1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
Evaluating the influence of surface soil moisture and soil surface
roughness on optical directional reflectance factors
Croft, H.a* Anderson, K.b and Kuhn, N.J.c
a
University of Toronto, Department of Geography, Toronto, M5S 3G3, Canada
*holly.croft@utoronto.ca
b
Environment and Sustainability Institute, University of Exeter, Cornwall Campus, Tremough,
Penryn, Cornwall, TR10 9EZ, UK
c
University of Basel, Department of Environmental Sciences, CH-4056 Basel, Switzerland
Abstract
Fine-scale information on soil surface roughness (SSR) is needed for calculating heat budgets,
monitoring soil degradation and for parameterising surface runoff and sediment transfer
models. Previous work has demonstrated the potential of using hyperspectral Hemispherical
Conical Reflectance Factors (HCRFs), to retrieve the SSR of different soil crusting states.
However, this was using dry soil surfaces generated in controlled laboratory conditions. The
primary aim of this study was therefore to test the impact that in situ variations in surface soil
moisture (SSM) content had on the ability of directional reflectance factors to characterise SSR
conditions. Five soil plots (20 x 20 cm in area) representing different agricultural conditions were
subjected to different durations of natural rainfall, producing a range of different levels of SSR.
SSM values varied from 8.70 to 20.05% across all soil plots. Point laser data (4 mm sample
spacing) were geostatistically analysed to give a spatially-distributed measure of SSR, giving sill
variance values from 3.15 to 22.99. HCRFs from the soil states were measured using a groundbased hyperspectral spectroradiometer for a range of viewing zenith angles from extreme
forwardscatter (θr = -60°) to extreme backscatter (θr = +60°) at a 10° sampling resolution in the
solar principal plane. The results showed that despite a large range of SSM values, forwardscattered reflectance factors exhibited a very strong relationship to SSR (R2 = 0.84 at θr = -60°).
These results demonstrate the operational potential of HCRFs for providing spatially-distributed
SSR measurements, across spatial extents containing spatio-temporal variations in SSM content.
Key words: Remote sensing; semi-variogram; erosion; laser; soil organic carbon; hyperspectral
Introduction
Soil surface roughness (SSR) at the centimetre scale is important in controlling surface runoff
and the ability of surface water flow to entrain, transport and deposit materials (Darboux et al.,
2001). Consequently, soil surface roughness is a key parameter in erosion and hydrological
models (Jetten et al., 1999; King et al., 2005), with its decline indicating soil structural
1
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
degradation (Lal, 2001). Accurate, spatially-distributed measurements of SSR are also needed
for calculating heat budgets and modelling land-atmosphere gaseous exchange (Bastiaanssen et
al., 1998; Morner and Etiope, 2002; Mosier, 1998). However, traditional methods of monitoring
SSR are resource heavy and time consuming. Of remote sensing methods, RADAR techniques
have shown the greatest developments and application for quantifying SSR (Anderson and Croft,
2009), however this is restricted to coarse classifications (smooth, moderately smooth and
rough; Baghdadi et al., 2008). Strong relationships between optical directional reflectance
factors and SSR have been found for laboratory-produced soil states, under controlled
conditions (Anderson and Kuhn, 2008; Croft et al., 2009; Croft et al., 2012a). However, in field
conditions the interaction between other soil properties may perturb the signal from SSR
(Mouazen et al., 2006). Soil surface moisture (SSM) and soil organic carbon (SOC) are known to
have a similar impact on reflectance as SSR, i.e. the lowering of baseline reflectance with
increasing water and SOC contents (Baumgardner et al., 1985; Bowers and Hanks, 1965; Lobell
and Asner, 2002). Accurate characterisation of SSR is also needed for quantifying spatiotemporal SSM and SOC content from remotely sensed data (Croft et al., 2012b); which are also
important soil properties and environmental variables and can be confounded by variations in
SSR. When predicting SSM and SOC in coarsely ground (<2 mm) soil samples, Dalal and Henry
(1986) found a much higher standard error compared to finely ground (<0.25 mm) samples.
