- Using the Web Site
- Review of Available Wave Data Sources
- The NMIMET Method of Analysis
- Wind Statistics
- Scatter Tables (Wave Height and Period)
- Extreme Wave Height Estimation
- Persistence of Storms and Calms
- Reliability and Validation
- Validation of the NMIMET Analysis
- Wave Height Data
- Wave Period Data
- Data Quality Control
- Comparison with 'Ocean Wave Statistics'
- Extreme Wave Height Validation
- Persistence Analysis Validation
- European Database
- Neil Hogben
Comparison with 'Ocean Wave Statistics'
Readers with experience of using data from 'Ocean Wave Statistics'  (or 'Hogben and Lumb' as it was colloquially referred to) may wish to assess the relative level of reliability, and to make comparisons with data in Global Wave Statistics Online. There are four main reasons why the data in the new database are more reliable than those in 'Ocean Wave Statistics'.
Firstly the number of observations now available is far greater. This has made it possible to derive data based on much larger samples of observations covering longer periods of years for more than twice as many areas.
The other three reasons are all concerned with features of the NMIMET method of analysis. The first is the use of wind observations as a basis for enhancing the reliability of the wave height data. The number of wind observations available is very much greater than the number of wave observations, and they also cover a much longer period of years. Secondly the NMIMET analysis used only the so-called '2-Group' observations where both sea and swell heights have been reported and the two heights H1 and H2 have generally been combined to yield a resultant height:
In 'Ocean Wave Statistics', however, all wave observations were included but where two groups were reported, only the higher was used. The data in 'Ocean Wave Statistics' is thus biased towards lower wave heights, especially in areas where there are high levels of swell. Thirdly, the NMIMET analysis derives the wave period statistics by modelling the joint probability with wave height which is based on analysis of measured data. It thus avoids any use of visual observations of wave period, which have been shown to be very unreliable.
It is considered therefore that the data in this database are much more reliable, and there may be quite large differences between the two in the statistics both of heights and periods. The comparisons shown for corresponding areas in the accompanying Figure 24 and Figure 25 confirms this assessment.
Figure 24 - Comparisons with 'Ocean Wave Statistics', Northeastern Atlantic: Wave Height Probabilities
Figure 25 - Comparisons with 'Ocean Wave Statistics', Northeastern Atlantic: Wave Period Probabilities
Figure 24 and Figure 25 show probability distributions of both wave heights and wave periods for an area of the North-Eastern Atlantic which is the same in both. In the case of the wave height comparison, it may be seen that, as expected due to the high incidence of swell in this area, the data from 'Ocean Wave Statistics' show a substantial bias towards lower wave heights in comparison with the data from this database. In the case of the wave period comparison, the data from 'Ocean Wave Statistics' have been interpreted as equivalent to statistics of modal period Tp, and converted to the base of zero crossing period by assuming Tp = 1.4Tz. It may be seen that they differ from the data from this book, to an extent which is comparable with the differences between the raw visual data and the NMIMET and instrumental data shown in Figure 15 to Figure 23.