The NOAA CPO Modeling, Analysis, Prediction, and Projections (MAPP) program will host a webinar on the topic of Reanalysis Research: Laying the Groundwork for the Next Generation of NOAA Reanalyses on Tuesday, June 24. The announcement is provided below; you are invited to remotely join the session.
Gil Compo -- The NOAA Climate Reanalysis Task Force: Activities and Examples from Stratospheric Ozone and the Pacific Walker Circulation -- An introduction to the new NOAA Climate Reanalysis Task Force will be given. The Task Force addresses outstanding issues in atmospheric, oceanic and land reanalysis and develops a greater degree of integration among Earth system reanalysis components. Two examples of outstanding issues will be discussed. The first example is the representation of stratospheric ozone in current and planned NOAA reanalyses spanning 1850 to present. The current stratospheric ozone representation in NOAA’s Global Forecast System atmosphere/land model uses a linearized parameterization of ozone production and loss from the Naval Research Laboratory. The linearization is taken about a climatology representing late 20th Century chemistry, which includes ozone-depleting chlorofluorocarbons (CFCs). Its success in representing ozone variability before CFCs, and plans to improve the parameterization for representation of 19th to 21st century ozone variability, will be discussed.
The second example addresses whether the Pacific Walker Circulation (PWC) has weakened or strengthened since 1900. Some researchers have suggested that observations of the PWC spanning the last century provide evidence that the global convective mass flux is decreasing. Global coupled climate models show a decrease in the PWC from the last century extending into the next. The debate surrounding this issue is complicated by different investigators using different indices to define the PWC, with some based on using both the rotational and divergent components of the tropical winds to diagnose what is in essence a divergent overturning circulation. The influence and effect of tropical sea surface temperatures (SST) is also a confounding issue. We find that, in contrast to coupled climate models, most observed aspects of the PWC show no trend or a strengthening over the last 120 years.
Arun Kumar -- Research Towards the Next Generation of NOAA Climate Reanalyses -- Reanalysis of various components of the Earth System provides global data sets to advance our understanding of climate variability and trends, and helps place the state of current climate system in a historical context. One important aspect of reanalyses efforts is to understand causes of spurious climate trends and discontinuities that arise due to changes in observational data platforms, and to develop data assimilation and bias correction techniques that can minimize those. The latest generation of NOAA’s reanalysis – the Climate Forecast System Reanalysis (CFSR) – was plagued by various such discontinuities. A joint research effort across NOAA involving NCEP, ESRL, and NCDC plans to advance data assimilation techniques to reduce the influence of changes in the observational data platforms in the context of climate reanalysis. A unique aspect of the proposed approach is to use a hierarchy of reanalysis of increasing complexity where analysis based on a simpler approach can be used to inform bias correction procedures for an analysis involving higher complexity.
Tom Hamill -- Medium-range reforecasts as a driver for a modern-era reanalysis capability -- Reforecasts are retrospective numerical simulations of the weather, typically initialized with reanalyses and typically using the same version of the forecast model that is used operationally. Reforecasts have been shown to be very useful for the statistical post-processing (calibration) of weather forecasts; the current forecast is adjusted using the discrepancies between past forecasts and observations. The large sample size of reforecasts is particularly helpful for the calibration of more rare events such as heavy precipitation and for the calibration of longer-lead forecasts, where the forecast is contaminated both by random noise due to growth of initial errors and by systematic errors from modeling system deficiencies. Ideally, the reforecasts are initialized with reanalyses that are very similar in accuracy and bias characteristics to the operational analysis. In this talk we will discuss some of what has been learned through sample-size sensitivity experiments using the most recent, 2nd-generation reforecast data set for the NCEP Global Ensemble Forecast System. They suggest that at least 20 years of reforecasts are optimal for calibration, meaning that a modern-era reanalysis spanning at least 20 years is needed. The talk will also touch on issues such as the changing accuracy characteristics of the reanalysis and its effect on reforecast accuracy.