Soil Science Society of America Journal 67:289-299 (2003)
© 2003 Soil Science Society of America
DIVISION S-6—SOIL & WATER MANAGEMENT & CONSERVATION
The Spectral Reflectance Properties of Soil Structural Crusts in the 1.2- to 2.5-µm Spectral Region
E. Ben-Dor*,a,
N. Goldlshlegerb,
Y. Benyaminib,
M. Agassib and
D. G. Blumbergc
a The Remote Sensing and GIS Laboratory, The Geography and Human Environment Dep., Tel-Aviv University, P.O. Box 39040 Ramat Aviv Tel-Aviv 69978 Israel
b Erosion Research Station, Soil Conservation and Drainage Division, Ministry of Agriculture, Israel
c Dep. of Geography and Environmental Development, Ben-Gurion University of the Negev, Israel
* Corresponding author (bendor{at}post.tau.ac.il)
 |
ABSTRACT
|
|---|
A controlled investigation of the spectral signature of soil's structural crust in the 1.2- to 2.5-µm spectral region was conducted to investigate the feasibility of accounting for the crust status solely from the spectral information. This research was conducted because there is no valid method for in situ assessment of soil-crust characteristics in agricultural fields while and after the crust is formed. Through the use of a laboratory rainfall simulator and a sensitive spectrometer, we showed that significant spectral differences occurred between the crust and the bulk soil for three selected soils from Israel. The study was divided into two parts: Part 1, in which qualitative observations of the three soils were conducted under one level of rainstorm energy, and Part 2, where a selected soil was further investigated under varying levels of rain energy. The spectral differences obtained for the crusted soil are related to the texture and mineralogy of the soil's surface. It was concluded that the relationship between structural crust and soil reflectance spectroscopy can be used as a tool for estimating soil properties governed by the physical crust formation, such as infiltration rate, soil runoff, and erosion. It was further suggested that this methodology be tested within a remote sensing domain, using field, air, or spaceborne hyperspectral sensors.
Abbreviations: IR, infrared • NIR, near infrared • SEP, standard error of prediciton • SWIR, short wave infrared • VIS, visible
|
INTRODUCTION
|
|---|
A STRUCTURAL CRUST is a thin layer formed on the soil's surface, caused by the action of rainfall. The crust is the result of a physical segregation and rearrangement of soil particles into a form that affects some soil properties, such as infiltration, runoff and soil erosion (Agassi et al., 1985). The structural crust is formed by two complementary mechanisms (Agassi et al., 1981): (i) physical disintegration of aggregates at the soil surface and rearrangement and compaction of the disintegrated soil particles forming at the "skin" seal, and physicochemical dispersion of the clay minerals and migration of the fine particles into the soil with infiltrating water, clogging pores immediately beneath the surface and forming a layer of low permeability termed the "washed in zone" (McIntyre, 1958; Onofiok and Singer, 1984).
The structural crust is generated within minutes, which reduces the soil's infiltration rate significantly, because the crust's hydraulic conductivity is lower by a few orders of magnitude than that of the underlying soil (McIntyre, 1958; Morin and Benyamini, 1977). When the hydraulic conductivity of the crust is decreased, even low-intensity rainfall will result in ponding, runoff, and soil erosion. This evidence stresses the idea that the crusting process is a matter of great concern, especially in arid and semiarid regions, and monitoring the soil crust condition is a crucial task for management of soils from both an agricultural and a land degradation perspective. To the best of our knowledge, there is no rapid, in situ method for monitoring, assessing, or mapping structural crust. Knowledge of the degree of crusting can improve decision-making (e.g., when to apply an agrotechnical activity) and can provide an assessment of potential soil erosion hazards.
Most of the available methods to account for the crust status use disturbed soil samples, which can be misleading (Keren and Singer, 1989, 1991). The available methods cannot exactly mirror field conditions, and hence the real crust status in the field remains uncertain. In the field, crusted soils are generally different in color than noncrusted soils. The color phenomenon is a cumulative product of the reflectance properties across the visible spectral region (VIS; 0.4–0.7 µm). The reflectance of soils across the entire spectral region of the solar illumination (0.4–2.5 µm) carries more information, as was reviewed by Baumgardner et al. (1985) and later also by Ben-Dor et al. (1998). In general, a wide range of information can be obtained from reflectance properties related to the nature and chemical composition of the soil material (Stoner et al., 1981). This is mainly based on specific absorption of spectrally active groups (known as chromophores), such as Fe, OH in water and minerals, CO3 in minerals, and many others in organic matter (Ben-Dor et al., 1997). Whereas the visible (VIS; 0.4–0.7 µm) information of soils and minerals is characterized by broader spectral features (typical of the electronic process at that range), the near infrared (NIR; 0.7–1.1 µm) and the short wave Infrared (SWIR; 1.1–2.5 µm) regions are characterized by intensive and strong absorption features that emerge from a combination mode and overtones of the fundamental processes in the infrared region (>2.5 µm). Rocks and minerals can be identified based on spectral features in the SWIR (Grove et al., 1992). Further, in soil science, it was shown that several important soil properties could be predicted using soil reflectance analysis (Dalal and Henry, 1986; Ben-Dor and Banin, 1994, 1995; Ben-Dor et al., 1997). As Ben-Dor and Banin (1995) pointed out, even small spectral features can hold significant information for a given matter, and thus careful assessment of spectral signals is essential in any soil spectral application.
