Purdue University/Laboratory for Applications of Remote Sensing Vegetation and Soils Research Data Summary: Eric Stoner Soils(781701)

(taken from project description)

Objective:

The general objective of this study was to define quantitatively the
relationships between soil reflectance and physiochemical properties
of soils of significance to agriculture and engineering. Selection of
soil samples with a wide range of important soil characteristics by
statistical stratification of continental United States climatic zones
permits the evaluation of climatic and genetic effects on the
relationships between multispectral reflectance and these soil
properties. A further objective is to define the relationships
sufficiently to design further research to quantify the contributions
which different soil components make to the multispectral
characteristics of specific soils.

Method:

Because of the need to provide a uniform moisture condition for
spectroradiometric analysis of the prepared soil samples, a procedure
was chosen which creates a one-tenth bar soil moisturetension on all
the soil samples (3,5). Two asbestos tension tables were constructed
and a 100 cm column of water was established to create a soil moisture
tension for up to 56 soil samples at one time. Sample holders were
designed and constructed of plastic rings 2 cm deep by 10 cm in
diameter with 60 mesh brass strainer cloth stretched taut and fastened
in a countersunk groove in one end. Sample holders were painted with
non-reflecting black paint to reduce unwanted reflection external to
the target of interest. After saturation of the soil filled, leveled
sample holders for about four hours, the samples were placed on the
tension tables for 24 hours in order to reach equilibrium. The
one-tenth bar moisture tension was desirable mainly for the ease with
which large numbers of samples could be prepared at uniform moisture
characteristics. Shortly after placement of each sample holder on the
sample table of the reflectometer for spectral readings, a portion of
the sample was transferred to a moisture tin, weighed, dried in a
forced air oven at 105 C, weighed again, and moisture content reported
as percentage of oven dry weight.

Quantification of Soil Properties

Modern soil classification systems emphasize the importance of
information about the quantitative compositions of soils. In order to
differentiate among soil groups, it is necessary to rely on laboratory
measurements of selected soil properties. Physical, chemical, and
engineering determinations of most soil properties follow well
established procedures of laboratory analyses. Certain of these soil
properties are selected as diagnostic criteria in the soil
classification process, based on their importance in understanding the
genesis of the soil. By a procedure of empirical correlation, critical
limits between sets of soils are established, designed to reflect the
influence of the soil forming factors of climate, parent material,
relief, biological activity, and time.

Quantitative measurements of soil spectral properties have become
available as a diagnostic tool for the soil scientist with the advent
of such instruments as the Exotech Model 20C
spectroradiometer. However, the climatic and genetic effects on the
relationships between measured spectral properties and specific
chemical, physical, and biological properties of the soil are not well
understood. Whereas soil color is used as diagnostic criterion in the
U.S. Soil Taxonomy (7), the determination of soil color by comparison
with a color chart continues to be a rather nonquantitative and
subject procedure. Spectral characterization of soil &color& by means
of quantitative spectroradiometric measurements may add to the
precision with which soils can be differentiated. With this increased
precision of soil spectral characterization, the relationships with
the more important diagnostic soil characteristics or qualities that
are not so easily and accurately observed may be better understood.

EXPERIMENTAL APPROACH

Stratification and Sampling

Approximately 250 soils, representing a statistical sampling of the
more than 10,000 soil series in the United States were selected for
this investigation. Selections were made from a list of the more than
1300 Benchmark soil series representing those soils with a large
geographic extent and whose broad range of characteristics renders
these soils so widely applicable for study. Stratification of soil
sampling was based on series type location within climatic
zones. Climatic strata included the frigid, mesic, thermic, and
hyperthermic soil temperature regimes as defined by the U.S. Soil
Taxonomy (2,6,7) as well as the perhumid, humid, subhumid, semiarid,
and arid moisture regions as identified by Thornthwaite's 1948
Moisture Index (8). A random selection procedure was used within each
stratified climatic zone to select a number of soils series
approximately in proportion to the geographic extent of that
region. Considerations were also made
to include soils which represent the major parent material categories
and the ten soils orders of the U.S. Soil Taxonomy (7).

The Soil Survey Investigation Division of the Natural Resource
Conservation Service (USDA) cooperated with LARS in the collection of
field samples from 39 states. Duplicate field samples were collected
for all Benchmark soil series requested: one sample from a site near
the type location for the current official series, and one sample from
a site located from one to 32 kilometers from the first site and in a
different mapping delineation. Natural Resource Conservation Service
field survey personnel were responsible for sample collection of
Benchmark soils in their locality. Of the original list of
approximately 250 Benchmark soils requested, the NRCS has collected,
properly identified, and forwarded 240 Benchmark soils, or 480
duplicate soil samples to LARS. This excellent response of over 95
percent of the requested samples forms an outstanding collection of
soil samples for detailed chemical, physical, and spectral
analysis. All samples conform to the central concept of each
individual soil series as each soil would be identified and mapped by
an experienced soil surveyor in the field.

Preparation of Soils for Analysis

After receipt of the soil samples and initial data logging, samples
were dried, crushed, and sieved to remove all particles larger than 2
mm diameter. Cardboard containers were used to store subsamples of
each soil sample for chemical, physical, spectral, and engineering
determinations.

Spectral Measurements

The Exotech Model 20C was used in an indoor configuration with a
bidirectional reflectance factor reflectometer (1,4) in order to
obtain spectral readings in the 0.52-2.3 um wavelength range. The
illumination source was a 1000 watt tungsten iodine coiled filament
lamp which transfers a highly collimated beam by means of a
paraboloidal mirror to the sample-viewing plane. Detector height above
the sample was 2.4 m, and a 3/4 field of view required that the sample
holder be approximately 10 cm in diameter.

see

&http://cobweb.ecn.purdue.edu/~frdata/FRData/data/&

for additional information and to download the dataset.