 |
| Titel |
Comparing soil moisture memory in satellite observations and models |
| VerfasserIn |
Tobias Stacke, Stefan Hagemann, Alexander Loew |
| Konferenz |
EGU General Assembly 2013
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 15 (2013) |
| Datensatznummer |
250076531
|
|
|
|
|
|
| Zusammenfassung |
| A major obstacle to a correct parametrization of soil processes in large scale global land
surface models is the lack of long term soil moisture observations for large parts of
the globe. Currently, a compilation of soil moisture data derived from a range of
satellites is released by the ESA Climate Change Initiative (ECV_SM). Comprising the
period from 1978 until 2010, it provides the opportunity to compute climatological
relevant statistics on a quasi-global scale and to compare these to the output of climate
models.
Our study is focused on the investigation of soil moisture memory in satellite observations
and models. As a proxy for memory we compute the autocorrelation length (ACL) of the
available satellite data and the uppermost soil layer of the models. Additional to the ECV_SM
data, AMSR-E soil moisture is used as observational estimate. Simulated soil moisture fields
are taken from ERA-Interim reanalysis and generated with the land surface model JSBACH,
which was driven with quasi-observational meteorological forcing data. The satellite data
show ACLs between one week and one month for the greater part of the land surface while
the models simulate a longer memory of up to two months. Some pattern are similar in
models and observations, e.g. a longer memory in the Sahel Zone and the Arabian Peninsula,
but the models are not able to reproduce regions with a very short ACL of just a few days.
If the long term seasonality is subtracted from the data the memory is strongly
shortened, indicating the importance of seasonal variations for the memory in most
regions.
Furthermore, we analyze the change of soil moisture memory in the different soil layers of
the models to investigate to which extent the surface soil moisture includes information about
the whole soil column. A first analysis reveals that the ACL is increasing for deeper layers.
However, its increase is stronger in the soil moisture anomaly than in its absolute values and
the first even exceeds the latter in the deepest layer. From this we conclude that the seasonal
soil moisture variations dominate the memory close to the surface but these are dampened
in lower layers where the memory is mainly affected by longer term variations. |
|
|
|
|
|
|