 |
|
| ESTAR Data | ESTAR Quick Looks | P-3 page | Washita92 | Washita94 |  Air remote sensing |
Microwave radiometry at long wavelengths can be used to measure and monitor surface soil moisture. A key issue in implementing this approach has been the inherent spatial resolution problem of long wavelength microwave radiometry at spacecraft altitudes. Synthetic aperture techniques can solve this problem. As part of SGP97, the Electronically Scanned Thinned Array Radiometer (ESTAR) was used to map soil moisture. Data were collected using this L band passive microwave mapping instrument over a 10,000 km2 for a one month period. The major objective of this investigation was to evaluate the extension of high resolution soil moisture retrieval algorithms to the coarser resolution of satellite sensors. Secondary objectives were the demonstration of the approach building from successful prior experiments and the collection of a unique data set to support collaborative efforts to understand surface and boundary layer interactions and the physical processes controlling the spatial and temporal variability at these scales. The data collection campaign resulted in an excellent data set for analysis. The ESTAR instrument performed very well at the high altitudes flown. Meteorological conditions were also excellent. Significant and spatially variable rainfall events were mixed with drydown sequences. Calibration of the ESTAR was verified using ground observations and results of previous campaigns in this region. An established soil moisture algorithm was implemented using ancillary data bases. This algorithm was validated using ground observations at several scales. Error levels were nominally 3% which was similar to previous investigations. The soil moisture images show consistent spatial structure. In SGP97, this structure was dominated by rainfall distribution with soil texture and vegetation levels having a secondary effect. Results clearly demonstrated the performance of both the ESTAR instrument and the soil moisture algorithm.
Three different data sets are provided; the ESTAR brightness temperature images for each date data were collected (subject to quality control), predicted soil moisture images, and parameter images used in the soil moisture algorithm.
Data were processed to produce a georeferenced gridded product for each day. This grid is exactly the same on each day. It has a pixel resolution of 800 m. The images are all 8 bit binary consisting of 206 pixels wide by 621 lines. The georeferencing information is:
Georeferencing information Projection Universal Transverse Mercator Zone 14S Earth Ellipsoid Clarke 1866 (NAD 27) Upper Left Corner 543600.000 E 4261000.000 N Upper Right Corner 708400.000 E 4261000.000 N Image Center 626000.000 E 4012600.000 N Lower Left Corner 543600.000 E 3764200.000 N Lower Right Corner 708400.000 E 3764200.000 N Pixel Size 800.000 E 800.000 N Upper Left Corner 98d30'00.05" W Lon 38d29'53.35" N Lat Upper Right Corner 96d36'39.90" W Lon 38d28'29.46" N Lat Image Centre 97d35'51.30" W Lon 36d15'06.87" N Lat Data type Byte File type Binary Dimensions 206 columns 621 rows Units Meters unit dist=1 Each type of data has been scaled to a range of 0-255 as follows
Data Scale Brightness Temperature (K) Brightness Temperature - 70 Soil Moisture (%) Soil Moisture Effective Soil Temperature (Deg. C) (Effective Soil Temperature - 10) * 10 Vegetation b Parameter Vegetation b Parameter * 1000 Vegetation Water Content (kg/m2) Vegetation Water Content * 100 Surface Roughness Parameter Surface Roughness Parameter * 100 Soil Bulk Density (g/cm3) Soil Bulk Density * 100 Soil Texture Code Soil Texture Code Percent Sand Percent Sand Percent Clay Percent Clay
The Files
FormatAll data are 8 bit binary images consisting of 206 pixels by 621 lines with no headers.
File name and directory information
Brightness Temperature sgptbMDD.raw Soil Moisture sgpsmMDD.raw Effective Soil Temperature sgpstMDD.raw Vegetation b Parameter sgp_b.raw Vegetation Water Content sgp_vwc.raw Surface Roughness Parameter sgp_h.raw Soil Bulk Density sgp_bd.raw Soil Texture Code sgp_tex.raw Percent Sand sgp_ps.raw Percent Clay sgp_pc.raw For brightness temperature, soil moisture, and soil temperature MDD refers to the month and day of the observations.
The directory path to the ESTAR Brightness Temperature and Soil Moisture ftp site is
/ftphttp://disc.sci.gsfc.nasa.gov/data/sgp97/air_remote_sensing/estar/sgpprodAll files are contained in four subdirectories of the path and are described in the table below.