Changes in soil color and infiltration rate that occur during the crusting process may be possibly captured by reflectance spectroscopy across the spectral regions of solar illumination (VIS-NIR-SWIR; 0.4–2.5 µm). The particle-size segregation process, which is part of the crusting mechanism, should affect spectral behavior, as reported in other materials (McIntyre, 1958; Chen et al., 1980; Gal et al., 1984; Onofiok and Singer, 1984). The particle segregation results in either chemical (minerals) or physical (particle size) differentiation, which are both correlated with reflectance properties of soils. Although initial evidence showed the feasibility of reflectance spectroscopy to provide information with respect to the crusting status (De Jong, 1992; Savin, 1996; and Metternicht, 1998), little work has been applied in this direction for real applications.
The scope of this paper is thus to examine the feasibility of the reflected radiation in the NIR-SWIR spectral region to predict crust-affected soil properties, such as infiltration rate, under controlled laboratory conditions.
|
MATERIALS AND METHODS
|
|---|
The study was conducted in two parts: Part 1, in which a preliminary study was conducted on three typical soils from Israel to examine the spectral feasibility approach under qualitative conditions. Part 2, in which a more systematic study was applied to quantitatively account for the soil infiltration status, using a selected soil from Part 1.
Part 1
Three typical soils from Israel were sampled for this study. The soils were selected to represent a variety of textures, ranging from clay to sand soils. This can be seen in Table 1, where some of the soils' basic properties and their definitions according to the USDA classification are presented.
The soil samples were air-dried and passed through 4-mm sieves (a size fraction that generally represents the soil-aggregate size in these soils). The soils were identically packed into 30 by 50 cm perforated trays, 2 cm deep, over a layer of coarse sand. The boxes were placed at a 5% slope and were subjected to a simulated rainstorm of 30 mm h-1 with energy of approximately 22.3 J mm-1 m-2, using distilled water and the "Morin Rotating Disk Rainfall Simulator" (Morin et al., 1967). The storm lasted until an equilibrium infiltration rate was achieved, at about 60 mm of cumulative rainfall. The boxes were then oven-dried at 35°C for 48 h and then left to dry in the room atmosphere for three more months. The dried soils were split into three fractions: the surface crust (crust), the soil at 5 mm below the surface layer (sole) and the noncrusted bulk soil (bulk). The sole fraction is defined as the bottom of the crust layer (about 5 mm below the surface) that is affected by percolating water, whereas the bulk fraction represents the original soil before application of the simulated rainstorm. These fractions were spectrally measured, using a standard sample holder of the Quantum 1200 Radiometer(LT Industry, Rockville, MD) that was optimized to the NIR-SWIR region (1.2–2.4 µm), with a bandwidth of 0.001 µm (1200 total spectral bands). The spectrometer was equipped with a concave holographic grating and a 100-W tungsten halogen lamp to irradiate a soil sample at a size of <2.7 cm in diameter. The diffuse reflected light was collected by two sulfide detectors, placed at about 45° from both sides of the samples. The reflectance of the soil samples were then measured against Halon powder in the same geometry, and the final spectrum of each measurement was presented relative to these references. For each soil fraction, an average spectrum, based on three replications, was calculated for further analysis.
