This page shows the validation plots for Samoa.


Back to the Samoa Wave Atlas Page



Samoa Wave Model Validation

Model validation was carried out in order to check the models accuracy against the measured waves.

In Samoa, waves were masured for a long period of time at 2 different locations. Location 1 had measurements conducted from 01/09/1989 - 11/04/1990 while location 2 was from 16/05/1990 - 18/07/1992.

Wave Buoy Location Map

The map below shows the locations of the wave buoys south of Samoa in a region exposed to south swells and trade wind waves.


WACOP

Validation statistics


The tables below shows the Root Mean Square Error, skill and Bias of the model.

Location RMS Skill Bias
Location 1 0.3271 0.9079 0.1312
Location 2 0.3037 0.8697 0.0861

The models skill is very close to 0.9 which means the model is accurate. The model has a relatively large bias probably due to unresolved bathymetry features. The RMS is within the expected error for this type of model.

The scatter of the comparison and the quantile show a more detailed analysis of the models' performance.


Model vs measured waves


The scatter plot shows the divergence between the model and measurements. The spread of the scatter is relatively uniform across all the measured wave heights suggesting that the model is valid for wave heights between 0.0m and 5.0m. The quantile/quantile plot shows how the bias varies with different wave heights. The bias is relatively small for wave heights below 5.0m which corresponds to all the trade wind waves and swells. Larger waves are often created by nearby cyclones which are not well resolved in the model.

Scatter plots

Location 1

WACOP

Location 2

WACOP

Quantile/Quantile plots


WACOP

WACOP

Back to the Samoa Wave Atlas Page