Remote Sensing

Home > Research Overview > Remote Sensing

Launch

The term remote sensing encompasses methods that employ electromagnetic energy as the means of detecting and measuring target characteristics, commonly with sensors mounted on aircraft or satellite platforms. 

SWRC remote sensing research has been directed toward four general activities:

  1. Calibration and signal processing
  2. Algorithm and model development
  3. Field experiments and inter-disciplinary programs
  4. Data archive

 

Calibration and signal processing

Sensor Calibration

Sensor Calibration

Arizona ARS scientists have conducted in-flight calibrations of airplane and satellite-based sensors in cooperation with scientists from the University Of Arizona Optical Sciences Center.

The results of over ten years of work have shown that:

  1. over time, sensor sensitivity degrades up to 27%
  2. sensor degradation depends on type and filter
  3. in-flight sensor calibration uncertainty is less than 5%, and
  4. regular calibration result in high-quality images

These results show that in-flight sensor calibration is both accurate and essential for proper interpretation of remote sensing images for monitoring temporal changes in crop and soil conditions.

 

Reflectance factor retrieval

Reflectance

Remote sensing products becomes infinitely more valuable if the digital numbers (dn) can be converted to a value that is independent of atmospheric and insulation variations, and is thus comparable over time for monitoring seasonal crop and soil 

conditions. The reflectance factor (the ratio of reflected and incident radiation at the surface) is a such a value, and has become the basic measurement required for most remote sensing algorithms and models. Retrieval of surface reflectance from image radiance can be accomplished through complicated measurements of atmospheric conditions, and modeling of atmospheric radiative transfer. ARS scientists have also explored the use of canvas reference tarps that could be deployed during each overpass and on-farm targets such as landing strips or dirt roads that could be used to normalize the images to a common reference.

 

Bidirectional reflectance distribution function

Spectral images are affected by interactions of the illumination source of the image and the pixel location within the image due to the bidirectional nature of most natural targets. These are often referred to as "geometric effects" or "viewing 

Bidirectional reflectance distribution function (BRDF)

geometry effects" because the effect is primarily related to geometric configuration of the sensing system. The need to correct for geometric effects has been recognized ever since remote sensing has been used for scientific, quantitative 

purposes, but only in the past decade have the computing tools been available to make the correction practical for large imagery. The bi-directional reflectance distribution function (BRDF) models that have been developed to correct for this effect tend to be complicated and often cannot be generalized over images acquired over different surface types. ARS scientists have been working to develop simplified methods for practical and operational corrections of geometric effect.

 

Algorithm and Model Development

Evaporation

Fields of a) air temperature, b) wind speed, c) water vapor mixing ratio, and d) precipitable water at 4-km resolution over a region of 100 by 100 km, derived from near real-time runs of the refined Regional Atmospheric Modeling System (RAMS') for the Upper San Pedro Basin on 9 June 1997. North is at the top, the scale is approximately one inch to 40 km.

Monitoring evaporation (E) at watershed scales is important for assessing the effect of climate and management on natural ecosystems. Techniques have been developed to evaluate E with remote sensing, which is the only technology that can efficiently and economically provide distributed estimates of E on a regional scale. These techniques are of three classes: empirical approaches, physically-based analytical approaches, and numerical models. Empirical techniques generally use the spatially-distributed multi-spectral image data to extrapolate a single (or multiple) in situ measurement(s) of E to a larger, heterogeneous surrounding region, or develop an empirical relation between a time series of in situ E measurements and multi-spectral measurements that can be applied to multi-spectral images to produce maps of E. Physically-based, analytical techniques directly evaluate energy balance (and thus E) through a combination of remotely sensed measurements of surface reflectance and temperature with in situ meteorological measurements. Numerical models use remotely sensed measurements as a source of intermittent grid-based information for soil-vegetation-atmosphere (SVAT) models to evaluate regional E continuously on an hourly or daily basis

Soil Moisture

Fields of a) air temperature, b) wind speed, c) water vapor mixing ratio, and d) precipitable water at 4-km resolution over a region of 100 by 100 km, derived from near real-time runs of the refined Regional Atmospheric Modeling System (RAMS') for the Upper San Pedro Basin on 9 June 1997. North is at the top, the scale is approximately one inch to 40 km.

