A. Hernandez, R.W. Saunders, A.P. McNally, G. Kelly, S. Uppala and P. Kallberg

European Centre for Medium-Range Weather Forecasts

Shinfield Park, Reading, RG2 9AX, U.K.

1. TOVS/ATOVS DATA FOR ERA-40

A new reanalysis, ERA-40, has been initiated at ECMWF (Uppala et al., 1999). It will cover the period from mid-1957 to 2001. Radiance data from the NOAA series of polar orbiting satellites are a major component of the ERA-40 observing system.

The TOVS system consists of three independent instruments:

• the High resolution Infra-Red Sounder/2 (HIRS/2), with 19 channels in the region 3.7 - 15 m m plus a channel at 0.7 m m (visible),

• the Microwave Sounding Unit (MSU), with 4 channels in the region 50 to 58 GHz,

• the Stratospheric Sounding Unit (SSU), with 3 pressure-modulated channels at 15 m m (not flown on NOAA-10 and NOAA-12)

The ATOVS system consists of:

• the High resolution Infra-Red Sounder/3 (HIRS/3), with 19 channels in the region 3.7 - 15 m m plus a channel at 0.7 m m (visible),

• the Advanced Microwave Sounding Units AMSU-A with 15 channels and a frequency range of 23 to 89 GHz, and AMSU-B with 5 channels and a frequency range of 59 to 183 GHz.

Figure 1. Data availability and sources of data

Fig. 1 illustrates the availability of data for each type of satellite. The data have been acquired from several sources. The sources of TOVS data are NCAR, the Laboratoire de Meteorologie Dynamique (LMD), NASA, and the ECMWF operational archive, and the sources of ATOVS data are NASA and the ECMWF operational archive. TOVS data are available from October 1978 to date, from satellites TIROS-N, NOAA-6, NOAA-7, NOAA-8, NOAA-9, NOAA-10, NOAA-11, NOAA-12 and NOAA-14. ATOVS data are available from June 1998 to date, from NOAA-15.

The TOVS/ATOVS dataset for ERA-40 is in level 1b form. The data are raw radiance counts to which Earth location and quality information has been appended and calibration information appended but not applied.

2. USE OF TOVS/ATOVS DATA IN ERA-40

There are considerable differences in the use of TOVS data between ERA-15 and ERA-40. The analysis scheme of ERA-15 (Gibson et al., 1997) was Optimum Interpolation and could only use observations of variables linearly related to the analysis variables. In ERA-15 a 1D-VAR scheme was used (Eyre et al., 1993) to produce profiles of temperature and humidity from the TOVS Cloud Cleared Radiances (CCR) data. ERA-40 will use a 3D-VAR analysis scheme (Courtier et al., 1998) and will use TOVS/ATOVS level 1c raw radiances (McNally et al., 1999).

The pre-processing of the level 1b data to prepare the TOVS data for the assimilation is considerably less complex in ERA-40 than it was in ERA-15 (pre-processing performed by NESDIS). In ERA-40

• each instrument is assimilated on its natural scan geometry
• no limb adjustments to nadir needed
• no emissivity adjustment needed
• no adjustment to cloud clear conditions needed

As a result of the elimination of complex (and changing) adjustment steps, the systematic differences between observations and background are now smaller and more stable, which allows a more effective bias correction to be applied (Kelly et al., 1999).

Channel HIRS 9 takes radiance measurements in the ozone absorption band at 9.7 m< m, related to the ozone content of the atmospheric column. Raw radiances from channel HIRS 9 will be part of the ozone observations in ERA-40.

3. PREPARATION OF TOVS/ATOVS DATA FOR THE ANALYSIS

3.1 Pre-processing

In ERA-40 the pre-processing of TOVS/ATOVS data consists basically of the calibration of the radiance counts, i.e., in the conversion of the data from level 1b to level 1c, and it is being performed at ECMWF with software developed in-house. It includes some basic checks and a quality filter. The 1c radiances are then encoded in FM94 BUFR format.

