Tải bản đầy đủ - 0 (trang)
III. Effects of Soil Constituents on Soil Reflectance

III. Effects of Soil Constituents on Soil Reflectance

Tải bản đầy đủ - 0trang







Wavelength (pm)

FIG.6. Soil spectral curve forms defined by Stoner and Baumgardner (1981); reflectance is

recorded as the bidirectional reflectance factor. curve a, Organic dominated; curve b, minimally

altered; curve c, iron affected; curve d, organic affected; curve e, iron dominated.

number of reflectance spectra of soils from other parts of the world, they

derived five soil spectral curve forms (Fig. 6 ) to which they established some

general genetic, physical, and chemical relationships to each spectral curve

form (Table I).

This capability to obtain calibrated data across the visible and infrared

reflectance spectrum provides an important new tool to soil scientists (Cipra

et al., 1971a). The remainder of this section will present the results of

observations made on those soil constituents which account for most of the

variation in soil reflectance.


It is a common observation that most soils appear darker when wet than

when dry. This results from decreased reflectance of incident radiation in the

visible region of the spectrum. Evans (1948), who presented reflectance curves

for three soils in both the wet and dry state, found that the wet samples

showed lower reflectance. Unfortunately, he provided no information about

the soil series or moisture contents of the soils. Brooks (1952) reported 10%

reflectivity of moist Yo10 fine sandy loam over the wavelength range of 0.4 to

2.5 pm but did not indicate the moisture content; for the dry condition, he

reported a reflectivity of 30%. Kojima (1958b) studied the effect of moisture

content on the color of 16 soils. His results, reported in Munsell color

notation, also showed a decrease in reflectance with an increase in moisture.

He made no reference to energy changes related to reflectance and moisture


In “Soil Taxonomy,” the Soil Conservation Service standard for soil

survey over much of the world, the range of change in soil color upon wetting

is given as varying between 1/2 and 3 Munsell color steps (Soil Survey Staff,

1975). No formulas are proposed for predicting change in color between the

Table I

Characteristics of Surface Samples of the Five Mineral Soils Represented by the Spectral Curve Forms in Fig. 6#

Reflectance curve form

Soil series

Horizon sampled

Soil subgroup

Sample location

Climatic zone

Parent material

Drainage class

Textural class

Moist soil color

Munsell color


Organic matter

Iron oxide

Moisture at 0.1

bar tension



















Typic Haplaquoll

Champaign Co.,

Ill., USA

Humid mesic

Loess over glacial


Poorly drained

Silty clay loam

lOYR 211




Typic Calciorthid

Lea Co., N. Mex.,


Semiarid thermic

Fine-textured alluvium

or lacustrine

Well drained

Loamy fine sand

lOYR 513




Typic Hapludalf

Rutherford Co.,

Tenn., USA

Humid thermic

Clayey limestone


Well drained

Silty clay loam

1.5YR 416

Strong brown



Alfic Haplorthod

Delta Co., Mich.,


Humid frigid

Glacial drift

(Not given)


Typic Haplorthox


Parana, Brazil

Humid hyperthermic


Well drained

Fine sandy loam

7.5YR 312

Dark brown

Excessively drained


2.5YR 316

Dark red








0.8 1%










5.61 %






From Stoner and Baumgardner (1981)






wet and the dry states. Soil surveyors correct for differences in moisture by

comparing soil color with Munsell standards at the two soil moisture levels

known as air dry andjeld capacity. The directions for a wet reading specify

color at field capacity as the estimated color observed after moistening a

sample and comparing the color with that of Munsell standards as soon as

the visible moisture films have disappeared.