This study aims to investigate the effects of SSM and SOC on the ability of directional reflectance
factors to retrieve accurate measurements of SSR. Five soil plots were used where reflectance
measurement acquisition was coincident with collection of SSM and SOC data. This study also
aims to take early steps to address the call of Lobell and Asner (2002) for additional research on
the relationship between SSM and soil reflectance with respect to surface roughness and
viewing geometry, in this case to investigate the opportunity of estimating soil moisture on an
operational basis. The specific research aims of this paper are to:
- assess if directional reflectance factors can be used to retrieve SSR in situ and with varying
SSM content;
- investigate if physical changes in soil structure, such as aggregate stability and SSR, are
represented by changes in directional reflectance factors;
- investigate if soil moisture shows a directional reflectance response.
2. Methods
2.1 Site and soil properties
The sites used in this experiment were located in two adjacent agricultural fields near the town
of Möhlin, located approximately 15 kilometres east of Basel, Switzerland (47.55°N; 7.83°E).
The climate is temperate, with approximately 1000 mm of rainfall annually and an average
annual temperature of 9.5°C (Egli et al., 2004). The sites were located on an upper Rhine river
terrace covered by loess, dating back to the last glaciation (Egli et al., 2004) and underlain by
gravel terraces and thin layers of moraine (Egli and Fitze, 1995). The region supports intensive
agriculture, and the soils are loess rich and well-drained, although are prone to erosion (Schaub
2
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
et al., 1997). The high base saturation and good drainage make Luvisols excellent for agriculture
(FAO, 2006). However, whilst clay accumulates in a subsurface horizon, the high silt content and
low carbon and clay content of surface horizons mean that they are prone to structural
degradation, particularly by slaking. For the study, a silt loam soil was used for two sites
representing agricultural management conditions (freshly tilled (TL) and seedbed condition
(SB)). From the two agricultural condition sites, five soil plots (each 20 x 20 cm in area) were
prepared (including one control and one degraded). The control plot was left covered for the
duration of the experiment, representing the original SSR conditions after agricultural treatment
(Table 1). The selected soil plots are assumed to be representative of the agriculture conditions.
[Insert Table 1 here]
The degraded plots were left uncovered and exposed to natural rainfall 49.3 mm over 9 days (T9
and S9). For the degraded seedbed plot, an intermediary stage was captured, after 4 days (T4) of
4.3 mm where SSR, SSM and SOC were measured concomitantly.
2.2 Soil tests
2.2.1 Aggregate stability
Aggregate stability is a measure of how resistant aggregates are to breakdown under external
forces and is the percentage weight of water-stable aggregates retained after wet-sieving of the
soil samples (Yoder, 1936; Bryan, 1968). Wet sieving was chosen to measure aggregate stability
because it is an efficient and relatively rapid technique which bears some relation to field
conditions (Blackman, 1992), representing the resistance of soil aggregates to raindrop impact
and water erosion (Chepil, 1962; López et al., 2007; Zobeck et al., 2003). The soil was air dried
to remove any moisture, because moisture content has a large influence on both aggregate
breakdown and soil erosion (Haynes and Swift, 1990; Le Bissonnais et al., 1989; Le Bissonnais et
al., 1995; Truman et al., 1990). To achieve comparability between different soil samples at
moisture levels, Le Bissonnais (1996) recommends air drying the samples. Soil aggregates with a
diameter >2 mm were used to allow comparison between soils with different initial soil
structures (Henin et al., 1958; Le Bissonnais et al., 1996). The soils were then sieved for 5 mins,
with a lift of 10 cm at a frequency of 30 lifts per minute.
2.2.2 Bulk density
Bulk density (ρb) is the weight of soil for a given volume, indicating the degree of compaction
and pore space within the soil and the degree of water and air movement, root growth and
seedling germination (Grable and Siemer, 1968; Vepraskas, 1988). Bulk density can also be used
3
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
to provide information on soil management strategies, for example changes in erodibility after
tillage (da Silva et al., 1997; Knapen et al., 2008). It also affected by inherent soil properties,
such as particle size distribution and organic matter content (da Silva et al., 1997). To collect the
soil sample, a metal bulk density ring was driven in to the soil until the top of the soil was level
with the top of the ring, giving a volume of 100 cm3. The soil sample was returned to the
laboratory and oven-dried to a constant mass at 105° C and weighed using a balance with a
precision of 0.0001 g.
2.2.3 Soil moisture
Water content was determined from mass measurements of soils by sampling during the in situ
reflectance measurement acquisition and expressed on a gravimetric basis. Previous studies of
soil reflectance have quantified moisture based on mass (Muller and Décamps, 2000; Narayanan
et al., 1993) or volume (Lobell and Asner, 2002; Weidong et al., 2002). The soil moisture content
is expressed by mass as the difference between the weights of wet and oven dry samples . The
soil sample was the same as that used for the calculation of bulk density. The wet soil sample
was weighed on a balance with a precision of 0.0001g and oven dried to constant weight at
105°C.