Subdirectory # Files File names Description Notes tb 16 sgptbMDD .raw Brightness temperature images Scaled values; Add 70 to digital number soilm 16 sgpsmMDD.raw Soil moisture images % soil moisture in sample soilt 16 sgpstMDD.raw Effective soil temperature Scaled values; To convert to Centigrade, divide by 10 then add 10 to result param Land cover 4 sgp_bd.raw bulk density Scaled values: divide by 100 sgp_b.raw vegetation Scaled values: divide by 1000 sgp_h.raw Surface roughness Scaled values: divide by 100 sgp_vwc.raw Vegetation water content Scaled values: divide by 100 Soil texture 3 sgp_pc.raw Clay content % clay in sample sgp_ps.raw Sand content % sand in sample sgp_tex.raw Texture categories DN : Category
0 : No Data
1 : Sand
3 : Sandy Loam
4 : Silt Loam
6 : Loam
8 : Silty Clay Loam
9 : Clay Loam
11 : Silty Clay
12 : Clay
14 : WaterTexture file (of 0-5 cm top layer of the soils) comes from the PSU web site in original grid of 1 km resolution. It was resampled to fit 800m grid GIS data layer. Later, based on soil texture samples and published data, percentage of clay and sand was assigned to texture categories.
The Science
Brightness Temperature Mapping
The electronically scanned thinned array radiometer (ESTAR) is a synthetic aperture, passive microwave radiometer operating at a center frequency of 1.413 GHz (21 cm) and width of 20 MHZ. For this experiment it was installed to provide horizontally polarized data. This instrument is the most efficient mapping device currently available.
Aperture synthesis is an interferometric technique in which the product (complex correlation) of the output voltage from pairs of antennas is measured at many different baselines. Each baseline produces a sample point in the Fourier transform of the scene, and a map of the scene is obtained after all measurements have been made by inverting the transform. ESTAR is a hybrid real and synthetic aperture radiometer which uses real antennas (stick antennas) to obtain resolution along-track and aperture synthesis (between pairs of sticks) to obtain resolution across-track (Le Vine et al., 1994). This hybrid configuration could be implemented on a spaceborne platform.
The effective swath created in the ESTAR image reconstruction (essentially an inverse Fourier transformation) is about 45o wide at the half power points. The field of view is restricted to 45o to avoid distortion of the beam but could be extended to wider angles if necessary. The image reconstruction algorithm in effect scans this beam across the field of view in 2o steps. The beam width of each step varies depending on look angle from 8 to 10o, therefore, the individual original data are not independent, since each data point overlaps its neighbors. Contiguous beam positions can be achieved by averaging the response of several of these data points. This results in approximately nine independent beam positions. For this experiment the swath was restricted to approximately 35o. Another approach to using the data, especially in a mapping mode, is to interpret each of the original nonindependent observations as a sample point and then use a grid overlay to average the data. The final product of the ESTAR is a time referenced series of data consisting of the set of beam position brightness temperatures at 0.25 second intervals.
Calibration of the ESTAR is achieved by viewing two scenes of known brightness temperature. By plotting the measured response against the theoretical response, a linear regression is developed that corrects for gain and bias. Scenes used for calibration include black body, sky, and water. During aircraft missions, a black body is measured before and after the flight and a water target during the flight. Water temperature is determined using a thermal infrared sensor when available.
The ESTAR instrument was flown on a P-3B aircraft operated by the NASA Wallops Flight Facility. ESTAR was installed in the bomb bay portion of the aircraft during this mission. Flights were conducted at an altitude of 7.5 km and, therefore, the aircraft was pressurized. It should be noted that radiometer calibration is based on its operating environment. At a particular aircraft altitude this is quite stable, however, operating at drastically different altitudes (and associated thermal environments) requires separate calibrations. All P-3B flights were conducted at a single altitude to avoid this problem.
The original flightline configuration called for four parallel lines and a water calibration line. After reviewing the first day of data from the SGP97 mission, a problem with radio frequency interference (RFI) was found. This was a localized problem in an area at the same latitude as Oklahoma City. The source may be associated with the Oklahoma City airport. This was a critical problem because the area affected included the El Reno study area. The flight plan was modified to include two east west lines in this area (these are flown as a deviation in the last of the four long parallel lines). The line sequence is shown in Figure 1. This reconfiguration eliminated the RFI.
An attempt was made to conduct the flights exactly the same way on a daily basis . For the most part this was accomplished, however, instrument, weather, and logistical constraints resulted in some deviations which are described in Table 1.
In addition to the SGP97 region, a secondary study area was flown on selected dates. This area was part of the Cooperative Atmospheric Surface Exchange Study (CASES) located in Kansas . The flightlines for this area were flown after the standard SGP97 flightlines. Data were collected on four dates; June 20, July 2, July 12, and July 16.
