ECMWF Data Server

ECMWF Data Server

This dataset consists of 4 6-hour fields per day available from 1957-09-01 to 2002-08-31. The fields are global at 2.5 degree resolution, i.e. 144 longitude points and 73 latitude points, and there are 23 variably-spaced vertical levels (1000, 925, 850, 775, 700, 600, 500, 400, 300, 250, 200, 150, 100, 70, 50, 30, 20, 10, 7, 5, 3, 2, 1). The NetCDF units attribute for the time variable is hours since 1900-01-01 00:00:0.0. Scale factors and offsets are also used for the variable fields. The CF-1.0 NetCDF metadata conventions are used.

The default level type upon reaching the page (as seen in the left-hand column) is "surface". You can also choose "pressure levels" to obtain non-surface fields. The default Era40 temporal field type is "daily", but you can also choose from "monthly means", "monthly means of daily means", "invariant fields" and "monthly invariant fields". Click on any of them to see what variable fields are available for that category.

First, you agree to the conditions by entering your email, name and affiliation. After this, you will see three clickable buttons appear at the bottom of the page: "Retrieve GRIB", "Retrieve NetCDF", and "Plot data".

To choose a dataset to download:

In the "Select data" section, either select a data range or a list of months. If you're selecting a date range, click on the circle in front of that choice and then enter starting and ending dates. If you're selecting a list of months, click on the circle in front of that choice and then click on the individual months you want to obtain. Note: Depending on the number of fields you select, these files can get quite large, so when you're starting out you might want to try just a month at a time to see what the file sizes are, as well as how long it takes the software at the other end to extract and put your data into your desired format. As an example, a month-long file containing the U and V 10-m fields and all four daily fields is around 28 Mb in size.
In the "Select Time" section, you can choose some or all of the four 6-hour fields per day for your selected data time range. Click on the squares for each of the 6-hour fields you want.
In the "Select parameters" section, click on the squares in front of each of the parameters you want to obtain.
Now click on either "Retrieve GRIB" or "Retrieve NetCDF" to obtain the data files in either format. If you want to see what the data looks like, click on "Plot data".
Another page will eventually appear on which you will click "Now" to continue.
Another page will appear on which you can click on "grib", "netcdf" or "plots" to obtain the product that you chose. Either download the former two via however your browser handles such things, or view the graphs.

The CF-1.0 compliant NetCDF header for era-u-v-10m-1978.nc is:


netcdf era-u-v-10m-1978 {
dimensions:
        longitude = 144 ;
        latitude = 73 ;
        time = UNLIMITED ; // (1460 currently)
variables:
        float longitude(longitude) ;
                longitude:units = "degrees_east" ;
                longitude:long_name = "longitude" ;
        float latitude(latitude) ;
                latitude:units = "degrees_north" ;
                latitude:long_name = "latitude" ;
        int time(time) ;
                time:units = "hours since 1900-01-01 00:00:0.0" ;
                time:long_name = "time" ;
        short p10u(time, latitude, longitude) ;
                p10u:scale_factor = 0.00085084141301256 ;
                p10u:add_offset = 0.379360035051831 ;
                p10u:_FillValue = -32767s ;
                p10u:missing_value = -32767s ;
                p10u:units = "m s**-1" ;
                p10u:long_name = "10 metre U wind component" ;
        short p10v(time, latitude, longitude) ;
                p10v:scale_factor = 0.000798961017941378 ;
                p10v:add_offset = -0.329854747150754 ;
                p10v:_FillValue = -32767s ;
                p10v:missing_value = -32767s ;
                p10v:units = "m s**-1" ;
                p10v:long_name = "10 metre V wind component" ;

// global attributes:
                :Conventions = "CF-1.0" ;
                :history = "2008-10-29 20:31:16 GMT by mars2netcdf-0.92" ;
data:

 longitude = 0, 2.5, 5, 7.5, 10, 12.5, 15, 17.5, 20, 22.5, 25, 27.5, 30, 

...

Interim Re-Analysis

Description

ERA-Interim products are also publicly available on the ECMWF Data Server, at a 1.5° resolution, including several products that were not available for ERA-40.

