly requires knowledge of which parameters have the greatest effect on model results, as well as the characteristics of the relationship between model output and changes to parameters. In this paper, we explore the functional relationship between model parameters and observations using a Markov chain Monte Carlo (MCMC) algorithm. We examine bulk cloud microphysics and radiation packages from a cloud resolving model that are similar to schemes used in modern regional and general circulation models, and demonstrate how the joint probability distribution returned from MCMC can be used to 1) map the functional relationship between changes in model physics parameters and changes in model output, 2) identify which parameters have the most significant effect on various model output fields, 3) describe the nature of nonlinearity in the parameter-state relationship, and 4) explore how changes in the characteristics of observations affect the model state. The results of the MCMC-based inversion shed light on the reasons behind the loss in �parameter identifiability� noted in previous parameter estimation studies, and suggest ways in which nonuniqueness can be avoided in the construction of an ensemble based data assimilation scheme that includes model physics parameters as control variables.
« Hide Abstract

Abstract:

The Atmospheric Infrared Sounder (AIRS) onboard the NASA Aqua satellite and the Infrared Atmospheric Sounding Interferometer (IASI) onboard the METOP-A satellite both are cutting-edge spectrometers for 3-dimensional mapping of air and surface temperature, moisture, greenhouse gases, and cloud properties. With several kilo-channel measurements in the infrared spectral coverage, AIRS and IASI has a spectral resolution more than 100 times greater than previous IR sounders, and provides more accurate information to improve weather forecasting and support climate research. To enjoy the greatest advantage of such high resolution infrared observations, high-performance-computing radiative transfer models for AIRS and IASI are needed to facilitate more effective applications in physical retrievals and data assimilation. Approximately every six months there is a doubling in the speed of Graphics Processing Units (GPUs). Currently, the flagship NVIDIA GPU has 240 computing cores, compared to the best INTEL CPU with just 6 cores. The computing performance of the GPU has significantly outpaced its CPU counterpart, with a theoretical peak performance of 1 TFlops per GPU in single precision and 345 GFlops per GPU in double precision. The combined features of general-purpose supercomputing, high parallelism, high memory bandwidth (102 GB per GPU), low cost, and compact size are what make a GPU-based desktop computer an appealing alternative to a massively parallel system made up of commodity CPUs (e.g. Beowulf clusters). This paper presents our effort to develop the GPU-based high-performance AIRS/IASI radiative transfer model running on NVIDIA GPUs via CUDA (Compute Unified Device Architecture), the compute engine in NVIDIA GPUs for massively multi-threaded parallel computation. Our GPU implementation of the forward model is tested on a low-cost (~$8,000) NVIDIA S1070 personal supercomputer with 4 Tesla GPUs (total 960 cores) delivering 4 TFlops peak performance. The result is compared with the native INTEL multi-core CPU implementation to show the significant speed-ups of computing the AIRS/IASI radiance spectrum using a GPU-based system. This is our first step towards the development of a GPU-based high-performance full-spectrum AIRS/IASI physical retrieval system for product applications and data assimilation.
« Hide Abstract

Abstract:

The variational data assimilation problem can be solved in either its primal (3D/4D-Var) or dual form (3D/4D-PSAS). As shown in El Akkraoui et al. (2008), both methods are equivalent at convergence but the dual method exhibits a spurious behavior at the beginning of the minimization which leads to less probable states than the background state. This is a serious concern when using the dual method in operational implementations when only a finite number of iterations can be afforded. Two minimization algorithms are examined: the Conjugate Gradient (CG) and the Minimum Residual (Minres) methods. The Minres algorithm insures a monotonic decrease of the gradient norm and when applied to the dual problem, it also leads to a monotonic decrease of the primal cost function. Moreover, it is shown that a new termination criteria, based on the error norm in model space, can be used in the dual case to achieve the same accuracy in the analysis state when only a finite number of iterations are completed. This is of great importance for an implementation of the dual form of a weak-constraint 4D-Var. This will be discussed in the presentation and preliminary results will be presented.
« Hide Abstract

