| Home > Newsevents > Training > Rcourse_notes > PARAMETRIZATION > SURFACE_ASSIMILATION > |
Land surface assimilation
March 2001
By Jean-François Mahfouf and Pedro Viterbo
European Centre for Medium-Range Weather
Forecasts
Table of contents
Next Section
Previous Section
3 Introduction to land surface assimilation
The major problem of land surface assimilation is the lack of routine observations of soil moisture and soil temperature. This is specially true in the case of soil moisture, where current methods cannot provide global coverage routinely (see Subsection 6.1 for more details). Furthermore, soil moisture observations show large variability in small spatial scales (see e.g. Wetzel and Chang 1988); not all scales are of relevance for the atmosphere, and the assimilation method has to take that into account. For soil temperature the climate network exists, with a coverage similar to the SYNOP stations and with most of the stations performing observations at least daily; unfortunately, those observations are not exchanged routinely at the time of measurement, so in practice they can only be used, in delayed mode, for verification purposes.
The nature and availability of the observations imposes the use of proxy variables for soil moisture. The amount of water in the root zone impacts on the evaporative fraction, which in itself determines, for a given amount of net radiation, midday summer screen level temperature and humidity (see previous section and the review by Betts et al. 1996). On the other hand, the time evolution of soil water depends on the rainfall intensity and on its timing. Three main types of data have been used in the past to infer soil moisture: (a) Screen level atmospheric temperature and humidity; (b) Rainfall rates; and (c) Radiometric surface temperature (infrared, microwave). Note that, since those observations already represent, to a certain extent, the impact of soil moisture on the atmosphere above, they have already filtered out the smaller scales in soil water.
There are two aditional difficulties in the assimilation of land surface observations. First, these observations are non-linearly related to soil moisture and soil temperature (through the equations of the land surface scheme). Secondly, the statistics of forecast errors, used to spatially distribute the local increments in atmospheric analyses, are not known for soil variables.
Training Course Notes Front Page
Table of contents
Next Section
Previous Section
 |
12.06.2002 |
 |
© ECMWF |