2.2.4 Soil organic carbon content
SOC content was measured at the University of Basel, Environmental Sciences laboratory using a
dry combustion method with a Leco RC-612 (Mönchengladbach, Germany) multiphase analyser.
Approximately 0.5 g of dried, ground and homogenized sample was placed into a clean, carbonfree combustion boat. The soil carbon content was determined by burning the soil sample in a
pure oxygen atmosphere at a regulated temperature of 550°C.
2.3 Modelling soil surface roughness
A calibrated laser-profiling instrument was used to characterise surface roughness, following the
method outlined in Anderson and Kuhn (2008) and Croft et al., (2009). The surface height (mm)
of a 20 cm x 20 cm area was measured at a sample spacing of 4 mm. Variations in spatial surface
structure associated with differences and changes in soil microtopography through agricultural
management practices, and aggregate breakdown through raindrop impact were measured.
From the laser data, digital surface models (DSM) were created to give 20 cm x 20 cm
representations of surface height. As plots T4 and T9 were dominated by steep topographic
gradients, the original digital soil model (DSMs) were filtered using a two-dimensional, finite
impulse response, linear filter function (Filter2) in Matlab, in order to extract aggregate-scale
roughness elements for comparison with the reflectance spectra. The filter calculates running
spatial mean values within a 10 x 10 mm moving window. This method created a coarse
resolution DSM of slope gradients, which was used to subtract from the original DSM to extract
fine-scale roughness. The laser-derived elevation data for all plots were then geostatistically
analysed using semi-variogram analysis to provide a quantitative, spatially distributed measure
of surface roughness (Anderson and Kuhn, 2008; Croft et al., 2013). This analysis was performed
on the original DSMs that were not influenced by slope gradients (S0, S9 and T0) and the slope
4
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
removed DSMs (T4 and T9) to ensure that all SSR values were related to aggregate-scale
roughness. A spherical model was used (Equation 1); because it was visually judged to best
describe the data and due to its suitability for representing properties with greater short-range
variability (Corwin et al., 2006).
3
 h
h 
 (h)  c1  1.5  0.5    c0
 a  
 a
[Eq. 1]
Where  is semi-variance for a given lag distance (h) and a is the variogram range. The total sill
variance (cT) is the nugget (c0) plus total sill variance (c1); for ha and (h)=c for h>a (Deutsch
and Journal, 1998). Semi-variogram analysis was performed using the Gstat geostatistical
software package (Pebesma, 2001).
2.4 Hemispherical Conical Reflectance Factor (HCRF) acquisition
Hyperspectral HCRF measurements were collected using an Ocean Optics USB2000
spectroradiometer (400 – 1000 nm), using an A-frame device (Anderson et al., 2012). This
allowed measurements to be collected at a fixed distance from the ground (60 cm), in the solar
principal plane (SPP) from θr = -60° (forwardscatter) to θr = +60° (backscatter) at a 10° sampling
resolution (Croft et al., 2009). A collimated fore-optic attachment with an 11° field of view gave
a measurement area on the ground of 12 cm diameter at nadir. Measurements were carried
out under ‘blue sky’ conditions and regularly referenced (< 2 minutes) to nadir measurements of
a calibrated Kodak grey card with nominal 18% reflectance at 700 nm (Anderson et al., 2012).
HCRF was derived as a ratio to the calibrated standard (Milton, 1987) according to Equation 2:
HCRF 
Vt
 K
Vs
[Eq. 2]
where Vt is the radiance from the target and Vsλ the radiance as measured from a calibrated
white standard Lambertian surface. Kλ is the measured reflectance of the standard reflector in
channel λ under the same illumination conditions (Duggin and Philipson, 1982).
3.
Results
3.1 Soil surface structure
The digital surface models (DSMs), created from point laser data, for all plots are shown in
Figure 1. The soil plots reveal different levels of roughness (i.e. spatial dissimilarity) at a range of
‘spatial scales’ (i.e. aggregate, orientated, slope).
