Printer friendly version

Direct methods

These use on-board calibration sources such as lamps or panels that may be positioned in the optical path. Limitations of this method include the susceptibility of the lamp or panel to degradation over time and the incomplete characterisation of the optical system due to the lamp or panel either not completely filling the field-of-view or being inserted part-way through the optical path.

Indirect methods (Vicarious calibration)

'Vicarious calibration' refers to the process by which a calibration is established using a method independent of that which was used for the primary calibration. Over the last 10-20 years vicarious calibration has become widely adopted as the means to provide independent assurance of the quality of remotely sensed data from spaceborne sensors. Much of the credit for this must go to researchers in the Remote Sensing Group in the College of Optical Sciences at the University of Arizona , who have devised and published various methodologies for vicarious calibration (Thome, 2004).

High altitude dry lake beds are widely used as vicarious calibration targets, for example White Sands Missile Range, New Mexico, Railroad Valley, Nevada, and Lunar Lake, Nevada in the US (Thome et al., 1998), Salar de Uyuni in Bolivia, Salar de Arizaro and Barreal Blanco in Argentina, Tuz Gölü in Turkey, and Lake Eyre in Australia . Seasonal changes in reflectance in response to fluctuations in groundwater levels have been observed at the White Sands site (Thome et al., 1993) and at Railroad Valley (Bannari et al., 2004), amongst others. Furthermore, those sites in which the evaporation of saline water has resulted in salt crusts may have significant microrelief as well as being subject to seasonal and episodic changes in reflectance. Polygonal cracks are also a feature of the dry lake beds which are commonly used for vicarious calibration (Abdou et al., 2000). Other terrestrial surfaces that have have been used for this purpose include La Crau sèche, an ancient river delta in southern France (Rondeaux et al., 1998), an extensive site near Dunhuang in the Gobi desert (Wu et al., 1997) and the Dome Concordia area in Antarctica (Six et al., 2004).

As an alternative to natural targets for this purpose, artificial targets can be installed, such as calibrated tarpaulin sheets. A practical limitation of this approach is the size of the sheets needed, which should be at least 5 x 5 pixels in size to allow for uncertainty in geometric correction and the point spread function of the detector.

Laying down a large plastic sheet for use in validating the radiometric calibration of an airborne sensor.

Artificial targets laid out during the NCAVEO 2006 Field Experiment.

The ‘reflectance-based' and ‘radiance-based' methods proposed by Slater and colleagues have also been combined to create a generalised approach to quality assurance and stability reference monitoring (QUASAR) for Earth observation data (Teillet et al., 2001a). The conceptual basis of this has been developed further in the proposal for a Global Instrumented and Automated Network of Test Sites (GIANTS) (Teillet et al., 2001b), each of which would be based upon automated sensor technologies such as that described in the Intelligent Sensorweb for Integrated Earth Sensing (ISIES) project (Teillet et al., 2005). The growing demand for well-characterised ground targets, for use both in vicarious calibration and for atmospheric correction (Moran et al., 1992; Moran et al., 2003) means that candidate type surfaces need to be studied in detail to assess temporal dynamics. Highly reproducible measurements using well-characterised field spectroradiometers are essential if uncertainty arising from the instrument and its method of use is to be separated from uncertainty introduced by the environment.

The considerable difficulty inherent in any vicarious calibration system relying upon terrestrial targets has encouraged some scientists to seek a space-based solution, whether the moon (Kieffer & Wildey, 1985), distant stars (Bowen, 2002), or a dedicated orbital standards laboratory such as is envisaged by a proposal known as TRUTHS (Traceable Radiometry Underpinning Terrestrial and Helio Studies; Fox et al., 2003).

An example of the reflectance-based method:
Thome, K.and Helder, D., Radiometric calibration of IKONOS using ground-reference test sites
http://ldcm.gsfc.nasa.gov/library/HSRCIW02/Ground_Ref_Test_SitesThome.pdf

Examples of vicarious calibration sites

Barreal Blanco, Argentina
Dunhuang, China
Ivanpah Playa, California
La Crau, France
Libyan Desert
Lunar Lake, Nevada
Newell County, Nevada
Railroad Valley, Nevada
Salar de Arizaro, Argentina
Tinga Tingana, Australia
White Sands, New Mexico

Vicarious calibration sites in Asia

Dunhuang
The Dunhuang test site is located in the Gobi desert in north-west China, about 35 km west of the city of Dunhuang (Gansu Province). Dunhuang was once an important settlement on the ancient Silk Road and is now a tourist centre. The calibration area is situated on a stabilised alluvial fan approximately 1160 metres above sea level. The area used for vicarious calibration is approximately 400 m x 400 m in the centre of the fan, and the surface comprises cemented gravels, with no vegetation.