Part 2
The soil from Qedma, Central Israel (Grumosol in Table 1) was selected for evaluating the effect of increasing raindrop energy on spectral measurements. A new soil batch from this region was collected and brought to the laboratory for this stage. In general, this soil has a clay texture and is characterized by many agrotechnical problems. The soil was prepared for the rain simulator as in Part 1. The boxes were placed on a carousel, five boxes per run, at a 5% slope, and were subjected to a simulated rainstorm, using distilled water. The infiltration rate of the soil was simultaneously measured during the rainstorm. At first, the simulated rainstorm provided a fog-type rain (no energy), with an intensity similar to the initial infiltration rate of the soil. When the measured rate of percolation reached the rainstorm intensity, the rainfall was stopped and the soil boxes were left until drainage from all the boxes ceased. One soil box was randomly selected for further treatments, and specific metal tubes, designed for this part, were placed on the wet soil to allow collection of noninterfere (dry) crust for later spectral measurements (the diameter of these tubes was smaller than the original LT-1200 sample holder). The four remaining soil boxes were subjected to a rainstorm intensity approximately similar to the initial infiltration rate of the soil, with an energy of approximately 22.3 J mm-1 m-2. The storm lasted until 6 mm of rainfall was applied (equal to approximately 134 J m-2). This procedure was repeated until 19, 25, and 93 mm of rainfall had accumulated (equal to approximately 424, 558, and 2074 J m-2, respectively). The five boxes were oven-dried for 48 h at 35°C and then left for seven more days in the room atmosphere. A daily spectral measurement enabled us to determine when the drying procedure had been accomplished. This appeared when minor spectral changes were observed, allowing us to start the systematic spectral measurements of each rainstorm treatment. Gravimetric measurements at that stage provided an average value of 6.8% with a standard deviation of 0.25% for the Grumosol samples exposed to the rainstorm, suggesting moisture in this soil is fully an hygorscopic type. Table 2 shows the equivalent infiltration rates that were measured for each corresponding cumulative level of energy. Five dry soil samples (replications) were taken from each box for spectral reflectance measurements using the Quantum 1200+ laboratory spectrometer. The reflectance of the soil samples was measured using the metal tubes against BaSO4, which was placed on the original Quantum 1200+ sample holder. An average spectrum for every cumulative level of rain energy was calculated, using the five replications. A spectral ratio manipulation technique and first derivative technique (e.g., Kosmas et al., 1984; Owen, 1987) were further used to emphasize spectral features.
|
RESULTS AND DISCUSSION
|
|---|
Figure 1
presents a general field overview of structural crust as developed on Hamra soil (Table 1) under field conditions, after a natural rainstorm event. The crusted soil (light color tone and not salt!) is clearly visible and differs from the noncrusted soil, which had been recently plowed.
View larger version (117K):
[in this window]
[in a new window]
|
Fig. 1. A field overview of Hamra soil after exposure to a natural rainstorm. Note the significant color differences between the crusted (light) and noncrusted (dark) plowed soils.
|
|
Generally, spectral changes result from changes in both soil chromophores and soil albedo. Chromophore is defined as a chemical group that absorbs electromagnetic radiation, and albedo is defined as the relative area under the reflectance curve. Albedo is related to physical characteristics (e.g., particle and aggregate sizes or soil moisture), whereas soil chromophore changes are dependent on the chemical composition (e.g., iron oxides, organic matter, and carbonates). Soil reflectance spectra are characterized with three broad spectral regions at around 1.4, 1.9, and 2.2 µm assigned to overtones and combination modes of OH- in adsorbed and structural water molecules. Other absorption features include CO2-3 in carbonate minerals at 2.33 µm (Gaffey, 1986), Fe3+ and Fe2+ in iron oxide minerals at around 0.5 to 0.9 µm (Hunt et al., 1971), and diverse functional groups in organic matter over the entire spectral region 0.4 to 2.5 µm (Ben-Dor et al., 1997). A comprehensive explanation for each of the absorption features can be found in Hunt (1982), Hunt et al. (1973), and Gaffey (1986).
Spectral libraries of known pure minerals make it possible to correlate the soil spectra with a given mineral or mixture of minerals. The major minerals in Israeli soils are secondary minerals, such as montmorillonite, kaolinite, and illite (Banin and Amiel, 1970), and other minerals, such as quartz, iron oxides, and calcite. A comparison of the soil reflectance spectra with the pure library spectra revealed that the previously mentioned features all contribute to clay and carbonate minerals. Figure 2 presents the reflectance spectra of bulk, crust, and sole samples as obtained in Part 1. In all soils, the spectral absorption features at around 1.4 and 1.9 µm are visible and assigned to combination modes of OH in adsorbed hygroscopic water (
W + 2
W and W + W, respectively; represents a stretching fundamental vibration mode and a bending fundamental vibration mode of OH), whereas the 2.2-µm absorption feature is assigned to a combination mode of OH in clay lattice (OH + ALOH). This absorption feature is known to be symmetric, in the case of montmorillonite clays (see Loess and Grumosols), and asymmetric, in the case of kaolinite clays (see Hamra; Kruse et al., 1991). Moreover, this feature could be slightly shifted toward the infrared region, depending on the lattice composition, that is, Al, Fe, or Mg (Ben-Dor and Banin, 1994). The weak spectral feature at 2.33 µm in both Hamra and Grumosol soils is assigned to carbonate (33) in the soil (Gaffey, 1986). Organic matter can also produce spectral features across the studied region (Ben-Dor et al., 1997), whereas quartz and iron oxides do not provide any spectral features across the SWIR region (Hunt, 1982).