Abstract:

Climate and weather prediction models are still quite inaccurate in representing clouds, turbulence, convection and the interaction between the ocean and the atmosphere. The problem is that these processes can occur in a variety of scales, from the planetary scale to very small scales that cannot be represented explicitly in any atmospheric model. A major challenge in climate and weather prediction is in how to improve the representation of these sub-grid scale physical processes: the parameterization problem. This talk will focus on two major issues in particular: (i) how to represent the sub-grid scale turbulent motions and (ii) how to represent clouds. Recent model results from an international evaluation effort (the GCSS Pacific Crossection Intercomparison - GPCI) will be presented in order to highlight major model deficiencies. New approaches for cloud parameterization that make use of probability density functions (pdfs) and unified parameterizations for turbulence and convection involving optimal combinations of Eddy-Diffusivity and Mass-Flux methods will be discussed in detail.
« Hide Abstract

Abstract:

The ocean plays a major role in regulating the atmospheric greenhouse gas concentrations produced by natural and anthropogenic processes. The oceanic drawdown/outgassing of CO2 is determined by the solubility pump, i.e. physical ocean processes (air-sea interaction, convection and the thermohaline circulation) and by the biological pump, i.e. the carbon removal from the atmosphere by the sea organisms through photosynthesis and respiration, its transport to the deep ocean and its sedimentation there. Changes in the atmospheric concentrations of CO2 due to natural or anthropogenic factors result in changes in the behavior of these pumps which in turn further adjust (positively or negatively) the atmospheric concentrations of CO2. Such feedbacks involve changes to the oceanic biodiversity, acidification and the ocean surface albedo. The correct estimation of the sign and magnitude of carbon climate feedbacks is featured as a major unresolved task in the upcoming IPCC AR5 report. The ocean carbon modeling effort at NASA-GISS combines GISS and GSFC modeling components in order to simulate the magnitude of the ocean carbon sink and the associated climate feedbacks, as well as its behavior under different climate change scenarios. In the current phase, the primary objective is to incorporate the NASA Ocean Biogeochemical Model (NBOM) developed at GSFC by Watson Gregg into the two operational NASA-GISS ocean models, the Russell ocean model and HYCOM, which employ different geometry and vertical coordinate systems, and to couple the ocean-carbon modules to the latest modelE atmospheric component. The fully coupled systems are run to examine ocean carbon feedbacks both in the context of AR5 and of the NASA biogeochemistry program. Preliminary results indicate that to first order, the fidelity of the simulated ocean carbon cycle depends on the realism of the background physical model. The model agrees with observations with regards to surface partial CO2 pressure and the air-sea flux, the surface and vertical distributions of DIC as well the nutrient concentrations. HYCOM presents a better depiction of ENSO and southern ocean ice coverage variability whereas the Russell ocean model has more realistic global SST. A problematic region common in both ocean models remains the North-East Pacific region with intense heat losses to the atmosphere which are not supported by observations. Surface albedo changes due to real chlorophyll distributions are small but may prove regionally important. Simulations with interactive alkalinity show that the current OCMIP scenarios may be overestimating the impacts on the global ocean.
« Hide Abstract

Abstract:

Most data assimilation algorithms are based on classical statistical methods that adopt pixel-oriented descriptions of natural phenomena. While these methods are convenient they can have difficulty capturing the distinctive spatial and temporal features of special interest in geoscience applications. Such features include storm systems, ocean currents, algae blooms, geological formations, and wildfires. The problem is that the spatial structure that is clearly evident in nature may not be adequately conveyed through the means and covariances of large pixel-oriented state vectors. In this talk we consider how methods from computer vision and machine learning might be applied to feature-based data assimilation problems. These methods typically adopt a Bayesian perspective that focuses on the population of possible features rather than the population of possible state vectors. The objective is to derive an ensemble of conditional features that are samples from the posterior Bayesian distribution (i.e. incorporate both observations and prior information). Prior information is conveyed by an ensemble of prior features generated from training images. These images may be postulated, derived from observations, or obtained from physically-based stochastic models. Posterior features can be generated with Monte Carlo sampling techniques that make relatively few assumptions but are currently computationally infeasible for large problems. The success of a feature-based data assimilation will depend on our ability to 1) characterize the distinctive properties of natural features in feature spaces that are significantly smaller than the original pixel-based space (e.g. using image compression techniques), 2) generate candidate samples (proposals) that incorporate measurement information while remaining physically realistic (i.e. consistent with the training image), 3) determine the prior probability and likelihood of candidate features (i.e. define probability measures over feature spaces) , 4) develop computationally efficient reduced-order models that make it possible to generate and test a large number of candidate features. In this talk we illustrate concepts and assess some of the major research challenges and opportunities posed by a feature-based perspective.
« Hide Abstract

