% IMPORTANT: The following is UTF-8 encoded. This means that in the presence
% of non-ASCII characters, it will not work with BibTeX 0.99 or older.
% Instead, you should use an up-to-date BibTeX implementation like “bibtex8” or
% “biber”.
@ARTICLE{Zhang:840115,
author = {Zhang, Hongjuan and Kurtz, Wolfgang and Kollet, Stefan and
Vereecken, Harry and Hendricks-Franssen, Harrie-Jan},
title = {{C}omparison of different assimilation methodologies of
groundwater levels to improve predictions of root zone soil
moisture with an integrated terrestrial system model},
journal = {Advances in water resources},
volume = {111},
issn = {0309-1708},
address = {Amsterdam [u.a.]},
publisher = {Elsevier Science},
reportid = {FZJ-2017-07678},
pages = {224-238},
year = {2018},
abstract = {The linkage between root zone soil moisture and groundwater
is either neglected or simplified in most land surface
models. The fully-coupled subsurface-land surface model
TerrSysMP including variably saturated groundwater dynamics
is used in this work. We test and compare five data
assimilation methodologies for assimilating groundwater
level data via the ensemble Kalman filter (EnKF) to improve
root zone soil moisture estimation with TerrSysMP.
Groundwater level data are assimilated in the form of
pressure head or soil moisture (set equal to porosity in the
saturated zone) to update state vectors. In the five
assimilation methodologies, the state vector contains either
(i) pressure head, or (ii) log-transformed pressure head, or
(iii) soil moisture, or (iv) pressure head for the saturated
zone only, or (v) a combination of pressure head and soil
moisture, pressure head for the saturated zone and soil
moisture for the unsaturated zone. These methodologies are
evaluated in synthetic experiments which are performed for
different climate conditions, soil types and plant
functional types to simulate various root zone soil moisture
distributions and groundwater levels. The results
demonstrate that EnKF cannot properly handle strongly skewed
pressure distributions which are caused by extreme negative
pressure heads in the unsaturated zone during dry periods.
This problem can only be alleviated by methodology (iii),
(iv) and (v). The last approach gives the best results and
avoids unphysical updates related to strongly skewed
pressure heads in the unsaturated zone. If groundwater level
data are assimilated by methodology (iii), EnKF fails to
update the state vector containing the soil moisture values
if for (almost) all the realizations the observation does
not bring significant new information. Synthetic experiments
for the joint assimilation of groundwater levels and surface
soil moisture support methodology (v) and show great
potential for improving the representation of root zone soil
moisture.},
cin = {IBG-3},
ddc = {550},
cid = {I:(DE-Juel1)IBG-3-20101118},
pnm = {255 - Terrestrial Systems: From Observation to Prediction
(POF3-255)},
pid = {G:(DE-HGF)POF3-255},
typ = {PUB:(DE-HGF)16},
UT = {WOS:000418592800016},
doi = {10.1016/j.advwatres.2017.11.003},
url = {https://juser.fz-juelich.de/record/840115},
}