View larger version (27K):
[in this window]
[in a new window]
|
Fig. 2. Loess, Hamra, and Grumusol soil reflectance spectra at crust, sole, and bulk positions (see text for details).
|
|
Based on measurements of the three soils and their corresponding spectral fractions (crust, sole, and bulk, as presented in Fig. 2), it is postulated that an albedo sequence occurs, moving from one fraction to another. All crusts are characterized by relatively high reflectance values across the entire spectral region, whereas the sole fractions show relatively lower values.
As observed by McIntyre (1958), Chen et al. (1980), Gal et al. (1984), and Onofiok and Singer (1984), the crust is composed of more fine (clay) materials than can be found in the bulk soils. Moreover, in heavily aggregated soils, such as Grumosol, the rain energy might also destroy the structure, decreasing the shadowing effects. Thus, the total reflectance of the crust is expected to be higher than that of the bulk or the sole fractions. Both the Grumosol and the Loess spectra suggest that the sole fraction is even coarser than the bulk fraction. The sole portion got wet during the rainstorm and thus was involved in the aggregation process, whereas the bulk portion of the soil was not subjected to this process. In the Hamra soil, the sole fraction is quite similar to the bulk fraction, because the wetting and drying processes are less pronounced, owing to lower clay content. Whereas, variation in the particle-size distribution in both the bulk and sole fractions is related to the soil type (based on the soil texture), the crust fraction in all of the soils is characterized by high reflectance values, resulting from the segregation process that enriched the soil surface with fine particles.
To minimize common and constant features, both the sole and the bulk fraction spectra were divided by the crust spectra (Fig. 3)
. Significant changes occurred at around 1.4 µm (assigned to OH in the adsorbed water) and 1.45 µm (assigned to OH in clay mineral lattice). In the bulk-crust spectra, spectral changes were observed at around 1.9 µm and attributed to OH in adsorbed water molecules (this is mainly because the soils were air dried and exposed to room temperature for long period of time, allowing only hygroscopic moisture to be active at these wavelengths). An intense feature indicates larger surface areas in the reference spectrum (crust), suggesting an enrichment of the soil surface with clay minerals, particularly with montmorillonite. The spectral signal near 2.2 µm is assigned to OH in clay lattice directly with no overlapping of other chromophores. Thus, this peak precisely indicates possible (clay) mineral enrichment within the crust cross-section. In the Hamra soil, the 2.2-µm peak shows a doublet feature (Fig. 3), which represents a kaolinite-montmorillonite mineral mixture, which agrees with the known clay mineralogy of Hamra soils in Israel (Ravikovitch, 1992). In the Loess and Grumosol soils, however, the 2.2-µm peak indicates the appearance of montmorillonite and illite type minerals (Fig. 3), which is also in agreement with the known soil mineralogy of both types of soil. The peak feature at 2.33 µm, which is assigned to CO3 in calcite minerals (Gaffey, 1986), is not significantly enhanced by the rationing technique. In Israeli soils, calcite is usually concentrated in the silt or sand fraction (Banin and Amiel, 1970). Because the crust is a clay-enriched layer, nonclay minerals such as calcite cannot produce significant spectral changes (around 2.33 µm).
View larger version (14K):
[in this window]
[in a new window]
|
Fig. 3. The ratio between (A) sole and crust and (B) bulk and crust spectra of Hamra, Grumosol, and Loess soils.
|
|
It can be summarized that clay enrichment processes, particle-size distribution, or aggregate destruction on the crusted zone, can be detected from the reflectance spectroscopy. Because clay minerals have a large fine particle-size fraction, and the nonaggregates provide micro shade to the measurements, the spectral (absorption) changes are correlated with the baseline (albedo) changes. The finer the particle size, the stronger are the relative absorption features of the clay chromophores and the higher is the baseline.
Part 2
Based on the significant spectral changes obtained in Part 1 for the three soils, we further investigated, quantitatively, a selected soil from Part 1, using varying rainstorm conditions. The selected soil was a Grumosol, which many farmers have noticed is severely affected by the structural crust. The effect of cumulative levels of rain energy on crust development and its influence on the Grumosol soil are shown in Fig. 4
. As the cumulative energy of the raindrops increases, the soil crusting level increases and hence the infiltration rate decreases accordingly (Agassi et al., 1985; Betzalel et al., 1995). Accordingly, the albedo shown in Fig. 4, increases from a low energy level to higher energy. It is assumed that this albedo sequence is a result of the finer particle enrichment on the soil surface (either destruction of aggregate or fine particle accumulation). Many studies have shown the above relationship, that is, as the (grain) particle-size decreases the overall albedo increases (e.g., Clark and Roush, 1984; Hapke, 1993) and hence confirms the crust spectral observation bellow. However, note that the reflectance of Grumosol in Part 1 is lower than that obtained in the Grumosol in Part 2. This difference is based on the different soil batch used, different measurement geometry, and different spectral standard used (see Materials and Methods for more information). This stresses the need for standardization in spectral crust measurements and calls for the separate analysis of each batch.