Abstract:

We analyze the surface energy balance and the estimation of fluxes of heat and moisture from land to the atmosphere. We introduce a variational data assimilation scheme that use sequences of radiometric surface temperature measurements to estimate both surface boundary effects as well as moisture-related surface control on the partitioning between turbulent heat fluxes. The objective is to develop a methodology based on remote sensing data that will enable mapping of the energy balance components over extended areas. The surface moisture control on evaporation is captured by the dimensionless evaporative fraction (ratio of latent heat flux to the sum of the turbulent fluxes) which is nearly constant for near peak radiation hours on days without precipitation. The parameter capturing the turbulent transfer characteristics (bulk scalar turbulent transfer coefficient) includes the impacts of surface roughness effects and atmospheric stability conditions. The approach is tested over the FIFE field experiment site (Kansas, USA, 1987 and 1988). It is shown that sequential radiometric surface temperature data contain useful information on the partitioning of available surface energy and may even be used to infer some key characteristics of surface turbulent transfer. The feasibility of extending the land data assimilation to use only sparse samples of radiometric surface temperature measurements (corresponding to overpass times of currently operational satellites) is demonstrated through an observing system simulation experiment. The data assimilation model is applied to the Southern Great Plains 1997 (SGP97) field experiment, using land surface temperature maps obtained from three satellites, namely, Advanced Very High Resolution Radiometer (AVHRR), the Special Sensor Microwave/Imager (SSM/I) and Geostationary Operational Environment Satellite (GOES). The application over the SGP domain gives reasonable estimates of surface fluxes, evaporative fraction and roughness-related parameters, confirming the potential of the land data assimilation scheme as an operational tool for the monitoring of surface energy balance and fluxes.
« Hide Abstract

Abstract:

In order to successfully validate and account for errors in remotely sensed estimates of near-surface soil water content, corresponding ground-based measurements must be accurate and representative. Complicating design of optimal sampling schemes for this purpose is the scale mismatch of the ground-based (point) measurements and the remotely sensed (areal average) footprint, and the fact that soil moisture patterns vary in both space and time. Our work attempts to identify underlying and time-invariant patterns in repeated soil moisture observations across and within successive seasons, and to characterize the drivers of those patterns. We are using traditional techniques like the Vachaud temporal stability analysis, along with emerging approaches like genetic algorithms and self-organizing maps. The overall goal is a methodology to understand the evolution of near-surface soil moisture patterns in a way that informs design of ground-based field campaigns of soil moisture monitoring. Preliminary results indicate that across-season variability is high, but that common patterns do exist and are to some extent predictable.
« Hide Abstract

Abstract:

Atmospheric deposition is a major source of mercury to aquatic and terrestrial ecosystems. Source attribution for mercury deposition is complicated by the presence of both natural and anthropogenic sources of mercury, and the cycling of mercury between land, ocean, and atmosphere. This study uses the GEOS-Chem global tropospheric chemistry model to quantify source contributions to mercury deposition and the role of long-range transport. The model is compared to observations to examine the consistency of a priori emission estimates with observed concentrations. This method provides improved constraints on modern mercury emissions from Asia and historic emissions from gold mining, both of which have large uncertainties.
« Hide Abstract

Abstract:

Ocean models commonly used as components of climate models, or coupled to models of ocean ecology, are implemented with resolution too coarse to resolve western boundary currents or eddies. Thus, signals present in observed data resulting from these phenomena cannot be usefully assimilated. The unmodeled part of the variability contributes to the model-data misfits and must be treated as error. This unmodeled variability is referred to as "representation error," since the physics of the model are inadequate to represent it. A method is proposed for estimation of the statistics of representation error, and results are presented to show that, at least in one case, our statistics are consistent with observation.
« Hide Abstract

Abstract:

Coupled climate model simulations reveal that a dipole like SST pattern and southward displacement of the ITCZ emerges in response to a slow-down of the Atlantic Meridional Overturning Circulation (AMOC). Using a 2-1/2-layer reduced gravity ocean model, we conducted a systematic investigation into oceanic processes controlling tropical Atlantic sea-surface temperature (SST) response to AMOC changes by varying the strength of northward mass transport at the open boundaries. It is found that a prominent equatorial warming occurs when the AMOC is weakened below a threshold value. A dynamic mechanism is proposed to explain how the AMOC change can affect the SST variability in the tropical Atlantic sector. The relative importance of oceanic versus atmospheric processes on AMOC-induced tropical Atlantic variability/change is further investigated within a framework of regional coupled model.
« Hide Abstract

Abstract:

Present-day shortcomings in the representation of clouds in general circulation models (GCMs) lead to errors in weather and climate forecasts as well as account for a source of uncertainty in climate change projections. An ongoing challenge in rectifying these shortcomings has been the availability of adequate, high-quality, global observations that discriminate and provide vertical structure of liquid and ice clouds, and related precipitating hydrometeors. In addition, the inadequacy of the modeled physics and the often-disjointed nature between model representation and the characteristics of the retrieved/observed values have hampered GCM development and validation efforts from making effective use of the measurements that have been available. Thus, even though parameterizations in GCMs accounting for cloud ice and liquid processes have, in some cases, become more sophisticated in recent years, this development has largely occurred independently of the global scale measurements. With the relatively recent addition of satellite- derived products (e.g., CloudSat, MLS, CALIPSO), there are now considerably more resources with new and unique capabilities to evaluate GCMs. In this presentation, we illustrate the shortcomings evident in model representations of cloud ice and liquid through a comparison of the simulations assessed in the IPCC Fourth Assessment Report, briefly discuss the range of global observational resources that are available, and describe the essential components of the model parameterizations that characterize their ice and liquid "clouds" and related fields. Using this information as background, we discuss some of the main considerations and cautions that must be taken into account in making model-data comparisons related to clouds from these satellite fields, illustrate present progress and uncertainties in applying these data -- specifically judiciously filtered versions CloudSat - to model diagnosis, and finally discuss a number of remaining questions and suggestions for pathways forward.
« Hide Abstract

Abstract:

Biophysical responses of the upper ocean to hurricanes in the Gulf of Mexico were examined using satellite observations and model simulations. It was important that both satellite observations and model simulations were used, since satellite sensors were sensitive to cloud interference, heavy rainfall, and only provided measurements near the ocean surface during hurricane periods. This study utilized 1/25 degree nested HYCOM simulations, 1/20 degree biophysical model simulations, and various satellite observations including QuikSCAT scatterometer winds, SeaWiFS and MODIS chlorophyll-a concentrations, and AVHRR and TMI sea surface temperatures. Such data were used to (1) assess the ocean surface response to Hurricanes Katrina, Rita, and Wilma of 2005 in the Gulf of Mexico, (2) examine the evolving three-dimensional (surface and subsurface) ocean response to Hurricane Katrina, and (3) analyze ecosystem dynamics, plankton biomass, and plankton distribution during Hurricane Katrina. Satellite observations of biophysical responses associated with Hurricanes Katrina, Rita, and Wilma of 2005 illustrated sea surface temperature changes of 6-7 degree C, 4-5 degree C, and 5-6 degree C, and chlorophyll-a enhancement of 3 mg m-3, 2 mg m-3, and 4 mg m-3. The degree and orientation of the responses exhibited were greatly affected by the oceanic processes that occurred within the Gulf of Mexico, as well as the translation speed of each hurricane. Satellite-detected surface responses associated with Hurricane Katrina occurred within a region from 23.5-25.5N and 85-83W. Analysis of model surface and subsurface dynamics in this region revealed strong upwelling/downwelling, wind-driven currents dominating the surface circulation, and near-inertial oscillations following Hurricane Katrina. The storm generated sea surface temperature cooling of 3-4 degree C and salinity freshening of 0.1-0.2. Analysis of heat-budget terms in the mixed layer indicated that wind-driven mixing dominated net upper-ocean cooling during hurricane passage, whereas at the mixed layer base temperature changes were largely due to vertical advection. Biophysical model simulations revealed that large phytoplankton were most responsive to hurricane-induced turbulent mixing and nutrient injection, with increases in biomass along the hurricane track. Small phytoplankton, microzooplankton, and mesozooplankton biomass primarily shifted in location and increased in spatial extent as a result of Hurricane Katrina. Hurricane passage disrupted the distribution of plankton biomass associated with mesoscale eddies.
« Hide Abstract

Abstract:

The nearby planet Mars represents an actual foreign climate system, yet Mars has a climate that is similar to Earth's in many ways. The Martian climate system has proven feasible to observe and model, and thus provides a rare opportunity for validation within the study of alternative climates. Earth and Mars have similar rotation rates and, somewhat coincidentally, similar atmospheric dynamical length scales (i.e., deformation radius, scale height). These factors have fostered a tradition of planetary atmospheric scientists borrowing and adapting models from their terrestrial counterparts. We have attempted to further borrow and adapt a data assimilation system - an ensemble Kalman filter - for use with a Martian GCM. The available observations appropriate for data assimilation are from remote sensing platforms in sun-synchronous orbits. These data sets are very sparse compared to the observational coverage available on Earth. After providing a brief introduction and overview to the Martian atmosphere and its existing observational record, we will report on our progress directly assimilating radiances and the challenges encountered.
« Hide Abstract

Abstract:

River not only plays a significant role to form a part of hydrological circulation on the Earth linking continents and oceans, but also it is the major part of available renewable freshwater resource. Although the flux of it (discharge) has received much attention, its storage does not. Most of previous studies in terrestrial water storage estimations such as GRACE applications implicitly assumed that soil moisture and snow water are the only major TWS components. In the present study, we investigated the temporal variation of three major TWS components, i.e., soil moisture, snow water, and river storage by land surface modeling approach. The role of river storage in total terrestrial water storage (TWS) variation was evaluated in 29 river basins globally. We quantified the contributions of major water storage components, and the GRACE data were used to validate the simulated total TWS variation. It is found that the simulated TWS variations are more accurate when river is taken into account for each target basin except in dry regions. Rivers take different roles in various climatic regions and are clearly classified through the indices devised here. River storage explains up to 73% of total TWS variation at Amazon, while it is negligible in dry basins. It also acts as a "buffer" between the signals of snow water and soil moisture seasonal variations in snow dominant basins. River is thus an important storage component of TWS for a major basin, and neglecting it will lead to insufficient amplitude and mismatch in TWS seasonal variation.
« Hide Abstract

Abstract:

The Madden-Julian Oscillation (MJO) remains an embarrassment to global atmospheric models: well resolved yet poorly simulated, presumably due to physical process parameterization shortcomings. MJO dynamics are delicate enough (with multiple processes including background state effects) that realistic modeling is the only hope for a true and robust "understanding". MERRA is a new opportunity: assimilation of observations keeps the model's state realistic, while access to analysis tendencies data sets permits diagnosis of model physics errors. Hypotheses about these errors, testable through re-reanalysis of selected time periods with adjusted physics, may guide model development usefully.
« Hide Abstract

Abstract:

A successful GCM simulation of the location and intensity of the ITCZ is vital to a successful simulations of tropical climatic events such as El Nino, of monsoons, and of realistic hurricane tracks. Although there has been some recent progress toward achieving such a simulation, this progress has been obtained through empirical means. As such, there is a need to understand why these empirical solutions are highly helpful. In this seminar, the speaker will present his contribution to such an understanding. The speaker will first review the basic mechanisms he has found that determine the latitudinal location of the ITCZ. (A comparison of the speaker's ITCZ theory with the previous CISK-based ITCZ theory will be presented.) The concepts behind these mechanisms will then be used to explain why a successful ITCZ simulation in GCMs is so elusive and how systematic errors, including the well-known double-ITCZ bias, arise. This new explanation provides some insight into the empirical results mentioned above. If time permits, an unrelated topic-how to better present the wavenumber-frequency spectrum of tropical large-scale waves-will be presented.
« Hide Abstract

Abstract:

Climate change and the threats of extremes to human life and natural ecosystems constitute a fundamental concern. Reliable regional climate change projection sufficient for the application to impact assessment and adaptation studies are increasingly required by the policy community.

In the talk, an attempt at dynamical downscaling of future hydrologic projection under global climate change in Asia is addressed. We conducted present and future Asian regional climate simulations which were nested in the results of Atmospheric General Circulation Model (AGCM) experiments. The regional climate model could capture the general simulated features of the AGCM. Also, some regional phenomena such as orographic precipitation, which did not appear in the outcome of the AGCM simulation, were successfully produced. As a result of the changes in the synoptic flow patterns and precipitation under global warming, the increases of annual mean precipitation and surface runoff were projected in many regions of Asia. However, both the positive and negative changes of seasonal surface runoff were projected in some regions which will increase the flood risk and cause a mismatch between water demand and water availability in the agricultural season.

In Japan, to add spatial resolution to accurately assess regional climate change impacts, the climate research project (S-5, "Getting a feel for climate change") has started (FY2007-2011). Dr. Dairaku would like to introduce the downscaling project in Japan and a part of our preliminary verification of simulations using two regional climate models with 20km horizontal grid spacing which used Japanese 25-year ReAnalysis (JRA-25) as lateral boundary conditions.

« Hide Abstract

Abstract:

Japanese GOSAT (Greenhouse gases Observing SATellite) was launched on January 23, 2009 and measurements will begin in late March. One instrument on GOSAT, the Thermal And Near infrared Sensor for carbon Observation Fourier-Transform Spectrometer (TANSO-FTS) will be used to derive columns and profiles of CO2 and CH4. It measures reflected solar radiation in three narrow bands (0.76, 1.6, and 2 micron) and thermal emissions from the ground and atmosphere in a wide band (5.5-14.3 micron), with 0.2 cm^-1 spectral resolution.

The a priori states for CO₂ and CH₄ retrievals will be obtained from simulations with the NIES TM (National Institute for Environmental Studies Transport Model) with forecast meteorological parameters and climatological flux data. The error covariance matrices are also derived from the NIES TM. Model-simulated CO₂ distributions often have biases. For example, the simulated amplitude of the seasonal cycle in Northern mid-latitudes tends to be smaller than the observed amplitude (Yang et al., 2007). However, the discrepancy between model and observation is not fully understood. The effort of the present study is to reduce the uncertainty of NIES TM though the comparison with the other simulated data and observational data.

I will first show the covariance matrices for CO2 and CH4 derived from NIES TM for the actual operational retrieval algorithm. Next I will show comparisons of CO2 between GEOS-5 and NIES05. The CO2 time series at each grid point in 2002 and 2003 are simulated in both models using the protocol of TransCom continuous experiment (TransCom04). Some climatological fluxes of CO2 (e.g. Fossil98, Takahashi02 and CASA) are imposed in the simulations. The seasonal and diurnal cycles of both models show similar features and both have smaller variability than the ground-based observations. It is found that vertical transport processes are quite different between GEOS-5 and NIES05, especially in the convective regions, due mainly to the different convective schemes. These differences have a clear signature in the CO2 covariance matrices computed using the two models.