During the SGP97 field campaign, a preliminary calibration was used for ESTAR. Data were processed into an image product within twelve hours of collection. This product provided valuable information for mission planning and quality control. The first step in quality control was the review of spatial and temporal features in these images.
Post processing of the ESTAR data consisted of refining the calibration, angular normalization, georegistration, temporal normalization, and RFI removal. Several different approaches to calibration were considered. These alternatives were developed based solely on calibration data sets. Following this, the alternatives were used to predict TB for the Little Washita watershed area. Results were compared to relationships that had been verified in the 1992 and 1994 experiments (Jackson et al., 1995). The approach selected was based on extensive water calibrations performed just prior to the field campaign.
As described in previous investigations with ESTAR (Jackson et al., 1995), because data are collected at varying observing angles and we wish to output a map product, it is necessary to normalize all of the data to a single angle. Here all data were normalized to nadir using the method described in Jackson et al. (1995).
These data are also reviewed for temporal consistency. It is possible that some instrument related drift can occur during the day and it is also possible that soil moisture and/or soil temperature may change enough during a flight that conditions may vary from the start to the end. Since we were interested in generating a snapshot of the region, the data are carefully reviewed for trends. During SGP97 the only time that such conditions were noted was preceding an instrument failure which resulted in the deletion of that days data from further analysis.
An ESTAR data record consists of the time and TB values for each beam position at that point in time. Global Positioning System (GPS) data collected during flight are used to georegister the center beam position of each data record. Then aircraft pitch, roll and yaw data were used to adjust the ground location of the center beam and all the footprint locations of that data record.
Criteria were established for identifying footprints that were contaminated with RFI. This was straightforward since RFI causes the TB to be unrealistically large. Once identified these footprints were dropped from the data set.
The product of all the steps described above was called the Level 1 ESTAR data and consisted of individual footprints described by a latitude, longitude, and TB. These footprints had a nominal size of 400 m. In order to produce map products and soil moisture estimates, these data were resampled to a standard 800 m grid as part of geographic information system (GIS). In this step, some spatial filtering is performed and small areas that may have had no observations due to RFI or flightline align ment were filled in. The resulting product is the Level 2 ESTAR data base provided here. The data for any given pixel can be compared on a temporal basis within the accuracy constraints of the process. These data are also fully georeferenced to all other data in the GIS.
The Level 2 ESTAR images are the input to the soil moisture algorithm described in Jackson et al. (1995). The algorithm is based on the inversion of the Fresnel Reflection Equation for horizontal polarization. For each pixel the following input data are required to apply the algorithm; soil temperature, vegetation type, vegetation water content, surface roughness, soil bulk density, and soil texture. Each of these inputs was generated as a plane of the GIS data base as described in the following sections .
Soil Temperature. The data collected as part of the Oklahoma Mesonet include observations of soil temperature observations of air temperature and soil temperature at 10 cm under sod every 5 minutes. A standard reference time for each flight was selected, usually 11:00 and all stations available were used to compute the effective soil temperature following the method described in Choudhury et al. (1982). Then a grid resampling program was used to generate an 800 m database of effective soil temperature for each day.
Vegetation Type. Vegetation type is used to define the functional form of the vegetation attenuation (Jackson and Schmugge, 1991). The SGP97 region is not very complex in terms of vegetation types. The data used are based on the land cover classification performed by Doraiswamy et al. (1998) and available through this site. Thematic Mapper (TM) data were used to perform a land classification. On location surveys were used as part of this supervised approach. For each land cover/vegetation type category a vegetation parameter, b, utilized in the retrieval algorithm was assigned based on published data. The TM data have a resolution of 30 m, in order to compute a value of b for each 800 m pixel the individual b values were averaged.
Vegetation Water Content. The strategy used for generating vegetation water content involved three components. First, vegetation characteristics were measured at various locations on the ground during the field campaign. Nearly all of the test sites were sampled. Next, Normalized Difference Vegetation Index (NDVI) values were computed for the entire region using the TM data collected on July 2 5, 1997. Values of the NDVI were extracted for all test sites and relationships to the vegetation water content were established for specific vegetation types. Next, for each TM pixel at 30 m the land cover and NDVI were used to estimate the vegetation water content. These were integrated to generate an 800 m pixel value and a data plane in the GIS.
Surface Roughness. The technique used is based on Choudhury et al. (1979). For each land cover category a roughness parameter was estimated based on previous investigations in this region. Values within each 800 m pixel were averaged to establish the surface roughness GIS data plane.