View larger version (29K):
[in this window]
[in a new window]
|
Fig. 4. Reflectance spectra of Grumusol soil, which was subjected to increasing levels of rainstorm energy.
|
|
The wavelengths that best represent the crust status in terms of reflectance were identified using an automatic-objective correlation between the reflectance spectra and the infiltration rate for each of the spectral wavelengths used (Fig. 5)
. The correlaogram in Fig. 5 was obtained by plotting the coefficient of correlation extracted by applying a linear correlation analysis between the soil reflectance at that wavelength and the corresponding infiltration rates.
View larger version (14K):
[in this window]
[in a new window]
|
Fig. 5. The correlogram between reflectance values at every spectral channel and the infiltration rate in Grumosol.
|
|
Selecting r = 0.86 as a threshold value reveals the two wavelengths with the highest correlations to be 1.833 and 2.143 µm (r = -0.8730 and r = -0.8688, respectively). Plotting the values of these two wavelengths in infiltration rate and reflectance domains showed that the relationship is better expressed by a logarithmic function as follows (Fig. 6)
: for 1.833 µm
 | [1] |
and for 2.1433 µm
 | [2] |
View larger version (20K):
[in this window]
[in a new window]
|
Fig. 6. The relationship between soil reflectance values at 1.833 and 2.143 µm and the corresponding infiltration rates in Grumosol.
|
|
The asymptotic behavior of these curves suggests that at the beginning of the crusting process (low cumulative rain energies and high infiltration rate values), the reflectance changes are minimal. However, when more energy has been applied to the soil and the crusting process is more developed (resulting in decreasing of the infiltration rate), the spectral changes are more significant and better detectable. These changes are mostly evident between 558 and 2074 J m-2, which is equivalent to 10.9 and 4.8 mm h-1 infiltration rates, respectively (Table 2).
When the crust is more developed, as expressed by the higher reflectance values, water percolation from the rainfall decreases and runoff increases. In this regard, it can be concluded that the albedo parameters can serve as tools for assessing potential soil degradation caused by rainstorms. The selected wavelengths (1.833 and 2.143 µm) are situated at the highest reflectance position of all treatment spectra, which is on the shoulder of two absorption peaks, assigned to OH in adsorbed water (the 1.83 µm is on the shoulder of the 1.92-µm peak) and to OH in clay lattice (the 2.143-µm is on the shoulder of the 2.2-µm peak).
Based on the observation made in Part 1 that the spectral changes within the crust can also occur as a result of mineral differentiation rather than particle-size distribution, the possibility of assessing the crust status from the chemical viewpoint was further examined. In this regard, the clay minerals, which are a major part of the soil's fine fraction, are expected to provide spectral signals relating to the crust's characteristics. To examine this, the albedo effect (scattering mechanism) should be eliminated and the chemical effect (absorption mechanism) should be enhanced. Thus, we applied a derivation manipulation technique to the reflectance data, assuming that this procedure would diminish the baseline (height) effect, as discussed by Owen (1987) and further used by others (Fig. 7)
. It is obvious that across most spectral regions, the albedo sequence disappeared (spectral derivative values are around or equal to zero), whereas many (absorption) peaks were enhanced. A band-by-band correlation technique, as applied to the raw spectral data, provided the correlogram shown in Fig. 8
. Setting a threshold of r = 0.99 revealed that two wavelengths, at 1.431 and 2.182 µm (r = 0.9941 and r = -0.9971, respectively) carried the highest linear correlation between the spectral derivative values and the infiltration rate data.
View larger version (35K):
[in this window]
[in a new window]
|
Fig. 7. The first derivative spectra of the reflectance response generated by increasing levels of cumulative energy at different infiltration rates in Grumosol.
|
|
View larger version (30K):
[in this window]
[in a new window]
|
Fig. 8. The correlogram between first derivative reflectance values and infiltration rates in Grumosol.
|
|
Plotting the values of these two wavelengths in an infiltration rate and a reflectance domain (Fig. 9)
showed that the relationship can also be expressed by a logarithmic function as follows:
View larger version (21K):
[in this window]
[in a new window]
|
Fig. 9. The relationships between soil first derivative reflectance values at 2.182 and 1.431 µm and the infiltration rates in Grumosol.
|
|
For 1.431 µm
 | [3] |
and for 2.182 µm
 | [4] |
These logarithmic functions determined an asymptotic behavior of the absorbance, similar to the curves presented in Fig. 6. The selected wavelengths are situated at the absorption positions of clay minerals (1.431 and 2.182 µm assigned to OH in clay lattice).
The various wavelengths captured for each spectral manipulation (either reflectance or its derivatives) stressed the idea that two mechanisms are involved: albedo sequences (governed by particle-size distribution; scattering effect), and spectral signals (governed by chemical chromophores; absorption effect). The selected wavelengths for the scattering effect are assigned to spectral positions where minimal absorption occurs (1.833 and 2.1433 m), whereas for the absorption effect, the selected wavelengths are assigned to OH in clay lattice (overtone of 2v at 1.431 µm and a combination mode of + at 2.182 µm).
