Shoshany, Maxim
Field and Satellite radiometry of Soil Erodibility along the climatic gradient of the
Judean Desert, Israel
Shoshany1, M., Sarah, P. 2, Jarmer, T. 3, Hill, J. 3 and Lavee, H. 2
1Remote Sensing & GIS Laboratory, Geography Department , Bar-Ilan University, Israel
2Geomorphology and Soils Laboratory, Geography Department , Bar-Ilan University, Israel
3Remote Sensing Department , Trier University, Germany
E-Mail : shosham1@mail.biu.ac.il
KEYWORDS: Soil erodibility, Soil erosion, Erodibility mapping, Desertification
ABSTARCT
Soil and vegetation loss are the most fundamental characteristics of desertification processes. The intensity and extent of land
degradation attract major attention as a global change phenomenon and as regional Mediterranean phenomenon. Satellite
remote sensing is most suitable therefore for monitoring such widespread environmental threat provided that there are
suitable inferential models. This study aims at the assessment of a TM band ratio index which allows mapping of soil
erodibility across areas representing transition between humid and arid climatic conditions such as existing at the margins of
the Judean Desert. Measurements of spectral and bio-chemical properties of soil samples where carried out for this purpose
both in the field laboratory. Relationships between the Landsat TM band ratio index and bio-chemical soil properties are
analyzed.
1 INTRODUCTION
Soil erosion is commonly assessed with regard to rainfall (average annual and storm frequencies), soil erodibility, slope
length, slope gradient and landuse. Between these parameters, soil erodibility as the “resistance to sediment detachment and
transport” (Kirkby, 1999) is determined by the physical, chemical and biological soil properties. Increase in the soil
erodibility is followed by higher potential of erosion to develop in the presence of certain combinations of environmental
conditions (storm event and slope gradient for example). Between these properties, soil aggregates size and organic matter
content are commonly regarded as having direct relationships with soil stability.
Several studies have indicated that a decrease in the organic matter content and in soil particle size causes increase of the
reflectance in the visible and the NIR spectral range. Stoner and Baumgardner (1981) and more recent studies by Escadfal
(1994) have shown that the reflectance sensitivity to organic matter content is high at the visible range. Relationships
between aggregates size and spectral reflectance of soils have been investigated by Bowers and Hanks (1965) and by Hunt
and Salisbury (1976) who have found that the integrated reflectance in the visible and the Near Infra-red is sensitive to the
aggregates’ size.
This work aims at assessing the relationships between the multi-spectral distribution of soil reflectance and its erodibility at
site scale using field radiometry and at regional scale using Landsat TM data..
2 FIELD STUDY
Field investigations included radiometric measurements with the FieldSpec radiometer with 2 nm resolution and soil
sampling (Jarmer et. al., 1996). Aggregate size, organic matter and chemical compositions were determined in the laboratory
for bare soil patches which were selected at the vicinity of four sites along a climatic transect running from the Judean
Mountains towards the Dead Sea. The four sites, representing Mediterranean, semi-arid, mildly arid and arid conditions are
as follow:
1. Giv’at Ye’arim (GIV). This site is located 11 km west of Jerusalem at an elevation of 650 m a.s.l. The average annual
rainfall is 620 mm and the mean annual temperature is 17ºC.
2. Ma’ale Adumim (MAL). This site is located 6.5 km east of Jerusalem at an elevation of 330 m a.s.l. The average annual
rainfall is 330 mm and the mean annual temperature is 19ºC.
3. Mishor Adumim (MIS). This site is located 6.5 km east of Jerusalem at an elevation of 330 m a.s.l. The average annual
rainfall is 330 mm and the mean annual temperature is 19ºC.
4. Kalia (KAL). This site is located 4 km west of the Dead Sea at an elevation of 70 m b.s.l. The average annual rainfall is
120 mm and the mean annual temperature is 23ºC.
The bedrock at all sites is calcareous (limestone or hard chalk). In spite of the similar lithology and topographical conditions,
different soil types have been developed in response to the long term effect of average climatic conditions during the last
millennia. At the Mediterranean site (GIV) a red soil, Terra Rossa, has developed which contains a relatively high amount of
International Archives of Photogrammetry and Remote Sensing. Vol. XXXIII, Part B7. Amsterdam 2000.
