source: trunk/adm/website/description.rst @ 137

.. +
..
..
.. DESCRIPTION
.. ===========
..
.. description
..
.. TODO
.. ====
..
.. (July 1983 to December 2009 at the time of writing), ... not anymore true
..
.. EVOLUTIONS
.. ==========
..
.. $Id$
..
.. $URL$
..
.. - jv 20120222
..
..   * typos 
..
.. - fplod 20120219
..
..   * fix links, add bold
..
.. - fplod 20120217
..
..   * creation from description.odt
..     transform to ReST using odt2sphinx + hand correction (diacritical,
..     ponctuations, short lines, links !!)
..
.. -

.. _description:

Description
===========

How are TropFlux fluxes estimated ?
+++++++++++++++++++++++++++++++++++

  * **The TropFlux heat and momentum fluxes are estimated from the COARE v3
    algorithm** (`Fairall et al. 2003 <http://dx.doi.org/10.1175/1520-0442(2003)016%3C0571:BPOASF%3E2.0.CO;2>`_),
    arguably one of the most suited algorithm for fluxes estimations in the
    tropical regions.
    This algorithm requests surface meteorological parameters (10m-winds,
    2m-air and sea temperature, 2m-air relative humidity) and downward
    radiative (shortwave and infrared fluxes) as input parameters.

  * The analysis of various data sources shows that, except for shortwave
    radiation, ERA-I data generally displays the best agreement to the
    `global tropical moored buoy array <http://www.pmel.noaa.gov/tao/global/global.html>`_
    data, despite systematic biases and underestimated variability
    `Praveen Kumar et al. (2011) <http://www.locean-ipsl.upmc.fr/~tropflux/papers/TropFlux_paper_accepted_r.pdf>`_.
    The `ISCCP dataset <http://isccp.giss.nasa.gov/projects/browse_fc.html>`_ provides the best shortwave data.
    **TropFlux hence uses bias and amplitude corrected ERA-I (10m-winds, 2m-air
    and sea temperature, 2m-air relative humidity and downward radiative
    fluxes) and ISCCP (shortwave radiation) fluxes.**
    All bias corrections are derived on the basis of comparisons with
    the
    `global tropical moored buoy array <http://www.pmel.noaa.gov/tao/global/global.html>`_
    data, and are described in detail in
    `Praveen Kumar et al. (2011) <http://www.locean-ipsl.upmc.fr/~tropflux/papers/TropFlux_paper_accepted_r.pdf>`_.

  * **An additional climatological surface temperature-dependent gustiness
    is applied to wind speed, in order to account for unresolved (sub-daily and subgrid-scale) variability.**

  * Since ISCCP shortwave data
    is not updated on a regular basis, we provide an alternative shortwave
    flux estimate that is less accurate, but available until present.
    This estimate uses the ISCCP shortwave climatology plus anomalies that are
    linearly related to `NOAA Outgoing Longwave Radiation <http://www.esrl.noaa.gov/psd/data/gridded/data.interp_OLR.html>`_ the anomalies with respect to the seasonal cycle.
    Such an estimate is almost as good as ISCCP in deep atmospheric convection
    regions such as the ITCZ or Indo-Pacific warm pool.
    It does not perform as well as ISCCP in regions of low clouds such as the
    eastern equatorial Pacific and Atlantic oceans, but is in any case better
    than most available re-analyses in those regions 
    (`Praveen Kumar et al. 2011 <http://www.locean-ipsl.upmc.fr/~tropflux/papers/TropFlux_paper_accepted_r.pdf>`_).
    The net surface shortwave that is provided in TropFlux data files is based
    on ISCCP data over the entire period for which it is available (July 1983
    to December 2009 at the time of writing), with a linear transition to
    OLR-derived shortwave over a three months window at the edges of the
    ISCCP-availability window.