These observations conform to other studies that showed finer particle enrichment of the crust, which the reflectance information suggests is probably involved with the clay-enrichment processes. It can be summarized that both mechanisms (physical and chemical) enable crust detection, using reflectance spectroscopy.
To enhance both chemical and physical effects, we applied a spectral ratio technique in which the fog (no energy) treatment was selected as a reference spectrum for each energy treatment used. Figure 10
presents the ration of crust-fog spectra, where the crust represents each of the raindrop-energy level. Significant and terraced changes occurred across the entire spectrum, both in the baseline height (albedo) and in the spectral feature intensity (absorption). The absorption signals at around 1.4 µm, 1.9 µm (assigned to OH in hygroscopic water), and 2.2 µm (assigned to OH in clay lattice) are clearly enhanced going from low (134 J m-2) to high (2074 J m-2) cumulative energy levels. The significant and systematic changes in the ratioed spectra confirm the previous discussion on the physical-chemical mechanisms. In the 134-J m-2 treatment, the baseline value is close to one (minor changes), whereas in the 2074-J m-2 treatment, the baseline is around 0.65 (35% change). The chemical (absorption) process is represented by the absorption peak changes at around 1.4, 1.9, and 2.2 µm. In the 2074-J m-2 treatment, for example, these changes reached values of approximately 0.55 (at around 1.9 µm).
View larger version (18K):
[in this window]
[in a new window]
|
Fig. 10. The ratio between the soil reflectance at each rain energy level divided by the fog treatment in Grumosol. The line at 1.0 represents no change of energy.
|
|
Whereas the previous three wavelengths are all assigned to the clay minerals, it is interesting to note that another spectral feature at around 2.33 µm emerged, especially in the 2074-J m-2 treatment. This wavelength is assigned to CO3 in carbonates (Gaffey, 1986), suggesting that at the highest energy level the crust in this soil may be enriched by carbonate-type minerals. Carbonates are known to be part of the soil's coarse fraction, however, in the soils of Israel, carbonates are reported to be present in the silt-fine fraction as well (Banin and Amiel, 1970). The spectral signature at 2.33 µm in the 2074-J m-2 treatment suggests that the carbonates (probably in the silt fraction) are affected by a relatively high energy level and migrate toward the soil surface, which eventually yields significant spectral features of carbonates on the surface.
An important application, based on the spectral analysis previously discussed, is that it is possible to assess structural crust-related properties (e.g., infiltration rate) solely from reflectance spectroscopy. To do so we selected both mechanisms (scattering and absorption) and used the four wavelengths chosen by the automatic correlation (Eq. 1–4) to yield a prediction equation for the infiltration rate (Fig. 11)
. The prediction ability (standard error of prediction [SEP]) was calculated according to the following equation:
 | [5] |
where Xp and Xm are predicted and measured IR values, respectively, and n is the number of samples.
The SEP calculation gave a favorable infiltration value of 12.9 mm h-1, which indicates that it is possible to estimate the crust status in the beginning of its formation stage (infiltration rates ranging from about 90 to 170 mm h-1) solely from the reflectance across the SWIR region. It is strongly assumed that additional combinations of measured infiltration rates and their respective soil spectral characteristics could improve such predictions.
Possible Applications: A Discussion
Although the results of this paper are only preliminary in view of practical uses, they show solid evidence that monitoring the soil crust status using reflectance spectroscopy can be a useful method. Knowledge of the spatial distribution of the soil crust status may improve soil conservation, soil cultivation, and soil drainage monitoring. To advance this methodology for practical uses, it is necessary to study additional soil types to archive possible spectral behaviors under varying rainstorm energies. High-quality reflectance data (from high signal/noise ratio spectrometers) taken from field, air, or space domains could serve as the data source from which such a methodology could be implemented for practical uses. To date, there are number of good portable spectrometers available for measuring crust status under field conditions. Use of these spectrometers with a wide-range database (both from soil type and rainstorm perspectives) might be a key factor in the creation of such a methodology for practical uses. Because at this point only a discrete selection of soils can be sensed, there is a great demand to enlarge this capability to encompass a broader spatial domain.