1385
Shoshany, Maxim
clay (50%), small amount of calcium carbonate (6%-8%), and medium amount of organic matter (6%-8%). A Brown
Rendzina soil is typical to the semi-arid site (MAL) which contains about 30% clay, 30% calcium carbonate and 2.5%-4.5%
organic matter. A non saline light brown lithosol has developed in the mildly arid site (MIS). This lithosol is a shallow
calcareous loamy soil with 20 % clay, 50% calcium carbonate and 2%-3% organic matter. At the arid site (KAL), a light
gypsic desert lithosol has developed which contains 10% clay, 80% calcium carbonate and very small amount of organic
matter (1%).
3 RESULTS
Application of clustering techniques on the bio-chemical properties of the sampled data allowed the identification of six
clusters (Figure 1). These clusters were found to have also characteristic spectral signatures except for the differentiation
between clusters 5 and 3. The main differentiating feature is the reflectance levels at the near and mid infra-red spectral
channels (Figure 2).
Assessment of relationships between the soil erodibility and its reflectance properties were carried out initially with reference
to Pickup and Chewings (1996) band ratios technique. Shoshany and Lavee (1998) have utilized the same technique using
TM band ratios of channel 2 / channel 4 and channel 3 / channel 4 for mapping soil erodibility. Empirical assessment of
these ratio data for four points representing the different erosion levels formed almost a straight line of positive correlation
between the three factors : changes in these two band ratios and in the erosion levels . This pattern of change is actually
orthogonal to the pattern expected according to Pickup and Chewings (1996). A new erodibility index was then formulated
by Shoshany and Lavee (1998) according to this pattern of band ratios change. Mapping the band ratios’ combinations of the
soil samples had shown that they highly correspond to the above described linear relationships (Figure 3). From the resulting
pattern it was possible to conclude that the linear combination of band ratios correlate with the organic carbon content , with
low organic carbon at the highly erodible sites and moderate and high organic carbon at the low erodible sites. Further
assessment of the data in Figure 3 suggest that those clusters below the line represent relatively lower ferum content and vice
versa. As discussed earlier soils are much more stable when their organic matter and ferum contents are high , and therefore
it is possible to suggest that the new erodibility index is of physical significance.
Figure 1. Clustering with ferrum, inorganic and organic carbon
4.0
JJ-1-1
Cluster 6
3.5
JJ-A-1
JJ-E-1
JJ-1-2
JJ-1-3
JJ-H-8
3.0
JJ-H-7
Cluster 5
Cluster 4
JJ-G-1
JJ-E-6
2.5
JJ-G-2
JJ-3-1
2.0
JJ-C-5
JJ-I-4
JJ-H-6
JJ-3-2
Ferrum
Cluster
JJ-H-9
6
JJ-2-1
JJ-I-2
JJ-G-3
JJ-A-3
JJ-C-1
JJ-3-3
JJ-H-5
5
JJ-4-2
JJ-I-3
JJ-F-1
JJ-4-1 JJ-I-5
JJ-C-2
JJ-G-6
JJ-G-7
JJ-G-4
JJ-F-2
JJ-E-5
JJ-G-5
JJ-C-3a
JJ-D-1
JJ-E-2
JJ-F-3
JJ-A-2
4
JJ-4-4
1.5
JJ-A-5
JJ-4-3
JJ-C-3b
Cluster 3
JJ-H-4
JJ-D-2
1.0
JJ-C-4
JJ-H-2
JJ-D-3
1 JJ-A-4
JJ-H-3
3
Cluster 2
JJ-I-1
JJ-E-4
2
JJ-H-1
Cluster 1
.5
1
.5
1.0
1.5
2.0
2.5
3.0
3.5
Organic carbon
1386
International Archives of Photogrammetry and Remote Sensing. Vol. XXXIII, Part B7. Amsterdam 2000.