[Insert Figure 1 here]
5
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
The surface structure in S0 clearly shows regular linear orientations, reflecting secondary tillage
treatment, along with overlying aggregate-scale roughness elements. The degraded SB plot (S9)
shows no indication of orientated roughness, with soil aggregates experiencing structural
breakdown through raindrop impact and slaking. This structural disruption resulted in lateral
movement of aggregate fragments and primary particles, through rainsplash and transfer by
runoff into the topographic depressions within the orientations. The control tilled plot (T0)
shows random roughness elements and a large furrow running across the top of the DSM. The
degraded plots (T4 and T9) show a smoother, more continuous surface structure, with fewer
and smaller aggregates present.
3.2 Changes in soil properties over time
In order to verify the visual observations from the DSMs to physical changes at the soil surface, a
number of soil properties were measured. These physical soil properties include SSR (as
represented by sill variance), bulk density and aggregate stability. SSM and SOC were also
measured in order to investigate their relationship to changes in soil structure (Table 2).
[Insert Table 2 here]
The aggregate stability values confirm that a structural decline has occurred from the control
plots (S0 and T0), with values declining from 0.55 to 0.46 and 0.83 to 0.41 and 0.37, respectively.
This decline is also reflected in the SSR measurements, as represented by geostatistical analysis
of the laser-derived elevation data, with a steady decrease from control plots (S0 = 22.99; T0 =
12.90) to their degraded counterparts (S9 = 10.14; T4 = 10.39 and T9 = 3.15). The corresponding
semi-variograms are shown in Figure 2.
[Insert Figure 2]
6
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
This corresponding decline in sill variance and aggregate stability demonstrates that the
geostatistical analysis of three-dimensional laser data reflects physical changes in soil surface
structure. This structural breakdown also corresponded to a decline in soil carbon values in the
tilled plots, from 1.15% to 1.05%, probably due to the release of soil carbon during aggregate
breakdown and preferential erosion along with lighter soil particles. The seedbed plots
displayed higher carbon contents which could reflect the physical differences in soil properties
for this agricultural treatment, for example, the less compacted nature of the soil (S0 = 1.31
g/cm3; T0 = 1.17 g/cm3). Despite differences in bulk density between the values for TL and SB
plots, they stayed relatively stable on exposure to rainfall, with the exception of the most
degraded tilled plot (T9 = 1.08 g/cm3). Soil moisture contents showed a positive relationship
with exposure time to rainfall, increasing progressively from the control plots (protected from
rainfall) to the degraded sites (SB = 17.07 to 20.05; TL = 8.70 to 10.12 and 14.08), reflecting the
49.3 mm of rainfall that fell during this period.
3.3 Reflectance spectra of the soil plots
It has been well-documented that SSR, SSM and SOC have a similar impact on soil reflectance at
nadir, i.e. the lowering of baseline reflectance factors and an overall ‘darkening’ of the soil (Irons
et al., 1989; Bowers and Smith, 1972; Baumgardner et al., 1985). Figure 3 shows on-nadir
reflectance factors for the five experimental soil plots, across all wavelengths.
[Insert Figure 3 here]
As the experimental plots are of the same soil type, there are no variations in reflectance
spectra as a result of different absorption features. Instead, differences in ‘baseline’ reflectance
exist across all wavelengths, which could be attributable to the different SSR, SSM and SOC
values for the soil plots (Table 1). This equifinality with on-nadir reflectance illustrates the
difficulty of separating the contribution of each soil variable to soil reflectance factors, and the
complexity that exists in attempting to retrieve SSR, SSM and SOC from optical remotely sensed
data.
3.3 Retrieving SSR from directional reflectance factors with varying SSM and SOC
In order to test potential of using directional reflectance factors for retrieving SRR across plots
containing different SSM and SOC values, directional reflectance factors are first regressed
against sill variance (SSR), alongside similar figures for SSM and SOC for comparison. Figure 4
shows the θr that represents the strongest relationship alongside the R2 values across all θr.
7
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
[Insert Figure 4 here]
Figure 4 shows very strong relationships for all three soil properties with directional reflectance
factors, although at different θr (SSM θr +20°, R2 = 0.99; SOC θr 0°, R2 = 0.85); indicating that
there is a directional component to their respective relationships. Despite varying SSM and SOC
contents, SSR (sill variance) still shows a strong relationship (R2 = 0.84) at extreme
forwardscattering θr, due to shadow-casting by soil aggregates, confirming previous findings
(Croft et al., 2009). The strength of this relationship is greater than previous results using dry
soils with a SOC value of 2.71% (R2 = 0.64; 870 nm; Croft et al., 2009). This could be because of
the lower SOC contents in the present study (Table 1), which had less of a perturbing effect on
reflectance factors. Figure 4 illustrates operational potential of using forward-scattered
reflectance factors for characterising SSR in situ and with variable soil properties.