It should be mentioned that a solid protocol for spectral measurements should be developed to allow a reliable comparison between one set and another and to make the findings of this study more standardized and general.
|
SUMMARY AND CONCLUSIONS
|
|---|
Significant spectral changes resulting from the structural crust formation were detected on the soil surface of three studied soils. These changes were attributed to both particle-size distribution and mineralogical composition effects (baseline height and peak position-intensity responses, respectively). Under rainstorm events of similar energy, the spectral measurements were found to provide different information about the crust's magnitude, depending on its soil type. Each of the soils examined produced different spectral behaviors, and thus it can be speculated that the ability to spectrally determine changes occurring during the formation of the structural crust could play a major role in field applications. Applying varying cumulative rain energy levels to one of three selected soils (Grumosol) showed that significant spectral changes, all of which can be characterized and calculated directly from the reflectance curve of the dry soil, emerged. These changes correlated significantly with infiltration values of the crusted soil. This suggests that there is a potential to develop, based on the reflectance spectral signals, a remote-sensing tool for studying soil degradation processes in almost real time. For that purpose, it is necessary to expand the area of spectral coverage, the number of soil types and the levels of rainstorm energy and water quality under both controlled (laboratory) and natural (field) conditions.
|
ACKNOWLEDGMENTS
|
|---|
This study was partially supported by Research Grant NO. IS-311-99 from the BARD, The United States Israel Binominal Agricultural Research and Development Fund. The work has been completed at The Center for the Study of Earth from Space, University of Colorado–Boulder, under a fellowship fund of the Cooperative Institute for Research in Environment Sciences (CIRES).
Received for publication August 6, 2001.
|
REFERENCES
|
|---|
- Agassi, M., J. Morin, and I. Shainberg. 1985. Effect of raindrop impact energy and water salinity on infiltration rates of sodic soils. Soil Sci. Soc. Am. J. 49:186–190.[Abstract/Free Full Text]
- Agassi, M., I. Shainberg, and J. Morin. 1981. Effect of electrolyte concentration and soil sodicity on infiltration rate and crust formation. Soil Sci. Soc. Am. J. 45:848–851.[Abstract/Free Full Text]
- Banin, A., and A. Amiel. 1970. A correlation study of the chemical and physical properties of a group of natural soils of Israel. Geoderma 3:185–198.
- Baumgardner M.F., L.F. Silva, L.L. Biehl, and E.R. Stoner. 1985. Reflectance properties of soil. Adv. Agron. 38:1–44.
- Ben-Dor, E., and A. Banin. 1994. Visible and near infrared (0.4–1.1 µm) analysis of arid and semiarid soils. Remote Sens. Environ. 48:261–274.
- Ben-Dor, E., and A. Banin. 1995a. Near infrared analysis as a rapid method to simultaneously evaluate, several soil properties. Soil Sci. Soc. Am. J. 59:364–372.[Abstract/Free Full Text]
- Ben-Dor, E., and A. Banin. 1995b. Near infrared analysis (NIRA) as a simultaneous method to evaluate spectral featureless constituents in soils. Soil Sci. 159:259–269.
- Ben-Dor, E., Y. Inbar, and Y. Chen. 1997. The reflectance spectra of organic matter in the visible near infrared and short wave infrared region (400–2500 nm) during a controlled decomposition process. Remote Sens. Environ. 61:1–15.
- Ben-Dor, E., J.A. Irons, and A. Epema. 1998, Soil Spectroscopy. p. 111–189. In A. Rencz (ed.) Manual of remote sensing. 3rd ed. J. Wiley & Sons Inc., New York.
- Betzalel, I., J. Morin, Y. Benyamini, M. Agassi, and I. Shainberg. 1995. Waterdrop energy and seal properties. Soil Sci. 159:13–22.
- Chen, Y., J. Trachitzky, and A. Banin. 1980. Scanning electron microscope observation on soil crusts and their formation. Soil Sci. 130:49–55.
- Clark, R.N., and T.L. Roush. 1984. Reflectance spectroscopy: Quantitative analysis techniques for remote sensing application. J. Geophys. Res 89:6329–6349.
- Dalal, R.C., and R.J. Henry. 1986. Simultaneous determination of moisture or organic carbon, and total nitrogen by near infrared reflectance spectroscopy. Soil Sci. Soc. Am. J. 50:120–123.[Abstract/Free Full Text]
- Dan, Y., and Z. Raz. 1970. Soil Association Map of Israel. (In Hebrew.) Volcani Institute for Agriculture Research, Bet Dagan, Israel.
- De Jong, S.M. 1992. The analysis of spectroscopical data to map soil and types and soil crusts of Mediterranean eroded soil. Soil Technol. 5(3):199–211.
- Gaffey, S.Y. 1986. Spectral reflectance of carbonate minerals in the visible and in near infrared (0.35–2.55 µm): calcite, aragonite and dolomite. Am. Mineral. 71:151–162.[Abstract]
- Gal, M., L. Arcan, I. Shainbereg, and R. Keren. 1984. Effect of exchangeable sodium and phosphogypsum on crust structure-scanning electron observations. Soil Sci. Soc. Am. J. 48:872–878.[Abstract/Free Full Text]
- Goldshleger, N., E. Ben-Dor, Y. Benyamini, D. Blumberg, and M. Agassi. 2000. The spectral reflectance of soil's structural crust in the SWIR region 1.2–2.5 µm. Terra Nova 7:23–55.