Soil Bulk Density. For each land cover category a bulk density value was estimated based on samples collected as part of SGP97. Values within each 800 m pixel were averaged to establish the soil bulk density GIS data plane.
Soil Texture. The algorithm utilizes the dielectric mixing model presented in Wang and Schmugge (1980) which requires estimates of the percent sand and clay. A multi-layer soil characteristics data set for the conterminous United States (CONUS- SOIL) has been developed at Penn State's Earth System Science Center (ESSC). The State Soil Geographic Database (STATSGO) developed by the USDA Natural Resources Conservation Service (NRCS) served as the starting point for CONUS-SOIL. Complete documentation of the elements of the data set, as well as the original STATSGO data, and the procedures used to generate each of the elements of CONUS-SOIL are described on a world wide web site. One of the products available from this source is the soil texture classification of the surface soil on a 1 km grid. Based on soil texture samples and published data, standard percent clay and sand values were established for each soil texture category. Table 2 summarizes the values used. The data were resampled to the 800 m grid to create two data planes (sand and clay for the GIS).
Soil Moisture Maps. Using a verified algorithm, we applied this and the ancillary data sets to each of the ESTAR data sets on a pixel by pixel basis on each date that data were available to produce a series of surface soil moisture. This soil moisture represents the 0 to 5 cm surface layer. Additional details on the data, processing and verification can be found in Jackson et al. (1998).
Data Access and Contacts
FTP SiteThe ESTAR data is in the following GES DISC ftp site:
Points of Contact
SGP97 ESTAR Brightness Temperature and Derived Soil Moisture Data
- The Principal Investigator for the ESTAR instrument is:
- Tom Jackson
- USDA ARS Hydrology Lab
- Beltsville, MD 20705 USA
- Bldg. 007, Rm. 104, BARC-West
- E-mail : tjackson@hydrolab.arsusda.gov
- Voice: (301) 5048511
- Fax: (301) 5048931
- For more information using ESTAR data from GES DISC, contact:
- Hydrology Data Support Team
- Goddard Earth Sciences
Data and Information Services Center (GES DISC)- Code 610.2
- NASA Goddard Space Flight Center
- Greenbelt, Maryland 20771
- E-mail : hydrology-disc@listserv.gsfc.nasa.gov
- Vocie 301-614-5165
- Fax: 301-614-5268
References
Choudhury, B.J., Schmugge, T.J., Newton, R.W., and Chang, A. 1979. Effect of surface roughness on microwave emission of soils. J. Geophysical Research, 84: 5699-5706.
Choudhury, B. J., Schmugge, T. J., and Mo. T. 1982. A parameterization of effective temperature for microwave emission. J. Geophysical Res., 87: 1301-1304.
Doraiswamy, P., Stern, A. J., and Cook, P. W. 1998. Classification techniques for mapping biophysical parameters in the U. S. Southern Great Plains. Proceedings of the Int. Geoscience and Remote Sensing Symp., IEEE Cat. No. 007803-4403: 862-865.
Jackson, T. J. and Schmugge, T. J. 1991. Vegetation effects on the microwave emission of soils. Remote Sensing of Environment., 36: 203-212.
Jackson, T. J., Le Vine, D. M., Swift, C. T., Schmugge, T. J., and Schiebe, F. R . 1995. Large area mapping of soil moisture using the ESTAR passive microwave radiometer in Washita'92. Remote Sensing of Environment, 53: 27-37.
Jackson, T. J., Le Vine, D. M., Hsu, A. Y., Oldak, A., Swift, C. T., Isham, J., and Haken, M., 1998: Soil moisture mapping at satellite spatial and temporal scales: the Southern Great Plains Hydrology Experiment. Will be submitted to IEEE Trans. on Geoscience and Remote Sensing.
Le Vine, D. M., Griffis, A. J., Swift, C. T., and Jackson, T. J. 1994. ESTAR: a synthetic aperture microwave radiometer for remote sensing applications. Proc. of the IEEE, 82: 1787-1801.
Wang, J. R. and Schmugge, T. J. 1980. An empirical model for the complex dielectric permittivity of soils as a function of water content. IEEE Trans. on Geosci. and Remote Sensing,. GE-18: 288-295.
 |
Last Revised: Aug 12, 2008 2:55 PM EST Hydrology Data Support Team -- hydrology-disc@listserv.gsfc.nasa.gov Web Curator: Anthony Drake NASA official: Steven Kempler, DAAC Manager -- Steven.J.Kempler@nasa.gov |