3.4 Separating SSR and SSM using directional reflectance
In order to further investigate the potential of using off-nadir reflectance factors to enhance the
separability of the three studied soil variables, their statistical relationship (R2) with directional
reflectance factors is shown across a θr range of -60 to +20 (Figure 5). The extreme
backscattering view zenith angles were omitted due to shadowing cast by the sensor with
varying θi across the different reflectance acquisition sequences.
[Insert Figure 5]
8
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
Surface roughness (sill variance) shows the strongest relationship at extreme forwardscattering
θr, progressively decreasing until backscattering measurement angles (θr = +20°). However, SOC
and SSM both show low R2 values over forwardscattering θr, with values progressively increasing
towards the nadir and backscattering θr. If only nadir reflectance factors were used to extract
values for SSR, SSM and SOC, the R2 values associated would be 0.20, 0.86 and 0.84,
respectively. Interestingly, SSM R2 values show a progressive increase from a low at
forwardscattering θr (θr = -60°; R2 = 0.20), to a maximum at θr = +20° (R2 = 0.99). Figure 5
indicates that SSM could have a directional signal, due to a change in the refractive index (n)
from air (n = 1) to water (n = 1.33), decreasing the contrast between soil particles (n = 1.5) and
their surrounding medium (Lobell and Asner 2002). This consequently results in increased
forward-scattered reflectance (Twomey et al., 1986) and an increased probability that the
radiation will be absorbed by the soil and resulting in less reflectance and a darker soil. Thus,
explaining the negative regression seen in Figure 4a, where the higher moisture contents
resulted in less reflectance at backscattering θr. It is recognized that only five data points were
used in this analysis and in order to test this hypothesis further, more measurements across a
greater number of soil surfaces would need to be taken.
3.5 Interdependency of soil properties
Whilst SSR, SSM and SOC all show different regression results across the measured θr, this could
be due to an inter-dependency between the soil variables. Figure 6 shows the relationships of
the soil properties to each other to test if it is each respective variable that is being directly
modelled in the R2 values seen in Figure 5, or if they are a result of complex associations with
the other variables.
[Insert Figure 6 here]
Figure 6 shows that SOC exhibits a stronger relationship with surface roughness and moisture
(R2 = 0.34 and 0.56, respectively), than that between SSR and SSM (R2 = 0.02). The presence of
carbon acts as a binding agent, reducing soil breakdown and contributing to the presence of
larger and more stable aggregates at the soil surface thereby increasing SSR. These results
suggest that the high R2 value between reflectance and SOC (Figure 5) is affected by its close
relationship to the other soil properties. However, the extremely weak relationship (R2 = 0.02)
between SSR and SSM in Figure 6a indicates that the results with directional reflectance at
9
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
different θr (Figure 5) are directly influenced by the measured variable, rather than a secondary
interaction.
3.6 Regression analysis at different wavelength channels
The relationship between reflectance factors and SSR and SSM at a θr range -60° - +20° across all
wavelengths are shown in Figure 7. The R2 values show reasonable consistency across all
wavelengths at a given viewing angle, for both soil variables, with the most variation occurring
in the shorter wavelengths (< 450 nm). This is due to the higher signal to noise ratio associated
with the ocean optics spectroradiometer at these wavelengths (Anderson et al., 2012) and
increased Rayleigh scattering.
[Insert Figure 7 here]
Figure 7 shows that there are two spectral features at approximately 670 nm and 750 nm which
affect the relationship between reflectance and both SSR and SSM. There is a decrease in R2
values for SSR at 670 nm and a slight increase at 750 nm, ascribed to oxygen presence. This sees
a corresponding increase and decrease for SSM at these wavelengths. The 670 nm absorption
feature is attributed to the presence of a chlorophyll absorption feature, present on biogenic
crusts (Karnieli et al., 1999). This feature is not visible on the spectra of the individual soil
surfaces (Figure 1), and is only present in the regressions between soil properties and
reflectance factors.