- Grove, C.I., S.J. Hook, and E.D. Paylor, II. 1992. Laboratory reflectance spectra of 166 minerals 0.4 to 2.5 Micrometer. JPL Publ., 92–2, Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA.
- Hapke, B.W. 1993 Theory of Reflectance and Emittance Spectroscopy. Cambridge University Press, New York.
- Hunt, G.R. 1982. Spectroscopic Properties of Rocks and Minerals. p. 295–385. In R.S. Carmichal (ed.) Handbook of Physical Properties of Rocks. CRC Press, Boca Raton, FL.
- Hunt, G.R., J.W. Salisbury, and A. Lenhoff. 1971. Visible and near-infrared spectra of minerals and rocks: III Oxides and hydroxides. Modern Geology 2:195–205.
- Hunt, G.R., J.W. Salisbury, and C.J. Lenhoff. 1973. Visible and near infrared spectra of minerals and rocks VI additional silicates. Modern Geology, 4: 85–106.
- Iron, J.R., R.A. Weismiller, and G.W. Peterson. 1989. Soil Reflectance. p. 66–106. In G. Asrar (ed.) Theory and application of optical remote sensing. J. Wiley & Sons, New York.
- Keren, R., and M.J. Singer. 1989. Effects of low electrolyte concentration on hydraulic conductivity of clay-sand-hydroxy polymer systems. Soil Sci. Soc. Am. J. 53:349–355.[Abstract/Free Full Text]
- Keren, R., and M.J. Singer. 1991. Hydroxy-aluminum's effects on permeability of clay-sand mixtures. Soil Sci. Soc. Am. J. 55:61–65.[Abstract/Free Full Text]
- Kosmas, C.S., N. Curi, R.B. Bryant, and D.P. Franzmeier. 1984. Characterization of iron oxides minerals by second derivative visible spectroscopy. Soil Sci. Soc. Am. J. 48:401–405.[Abstract/Free Full Text]
- Kruse, F.A., M. Thiry, and P.L. Hauff. 1991. Spectral identification (1.2–2.5 µm) and characterization of Paris basin kaolinite/smectite clays using a field spectrometer. Proceedings of the 5th International Colloquium, Physical Measurements and Signatures in Remote Sensing, Courchevel. France I:181–184.
- McIntryre, D.S. 1958. Permeability measurement of soil crusts formed by raindrop impact. Soil Sci. 85:185–189.
- Metternicht, G.I. 1998. Analyzing the relationship between ground-based reflectance and environment indicators of salinity processes in the Cochabamba Valley (Bolivia). Inter. J. Eco. Environ. Sci. 24:359–370.
- Morin, J., and Y. Benyamini. 1977. Rainfall infiltration into bare soil. Water Resour. Res. 13:813–817.
- Morin, J., D. Goldberg, and I. Singer. 1967. Rainfall simulator with rotating disk. Trans ASAE 10:74.
- Onofiok, O., and M.J. Singer. 1984. Scanning electron microscope studies of surface crusts formed by simulated rainfall. Soil Sci. Soc. Am. J. 48(5):1137–1143.[Abstract/Free Full Text]
- Owen, T. 1987. Advances in UV-VIS spectroscopy. Derivative Spectroscopy. Int. Lab. I:58–64.
- Ravikovitch, S. 1992. The Soil of Israel formation, matured and properties. (In Hebrew.) Hakibbuz Hameuchad, Tel Aviv, Israel.
- Savin, I.YU. 1996. The effect of heavy rain on the net surface reflectivity of Cherozems. Euras. Soil Sci. 28: 168–174.
- Soil Survey Staff. 1999. Soil taxonomy: A basic system of soil classification for making and interpreting soil surveys. 2nd ed. USDA-NRCS Handb. 436. U.S. Gov. Print. Office.
- Stoner, E.R., and M.F. Baumgardner. 1981. Characteristic variations in reflectance of surface soils. Soil Sci. Soc. Am. J. 45:1161–1165.[Abstract/Free Full Text]
This article has been cited by other articles:
|
|
|
|
 
D. G. Sullivan, J. N. Shaw, and D. Rickman
IKONOS Imagery to Estimate Surface Soil Property Variability in Two Alabama Physiographies
Soil Sci. Soc. Am. J.,
September 29, 2005;
69(6):
1789 - 1798.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
G. Eshel, G. J. Levy, and M. J. Singer
Spectral Reflectance Properties of Crusted Soils under Solar Illumination
Soil Sci. Soc. Am. J.,
November 1, 2004;
68(6):
1982 - 1991.
[Abstract]
[Full Text]
[PDF]
|
|
|
TOP
ABSTRACT
INTRODUCTION
