MINIMIZING REANALYSIS JUMPS DUE TO NEW OBSERVING SYSTEMS

Loading...
Thumbnail Image

Files

Publication or External Link

Date

2014

Citation

Abstract

A major problem with reanalyses has been the presence of jumps in the climatology associated with changes in the observing system. Such changes are common in reanalysis products. These jumps became especially obvious when satellites were first introduced in 1979. After 1979, however, during the "satellite era" jumps have continued to appear whenever a new observing system was introduced. To explore possible solutions to this problem, we develop and test new methodologies to minimize these reanalysis jumps in the reanalyses time series due to new observing systems.

In the first part of this dissertation, we study a state-of-the-art reanalysis, NASA's Modern Era Retrospective-analysis for Research and Applications (MERRA thereafter). Analysis increments from the MERRA and from one reanalysis without SSM/I observations (NoSSMI thereafter) are compared and their differences are defined as correction terms. The correction terms are then introduced into the tendency equation of the forecast model, i.e., GEOS-5. The debiased reanalysis without SSM/I observation shows improvements in almost all fields, even in the precipitation field, which is generally considered to be significantly uncertain on all time and space scales. However, the difference between the analysis increments of MERRA and NoSSMI is not just due to the assimilation of SSMI, but to the accumulated effect of the assimilation of previous SSMI observations. These produce a change in the model climatology and nonlinear interactions between the variables currently observed by SSM/I, and the variables that have been modified by previous assimilations of SSMI. The nonlinear interactions introduce an additional accumulated impact during the 2-year training period.

In the second part of this dissertation, we test a new methodology in a simpler data assimilation system, SPEEDY-LETKF, because it would be unfeasible for our computational resources to apply this method to the complex MERRA system. The new method defines the correction terms by calculating the difference of analysis increments from the following two analyses, 1) assimilating both rawinsondes (RAOB) and AIRS observations, named RaobAirs, and 2) assimilating only RAOB but with its background coming from the RaobAirs analysis at every 6-hour analysis cycle. This new method limits the growth of nonlinear interactions between variables observed by AIRS and the variables that have been modified by previous assimilation of AIRS. The results show that the new method is significantly more effective in minimizing reanalysis "jumps" compared with the method applied to MERRA system.

In the third part of this dissertation, we explore a spectral model instability problem. Imperfect SPEEDY-LETKF OSSEs are unstable when assimilating RAOB observations only. Data assimilation processes worsen this problem. We found two methods to stabilize the imperfect SPEEDY-LETKF OSSEs. Traces of the spectral waves are also clearly present in other spectral reanalyses such as the NCEP and the ERA15, but since their resolutions are higher than that of the SPEEDY model, their impact is smaller.

Notes

Rights