![Hier klicken, um den Treffer aus der Auswahl zu entfernen](images/unchecked.gif) |
Titel |
Assessment of soil moisture drought uncertainty using mHM and TERRA-ML simulations in Germany |
VerfasserIn |
L. Samaniego, H. Feldmann, R. Kumar, G. Schaedler |
Konferenz |
EGU General Assembly 2012
|
Medientyp |
Artikel
|
Sprache |
Englisch
|
Digitales Dokument |
PDF |
Erschienen |
In: GRA - Volume 14 (2012) |
Datensatznummer |
250064486
|
|
|
|
Zusammenfassung |
Currently, the hydro-meteorologic mechanisms originating droughts are relatively well
understood but our ability to predict them remains unsatisfactory despite the fact that
substantial amount of research has been carried out during the previous decades. As a result,
drought which is among the most costly natural disasters, remain as one of the least
understood natural hazards. The main reasons for the lack of skill of the predictive models are
the uncertainty related with the forcings, the conceptualization of dominant processes,
and their parametrization at a given spatio-temporal scale. As a result, modeling
the soil moisture dynamics at large-scales (from 500Â m to 50Â km) is becomes
extremely difficult as was originally demonstrated by the PILPS project and subsequent
studies.
The main goal of this study was to quantify the degree of statistical dependence between
monthly soil moisture fields obtained with two land surface models (LSM) (i.e TERRA-ML
and mHM) for Germany from 1971 to 2001. A particular goal was to understand how the
parametric uncertainty of mHM affect a monthly soil moisture index and derived statistics
such as area under drought, severity, magnitude. The effects of modelling scale were also
investigated. TERRA-ML is a LSM implemented within the nonhydrostatic regional climate
model (RCM) COSMO-CLM (www.clm-community.eu). In this study, COSMO-CLM was
forced with ERA40 reanalysis data (www.ecmwf.int) to generate finally aggregated monthly
soil moisture fields at 7Ã7Â km resolution covering Germany and its surrounding during the
period 1971 to 2001. Due to the computational burden of this model, only one run
could be completed for this study. mHM, on the contrary, is a mesoscale distributed
hydrological model operated here as a LSM (Samaniego et al. 2010). It was forced
with grided daily precipitation and temperature data at 4Ã4Â km resolution from
1950 to 2010. Daily time series for more than 5600 rain gauges and about 1120
meteorological stations (DWD) were interpolated with external drift Kriging to produce
highly consistent fields of meteorological variables. Land cover changes during this
period were also considered. It should be noted that, mHM aims at a consistent
analysis of soil moisture based on available observations whereas the COSMO-CLM
simulation describes the atmospheric and the soil processes consistent with the in-
and outflows at the outer boundaries of the modelling domain from the ERA40
re-analysis.
The best hundred parameter sets obtained for mHM were employed to generate a
100-member ensemble of daily soil moisture fields. Using these ensemble, the effects of
parameter uncertainty on the soil moisture index (SMI) and related statistics were estimated.
Results indicated that the ensemble mean of SMI exhibited an excellent agreement with
extreme drought events reported in the literature. Parametric uncertainty of the SMI is,
however, considerably during some periods. The uncertainty of the area under drought and
severity, however, indicate that a single simulation is not enough to draw conclusive results. A
non-parametric test was be applied to investigate the existence of significant trends in
soil moisture simulations. Preliminary results indicated the existence of a negative
significant trend (p-value 5%) in soil moisture during summer months which is the
consequence of observed downward trend in precipitation and upward trend in
temperature. On the contrary, soil moisture simulations in winter months did not exhibit
significant trends. Finally, canonical correlation analysis will be employed to identify the
maximum correlation between the monthly soil moisture fields obtained with both
LSMs. |
|
|
|
|
|