Strengths and Limitations of First-Generation Reanalyses for Understanding Climate Variability and Trends II Posters
Presiding: R S Webb, NOAA Climate Diagnostics Center; J D Laver, NOAA Climate Prediction Center; R M Dole, NOAA Climate Diagnostics Center; P Arkin, Earth System Science Interdisciplinary Center, University of Maryland at College Park
A41C-01 0830h
North Atlantic Oscillation: The Non-Linear Aspect and Its Long-Range Prediction
The current prediction skill on seasonal to inter-annual time-scales is only marginal, yielding around 0.15 in anomaly correlation metric. The climate variability models are not performing better. How to achieve a breakthrough in long-range prediction is the goal of this investigation. Instead of the conventional linear and statistical thinking, our new approach focuses on non-linear characteristics and physical understanding. A reduction of the intertwined high-dimensional atmospheric state to a few manageable climatic circulation regimes is found to be a crucial consideration in improving the long-range predictions. For 1-yr lead forecast of the North Atlantic Oscillation, which is known to be extremely difficulty owing to its being composed of mostly natural variability, the present innovative approach is able to yield an average anomaly correlation skill of about 0.27. Cross-validation has been carefully exercised during this study.
A41C-02 0830h
Using Reanalyses to quantify sensitivities in the simulation of Microwave Sounding Unit temperatures.
Considerable scientific attention has been devoted to comparisons of observed and simulated Microwave Sounding Unit (MSU) temperatures. When simulating synthetic MSU temperatures, most studies apply a single static weighting function to the profiles of temperature. A variant of this procedure relies on different weighting functions for land and ocean regions to account for differences in surface emissivity values. A small number of investigations have also used full radiative transfer codes to derive synthetic MSU temperatures from model and/or reanalysis data. To date, few studies have rigorously quantified the sensitivity of simulated MSU temperatures to the choice of simulation method. Here, we apply both the static weighting function and full radiative transfer code approaches to atmospheric temperature data from multiple reanalyses. Our motivation is to document the sensitivity of estimated synthetic MSU temperatures to a variety of different processing options. The static weighting function approach is applied directly to monthly-mean pressure level data. Surface pressure is used to mask values that reside below the material surface. In addition to pressure level temperatures, the view-angle resolving radiative transfer code also utilizes skin temperatures and profiles of specific humidity. Regions of high topography are masked with surface pressure data, and there is explicit treatment of surface temperature data when surface pressure exceeds the lowest standard pressure level output (e.g., 1000hPa). Additionally, a land/sea mask is used to specify differing surface emissivity values. Further processing choices relate to the vertical resolution of the input temperature data and the treatment of upper-level moisture. Results shown will highlight observed and simulated temperatures from the NCEP/NCAR, NCEP/DOE and ERA40 reanalyses data for MSU 2, MSU2LT and MSU4. This work is supported under the auspices of the Office of Science, U.S. Department of Energy at the University of California Lawrence Livermore National Laboratory under contract No. W-7405-Eng-48
A41C-03 0830h
The Thermodynamic Structure of the Saharan Air Layer Simulated in the NCEP/NCAR Reanalysis
The propagation of the Saharan air layer (SAL) across the tropical North Atlantic Ocean during boreal summer is related to the atmospheric variability associated with the African easterly wave. There is evidence showing the suppression of tropical cyclone (TC) activity that encounters the SAL. A model that does not appropriately simulate the SAL will therefore tend to overestimate the intensity of TCs that are interacting with the SAL. In this study, we apply the dust amount observed in the Moderate Resolution Imaging Spectrometer (MODIS) as a proxy to track the location of the SAL. Composites of NCEP/NCAR reanalysis temperature and moisture profiles are constructed to illustrate the thermodynamic structure of the SAL simulated in the reanalysis data. Comparison of the thermodynamic structure of the simulated SAL in the eastern and central tropical North Atlantic with the known sounding structure of the SAL can evaluate the ability of using reanalysis data to study the climatic effects of the SAL. Finally, principle component analysis (PCA) is applied to the reanalysis temperature at 850 hPa for August-September 2002 to identify the low-level temperature variation associated with the African easterly wave. The geographical pattern of the warm anomalies will then be compared with available MODIS daily observations of dust distribution for the same period as an evaluation of the propagation of the SAL simulated in the reanalysis data.