3.2 Pre-assimilation quality monitoring of TOVS/ATOVS

To make the best possible use of the data in the re-analysis, it is necessary to know as much as possible about the behaviour of the data before the assimilation. In particular, it is essential to detect operational problems concerning specific channels, on specific satellites, for specific periods, and to address these problems before the assimilation. Several tools can be used for this purpose:

• The NOAA-NESDIS Polar Orbiter Archived TOVS Sounding Data Change and Problem Record (Kidwell, 1998) can provide information on periods of bad Earth location, noisy channels, channel or instrument failures, calibration problems, adjustments of calibration parameters, etc.

• The ERA-15 record and the tools developed for ERA-15 are expected be of some help (in identifying periods of bad Earth location, for instance), although as the TOVS dataset used in ERA-15 was CCR this help is likely to be quite limited.

• Time series of 6-hour mean raw radiances will be plotted (fig. 2) as part of the preparation of the data for production, with the aim of knowing the actual availability of data and helping in the detection of problems like gaps in the data, noisy or erratic behaviour of channels or instruments, or changes in the response of instruments. The final form of this tool is not yet decided, but the aim is to build a flexible system that produces some routine plots and additional information when required. The possible choice of soundings for the time series could be

  • latitude bands: choice of global, tropics, possibly others

  • scan positions: central, possibly others

  • only soundings over sea for channels sensitive to surface, possibly choice of underlying surface for the rest

  • only clear soundings for HIRS channels that peak in the troposphere

The time series plots will also include the satellite's height, to enable the observation of the orbital decay.

Figure 2. Example of monthly series of mean BT for some channels

4. ERA-40 TOVS/ATOVS ARCHIVE

The TOVS/ATOVS archive will be part of the ERA-40 archive (Gibson et al., 1999). It will contain:

• data as received from sources

  - size ~ 1 Tbytes

  - file structures and dataset naming conventions not entirely consistent

• ERA-40 1b-level structured archive

  - consistent file structures and formats (NESDIS)

  - consistent naming conventions (NESDIS)

  - size ~ 1 Tbytes

• ERA-40 1c-level structured archive

  - encoded in FM94 BUFR format

  - size ~ 0.5 Tbytes

• Quality monitoring plots

  - time series of 6-hour mean brightness temperatures

  - subset will be available from ERA-40 web pages

• Feedback statistics

REFERENCES

Courtier, P., E. Andersson, W. Heckley, J. Pailleux, D. Vasiljevic, M. Hamrud, A. Hollingsworth, F. Rabier, and M. Fisher, 1998: The ECMWF implementation of three dimensional variational assimilation (3D-Var). Part I: Formulation. Q. J. R. Meteorol. Soc., 124, 1783-1808.

Eyre, J.R., G. Kelly, A.P. McNally, E. Andersson and A. Persson, 1993: Assimilation of TOVS radiance information through one-dimensional variational analysis. Q. J. R. Meteorol . Soc., 119, 1427-1463.

Gibson, J.K., P. Kallberg, S. Uppala, A. Hernandez, A. Nomura and E. Serrano, 1997: ERA Description. ECMWF Re-Analysis Project Report Series, 1.

Gibson, J.K. , S. Uppala, P. Kallberg, M. Fiorino, A. Hernandez, K. Onogi and X. Li, 1999: ERA-40 Archive Plans. Proceedings of the Second International Conference on Reanalyses. Wokefield Park, Mortimer, Reading, UK, 23-28 August 1999.

Kelly, G., and A. Hernandez, 1999: The Problem of how to bias correct satellite radiances in ERA-40. Proceedings of the Second International Conference on Reanalyses. Wokefield Park, Mortimer, Reading, UK, 23-28 August 1999.

Kidwell, K.B., 1998: NOAA Polar Orbiter Data (POD) User's Guide, November 1998 revision. http://www2.ncdc.noaa.gov/docs/podug/

McNally, A.P., E. Andersson, G. Kelly, and R.W. Saunders, 1999: The Use of Raw TOVS/ATOVS Radiances in the ECMWF 4D-Var Assimilation System. ECMWF Newsletter 83, Spring 1999.

Uppala, S., J.K. Gibson, M. Fiorino, A. Hernandez, P. Kallberg, X. Li, K. Onogi and S. Saarinen, 1999: ECMWF's Second Generation Reanalysis - ERA-40. Proceedings of the Second International Conference on Reanalyses. Wokefield Park, Mortimer, Reading, UK, 23-28 August 1999.