| Knowledge of the distribution of soil moisture (SM) in semi-arid Mediterranean ecosystems,
and of the environmental factors influencing it, will enable the acquisition of in situ data
simultaneous to the observations from SMOS in the area of the Valencia Anchor Station
(VAS). The variability of SM depends on intrinsic and extrinsic soil factors. It is necessary to
define a sampling strategy that integrates the relationships between the hydrological
variables, particularly SM, and the parameters of the landscape at different spatial
scales
SM data acquisition for the CAROLS (Combined Airborne Radio - instruments for Ocean
and Land Studies) Campaign extended over an area of 27 x 35 km2, within the VAS site,
demands a selection of sampling points which are representative of larger areas. The total
area was sub-divided in Environmental Units which were homogeneous with respect to
landscape, geological material, land use (or vegetation cover) and soil type (mainly soil
texture and geomorphologic aspects).
The most representative units of the area correspond to (i) vineyards over stony-sandy
soil, (ii) vineyard over clayey soil, (iii) cereal with partly fallow land, (iv) forest
areas with sub-arbustive and dense shrub-land divided into Northern and Southern
exposure, and (v) mixed vineyards and forest vegetation. The sampling strategy
during the CNES CAROLS campaign was designed with the following threefold
criteria:
(i) Within each Environmental Unit, an area of 1 x 1 km2 was selected where a
simple random sampling of about 35 plots/km2 was defined and where volumetric
SM samples were obtained with small cylinders, together with ThetaProbe and
surface temperature measurements. Soil samples were processed to obtain volumetric
soil moisture, texture, bulk density and organic material for each measurement
plot
(ii) A selection of 10 stationary points, each one respectively and nearly close to 10
thermopluviometric stations, representative of the different units of the area, was also made
with the aim to reflecting all of the variability of vegetation types and of the different
roughness parameters
(iii) A transect sample strategy in the 10 x 10 km2 ESA SVRC (SMOS Validation
Rehearsal Campaign) control area of 2008 was kept in the CNES CAROLS campaign but at
a less dense point sampling
The analysis of the data gathered show that the water content variation at any given point
of the soil depends partially on the degree of variability of the intrinsic edaphic properties that
most influence water retention capacity. Soil moisture (m3/m3) throughout the three days of
the campaign was significantly different, being 0.10 for the first day, 0.081 for the second and
0.048 for the third in average. With respect to vegetation cover, units with natural vegetation
presented significant differences with those with agricultural use, stratified by density cover
and plot orientation. Plots with fallow land and cereal showed heterogeneous soil moisture
content throughout the campaign due to their irrigation state. No differences were found
between SM of the stationary plots and SM of the respective units in which these
plots were located; therefore, these points should be representative of the average
SM of the unit where they are found. A clear negative correlation was also seen
between the probe mV and surface temperature. SM data from the transects of the
10 x 10km2 control area corroborate their similarity to the units where they are
located, and the stationary measurements are representative of the situations they
typify.
In the CAROLS campaign, SM data corresponding to each km2, representative of the
environmental units present certain homogeneity in relation to the interior variability of each
unit. In order to be able to extrapolate the SM value to the whole area and draw the SM map
corresponding to each campaign day, it is necessary to lean on other meteorological
parameters such as antecedent precipitation fields, for example, which is the forthcoming task
in this work. |