Figure 5. WIS results compared to NDBC 42036 for February 1995.

the low concentration statistic.

and skill score statistics look slightly better for the

WW3 results for Jan-Mar. WIS and WW3 both show

Figures 5-7 show some similarities between all three

the lowest mean correlation.

hindcasts. Hindcast wave period results are low in

times of low wave height as shown in all plots for

Tables 38-43 in the Appendix show comparison mean

February 17-19. A major front passed over the Gulf

period and peak period statistics for the three models

on February 4 and high winds on the back side of a

at NDBC 42036. Using Table 4, WAM shows a mean

LOW are noted on February 4. WIS does a good job

positive bias of 0.001 sec in comparison to a negative

of tracking the peak of this event while WW3 and

WIS bias of 0.05 sec. WW3 shows a negative bias

WAM are low. WIS and WAM both do a good job on

of 0.65 sec for Jan-Mar indicating under-prediction

the wave period associated with this event. A LOW in

of the mean period. WW3 shows the lowest mean

the southeastern Gulf around February 21 produced

RMS for Jan-Mar, 0.59sec. WW3 shows the highest

another event with high winds and a direction change.

correlation for Jan-Mar; all three models show similar

WIS shows a little over-prediction of this event but

skill score statistics although WAM shows some big

captures the before and after characteristics

differences with measured results in September (see

Table 38 in the Appendix). WW3 has slightly better

scatter index statistics for Jan-Mar. Peak period

correlations shown in Table 5 indicate WW3 has the

best statistics for Jan-Mar with a mean value of 0.68.

All bias values are negative indicating under-

prediction of the peak period. RMS values are all

about 1 sec. WW3 and WAM show equal mean

values for the lowest scatter index for Jan-Mar. All

the models show skill scores in excess of 0.96. Again,

WAM shows some problems in the September

hindcast of peak period.

6.4. DIRECTIONAL COMPARISONS AT 42036

Figures 5-7 show the line plots from the hindcasts for

February 1995 at NDBC 42036. Model results are

shown in blue and measured results are shown in red.

Figure 5. WIS results compared to NDBC 42036

Directional information was available for this month

for February 1995. Plots from top to bottom

and Table 7 shows the directional statistics for the

include significant wave height, wave period, and

three models using wave direction comparisons for

wave direction.

waves 1 m and above. WW3 shows the best

directional comparison with a mean directional

difference of 7.32 deg, a concentration statistic of

11.08 indicating that both the measured and

hindcasted directions represent similar distributions.

Circular correlation for WW3 is 0.906. WIS shows

results almost as good as WW3. WAM shows a

concentration statistic below 5.0 indicating that the

WAM directional distribution is different from the

measured distribution. The WAM wave direction

comparison plot in Figure 6 shows major directional

differences close to February 25, and this accounts for

Table 7. Mean Wave Direction Statistics for Gulf

Level 3 February 1995 at NDBC 42036

Model x (deg)

Circor

Number

WIS

-9.23

10.92

0.861

296

Figure 6. WAM results compared to NDBC 42036

for February 1995. Plots from top to bottom

WW3 -7.32

11.08

0.906

296

include significant wave height, wave period, and

WAM -12.92

3.50

0.712

296

wave direction.