« Hide Abstract

Abstract:

Stratospheric-tropospheric coupling in the polar region, through the vertical propagation of planetary waves, is well known as a cause of stratospheric sudden warmings. For a long time it is has been known that the high-latitude sinking motion produces warming in the polar region, while cooling occurs in the tropical stratosphere by a compensating upward motion. What is not well known is that whether such a stratospheric meridional circulation or Brewer-Dobson circulation change can further affect equatorial troposphere. Here we will show some results of our work suggesting a measurable impact of the stratospheric circulation change on the equatorial convective activity.

A case study of the unique major warming of the southern hemisphere in September 2002 revealed that convective activity in the equatorial southern hemisphere (SH) increased following the stratospheric warming, while that in the northern hemisphere (NH) was suppressed. A composite analysis of mid-winter sudden warmings in the NH also showed an impact on the tropical convection. The result of a preliminary general circulation model forecast study of the sudden warming in December 2001 also supports the notion of an impact on the tropical convection. For further understanding, it is necessary to investigate how the modification of cloud processes takes place in the lower stratosphere/upper troposphere region.

« Hide Abstract

Abstract:

Quantitative assessments of large-scale precipitation over the world's oceanic regions are problematic, particularly for significant regions of the data-sparse Southern Hemisphere. Available data sets are based on the assimilation of land-based measurements, satellite radiance values, numerical weather forecast models, or some combination of the three. In this study we examine several products that cover most or all of the satellite era 1979-2007 over the Southern Ocean and surrounding mid-latitudes to 45S. These include CMAP, the NCEP Reanalysis II, ERA-40, GPCP version 2, and the Japanese Re-analysis. Averaged fields from these data show large discrepancies in the mean spatial depiction and the annual cycle. Comparisons with unique in situ snowfall measurements are presented. The available record of oceanographic measurements in the Ross Sea indicates that salinity below 200 m in the Ross Sea has decreased by 0.03 per decade since 1958, with the highest (lowest) values in 1967 (2000). The fields examined here suggest that precipitation is likely not providing a directly influence on oceanic freshening in coastal seas adjacent to Antarctica. The salinity anomaly is consistent with increasing attrition of continental ice. Potential contributions to oceanic freshening from changes in sea ice extent, transport, and thickness are discussed.

« Hide Abstract

Abstract:

A high-resolution global climate model with a horizontal resolution of about 20 km has been developed and utilized to simulate seasonal mean anomalies, Madden and Julian oscillation (MJO), synoptic transients, Typhoon, and diurnal cycle. The six ensemble simulations for 1997 (El-Nino) and 1999 (La-Nina) summers show that the tropical anomalies with time scales longer than few weeks such as ENSO anomalies and MJO are not much affected by the change of horizontal resolution, indicating that those anomalies are mainly controlled by the physics. However, shorter-term variations appear to be very much affected by the resolution. The transients are well represented by the high-resolution model and the streamfunction tendency due to transient vorticity flux divergence is greatly improved by increasing the resolution, which affects the seasonal mean anomalies in the extratropics. The benefit of high resolution model is mainly in the simulation of extreme events such as Typhoon. The typhoon structure and its statistics within the seasons are reproduced very well by the high-resolution model. In particular, the ENSO differences in the location of Typhoon generation over the western Pacific are remarkably well reproduced. Also, preliminary results of high-resolution ocean-atmosphere coupled model will be presented in the seminar.
« Hide Abstract

12/22/2008

12/12/2008

12/09/2008

08/05/2008

07/28/2008

07/24/2008

07/16/2008

04/10/2008

04/07/2008

03/31/2008

03/10/2008

03/10/2008

03/03/2008

02/25/2008

12/06/2007

06/4/2007

01/19/2007

11/28/2006

Xiaohong Liu

11/02/2006

Brian Mapes

10/05/2006

Dara Entekhabi

09/07/2006

Rui Xin Huang

07/20/2006

Shaoqing Zhang

05/11/2006

Roger Saunders