![Hier klicken, um den Treffer aus der Auswahl zu entfernen](images/unchecked.gif) |
Titel |
Robust spatialization of soil water content at the scale of an agricultural
field using geophysical and geostatistical methods |
VerfasserIn |
Hocine Henine, Julien Tournebize, Laurent Gourdol, Christophe Hissler, Paul-Henry Cournede, Clement Remi |
Konferenz |
EGU General Assembly 2017
|
Medientyp |
Artikel
|
Sprache |
en
|
Digitales Dokument |
PDF |
Erschienen |
In: GRA - Volume 19 (2017) |
Datensatznummer |
250149662
|
Publikation (Nr.) |
EGU/EGU2017-14035.pdf |
|
|
|
Zusammenfassung |
Research on the Critical Zone (CZ) is a prerequisite for undertaking issues related to
ecosystemic services that human societies rely on (nutrient cycles, water supply and quality).
However, while the upper part of CZ (vegetation, soil, surface water) is readily accessible,
knowledge of the subsurface remains limited, due to the point-scale character of conventional
direct observations. While the potential for geophysical methods to overcome this limitation
is recognized, the translation of the geophysical information into physical properties or states
of interest remains a challenge (e.g. the translation of soil electrical resistivity into soil water
content).
In this study, we propose a geostatistical framework using the Bayesian Maximum
Entropy (BME) approach to assimilate geophysical and point-scale data. We especially focus
on the prediction of the spatial distribution of soil water content using (1) TDR point-scale
measurements of soil water content, which are considered as accurate data, and (2) soil water
content data derived from electrical resistivity measurements, which are uncertain data but
spatially dense.
We used a synthetic dataset obtained with a vertical 2D domain to evaluate the
performance of this geostatistical approach. Spatio-temporal simulations of soil water content
were carried out using Hydrus-software for different scenarios: homogeneous or
heterogeneous hydraulic conductivity distribution, and continuous or punctual infiltration
pattern. From the simulations of soil water content, conceptual soil resistivity models were
built using a forward modeling approach and point sampling of water content values,
vertically ranged, were done. These two datasets are similar to field measurements of soil
electrical resistivity (using electrical resistivity tomography, ERT) and soil water content
(using TDR probes) obtained at the Boissy-le-Chatel site, in Orgeval catchment (East of
Paris, France). We then integrated them into a specialization framework to predict the soil
water content distribution and the results were compared to initial simulations (Hydrus
results). We obtained more reliable water content specialization models when using the
BME method. The presented approach integrates ERT and TDR measurements, and
results demonstrate that its use significantly improves the spatial distribution of
water content estimations. The approach will be applied to the experimental dataset
collected at the Boissy le Ch-tel site where ERT data were collected daily during
one hydrological year, using Syscal pro 48 electrodes (with a financial support of
Equipex-Critex) and 10 TDR probes were used to monitor water content variation.
Hourly hydrological survey (tile drainage discharge, precipitation, evapotranspiration
variables and water table depth) were conducted at the same site. Data analysis and the
application of geostatistical framework on the experimental dataset of 2015-2016 show
satisfactory results and are reliable with the hydrological behavior of the study site. |
|
|
|
|
|