 |
| Titel |
Assimilation of soil moisture on multiple spatial scales into the mesoscale hydrologic model (mHM) |
| VerfasserIn |
Diana Spieler, Oldrich Rakovec, Matthias Zink, Martin Schrön, Luis Samaniego |
| Konferenz |
EGU General Assembly 2014
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 16 (2014) |
| Datensatznummer |
250091741
|
| Publikation (Nr.) |
 EGU/EGU2014-6051.pdf |
|
|
|
|
|
| Zusammenfassung |
| Soil moisture observations are often not acquired at scales suitable for mesoscale
hydrological modelling. While in situ measurements and observational networks provide
measurements for the local scale at fine spatio-temporal resolutions, spaceborne sensors
deliver global soil moisture observations in coarse spatio-temporal resolutions. None
of these available measurements can be used directly for the assimilation into a
mesoscale hydrological model. Nevertheless, these measurements provide valuable
information, which have been proven to enhance model performance. Therefore, the
application of scaling techniques is a common way to deal with the discrepancy between
observational and modelling scale before the observations are assimilated into the model.
However, this will always introduce uncertainties into the dataset. This ongoing
study aims to quantify the potential gain of assimilating soil moisture data on its
original scale. It will be analysed how mesoscale modelling can benefit from the
assimilation of soil moisture observations without the use of prior scaling techniques.
For this purpose, the mesoscale Hydrological Model (mHM) is used. It employs a
Multiscale Parameter Regionalization (MPR) technique that allows the estimation
of quasi-scale invariant parameters. Thus, the model is able to run on different
scales simultaneously while preserving model fluxes like for example soil moisture
dynamics, infiltration, surface runoff and discharge generation. (For a more detailed
description of the model see: http://www.ufz.de/index.php?en=31389.) This unique
feature of mHM is used to conduct a multiscale synthetic experiment to estimate the
potential benefits of multiscale data assimilation. A proxy soil moisture dataset at a
mesoscale resolution (e.g. 4x4 km2) is created and averaged to the size of typical
coarse remote sensing pixels (e.g. 25x25 km2). This proxy remote sensing dataset is
assimilated into mHM by the ensemble Kalman filter (EnKF) to update the global model
parameters that are common for all scales. The updated parameters are then used for
modelling different spatial scales, e.g. the scale of the proxy dataset before it was
averaged on remote sensing pixel size. In this way, the information gained through the
assimilation of large scale data is preserved and used for modelling smaller scales.
Thus, it will be possible to assimilate data at the observational scale and transfer
the gained information (updated model parameters) to other scales without the
necessity of using prior scaling techniques. The study shows first results for the Saale
catchment (23 700 km2) and demonstrates the positive and negative aspects of
assimilating large scale data, including how its effect propagates to smaller scales and how
the resolution of the assimilated data influences the potential gain for the model. |
|
|
|
|
|
|