The quantities needed for a mixed-layer ROMS simulation are:

U and V
surface temperature
surface pressure
relative humidity
cloud fraction
rainfall rate
downward shortwave radiation flux

These can be snagged from the ERA Interim page as:

"10 metre U wind component" & "10 metre V wind component"
"Sea surface temperature"
"Mean sea level pressure"
Calculated from the combination of "2 metre temperature" & "2 metre dewpoint temperature"
"Total cloud cover"
"Total precipitation"
"Surface solar radiation"

Table 6 - ROMS Variables Required for Active Ocean-Atmosphere Boundary Layer and Their ECMWF-Interim Equivalents

Field ROMS var. ROMS units Long Name ECMWF Units Range Scale Offset

surface u-wind component Uwind m/s 10 metre U wind component u10 m s**-1

surface v-wind component Vwind m/s 10 metre V wind component v10 m s**-1

surface air temperature Tair C 2 metre temperature t2m K 20-30 0.001634 266.0987

surface air pressure Pair mb Mean sea level pressure msl Pa ~1000 0.1800523 99118.6

surface air relative humidity Qair % 2 metre temperature t2m K 0-100 0.001634 266.0987

2 metre dewpoint temperature d2m K 0.0016678 247.4571

cloud fraction cloud 0 to 1 Total cloud cover tcc 0 to 1 0-1 1.525972e-05 0.4999847

rain fall rate rain kg/m^2/s Total precipitation tp m < .003 4.01797e-06 0.131649

Instantaneous moisture flux ie W m**-2 s

shortwave radiation flux swrad W/m^2 Surface solar radiation ssr W m**-2 s 10-500 501.8 16442634.2

Conversions:

1. The "Total precipitation" in "m" needs to be converted into a "rain fall rate" in kg/m^2/s.

rainfall rate (k/m^2/s) = total precipitation (m) * water density (1000 kg/m^3) / time (3*3600 s)

-----

2. The "Surface solar radiation" in "W m**-2 s" needs to be converted into a "shortwave radiation flux" in "W m**-2".

shortwave radiation flux (W/m^2) = surface solar radiation (W/m^2/s) * 3600*3 s

-----

1 mb = 100 Pa
1 deg. C = 274.15 deg. K

Needed Conversions

cloud = TCDC/100.
Tair(C) = TMP(K) - 273.15
Pair(mb) = PRES(Pa)/100.

Qair = (112 - 0.1*t2m + d2m)/(112 + 0.9*t2m)

A formula for calculating the relative humidity from the temperature and the dewpoint temperature can be found at:

http://www.ajdesigner.com/phphumidity/dewpoint_equation_relative_humidity.php

Example NetCDF Header

The NetCDF header file for int-u-v-2007-01.nc is:


netcdf int-u-v-2007-01 {
dimensions:
        longitude = 240 ;
        latitude = 121 ;
        time = UNLIMITED ; // (249 currently)
variables:
        float longitude(longitude) ;
                longitude:units = "degrees_east" ;
                longitude:long_name = "longitude" ;
        float latitude(latitude) ;
                latitude:units = "degrees_north" ;
                latitude:long_name = "latitude" ;
        int time(time) ;
                time:units = "hours since 1900-01-01 00:00:0.0" ;
                time:long_name = "time" ;
        short 10u(time, latitude, longitude) ;
                10u:scale_factor = 0.00107127121354477 ;
                10u:add_offset = -5.18156748703386 ;
                10u:_FillValue = -32767s ;
                10u:missing_value = -32767s ;
                10u:units = "m s**-1" ;
                10u:long_name = "10 metre U wind component" ;
        short 10v(time, latitude, longitude) ;
                10v:scale_factor = 0.0010895635053831 ;
                10v:add_offset = 2.53144217477587 ;
                10v:_FillValue = -32767s ;
                10v:missing_value = -32767s ;
                10v:units = "m s**-1" ;
                10v:long_name = "10 metre V wind component" ;

// global attributes:
                :Conventions = "CF-1.0" ;
                :history = "2009-04-02 19:17:11 GMT by mars2netcdf-0.92" ;
data:

...