Angstrom (1925) attributed this darkening effect of moisture in soils to

internal total reflections within the thin water film covering soil particles. It

was felt that a portion of the energy would not be reflected to space but would

be re-reflected between the surface of the particle and the surface of the water

film. Planet (1970) indicated that the reflectance difference of a soil between

its dry and wet states could be determined if the following factors were taken

into account: (1) variations in the index of refraction of the water due to

dissolved soil constituents, reflectance being known to decrease with an

increase in the index of refraction of the transmitting medium; (2) changes in

the physical nature of soil particles by the presence of water; and (3)

similarities in the indices of refraction of the soil and water leading to the

Christiansen effect.

Hoffer and Johannsen (1969) showed that moist soils had an overall lower

reflectance than their dry counterpart in the 0.4-2.5-pm wavelength region.

Bowers and Hanks (1965) noted a lowering in reflectance for Newtonia silt

loam (Typic Paleudoll) at six increasing soil moisture contents over the

wavelength range from 0.5 to 2.5 pm. Hunt and Salisbury (1971) found the

spectral reflectance curve of montmorillonite to be similar to that of the

Newtonia silt loam used in the Bowers and Hanks study. The spectral

reflectance curve of montmorillonite was found to be dominated by very

strong absorption bands at 1.4 and 1.9 pm, which were assumed to be caused

by “bound” water typical of montmorillonite. Usually a weaker band was

noted at 1.16 pm and was thought to be due to adsorbed water.

The major spectral reflectance features of kaolinite were found to be very

strong hydroxyl bands in the “near” infrared centered near 1.4 and 2.2pm

(Hunt and Salisbury, 1971). As would be expected, the low amount of bound

water present resulted in weakness of the band at 1.9 pm.

Bowers and Smith (1972) showed that soil water content could be

measured by transmitting a beam of 1.94-pm energy through a methanol-soil

extract. Later Bowers et al. (1975) improved the design of a spectrophotometer to make such measurements.

Obukhov and Orlov (1964) observed that the spectral curve does not

change in appearance upon wetting of soil and that the ratio of the reflectance

of moist soil to that of dry soil remained practically constant in the visible

portion of the spectrum. It was also noted that the decrease in reflectance was

greater upon wetting of forest soils containing little organic matter than upon



wetting of prairie soils high in organic matter. Condit (1970, 1972), in his

study of 160 soils from 36 states in the United States, was able to identify

three characteristic shapes of reflectance curves in the 0.32- 1.0 pm wavelength range. Although the curve shape was not noted to change between dry

and wet soil reflectance readings, the soil moisture content was not reported

for any of the soil samples. The need for carefully controlled moisture tension

equilibria and soil moisture content determination in soil reflectance studies

was emphasized by Beck et al. (1976).

The shape of soil reflectance curves is affected by the presence of strong

water absorption bands at 1.45 and 1.95 pm, and occasionally weaker water

absorption bands at 0.97, 1.2, and 1.77pm. Specifically, these bands are

overtones and combinations of the three fundamental vibrational frequencies

of the water molecule which occur beyond 2.5 pm (Bowers and Smith, 1972).

The band at 1.94pm, a combination of the v 2 and v 3 fundamental frequencies,

is the most sensitive to water and has been found best for relating reflectance

measurements to soil moisture content (Bowers and Hanks, 1965). An

absorption band at 2.2pm was not identified in early studies but was later

identified as a vibrational mode of the hydroxyl ion (Hunt and Salisbury,

1970). Absorption due to the hydroxyl ion also gives rise to a band at

1.45 pm, the same as that of liquid water. The appearance of the 1.45-pm

band without the 1.95-pm band indicates that hydroxyl groups and not free

water are present in the material. Sharp bands at 1.45 and 1.95-pm indicate

that the water molecules are located in well-defined, ordered sites, while

broad bands at these wavelengths indicate that they are relatively unordered,

as is often the case in naturally occurring soils (Fig. 7). Weak absorption

bands at 0.97, 1.2, and 1.77 pm correspond to the absorption bands observed

in transmission spectra of water of a few millimeters in thickness (Lindberg

and Snyder, 1972). These weak water absorption bands can be assigned in

terms of water vibrations, with the additional combination of librations of the

water molecule in the 1.77-pm band (Hunt et al., 1971a).





