WAVEWATCH III® 30-year Hindcast Phase 2

Summary

A 30-year wave climatology has been generated with the NOAA WAVEWATCH III® using the Ardhuin et al (2010) physics package, 15 regular lat-lon grids, and the NCEP Climate Forecast System Reanalysis and Reforecast (CFSRR) homogeneous dataset of hourly high-resolution winds. The time period covers from 1979 through 2009.

Data Set Access and Validation

• Phase 2 Data*
  
• Phase 2 Validation

* Important Note: The Phase 2 Hindcast data set was recently moved from our FTP server to our web server. Please update your bookmarks and links as needed with the new location.

Some browsers (Firefox, Chrome) may try to display the contents of a GRiB file in the browser window instead of downloading it. If that happens, right-click on the GRiB file link and select "Save Link As..." to download the file.

Model Setup

• Propagation scheme: Higher-order schemes with Tolman (2002a) averaging technique (PR3)
• Linear input: Cavaleri and Malanotte-Rizzoli with filter (LN1)
• Nonlinear interactions: Discrete interaction approximation (NL1)
• Bottom friction: JONSWAP bottom friction formulation (BT1)
• Depth induced breaking: Battjes-Janssen (DB1)
• Use Miche-style shallow water limiter in equation for maximum wave energy (MLIM)

The model was run with the Ardhuin et al (2010) source term package (ST4) which includes the flux computation in the sources (FLX0, STAB0). Additionally, the model uses a third order propagation scheme (UQ), with no damping or scattering by sea ice (IC0, IS0), and no reflection (REF0).

Model Grid Descriptions

The wave model suite consists of global and regional nested grids. The rectilinear grids were developed using ETOPO-1 bathymetry (Amante and Eakins, 2009), together with v1.10 of the Global Self-Consistent Hierarchical High Resolution Shoreline (GSHHS) Database. The higher-resolution grids have been masked to improve model efficiency, and only provide data near shore. See Chawla and Tolman (2007,2008) for details on the software used for developing these grids.

• Grid bathymetry data sets (NetCDF)

Note: Model was run using the Global (glo_30m) and Arctic Ocean (aoc_15m) grids, but the results are interpolated onto the glo_30m_ext (and not given on the other two grids).

Name (click for map)	ID	Nx	Ny	dx	dy	Lon0	Lat0
Global 30 min	glo_30m	720	361	1/2	1/2	0	-90.0
Arctic Ocean curvilinear	aoc_15m	unstructured grid
Gulf of America and NW Atlantic 10 min	ecg_10m	301	331	1/6	1/6	260	0.0
US West Coast 10 min	wc_10m	241	151	1/6	1/6	210	25
Pacific Islands 10 min	pi_10m	511	301	1/6	1/6	130	-20
Alaskan 10 min	ak_10m	401	187	1.5/6	1/6	140	44
North Sea Baltic 10 min	nsb_10m	237	157	3/12	2/12	-28	42
Mediterranean 10 min	med_10m	301	109	1/6	1/6	-7	30
North West Indian Ocean 10 min	nwio_10m	241	307	1/6	1/6	30	-20
Australia 10 min	oz_10m	361	301	1/6	1/6	105	-50
Gulf of America and NW Atlantic 4 min	ecg_4m	586	481	1/15	1/15	261	15
US West Coast 4 min	wc_4m	736	526	1/15	1/15	195	15
Alaskan 4 min	ak_4m	548	391	2/15	1/15	165	48
North Sea Baltic 4 min	nsb_4m	443	391	2/15	1/15	-28	42
Australia 4 min	oz_4m	901	751	1/15	1/15	105	-50

Model input

• High resolution global winds at 10m height from the NCEP Climate Forecast System Reanalysis (CFSR) - a coupled reanalysis of the atmospheric, oceanic, sea-ice, and land data - were used for this study. A bias-correction based on collocated altimeter data was implemented to reduce the bias in the strongest winds, especially over the Southern Ocean. The wind fields used are 1/2 degree resolution at 1 hour intervals.
• Hourly air-sea temperature difference with 1/2 degree spatial resolution, also from the CFSRR.
• Ice concentrations are obtained from passive microwave sea ice concentration from the SMMR and SSMI, using the NCEP high resolution algorithm based on NASA Team2. The ice fields used are daily with 1/2 degree resolution.

Model output

• Requested WW3 output field parameters, every 3 hours, in grib2 format
  .wind. = U-component of input wind, V-component of input wind (m/s). The vector described by these components shows the direction in which the wind blows (oceanographic convention)
  .hs. = significant height of combined wind waves and swell (m)
  .tp. = primary wave mean period (s)
  .dp. = primary wave direction (degrees true, i.e. 0 deg => coming from North; 90 deg => coming from East)
• Partition output in NetCDF format
  These provide bulk spectral estimates for each wave system.
  Available hourly for each individual grid (where there is data).
• Point output
  Hourly spectra, bulk parameters, and partition data is saved at over 2000 points (buoy locations and virtual locations).

Here is information about what each buoy file type contains.

Technical notes

• An initial look at the CFSR Reanalysis winds for wave modeling
• WAVEWATCH III^® Hindcasts with Re-analysis winds. Initial report on model setup
• 30 Year Wave Hindcasts using WAVEWATCH III^® with CFSR winds: Phase 1

Help and hints

There is a problem extracting data from the Mediterranean GRiB files, where the longitude values are incorrect.
Here is an example Python script to work around the problem.

Here is an example Python script to read the NetCDF partition files.

Here is an example Matlab script to read the NetCDF partition files.

Here is an example Matlab script to subset the NetCDF partition files.