4.0 Summary and conclusions
It is well recognised that distinguishing between SSR and SSM using remotely sensed data is
problematic (Anderson and Croft, 2009). These soil variables, along with SOC, can lead to an
overall change in baseline reflectance factors. This difficultly in separating SSR and SSM has also
been experienced by studies using RADAR techniques (Baghdadi et al., 2008). Multiple view
angle research has been shown to strongly represent changes in SSR (Croft et al., 2009; 2012a),
but has not yet been tested on soils of varying SSM content. This study found that forwardscattered reflectance showed a strong relationship with SSR (R2 = 0.84 at θr = -60°), despite the
range of SSM values (8.70% to 20.05%). The presence of varying SOC content had little impact
on this relationship, but this was probably due to the small amount of SOC present and low
range of values (1.05-1.33%). An interesting result to arise from the study was the apparent
directional response of SSM, showing very high R2 of 0.99 (θr = +20°; 0.01). However, as this
result was for only five data points, further work is needed to test the accuracy of these findings,
10
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
for example, over soils with unchanging SRR and the same measurement conditions (θi).
Nonetheless, the accurate characterisation of SSR in situ and with changing SSM conditions
illustrates the operational potential of optical directional remote sensing techniques, for
retrieval over coarser spatial extents and possible integration with other RS techniques (i.e
RADAR).
References
Anderson, K. & Kuhn, N. 2008. Variations in Soil Structure and Reflectance During a Controlled
Crusting Experiment. International Journal of Remote Sensing, 29, 3457 - 3475.
Anderson, K. & Croft, H. 2009. Remote sensing of soil surface properties. Progress in Physical
Geography, 33, 457-473.
Baghdadi, N., Cerdan, O., Zribi, M., Auzet, V., Darboux, F., El Hajj, M. & Kheir, R.B. 2008a.
Operational performance of current synthetic aperture radar sensors in mapping soil surface
characteristics in agricultural environments: application to hydrological and erosion modelling.
Hydrological Processes, 22, 9–20.
Baumgardner, M.F., Silva, L.F., Biehl, L.L. & Stoner, E.R. 1985. Reflectance Properties of Soils.
Advances in Agronomy, 38, 1-44.
Blackman, J.D. 1992. Seasonal variation in the aggregate stability of downland soils. Soil Use and
Management, 8(4), 142-150.
Bowers, S. & Hanks, R. 1965. Reflection of Radiant Energy from Soils. Soil Science, 100, 130-138.
Bowers, S. & Smith, S. 1972. Spectrophotometric Determination of Soil Water Content. Soil
Science Society of America Journal, 36, 978-980.
Bryan, R.B. 1968. The Development, Use and Efficiency of Indices of Soil Erodibility. Geoderma,
2, 5-26.
Chepil, W.S. 1962. A Compact Rotary Sieve and the Importance of Dry Sieving in Physical Soil
Analysis. Soil Science Society of America Journal, 26, 4-6.
Corwin, D.L., Hopmans, J. & de Rooij, G.H. 2006. From Field- to Landscape-Scale Vadose Zone
Processes: Scale Issues, Modeling, and Monitoring. Vadose Zone Journal, 5, 129-139.
Croft, H., Anderson, K. & Kuhn, N.J. 2009. Characterizing Soil Surface Roughness Using a
Combined Structural and Spectral Approach. European Journal of Soil Science, 60, 431-442.
11
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
Croft H., Anderson K. & Kuhn N.J. 2012a. Reflectance anisotropy for measuring soil surface
roughness of multiple soil types. Catena, 93, 87-96.
Croft, H., Anderson, K. & Kuhn, N.J. 2012b. On the use of remote sensing techniques for
monitoring spatio-temporal soil organic carbon dynamics in agricultural systems. Catena, 94, 6474.
Croft, H., Anderson, K., Brazier, R.E., & Kuhn, N.J. 2013. A multi-scale approach to modelling soil
surface structure using geostatistical analysis. Water Resources Research, 49(4), 1858-1870.
da Silva, A.P., Kay, B.D. & Perfect, E. 1997. Management Versus Inherent Soil Properties Effects
on Bulk Density and Relative Compaction. Soil and Tillage Research, 44, 81-93.
Dalal, R.C. & Henry, R.J. 1986. Simultaneous Determination of Moisture, Organic-Carbon, and
Total Nitrogen by near-Infrared Reflectance Spectrophotometry. Soil Science Society of America
Journal, 50, 120-123.
Darboux, F., Davy, Gascuel-Odoux, C. & Huang, C. 2001. Evolution of Soil Surface Roughness and
Flowpath Connectivity in Overland Flow Experiments. Catena, 46, 125-139.