Wavelength (pn)


FIG.7. Spectral curves from duplicate samples of a highly reflective soil annotated with

prominent iron and water absorption bands; reflectance is recorded as BRF. -,

sample 1;

---, sample 2.



Beck et al. (1976) found that of several factors studied, soil moisture had

the greatest influence on soil reflectance at the one-third bar moisture tension

level. Because of the impracticality of measuring soil reflectance in the field in

the 1.95-pm water absorption band (also a region of strong atmospheric

water absorption), reflectance in the 1.50-1.73-pm wavelength region was

suggested as the best possibility for mapping water content in surface soils.

Although the importance of soil moisture to reflectance was recognized by

Montgomery and Baumgardner (1974) and Montgomery (1976), the contribution of this parameter to soil reflectance was not evaluated quantitatively

because of the air dry state in which all of the soil samples were measured.

Stoner (1979) found soil moisture content to be the most important variable

for explaining reflectance differences in the 2.08-2.32-pm band, similar to one

of the middle infrared bands of the thematic mapper scanner of Landsats 4

and 5.

Bowers and Smith (1972) showed that a linear relation between absorbance and percentage soil water was adequate for moisture determination

from air dry to the moisture equivalent. Peterson et al. (1979) assumed that

the integrated effect of all factors contributing to soil reflectance would

exhibit a constant shift in reflectance for a given soil observed under different

moisture tensions. They compared the reflectance values at 0.71 pm for 15

different surface samples of Alfisols and Mollisols in Indiana. A plot of the

bidirectional reflectance factor of soils equilibrated at 15 bar moisture

tension, R,,, bar), versus the bidirectional reflectance factor of oven dry soils,

Rs(ovendry), for these 15 soils gave an r2 value of 0.95. The equation for

predicting RS(,bar) from Rs(ovendry) data at 0.71 pm is

Rs(15 bar)

= 1.685 +




The equation for calculating the reflectance at 0.71 pm for soils at 0.3 bar

moisture tension, R s ( 0 . 3 bar), from

dry) data is

& ( 0 . 3 bar)

= Oe709-k 0*487Rs(oven



This work was followed by a study of the relation of wetting to soil

reflectance for 57 soils supplied by the Soil Conservation Service (SCS) from

selected sites of soil series benchmark soils within the 48 contiguous states in

the United States (Peterson, 1980). Samples were selected to give a wide

range of soil reflectance values. Regression values of loss in reflectance upon

wetting to 0.1 bar versus oven-dry reflectances varied with the wavelength

band used. The r2 value for reflectance at 0.76 to 0.90 pm was 0.65, that for

0.45 to 0.52 pm was 0.79, that for 1.42 to 1.52 pm was 0.92, and that for 1.92 to

2.02 pm was 0.96 (Fig. 8).

Thus, it was found that the effect of soil moisture on reflectance among

soils with different chemical and physical properties could be generalized if






0 40













g 10








RS(oven dry) (%)

FIG. 8. Relationship between middle-infrared (1.92-2.02 pm) reflectance of oven-dry soil

and middle-infrared reflectance of oven-dry soil minus middle-infrared reflectance of soil at 0.1

bar moisture tension. 2 = 1.92-2.02 pm; r2 = 0.96.

the soil moisture content were related to the energy with which it is held by

the soil rather than expressed in percentage of the dry weight of the soil

(Peterson et al., 1979). This approach has long been used by soil and plant

scientists to relate in a universal way the relationships of soil moisture to

plant growth and plant behavior. Soil scientists have capitalized on the fact

that as soils become progressively drier below the field capacity level, the soil

water is held predominantly in an ever thinner film on the particle surfaces

and with an ever greater force. Awareness of this phenomenon has led soil

scientists to describe soil water in terms of the work per unit mass of water

that has to be done to change the soil water’s energy status to that of pure free

water (Kohnke, 1968).