Knowledge of distributed surface soil moisture content (~5cm depth) is important for many hydrologic applications including mapping rainfall events, monitoring differential drying patterns, and assessing water availability for plant growth. Surface soil moisture can also be used to parameterize soil water simulation models that estimate soil moisture content with depth in the plant rooting zone. Though the demand for distributed surface soil moisture information is high, the means for obtaining such information are few. There is some evidence that satellite-based Synthetic Aperture Radar (SAR) sensors could provide a regional assessment of surface volumetric soil moisture content. Theoretically, SAR backscatter detected by orbiting satellite-based sensors is directly related to the target dielectric constant (e'), where e' is the real part of a complex parameter that describes the electrical properties of a medium relative to the dielectric constant of "free space". For water, e'~80; for dry soil, e'~2. Consequently, an increase in soil moisture content changes e' markedly, and results in a strong sensitivity of the SAR signal to volumetric soil moisture content. In practice, SAR backscatter is also highly influenced by topographic features, vegetation density, and variations in small-scale surface roughness. ARS scientists have addressed the difficult task of converting single-channel SAR images directly into maps of regional soil moisture content for heterogeneous terrain.

 Vegetation

Vegetation density and distribution is one of the most important physical parameters controlling hydrological processes across geosphere-biosphere-atmosphere boundaries. Estimation of this parameter using remote sensing techniques has been associated with computation of vegetation indices, inversion of bidirectional reflectance distribution function (BRDF) models, and calibration of plant growth models with remotely sensed images. With advances in space technology, more remote sensing platforms are to launched with sensors suitable for such numerical and analytical approaches.Vegetation density and distribution is one of the most important physical parameters controlling hydrological processes across geosphere-biosphere-atmosphere boundaries. Estimation of this parameter using remote sensing techniques has been associated with computation of vegetation indices, inversion of bidirectional reflectance distribution function (BRDF) models, and calibration of plant growth models with remotely sensed images. With advances in space technology, more remote sensing platforms are to launched with sensors suitable for such numerical and analytical approaches.

 

Field Experiments and Inter-Disciplinary Programs

University of Arizona scientists making atmospheric measurements with a solar radiometer (upper) and ARS scientist making surface reflectance measurements (lower) in support of satellite sensor calibrations at White Sands, New Mexico

Monsoon '90

The Monsoon'90 multidisciplinary field campaign was conducted at the USDA ARS Walnut Gulch Experimental Watershed (WGEW) in SE Arizona during June-September 1990. The objective of this combined ground, aircraft and satellite campaign was to assess the feasibility of utilizing remotely-sensed data coupled with water and energy balance modeling for large area estimates of fluxes in semiarid rangelands.The Monsoon'90 multidisciplinary field campaign was conducted at the USDA ARS Walnut Gulch Experimental Watershed (WGEW) in SE Arizona during June-September 1990. The objective of this combined ground, aircraft and satellite campaign was to assess the feasibility of utilizing remotely-sensed data coupled with water and energy balance modeling for large area estimates of fluxes in semiarid rangeland.

 

Walnut Gulch '92

The Walnut Gulch '92 field campaign was conducted during the dry, early-monsoon, mid-monsoon, post-monsoon and "drying" seasons from April through November 1992 at WGEW. The overall research goal was to investigate the seasonal hydrologic dynamics of the region and to define the information potential of combined optical-microwave remote sensing.

Semi-Arid Land Surface Atmosphere (SALSA) Program

Semi-arid lands are especially vulnerable to the impacts of natural and human induced environmental stresses. The Semi-Aridland-Surface-Atmosphere (SALSA) program is an ARS-led,multi-agency research program that seeks to understand, model, and predict the consequences of widespread environmental (global) change on the water balance and ecological complexity of semi-arid river basins over a range of time scales. Combining both ground-based and remote sensing technologies, SALSA scientists from several countries are intensively studying the Upper San Pedro Basin of southeastern Arizona and northeastern Sonora, Mexico. The results of this effort will not only help the local community better manage resources in their basin but will also provide the scientific and technological understanding needed to resolve similar problems in other semi-arid regions of the USA and the world.

