The ERA-40 ocean wave product has been intensely validated against observations and other
reanalysis products. The data is of high quality, especially in terms of monthly means and
other longer-term statistics, when compared with competing datasets.
The assessment of the data focused on the 10-metre wind speed, the significant wave height,
the mean wave period and the mean wave direction. Comparisons with observations were
done mainly from 1978 onwards when buoy observations start being available but the
comparisons are restricted to the Northern Hemisphere. Significant wave height and wind
speed data from 1987-1989 and from 1993 onwards were also assessed against global
altimeter observations, from the Geosat and TOPEX altimeters respectively. Comparisons for
the period before 1978 were done just at one location in the Northern Hemisphere.
In terms of their error characteristics the ERA-40 wave data needs to be divided into 4 sets:
- Data for the periods from 9-1957 to 11-1991
and from 6-1993 to 12-1993 had no altimeter
wave height data assimilation.
- ERS-1 Fast Delivery Product (FDP) altimeter wave-height data were assimilated in ERA-
40 from December 1991 onwards. The data are, however, of poor quality during the first
two years due to an external processing error. Assimilation of the FDP data was halted as
soon as this problem was realised, production having reached May 1993.
- Assimilation was resumed in January 1994 using
good but uncalibrated ERS-1 FDP data
up to May 1996.
- FDP ERS-2 measurements of wave height have been
assimilated in ERA-40 from June
1996 onwards.
In the first period the monthly mean wave fields compare well with observations, but the
analyses exhibit peaks in synoptic significant wave heights timeseries that are lower than
observed and there tends to be also some overestimation of the low troughs. The mean wave
period data has in general a root-mean square error below 2 seconds with the higher errors
occurring in the Tropics where along with the North Pacific the periods are overestimated; in
the North Atlantic there is a tendency for underestimation of the mean wave periods. In the
second period the significant wave heights below 3 metres are overestimated and those
above underestimated. The quality of the waves with heights above 3 m is similar to that in the
period with no wave data assimilation. The mean wave periods are overall overestimated. In
the third period the known calibration correction to the ERS-1 FDP data was not applied
because, although it would have improved analysed wave heights, it would have given poorer,
too high mean wave periods. The quality of the wave height data is therefore similar to the one
of the data in period one. The quality of the mean wave period data is quite high, with monthly
root-mean-square errors often below 0.5 seconds. The assimilation of the ERS-2 FDP
measurements of wave height has improved the wave-height analyses, especially in the
region of the tropics. The mean wave periods do not compare as well with observations as
they do for the third period.
Motivated by deficiencies in the ERA-40 significant wave height dataset, the data was
corrected through a non-parametric method that predicts the bias between significant wave
height ERA-40 data and TOPEX altimeter measurements. Thus creating a new 45-year global
6-hourly dataset--the C-ERA-40 dataset. Comparison of the corrected data with significant
wave height measurements from in-situ buoy and global altimeter data shows clear
improvements in both bias, scatter and quantiles in the whole range of values.
More info in
article: "Validation and non-parametric correction of significant wave height data from the
ERA-40 reanalysis" (reprint available), Sofia Caires and Andreas Sterl, KNMI, De Bilt, The
Netherlands, Submitted to "Journal of Atmospheric and Oceanic Technology"