The most common term used to express the free energy level of soil

moisture is soil moisture tension, expressed in centimeters or bars, where 1

bar = 1 x lo6 dyn/cm2 or 0.98692 atm. Using such measurements the

relation of soil moisture content to plant growth can be readily generalized.

For example, all plants will vary closely to the soil water tension a t the

“wilting point” or at approximately 15 bar, regardless of the kind of soil or

the amount of water present by volume or weight. A basis for this assumption

is the long-held view of soil scientists, deduced from indirect evidence, that in

soils drier than field capacity the water films around the particles are of

uniform thickness for all soils when they are at a uniform tension. Evidence of

this was presented by Low (1980), who showed that at uniform spacing of

clay particles in a water suspension the suction (tension) on the clay is the

same for all clays. This has lately been verified by further evidence now being

prepared for publication. This, together with the view that the thickness of












2.0 '





Wavelength (urn)


FIG.9. Spectral (BRF) curves of a Typic Hapludalf soil at four different moisture tensions:

oven dry; --, 15 bar; ---,0.3 bar; * - - , 0.1 bar.

the water on the surface of a particle influences reflectance (Rao and Ulaby,

1977; Strandberg, 1968), provides reason for expecting the influence of soil

moisture on reflectance to be readily generalized on the basis of soil moisture

tension (Fig. 9).



Soil organic matter content and thL composition of organic constituents

are known to have a strong influence on soil reflectance. A general observation has been that as organic matter content increases, soil reflectance

decreases throughout the 0.4-2.5-pm wavelength range (Hoffer and Johannsen, 1969). Baumgardner et al. (1970) found that organic matter content plays

a dominant role in bestowing spectral properties to soils when the organic

matter content exceeds 2.0%. As the organic matter content drops below

2.0%, it becomes less effective in masking the effects on reflectance of other

soil constituents. Although it was not elaborated by Condit (1970, 1972), his

Type 1 and Type 2 curves corresponded, respectively, to the reflectance

curves of high surface organic content Mollisols and low surface organic

content Alfisols (Cipra et al., 1971b). In a similar manner, curve forms

described by Stoner and Baumgardner (1981) as organic dominated and

organic affected owe their character to the elevated content (>2%) and

decomposition state of organic matter (Fig. 6).

Organic constituents, including humic and fulvic acid and nonspecific

compounds including decomposing plant residues, are known to influence

soil reflectance to differing degrees (Obukhov and Orlov, 1964; Vinogradov,

1981), although the contribution of each has been difficult to quantify. The

decomposition state in organic soils has been observed to alter drastically

nonvisible reflectance (Fig. $0).The high reflectance of fibric soil materials in

the infrared region resembles the infrared reflectance of senesced leaves



Wavelength (pm)

FIG.10. Spectral (BRF) curves of three organic soils exhibiting significantly different levels

of decomposition. Curve a, fibric; curve b, hemic; curve c, sapric.

(Gausman et al., 1975). This increased infrared reflectance has been attributed to tissue morphology in which an increased number of air voids

provides more air-cell interfaces for enhanced reflection.

Oxidation of organic matter in a soil sample with H z O z resulted in

increased reflectance from 0.44 to 2.4 pm, although the difference in reflectance beyond 1.3 pm became very small (Bowers and Hanks, 1965). Mathews

et al. (1973a) destroyed the organic matter in a 12.8% organic matter silty

clay soil, with the resulting reflectance being increased greatly in the spectral

region from 0.4 to 1.3 pm, while the reflectance actually decreased slightly in

the region from 1.5 to 2.4 pm.

Regression studies indicated that organic matter content could be related

to soil reflectance by a curvilinear exponential function (Schreier, 1977).

Mathews et al. (1973b) found that organic matter correlated most highly with

reflectance in the 0.5-1.2-pm range, while Beck et al. (1976) suggested that the

0.90-1.22-pm range was best for mapping organic carbon in soils. Montgomery (1976) indicated that organic matter contents as high as 9.0 % did not

appear to mask the contributions of other soil parameters to soil reflectance.