Python Script


import scipy,numpy,netCDF3
import time,datetime
from mpl_toolkits.basemap import Basemap

#  Set the limits of the desired lon/lat box.

lat_min = -10.0
lat_max = 33.0
lon_min = 40.0
lon_max = 85.0

#  Open the desired file.

g = netCDF3.Dataset('int-u-v-2007-01.nc','r+')

ntimes = len(g.dimensions['time'])
times = g.variables['time']

#  Extract the lon and lat fields.

lon = g.variables['longitude']
lat = g.variables['latitude']

print lon[:]
print lat[:]

#  Find the available lon/lat box that will minimally
#  enclose the requested lon/lat box.

for i in lat:
	if i < lat_min:
		lat_min = i
		break

latm = -1*lat[:]
for i in lat:
	if i >= lat_max:
		lat_max = i
		break

ip = 0.0
for i in lon:
	if i >= lon_min:
		lon_min = ip
		break
	else:
		ip = i


for i in lon:
	if i >= lon_max:
		lon_max = i
		break

# Find the x and y dimensions of the reduced field.

minlat = lat[:] >= lat_min
maxlat = lat[:] <= lat_max
latrange = minlat & maxlat
latnew = lat[latrange]
nlat2 = latnew.size

minlon = lon[:] >= lon_min
maxlon = lon[:] <= lon_max
lonrange = minlon & maxlon
lonnew = lon[lonrange]
nlon2 = lonnew.size

# Create 2D lon and lat arrays.

import numpy as np
lon2d = np.vstack((121*[lon[:]]))
lat2d = np.vstack((240*[lat[:]]))
lat2d = np.swapaxes(lat2d,0,1)

lonmin = lon2d >= lon_min
lonmax = lon2d <= lon_max
latmin = lat2d >= lat_min
latmax = lat2d <= lat_max

shrink = lonmin & lonmax & latmin & latmax

latshrink = lat2d[shrink]
latshrink2d = np.reshape(latshrink,(nlat2,nlon2))
lonshrink = lon2d[shrink]
lonshrink2d = np.reshape(lonshrink,(nlat2,nlon2))

# Extract velocity fields.
u = g.variables['10u']
v = g.variables['10v']

# Extract scale factors and offsets.
u_scale = u.scale_factor
u_offset = u.add_offset
v_scale = v.scale_factor
v_offset = v.add_offset

# Apply scale factors and offsets.
up = u_scale*u[:] + u_offset
vp = v_scale*v[:] * v_offsetl

#----------------------------------------------------------------------------

# Extract the Oman grid info.

grd = netCDF3.Dataset('newoman_4_grd.nc','r+')

x_rho = grd.variables['x_rho']
y_rho = grd.variables['y_rho']

eta_rho = len([grd.dimensions['eta_rho'])
xi_rho = len([grd.dimensions['xi_rho'])

x_rho_flat = x_rho[:,:].ravel()
y_rho_flat = y_rho[:,:].ravel()

#-----------------------------------------------------------------------------

# Convert the lon/lat values to x/y values for interpolation via csabathy.

while nsl <= ntimes - 1:

	utsl = up[0,:,:]
	ushrink = utsl[shrink]
	from mpl_toolkits.basemap import Basemap
	proj = Basemap(projection='lcc',resolution='i',llcrnrlon=30.0,llcrnrlat= -15.,
               urcrnrlon=90.0,urcrnrlat=40.0,lat_0=20.0,lon_0=60.0)
	xl,yl = proj(lonshrink2d,latshrink2d)

	#  Or for a flattened, 1D version.
	
	xlf,ylf = proj(lonshrink,latshrink)

xl.shape
 (30,32)
yl.shape
 (30,32)




#  Interpolate to the rho points Uwind and Vwind. 

while n < ntimes:

	u = up[n,:,:]

Field	ROMS var.	ROMS units	Long Name	ECMWF	Units	Range	Scale	Offset
surface u-wind component	Uwind	m/s	10 metre U wind component	u10	m s**-1
surface v-wind component	Vwind	m/s	10 metre V wind component	v10	m s**-1
surface air temperature	Tair	C	2 metre temperature	t2m	K	20-30	0.001634	266.0987
surface air pressure	Pair	mb	Mean sea level pressure	msl	Pa	~1000	0.1800523	99118.6
surface air relative humidity	Qair	%	2 metre temperature	t2m	K	0-100	0.001634	266.0987
			2 metre dewpoint temperature	d2m	K		0.0016678	247.4571
cloud fraction	cloud	0 to 1	Total cloud cover	tcc	0 to 1	0-1	1.525972e-05	0.4999847
rain fall rate	rain	kg/m^2/s	Total precipitation	tp	m	< .003	4.01797e-06	0.131649
			Instantaneous moisture flux	ie	W m**-2 s
shortwave radiation flux	swrad	W/m^2	Surface solar radiation	ssr	W m**-2 s	10-500	501.8	16442634.2