Comparison of 4 Hindcast/Nowcast Results for 2010 Using PROFS (Princeton Regional Ocean Forecast System) # F.-H. Xu & L.-Y. Oey* Princeton University *

Comparison of 4 Hindcast/Nowcast Results for 2010 Using PROFS (Princeton Regional Ocean Forecast System)#

F.-H. Xu & L.-Y. Oey*Princeton University* [email protected]

# This document is available online at ftp://aden.princeton.edu/pub/lyo/gomex/XuOeyPROFS4hcastComparisons.pdf

Fig. 1a mpiPOM (Princeton Ocean Model) domain of the northwestern Atlantic Ocean including the Gulf of Mexico. Color is sea-surface height (SSH; in meters) on Jun/19/2010 – red indicates high SSH = 0.5m, and blue is low SSH = -0.5m. Squiggly black lines indicate the corresponding surface currents velocity in m/s (2 scales are shown for water depths deeper and shallower than 200m). Blue vectors indicate kinematic wind stress vectors in (m/s)2 (i.e. wind stress in N/m2 divided by water density ~ 1000 kg/m3). Grey contour is the 50m isobath, and white contour is 200m. See Fig.1b for main model features.

Fig.1b Highlights of the mpi-version of the Princeton Ocean Model now in use for GOMEX-PPP research in the Gulf of Mexico and the Mid-Atlantic Bight.

Note that we found some problems with GFS wind data from Jan/01~Mar/15/2010, and were too late to download the corrected wind from NCEP site at that time before they were permanently removed.

Therefore, ALL PROFS results for that period (Jan~Mar/15/2010) have no wind forcing. The effects for Loop Current and rings appear to be small when data assimilation is turned on. But, as we will show, effects on transport fluctuations through the Straits of Florida are significant.

Experiments:

Parameters:

hcastNoT

(Exp.1)

hcastT

(Exp.2)

hcast.horcon0p2

(Exp.3)

hcastT.horcon0p2

(Exp.4)

Tides* No Yes No Yes

Horcon# (eddy viscosity)

0.1 0.1 0.2 0.2

Horcon# (eddy diffusivity)

0.02 0.02 0.02 0.02

* Tides (if turned on in the model) include M2, S2, K1 & O1# Horcon = Smagorisnky coefficient of horizontal eddy viscosity and diffusivity (dimensionless)

Table 1. List of 4 hindcast/nowcast experiments conducted with and without tides and with and without increased values of the Smagorinsky horizontal mixing coefficient.

(A) hcastNoT (B) hcast.horcon0p2

Fig. 2a,b Effects of increased horizontal mixing. Daily-averaged SSH (colors in m), surface currents (squiggly black lines) and wind stress vectors (blue) shown for the 15 th day of each month from Jan-Jun in 2010 for (a) the hindcast experiment without tides and with horcon=0.1 (see Table 1), and (b) the experiment without tides and but with horcon=0.2. Magenta contour=AVISO SSH=0 and white contour=200m isobath.

(C) hcastNoT (D) hcast.horcon0p2

Fig. 2c,d Effects of increased horizontal mixing. Daily-averaged SSH (colors in m), surface currents (squiggly black lines) and wind stress vectors (blue) shown for the 15 th day of each month from Jul-Dec in 2010 for (c) the hindcast experiment without tides and with horcon=0.1 (see Table 1), and (d) the experiment without tides and but with horcon=0.2. Magenta contour=AVISO SSH=0 and white contour=200m isobath.

(A) hcastNoT (B) hcastT

Fig. 3a,b Effects of tides. Daily-averaged SSH (colors in m), surface currents (squiggly black lines) and wind stress vectors (blue) shown for the 15th day of each month from Jan-Jun in 2010 for (a) the hindcast experiment without tides and with horcon=0.1 (see Table 1), and (b) the experiment with tides and with the same horcon=0.1. Magenta contour=AVISO SSH=0 and white contour=200m isobath.

(C) hcastNoT (D) hcastT

Fig. 3c,d Effects of tides. Daily-averaged SSH (colors in m), surface currents (squiggly black lines) and wind stress vectors (blue) shown for the 15th day of each month from Jul-Dec in 2010 for (a) the hindcast experiment without tides and with horcon=0.1 (see Table 1), and (b) the experiment with tides and with the same horcon=0.1. Magenta contour=AVISO SSH=0 and white contour=200m isobath.

(A) SSHA (B) (Ug,Vg)

Fig. 4 Comparison between AVISO and modeled (a) SSHA (anomaly) and (b) surface geostrophic velocity (Ug,Vg). The “(a)” panels on left show top: spatial correlation coefficient between the model and AVISO SSHA’s in the region north of 23oN and west of 84oW, and over water region deeper than 500 m in the Gulf of Mexico for time period Jan-Dec/2010; and bottom: the corresponding RMS error for the same region. The “(b)” panels on right compare the corresponding (Ug,Vg): top & middle are magnitude and angle of the complex correlation and bottom is the RMS error. In each case, the 2010-mean’s are also shown. See Table1 for experiment names.

Fig. 5 Comparison between modeled transport at the Straits of Florida off the eastern coast of Florida between West Palm Beach, FL to Eight Mile Rock, Grand Bahama Island and cable-estimated observed transport (http://www.aoml.noaa.gov/phod/floridacurrent/data_access.php) for the time period Jan-Dec/2010. Note that model was not forced by wind from Jan/01 through Mar/15. The Table on left shows the 2010-mean’s for observed and modeled transports, and their zero-lag correlations and 95% significances computed for the entire 2010 from Jan-Dec (3rd column in the Table) and for Apr-Dec only (4th column in the Table).

Mean(Sv)

R /SJan-Dec

R /SApr-Dec

Obs. 30.5 - -

Exp.1 31.2 .32/.16 .52/.17

Exp.2 30.7 .34/.12 .57/.12

Exp.3 30.2 .39/.14 .71/.19

Exp.4 30.9 .38/.13 .60/.13

Exp.1= hcastNoTExp.2= hcastTExp.3= hcast.horcon0p2Exp.4= hcastT.horcon0p2

R=Correlation; S=95% Significance