A41C-04 0830h
The Use of NCEP Reanalysis in Improving our Understanding of African Climate Variability
The base state and the variability of the African climate on intraseasonal, interannual and interdecadal time scales as seen from the NCEP reanalysis are presented. The focus is on the West African monsoon, the Greater Horn of Africa, and the southern Africa climate systems. Some aspects of our current level of understanding of these climate systems are reviewed and compared with earlier work from the pre-reanalysis era. Limitations in the NCEP reanalysis in understanding the underlying physics associated with the African climate variability are also covered.
A41C-05 0830h
Evaluation of soil moisture in the FSUCLM using two Reanalyses (R2 and ERA40) and in-situ observations
The simulated soil moisture from the Florida States University (FSU) climate model coupled to the community land model (CLM2) is evaluated using two reanalyses (NCEP/DOE R2 and ERA40) and in-situ observations over Illinois and China from the Global Soil Moisture Data Bank. While the soil moisture was prescribed in the previous FSU climate model, the implementation of the CLM2 within the FSU model (hereafter, FSUCLM) provides a prognostic 10 layer soil moisture parameterization (both soil liquid water and soil ice). The comparison of soil moisture annual cycle between the FSUCLM and the two reanalyses over the period of 1992-1996 for the top 10cm shows that, in the Northern Hemisphere (NH), the coupled model and the ERA40 are in phase, but out of phase with the R2. In the Southern Hemisphere (SH), all of them are in phase. This result might be due to the fact that soil ice plays a key role in determining the soil moisture variation over the NH. An accurate soil moisture simulation requires the model to account for freezing process. While a parameterization of soil ice process is included in both the FSUCLM and the ERA40, the R2 does not. The two reanalyses and the FSUCLM are also compared to in-situ observations in China and Illinois. The soil liquid water component from the FSUCLM is used here, since not soil ice but soil liquid water soil is observed. The FSUCLM turns out to be following the observed variations better than both reanalyses. It is thus shown that the inclusion of explicit soil ice treatment in a land surface model has an impact on soil moisture annual variability.
A41C-06 0830h
Comparison and Validation of Latent Heat Fluxes From NWP Model Analyses Over the Tropical Pacific
Latent heat fluxes estimated by the NCEP/NCAR and NCEP/DOE reanalyses, ECMWF reanalyses and operational analyses are assessed by comparison with the fluxes derived from buoys measurements in the tropical Pacific covering the period 1991-2002. The buoys latent heat fluxes are obtained using the bulk flux algorithm COARE3.0 (Fairall et al., 2003). The mean differences between the buoys fluxes and the fluxes estimated from NCEP/NCAR, NCEP/DOE, ECMWF reanalyses and operational analyses are about 4, 22, 11 and 30W/m2, espectively, indicating the NCEP and ECMWF models tend to overestimate latent heat loss to the tropical ocean. Furthermore, NCEP/DOE and ECMWF reanalyses reveal a significant upward trend of the regional mean latent heat fluxes during 1991-2002, whereas the other two analyses as well as the buoys fluxes indicate a weak upward trend. Temporal correlation analysis shows ECMWF reanalyses is the best in comparison with the buoys fluxes. Using the NWP analyses surface layer meteorological fields and COARE3.0 algorithm, the latent heat fluxes were recalculated and the revised latent heat fluxes are found to be significantly reduced in comparison with the original ones. However, temporal correlations between the revised and the original fluxes are extremely high for all the cases except for ECMWF reanalyses, of which the revised fluxes before 1996 are much higher than the average while the original fluxes are lower than the average. In comparison with the buoys fluxes, the mean revised NCEP/NCAR fluxes is found to be underestimated due to a weak wind. On the other hand, the mean revised ECMWF operational fluxes is still overestimated due to a dry bias. The mean revised NCEP/DOE latent heat fluxes is closed to the buoys average. And the revised ECMWF reanalyses fluxes agree very well with the buoys fluxes after 1996.