Deutsch, C. & Journal, A. 1998. Gslib: Geostatistical Software Library and Users Guide. New York:
Oxford University Press.
Duggin, M.J. & Philipson, W.R. 1982. Field Measurement of Reflectance - Some Major
Considerations. Applied Optics 21, 2833-2840.
Egli, M., Durrenberger, S. & Fitze, P. 2004. Spatio-Temporal Behaviour and Mass Balance of
Fluorine in Forest Soils near an Aluminium Smelting Plant: Short- and Long-Term Aspects.
Environmental Pollution, 129, 195-207.
Egli, M. & Fitze, P. 1995. The Influence of Increased NH4+ Deposition Rates on Aluminum
Chemistry in the Percolate of Acid Soils. European Journal of Soil Science, 46, 439-447.
FAO, I.W.G.W. 2006: World Reference Base for Soil Resources 2006. World Soil Resources
Reports Rome: FAO.
Grable, A.R. & Siemer, E.G. 1968. Effects of Bulk Density, Aggregate Size, and Soil Water Suction
on Oxygen Diffusion, Redox Potentials, and Elongation of Corn Roots. Soil Science Society of
America Journal, 32, 180-186.
Haynes, R.J. & Swift, R.S. 1990. Stability of Soil Aggregates in Relation to Organic-Constituents
and Soil-Water Content. Journal of Soil Science, 41, 73-83.
12
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
Henin, S., Monnier, G. & Combeau, A. 1958. Methode Pour I’ettude De La Stabilite Structurale
Des Sols. Annales Agronomiques, 9, 73-92.
Irons, J.R., Weismiller, R.A. & Petersen, G.W. 1989. Soil Reflectance. In Asrar, G., editor, Theory
and Applications of Optical Remote Sensing, New York, USA: John Wiley and Sons, 66-106.
Jetten, V., de Roo, A. & Favis-Mortlock, D. 1999. Evaluation of Field-Scale and Catchment-Scale
Soil Erosion Models. Catena 37, 521-541.
Karnieli, A., Kidron, G.J., Glaesser, C. & Ben-Dor, E. 1999. Spectral Characteristics of
Cyanobacteria Soil Crust in Semiarid Environments. Remote Sensing of Environment, 69, 67-75.
King, C., Baghdadi, N., Lecomte, V. & Cerdan, O. 2005. The Application of Remote Sensing Data
to Monitoring and Modelling of Soil Erosion. Catena, 62, 79-93.
Knapen, A., Poesen, J. & Baets, S.D. 2008. Rainfall-Induced Consolidation and Sealing Effects on
Soil Erodibility during Concentrated Runoff for Loess-Derived Topsoils. Earth Surface Processes
and Landforms 33, 444-458.
Lal, R. 2001. Soil Degradation by Erosion. Land Degradation & Development, 12, 519-539.
Le Bissonnais, Y. 1996. Aggregate Stability and Assessment of Soil Crustability and Erodibility .1.
Theory and Methodology. European Journal of Soil Science, 47, 425-437.
Le Bissonnais, Y., Bruand, A. & Jamagne, M. 1989. Laboratory Experimental Study of Soil
Crusting: Relation between Aggregate Breakdown Mechanisms and Crust Stucture. Catena, 16,
377-392.
Le Bissonnais, Y., Renaux, B. & Delouche, H. 1995. Interactions between Soil Properties and
Moisture-Content in Crust Formation, Runoff and Interrill Erosion from Tilled Loess Soils.
Catena, 25, 33-46.
Lobell, D.B. & Asner, G.P. 2002. Moisture Effects on Soil Reflectance. Soil Science Society of
America Journal, 66, 722-727.
López, M.V., de Dios Herrero, J.M., Hevia, G.G., Gracia, R. & Buschiazzo, D.E. 2007.
Determination of the Wind-Erodible Fraction of Soils Using Different Methodologies. Geoderma,
139, 407-411.
13
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
Mouazen, A.M., Karoui, R., De Baerdemaeker, J. & Ramon, H. 2006. Characterization of Soil
Water Content Using Measured Visible and near Infrared Spectra. Soil Science Society of America
Journal, 70, 1295-1302.
Muller, E. & Décamps, H. 2000. Modeling Soil Moisture-Reflectance. Remote Sensing of
Environment, 76, 173-180.