Landsat-7 Science Team Program

Landsat-7 Enhanced Thematic Mapper Plus (ETM+) color composite image (Upper San Pedro Basin, May 14, 1999)

A NASA-funded project has been underway for three years to develop operational techniques for the use of Landsat TM and ETM+ data for agricultural and natural resource management. The approach will include (1) supplementing Landsat TM data with synthetic aperture radar (SAR) data from other currently operating and proposed sensors to compensate for the infrequent coverage provided by Landsat, and (2) using the TM-derived determinations of surface plant and soil conditions to update modeled plant growth and evaporation based on physical models and meteorological data. The project is in the third stage of a three-stage schedule, where the first year's work was 

Landsat-7 Enhanced Thematic Mapper Plus (ETM+) thermal image (Upper San Pedro Basin, May 14, 1999)

based solely on existing data sets. The products were algorithms to allow "interchangeable" usage of SAR data with TM data for vegetation status and moisture condition determination, and a plant growth model designed to assimilate infrequent inputs of plant parameters from remotely sensed data. In years 2-3, Landsat TM and ERS-2 SAR images were acquired at two sites in Arizona to validate these techniques at the regional scale for heterogeneous landscapes. Current work is focused on using the combined remote sensing/modeling approach for investigations of the daily to interannual changes of rangeland vegetation in the semiarid Southwest, and for aiding such farm management decisions as scheduling applications of fertilizer and irrigation water. This project has been conducted as part of the NASA Landsat-7 Science Team which provides expertise and information on such activities as instrument pre-development, science planning, instrument calibration, field validation, and algorithm development.

Ranges

One half of all U.S. land and 60% of the landmass of the world is classified as rangeland and many of these rangelands have suffered severe degradation over the years. Ecological degradation and an increase in bare soil negatively impact livestock production, erosion control, biodiversity, and CO2 sequestration. To effectively manage these rangelands, managers need spatially and temporally distributed information on vegetation cover, biomass production, and green-up time. The RANGES program proposes to use NASA's EOS (Earth Observation System) products to provide critical information to end-users on an operational basis for livestock management, fire-fuel estimation, wildlife habitat analysis and rangeland health assessment. RANGES offers a) an operational procedure to transform EOS data into information products and b) a protocol that involves range managers, remote sensing scientists, and commercial firms to apply, and then to assess the value-added EOS products for management applications. Scientists will work hand in hand with range managers and ranchers to define, develop, evaluate, and prototype EOS-derived products. At the end of the project, an assessment of the EOS-derived products for rangeland management will be made by our end-users to determine their willingness to pay for the ongoing provision of EOS-derived products.

Earth Observation 1 (EO-1) Validation Team Program

The EO-1 Satellite at NASA Goddard Space Flight Center in the Clean Tent

ARS scientists at SWRC secured a grant from the National Aeronautics and Space Administration (NASA) to participate in validation of the Earth Observation-1 (EO-1) mission. The EO-1 satellite is scheduled for launch in August 2000 with a primary objective of evaluating selected technologies for maintaining the 27-year continuity of the Landsat data stream. The primary objective of the SWRC study is to determine the suitability of the EO-1 image quality and data products to meet the needs of Landsat-class observations, termed Landsat data continuity. One study site will be the USDA ARS Walnut Gulch Experimental Watershed (WGEW) near Tucson, Arizona. Using on-site measurements of atmospheric conditions and measurements of surface reflectance of large, uniform targets, it will be possible to determine the accuracy of image reflectance factor retrieval. For the same images, on-site measurements of soil and plant characteristics will be used to determine the accuracy of image-derived data products suitable for agricultural and rangeland applications.

 

Data Archive

Water Conservation Laboratory Image and Ground Data Archive (WIGDA'99)

Illustration of the image coverage (center) of the Upper San Pedro Basin and Walnut Gulch Experimental Watershed leading to the 100s of images and ground data files archived in WIGDA'99.

Since 1984, ARS scientists at SWRC and U.S. Water Conservation Laboratory (WCL) have been conducting large-scale, multidisciplinary remote sensing experiments at two Arizona field sites : Maricopa Agricultural Center (MAC) and the Upper San Pedro Basin (USPB). This continuing work has resulted in an accumulation of hundreds of spectral image files from a variety of satellite- and aircraft-based sensors (the "images"), and the association of those images with data files containing high-quality ground-based measurements of soil, plant and atmospheric conditions (the "ground data"). These images and the supporting ground data have been compiled in one location, and transferred in an orderly fashion to compact disks (CD-ROM). Each image on CD-ROM includes a companion "readme" file containing information on the acquisition data and location, processing level, file size and format, and any relevant comments about the image or the archiving procedure. Supporting files of ground, atmospheric and low-altitude aircraft measurements - collectively referred to in WIGDA'99 as "ground data" - were archived with an internal header describing techniques, instrumentation, location and other relevant information. Metadata on all archived images and ground data were entered into an Access database to link the information in the two data sets and to allow easy queries of either image or supporting ground data files. This work will allow long term archiving of and easy access to an exceptional remote sensing data set.