Montgomery differed from Beck in recommending the visible wavelength

region as the best for spectral measurement of organic matter content in soils.

Stoner (1979) showed organic matter to be the single most important

variable to explain reflectance differences in the spectral region 0.52-1.75 pm,

while the strongest correlations occurred in the visible wavelengths.

In laboratory measurements with a color-difference meter, Page (1974)

found that nearly 80% of the total variation in organic matter content of 96

Atlantic Coastal Plain soils could be accounted for by reflectance.



In the literature one of the early reports of the effect of soil particle size on

reflectance was made by Zwerman and Andrews (1940). Working with

enameled surfaces, they stated that at a given wavelength a material of given



refractive index reflects light with an intensity that varies inversely as the

particle diameter.

Kojima (1958a), using a photocolorimeter, measured the change in soil

color with change in particle size. He reported results in tristimulus coordinates and, in general, indicated an increase in the Y coordinate-the luminosity function-as particle size decreased.

Aside from the reflectance differences which can be accounted for by

differences in surface roughness and soil structure, soil particle size and shape,

as well as the size and shape of soil aggregates resulting from mild crushing,

appear to influence soil reflectance in varying manners. Rowers and Hanks

(1965) measured the reflectance of pure kaolinite in size fractions from 0.022

to 2.68 mm diam. (coarse silt to very coarse sand particle size classes) and

found a rapid exponential increase in reflectance at all wavelengths between

0.4 and 1.0 pm with decreasing particle size. The most notable increases in

reflectance occurred at sizes less than 0.4 mm diam. (approximately medium

sand particle size class and finer). It was felt that particles or aggregates larger

than 2-3 mm diam. would have little influence on additional absorption of

solar energy.

Orlov (1966) found the reflectance of aggregates from 0.25 to 10 mm in

diameter to vary little for Mollisol-type soils. However, for the fraction less

than 0.25 mm diam. (fine sand particle size class and finer) reflectance

increased, a fact that Orlov attributed to sharp changes in chemical composition of aggregates less than 0.25 mm diam. compared with coarser aggregates.

Averaged reflectance spectra from a broad range of sandy soils exhibit the

trend of increasing reflectance with decreasing particle size (Fig. 11).

Surface roughness on a micro scale may be the determining factor in

explaining changes in reflectance as a function of particle or aggregate

diameter. Bowers and Hanks (1965) observed that as particle size decreased,

the surface of kaolinite became smoother. Similarly, Orlov (1966) found that

fine particles filled a volume more completely and gave a more even surface.

Coarse aggregates, having an irregular shape, formed a complex surface with




' ' 1.2


' 1.6

' ' ' ' 2.0





' 214


FIG.11. Spectral (BRF) curves of soils differentiated by predominant particle size. -,

sands; - -, Fine loamy sands; ---, loamy sands; * .,loamy coarse sands.





a large number of interaggregate spaces. As light falls on large, irregularly

shaped aggregates, most of the incident flux penetrates into light traps and is

completely extinguished there.

Hunt and Salisbury (1971, 1976a,b) and Hunt et al. (1971a,b, 1973a-c,

1974) measured the reflectance of a large number of minerals and rocks in

four size fractions: 0-0.005,O-0.074,0.074-0.25, and 0.25- 1.2 mm. For silicate

and carbonate minerals it was noted that the general effect of decreasing the

particle size of the samples was to increase the reflectance at all wavelengths

and to decrease the contrast of any well-resolved spectral features. Conversely to previously mentioned studies, in the case of oxides and sulphides the

reflectance as a function of wavelength sometimes actually decreased with

decreasing particle size. This phenomenon appeared to occur in materials of

very low reflectance. As in other studies, however, it was found that only

integral reflection varied with particle diameter, whereas the shape of the

spectral curve remained the same.

