Radar and field measurement campaign in the Weisseritz catchment area
Figure 1: Interdisciplinary team of scientists of University of Potsdam: Institute for Geoecology and Institute for Geosciences, GFZ Potsdam: Section Remote Sensing and Section Engineering Hydrology, University of Stuttgart, University of Bern and the German Aerospace Centre (flight crew not present).
One of the main goals of OPAQUE
is the identification of critical states of
catchments such as high soil moisture conditions. Those
conditions are important reasons for uncertainties in
operational flood forecasting, especially in the early warning
stage. Therefore soil moisture and its scale-dependent
variability were determined by a combination of innovative
field and airborne measuring techniques. The combination of
measurement at different scales makes a consistent up scaling
from the point scale (TDR, FDR) over field scale to the
regional scale of a hydrological catchment (space borne RADAR)
possible. In order to test the feasibility of that approach
two field campaigns were planed to measure the catchment state
of the Weisseritz river at two
different states, in early summer and spring. The first
campaign was carried out by an interdisciplinary team of
scientists (May, 29th to June,1st of 2007).
Measuring concept of the field campaign
Soil moisture is estimated on four different scales with different retrieval methods:
- Point scale: Volumetric water content from soil samples, FDR
(Frequency Domain Reflectometry), S-TDR (Time Domain Reflectrometry),
S-TDR (Spatial-Time Domain Reflectrometry);
- Field scale: GPR (Ground Penetrating Radar);
- Sub-catchment scale: Airborne RADAR remote sensing;
- Catchment scale: Spaceborne RADAR remote sensing;
The combination of the different techniques makes it possible to compare and validate the different methods of soil moisture measurement. Additionally, the influence of the surface roughness and the phenological characteristics of the vegetation layer on the airborn soil moisture measurement were studied. All measurements were surveyed by GPS or by Differential-GPS for subsequent integration of all data in a GIS.
Measurements of soil moisture:
Figure 2: Measurement grid for the high spatial resolution measurement.
High spatial resolution soil moisture measurements
Four fields of the size 50 by 50 m² on pasture and ploughed soil were investigated intensively.
Measurements with FDR probes were carried out on a soil moisture monitoring grid with 1m spacing (Figure 2) to investigate the upper top soil layer. At the corners of the inner raster soil cores of 1m depth were taken to investigate the vertical soil moisture distribution in 20cm-intervalls.
Two S-TDR clusters continuously recorded the soil moisture down to a depth of 60 cm at two sites in the catchment area with high temporal as well as spatial resolution.
In addition two GPR-teams (University of Potsdam, GFZ Potsdam) investigated the electric permittivity by means of the velocity of the ground wave. A frequency range from 100MHz to 400MHz was used. Common Offset (CO) and Wide Angle Reflection and Refraction (WARR) measurements were performed.
Intermediate spatial resolution soil moisture measurements
On eight fields with different land use (summer barley, winter barley, winter wheat, winter rape, silage maize, triticale, seed grass and forest) the soil moisture in the upper soil layer was monitored with a coarser spacing (15m).
Measurements of vegetation cover:
Figure 3: Vegetation measurements
In order to analyse the vegetation cover the following parameters were measured: stage of phenology, vegetation height, density, orientation and biomass. The phenological information is used to study the influence of the vegetation layer on the radar waves and their penetration capability through the vegetation cover.
Measurements of surface roughness:
Another parameter which influences the estimation of soil moisture by radar measurements is the surface roughness, which also contributes to the recorded backscatter. Roughness has to be measured on-site to validate the separation of roughness information from the soil moisture. Roughness was measured with two techniques: Laser scanning (GFZ Potsdam) and stereoscopic photography (University of Bern, Switzerland).
Figure 4: Roughness measurements: Stereoscopic photography and laser scanning.
Figure 5: E-SAR system on Dornier 228, Horn (2000).
The Experimental-SAR (E-SAR) system
mounted on a Dornier 228 of the German Aerospace center was used to collect high resolution (2m x 2m per pixel) Radar data. A strip of 16.3km x 3km was recorded twenty times in different acquisition modes.
Different frequencies were used in order to vary the penetration depth of the waves into the target medium: X-Band (9.6 GHz), C-Band (5.3 GHz), L-Band (1.3 GHz), P-Band (0.35 GHz). For each overflight the acquisition geometry was altered perpendicular to the flight direction to generate different baselines (0 m to 120 m) for interferometric analysis.
Figure 6: Acquisition geometry for subsequent interferometric analysis
Figure 7: Corner reflector with 1.5m³ size
Interferometry is used to locate the different phase centres of the target media. C-, L- and P-Band were recorded in full-polarimetric mode (HH, HV, VV, and VH) and X-Band at single polarimetry (VV) (for the generation of a digital elevation model). Polarimetry is used to distinguish different scattering mechanisms. Three corner reflectors were deployed for subsequent calibration of the RADAR signal. These calibration targets reflect 100% of the incoming signal.
The weather before and during the campaign was favourable for
studying soil moisture. Shortly before the campaign heavy rain
moistened the soil in the upper layers. During the campaign no
precipitation was recorded and low evapotranspiration could be
suspected due to strong cloud cover and moderate temperature
Figure 8: Temperature and precipitation during the period of May at Rehefeld-Zaunhaus
Data of the ENVISAT –ASAR system, acquired on June 4th, makes it possible to analyze the observation of the soil moisture on the catchment level. The sensor operates at C-Band, the data was recorded with an approximate spatial resolution of 100by 100 m².