Sources of meteorological data for the site include the Dunhuang-PAM automated weather station which is part of the Asian Automatic Weather Station Network. The site was one of three in China identified by Wu et al. (undated) for satellite radiometric calibration purposes. The radiometric characteristics of the site were described by Wu et al. (1997), who also made some atmospheric measurements using sunphotometers. The atmospheric aerosols over the site have also been studied by Xianjun and Yonghao (1995), and found to be typical of a rural continental location, although some larger particles were observed, possibly influx from the sand dunes to the north-west. Sandstorms affect the site on around 8 days per annum and atmospheric dust is a significant factor on around 60 days per annum (Wu et al., 1997).

Figure 1. Researchers collecting data at the Dunhuang calibration site.

Return to top of page

Vicarious calibration sites in Australia

The CSIRO hosts an excellent website describing a number of test sites in Australia, some of which have been used for radiometric calibration of aircraft and satellite sensors, while others are used for validation purposes. The website includes descriptions of the sites and characteristic meteorological data. Tinga Tingana is a highly reflective site north of the Flinders mountain range at latitude 29.00 deg. S, longitude 139.75 deg. E. The CSIRO website notes that it is difficult to access on the ground.

Figure 2. Tinga Tingana from CHRIS-PROBA, 16th Feb 2004

Return to top of page

Vicarious calibration sites in Europe

La Crau

The La Crau calibration site is 400 m x 400 m in size, and located in the centre of "La Crau Seche" (latitude 43.50 deg. N, longitude 4.87 deg. E), about 50 km north-west of Marseille. The site was selected from a survey of satellite images in preparation for calibration of the SPOT 4 and 5 sensors (Santer et al., 1992). The area has a thin pebbly soil with sparse vegetation cover. The radiometric properties of La Crau have been studied by Gu et al. (1990, 1992) and Santer et al. (1992), whilst Rondeaux et al. (1998) have assessed the suitability of the area immediately around the central area for its suitability as a calibration site for large field-of-view sensors such as ATSR-2.

Figure 3. The La Crau calibration site. (source : CNES, 1998).

Return to top of page

 

Vicarious calibration sites in North America

Ivanpah Playa, California
Lunar Lake, Nevada
Railroad Valley, Nevada

Figure 4. AirMISR image of Lunar Lake, Nevada, 30 Jun 2001 (source : http://www-misr.jpl.nasa.gov/mission/valwork/val_reports/010630_lunar_lake/airmisr.html)

Time series of images of the Railroad Valley cal/val site from CHRIS-PROBA:

30th April 2004

9th Sept 2004

16th Sept 2004

22nd Nov 2004

17th Dec 2004

 

White Sands, New Mexico

White Sands Missile Range (WSMR) has a long history of use as a calibration target for Earth observing sensors. The site is located in a mountain basin in south-central New Mexico (latitude 32.23 deg. N, longitude 106.28 deg. W). The area is extensive (35 km x 50 km) and comprises a mixture of salt pans and dunes at an approximate altitude of 1250 m above sea level. Although the area has very suitable weather conditions, with many clear days, the calibration site is not ideal because its reflectance is greatly affected by changes in soil moisture. The surface also deviates considerably from a lambertian response at high solar and viewing zenith angles (Wheeler et al., 1994). However, the University of Arizona group report that the 500 m x 500 m area within the WSMR known as Chuck Site (latitude 32.92 deg. N, longitude 106.35 deg. W), has a near-lambertian response.

Figure 10. White Sands National Park, New Mexico (source : http://digilander.libero.it/sarodb/usaparks/foto/nm/ws1.jpg)

Newell County, Alberta

This is an unusual vicarious calibration site, because it is vegetated. It comprises an area of rangeland 7 km x 7 km, at an altitude 750 m above sea level. (Latitude 50.30 deg. N, longitude 111.64 deg. W).

Return to top of page

Vicarious calibration sites in South America

Barreal Blanco, Argentina

Salar de Arizaro, Argentina

Return to top of page

Vicarious calibration sites in North Africa

Libyan Desert

Figure 11. Libyan Desert test site from CHRIS-PROBA, 30th Aug 2004

The same area from CHRIS-PROBA on 8th Feb 2004, presumably during a sandstorm.

 

Return to top of page

Last updated 21/08/2007
Partners