Regression studies by Montgomery (1976) and Montgomery and Baumgardner (1974) found silt content to be the single most significant parameter

in explaining the spectral variations in soils. It was felt that the significance of

silt content may have been attributable to the size of the silt particles relative

to the reflective wavelengths. Beck et al. (1976), studying predominantly silty

soils, concluded that the wavelength region from 1.50 to 1.73 pm was best for

mapping clay content in surface soils. While fine silt content was found to

contribute significantly to explaining reflectance differences in the 0.52-0.62pm band, contents of other particle sizes were less important in prediction

equations for near- and middle-infrared wave bands (Stoner, 1979).

Gerberman and Neher (1979) artificially added increasing increments of

sand to a Harlingen clay soil (Entic Chromuslesterts). They measured the

reflectance of a set of clay soil-sand mixtures consisting of samples containing from 0 to 100% sand. They found a linear relationship between the

increasing sand content and increasing reflectance for wavelengths of 440,

540,640,720, and 860 nm.


The type and relative amotm of constituent iron oxides are known to

influence the colors of red and yellow soils high in sesquioxide clays (Bigham

et al., 1978). Predominantly yellow soils high in goethite were found to

absorb more phosphate per unit weight than did otherwise similar red soils

high in hematite. Soil spectral reflectances may be meaningful criteria for

both taxonomic and management separations in highly weathered soils.



Obukhov and Orlov (1964) reported that soils with an elevated content of

iron could be easily distinguished by the inflection characteristic for pure

Fe,O,. They found the intensity of the reflection in the region from 0.5 to

0.64 pm to be inversely proportional to the iron content. Karmanov (1970)

noted that the reflection intensity of iron hydroxides containing little water

and having a dark brown-red color increased most strongly in the wave

interval from 0.554 to 0.596 pm, while that of hydrous iron oxides increased

most strongly in the spectral range from 0.50 to 0.54pm. Neither of these

studies investigated iron oxide reflectance beyond the visible wavelengths.

Most of the well-resolved electronic features of iron oxides in minerals and

rocks can be attributed to transitions in the iron cations (Hunt et al., 1971a).

Typically, the ferrous ion produces the band near 1.0pm due to the spinallowed transition between the E , and T2@quintet levels into which the D

ground state splits in an octahedral crystal field. For the ferric ion, the major

bands produced in the spectrum are a result of transition from the GA1,

ground state to 4T1,at about 0.87 pm and to 4T2, at about 0.7 pm. Whereas

only 1 % by weight of finely powdered hematite was found to alter a clayey,

yellow Oxisol from lOYR to 5YR in color (Resende, 1976), as little as

0.0005 % of iron by weight was capable of producing a perceptible iron band

at 0.87 pm in a highly transparent calcite mineral (Hunt and Salisbury, 1971).

In addition to the ferrous iron band at 1.0 pm, another absorption band near

1.0 pm has been identified in a sample of gibbsite as a second overtone and

combination of stretching modes of the hydroxyl radical (Hunt et al., 1971a).

The iron absorption band at 0.87 pm is evident even in fine sandy soils with

iron oxide coatings on sand grains (Fig. 12). Higher iron content soils reveal a

broader absorption band at 0.87 pm contrasted with the narrow, yet distinct

band in soils of lower iron content. Soils such as the Cecil series have been

described as demonstrating an iron-affected curve form, typical of soils with

40 r







I .6

Wavelength ( p m )


2 .o


FIG.12. Spectral (BRF) curves of soils of different textures but exhibiting iron absorption

bands. -.-, Fine sand, 0.20%Fe,O,; -..-, sandy loam, 0.64%Fe,O,; -, silty clay loam,

0.76%Fe,O,; --, clay, 25.6% Fe,O,.

Tài liệu bạn tìm kiếm đã sẵn sàng tải về

III. Effects of Soil Constituents on Soil Reflectance

Tải bản đầy đủ ngay(0 tr)