Mise à jour:
Daily (with 24 hour delay)
4 times per day, from 0:00, 06:00, 12:00 and 18:00 UTC
SST:
A daily, high-resolution, real-time, global, sea surface temperature (RTG_SST) analysis has been developed at the National Centers for Environmental Prediction/Marine Modeling and Analysis Branch (NCEP / MMAB). The analysis was implemented in the NCEP parallel production suite 16 August 2005. It became fully operational on September 27, 2005.
The daily sea surface temperature product is produced on a twelfth-degree (latitude, longitude) grid, with a two-dimensional variational interpolation analysis of the most recent 24-hours buoy and ship data, satellite-retrieved SST data, and SST's derived from satellite-observed sea-ice coverage. The algorithm employs the following data-handling and analysis techniques:
Satellite retrieved SST values are averaged within 1/12 o grid boxes with day and night 'superobs' created separately for each satellite;
Bias calculation and removal, for satellite retrieved SST, is the technique employed in the 7-day Reynolds-Smith climatological analysis;
Currently, the satellite SST retrievals are generated by a physically-based algorithm from the Joint Center for Satellite Data Assimilation. Retrievals are from NOAA-17 and NOAA-18 AVHRR data;
SST reports from individual ships and buoys are separately averaged within grid boxes;
The first-guess is the prior (un-smoothed) analysis with one-day's climate adjustment added;
Late-arriving data which did not make it into the previous SST analysis are accepted if they are less than 36 hours old;
Surface temperature is calculated for water where the ice cover exceeds 50%, using salinity climatology in Millero's formula for the freezing point of salt water:
t(S) = -0.0575 S + 0.0017 S3/2 - 0.0002 S2,
with S in psu.
An inhomogeneous correlation-scale-parameter l, for the correlation function: exp(-d2/l2) , is calculated from a climatological temperature gradient, as
l = min ( 450 , max( 2.25 / |grad T| , 100 )),
with d and l in kilometers. "grad T" is in oC / km
Evaluations of the analysis products have shown it to produce realistically tight gradients in the Gulf Stream regions of the Atlantic and the Kuroshio region of the Pacific, and to be in close agreement with SST reports from moored buoys in both oceans. Also, it has been shown to properly depict the wintertime colder shelf water -- a feature critical in getting an accurate model prediction for coastal winter storms.
NWP:
La prévision numérique du temps (PNT) est une application de la météorologie et de l'informatique. Elle repose sur le choix d'équations mathématiques offrant une proche approximation du comportement de l'atmosphère réelle. Ces équations sont ensuite résolues, à l'aide d'un ordinateur, pour obtenir une simulation accélérée des états futurs de l'atmosphère. Le logiciel mettant en œuvre cette simulation est appelé un modèle de prévision numérique du temps.
Prévision numérique du temps. (2009, décembre 12). Wikipédia, l'encyclopédie libre. Page consultée le 20:48, février 9, 2010 à partir de
http://fr.wikipedia.org/w/index.php?title=Pr%C3%A9vision_num%C3%A9rique_du_temps&oldid=47652746.