Narayanan, R.M., Green, S.E. & Alexander, D.R. 1993. Mid-Infrared Laser Reflectance of Moist
Soils. Applied Optics, 32, 6043-6052.
Pebesma, E.J. 2001. Gstat User's Manual. http://www.gstat.org/gstat.pdf, Amsterdam: Dept. of
Physical Geography, Utrecht University, Utrecht, The Netherlands.
Schaub, D., Goetz, R.U. & Meier-Zielinski, S. 1997. How Landscape Features Influence Sediment
Yields (a Comparison of Two Investigation Areas in North West Switzerland). Supplementi di
Geografia Fisica e Dinamica Quaternaria III, 1, 343.
Truman, C.C., Bradford, J.M. & Ferris, J.E. 1990. Antecedent Water Content and Rainfall Energy
Influence on Soil Aggregate Breakdown. Soil Science Society of America Journal, 54, 1385-1392.
Twomey, S.A., Bohren, C.F. & Mergenthaler, J.L. 1986. Reflectance and Albedo Differences
between Wet and Dry Surfaces. Applied Optics, 25, 431-437.
Vepraskas, M.J. 1988. Bulk density values diagnostic of restricted root growth in coarse-textured
soils. Soil Science Society of America Journal, 52, 1117-21.
Weidong, L., Baret, F., Xingfa, G., Qingxi, T., Lanfen, Z. & Bing, Z. 2002. Relating Soil Surface
Moisture to Reflectance. Remote Sensing of Environment, 81, 238-246.
Yoder, R.E. 1936. A Direct Method of Aggregate Analysis of Soils and a Study of the Physical
Nature of Erosion Losses. Journal of American Society of Agronomy, 28, 337-351.
Zobeck, T.M., Popham, T.W., Skidmore, E.L., Lamb, J.A., Merrill, S.D., Lindstrom, M.J., Mokma,
D.L. & Yoder, R.E. 2003. Aggregate-Mean Diameter and Wind-Erodible Soil Predictions Using Dry
Aggregate-Size Distributions. Soil Science Society of America Journal, 67, 425-436.
14
Sites
Agricultural
condition
Sand
(%)*
Silt
(%)*
Clay
(%)*
Soil
series
Soil
type
Seedbed
(SB)
30.15
54.67 15.18
Luvisol
Silt
loam
Luvisol
Silt
loam
Tilled (TL)
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
Plots
9.76
75.41 14.83
Notation
Rainfall
(mm)
Exposure
duration (days)
S0
0
0
S9
49.3
9
T0
0
0
T4
4.3
4
T9
49.3
9
*Number of replicates = 6
Table 1: Properties of experimental plots.
Site
Solar
zenith
angle
Surface
roughness
(Sill variance)
Aggregate
stability
Bulk
density
(g/cm3)
Moisture
content
(%)
Carbon
content
(%)
S0
S9
58°
45°
22.99
10.14
0.55
0.46
1.31
1.30
17.07
20.05
1.27
1.33
T0
T4
T9
35°
34°
35°
12.90
10.39
3.15
0.83
0.41
0.37
1.17
1.22
1.08
8.70
10.12
14.08
1.15
1.09
1.05
Table 2: Soil properties of plots used for in situ regressions with soil roughness (sill variance), moisture
content (%) and carbon content (%).
15
S0
597
598
599
600
601
602
603
604
605
606
S9
T0
T4
T9
Figure 1: DSMs of all five soil plots, constructed using point laser data.
Figure 2: Empirical semi-variograms of the laser-derived surface height data for each
sampled soil plot.
16
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
Figure 3: Soil spectra for studied soil plots (θr = 0°).
a)
b)
c)
Figure 4: Regressions for reflectance factors and a) surface roughness (θr -60°; λ – 870 nm); b)
moisture content (θr +20°; λ – 870 nm); c) carbon content (θr 0°; λ – 870 nm).
17
626
627
628
629
Figure 5: R2 values between reflectance factors and surface roughness (sill variance), moisture
630
content (%) and carbon content (%) across θr -60° - +20° (λ – 870 nm).
631
632
633
634 a)
b)
c)
635
636
637
638
639
640
641
642
643
Figure 6: Inter-dependency of the soil variables, where a) SSR and SSM; b) SSR and SOC; c) SSM
and SOC).
18
644
645
646
647
648
649
a)
b)
Figure 7: R2 values between reflectance factors a) sill variance and b) moisture content, for θr = 60° to +20° and at all sampled wavelengths.
19