- 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
Wave Period Data
Results for wave period are shown in Figure 15 to Figure 23. In this case there are only seven comparisons because relevant period data were not available from the NOAA buoy. The interpretation of these comparisons is complicated by the number of different ways in which wave period can be defined.
It will be seen that the comparisons are plotted on a base of 'zero crossing period'. This is usually denoted by the symbol Tz and is defined as the mean time interval between upward or downward zero crossings on a wave record. Most published instrumental data are presented in terms of Tz and the NMIMET period data given by the program are intended to be equivalent to statistics of zero crossing period.
With the raw visual data, account was taken of the fact that visual observations of wave period are widely regarded as corresponding most nearly to a period known as the 'modal period' . This is the period associated with the peak of the wave energy spectrum, and may be denoted by the symbol Tp. On the basis of data which may be found for example in  it was thus assumed, for the purpose of the comparison, that visual observations of period are equivalent to estimates of Tp, and that Tp = 1.4 Tz. In most of the Figure 15 to Figure 23 it may be seen that the NMIMET results agree quite well with the instrumental data but the raw visual data do not. These results are typical of all the comparisons made so far, further details of which may be found in  - .
There are a number of points to which attention should be drawn.
Instrumental data are generally regarded as the most reliable, but the quality can vary quite widely and, in particular, estimates of zero crossing period can be very sensitive to differences in methods of analysis and to the distorting effects of 'noise' in the records. It is therefore considered preferable to use modal periods as a basis for comparison when they are available. The comparisons published in  -  are, in fact, presented in terms of modal period using the assumed relation Tp = 1.4Tz, for conversion of instrumental data available in terms of zero crossing period.
The NMIMET data have been derived from visual observations by methods based on parametric modelling of the joint probability distribution of wave heights and periods, and make no use of visual observations of period. The parameters used have been derived by regression analysis of over 20 sets of measured wave height and period data, mostly from sites in North-West European waters. It is thus particularly reassuring to note the good agreement between the NMIMET and instrumental data for the Cobb Seamount location, which is in the North-East Pacific.
In the case of the Sevenstones Lightvessel (Figure 19 and Figure 20) the results shown using instrumental data for the seven-year duration 1968 to 1974 are in better agreement than those using the data for the one-year period 1962 to 1963.
The comparison for the Slangkop Point location is also of special interest and calls for some comment. Here instrumental data were available in terms of both zero crossing and modal period, and it may be seen that the NMIMET data agree well when compared on the basis of modal period (assuming Tp = 1.4Tz), but not on the basis of zero crossing period. The good agreement in terms of modal period may seem surprising, in view of the relatively poor agreement with the instrumental data of the NMIMET wave height distribution used for deriving the period statistics. This may be explained however by noting that the modelling used depends heavily on the NMIMET value of the mean wave height, which is in reasonable agreement with the instrumental data.
The raw visual data are based on the assumptions that visual estimates of wave period should be comparable with modal period, and that Tp = 1.4 Tz. There is evidently very wide deviation from the instrumental data, and results presented in  show that, when other assumptions are made regarding the interpretation of visual estimates of period, such as equating them to Tz, the comparison with instrumental data still shows large deviations.
Figure 15 to Figure 23 thus underline the importance of the use of the NMIMET analysis by demonstrating not only that the wave period data so derived agree well with instrumental statistics of zero crossing period, but also that visual observations of wave period are unreliable. The validation results illustrated by the comparisons with instrumental data given in Figure 6 to Figure 23 offer considerable confidence in the reliability of the statistics of wave height and of wave period contained in this database. It should perhaps be noted, however, that the good agreement of the respective separate distributions of height and period does not guarantee a correspondingly good agreement of the joint probability distributions.