Shoshany, Maxim
Figure 2. Mean reflectance using reduced clusters (TM channels)
0.6
cluster 1
0.5
Reflectance
cluster 2
0.4
cluster 3&5
Mean reflectance cluster 1 (TM)
Mean reflectance cluster 2 (TM)
Mean reflectance cluster 3 (TM)
cluster 4
0.3
Mean reflectance cluster 4 (TM)
Mean reflectance cluster 5 (TM)
Mean reflectance cluster 6 (TM)
cluster 6
0.2
0.1
0
0
500
1000
1500
2000
2500
Wavelength
Figure 3. Biochemical clusters in relation to band ratio combinations
.86
JJ-A-5
JJ-3-1
.84
JJ-4-4
JJ-A-4
JJ-G-4
Clusters 2& 5
JJ-4-1
JJ-4-2
JJ-C-5 JJ-4-3
JJ-G-5JJ-C-2
JJ-G-6 JJ-C-3a JJ-H-2
.82
JJ-I-4
JJ-E-6
.80
JJ-G-3
JJ-I-2
JJ-G-2
JJ-H-8
JJ-H-5
JJ-H-6JJ-I-3
JJ-3-3
JJ-H-9
Clusters 4& 6
.78
Cluster
JJ-3-2
JJ-H-3
JJ-F-2
JJ-G-7 JJ-D-2
JJ-I-5
JJ-D-3
JJ-D-1 JJ-E-4
JJ-C-1
JJ-E-3
JJ-E-5 JJ-H-4 JJ-I-1
JJ-F-1
Cluster 1
6
JJ-H-1
5
JJ-G-1
.76
JJ-H-7
JJ-1-1
.74
JJ-1-3
JJ-A-1
JJ-2-1
JJ-E-1
JJ-F-3
JJ-A-3 JJ-C-3b
JJ-A-2JJ-C-4
4
Cluster 3
3
2
JJ-E-2
1
JJ-1-2
.72
Total Population
TM3/TM4
.70
.4
.5
TM2/TM4
.6
.7
.8
R-Qu. = 0.7626
International Archives of Photogrammetry and Remote Sensing. Vol. XXXIII, Part B7. Amsterdam 2000.
1387
Shoshany, Maxim
4 CONCLUSIONS
Application of satellite remote sensing techniques for mapping soil erosion is widespread (see for example De Jong , 1994
and Shoshany et. al., 1995). This study was aimed at strengthening the relationships between band ratios data and the soil
physical properties. Field assessment of these relationships at the soil patch scale was reported here. It was shown that
Landsat TM band ratios may serve as good indicators for organic matter content which govern soil erodibility. Using this
band ratio combination it is possible to provide regional mapping of soil erosion potential and the progress of desertification.
ACKNOWLEDGMENTS
This work was partly supported by the MEDALUS III program.
REFERENCES
Ben-Dov., I., 1987, Remotely Sensed reflectance from soils as a means of mapping according to organic matter content.
MSc. Thesis, Faculty of Agriculture, Hebrew University, Jerusalem. 153 pp.
Borggaard, O.K. (1983): Iron oxides in relation to aggregation of soil particles. Acta Agricultura Scandinavica, 33: 257-260.
Bowers, S., A., and Hanks, R.J., 1965, Reflection of radiant energy from soils. Soil Science, 100, 130-138.
De Jong , S. M., 1994, Derivation of vegetation variables from a Landsat TM image for modelling soil erosion. Earth Surface
Processes and Landforms, 19, 165-178
Hunt, G., R., and Salisbury, J. W., 1976, Visible and Near- Infrared of Minerals and rocks : XI Sedimentary Rocks, Modern
Geology, 5, 219-228.
Imeson, A., C., 1996, Desertification research : thematic issues and spatial and temporal scaling. In The use of remote
sensing for land degradation and desertification monitoring in the Mediterranean basin (Edited by Hill, J., and Peter, D.,).
Proceedings of an experts workshop , Valencia, Spain, 13-15 June 1994. Pp 1-7.
Jarmer, T., Lavee, H. and Hill, J., (1996). Field spectroscopy of soils, rocks and vegetation: Effects of climate gradients from
Mediterranean to desert environments, Israel, The Judean desert – transect. ERMES-2.
Lavee, H., and Sarah, P.,, 1995, Spatial Variability of soil properties along a climatic gradient, Proceedings of the
International Conference on Geomorphic Response of Mediterranean and Arid areas to Climatic Change. May 13-22, 1995,
Israel. 16- 30.
PCI, 1996, Software Manual, PCI Inc. Ontario Canada.
Perez-Trejo, F., 1994, Desertification and land degradation in the European Mediterranean. Report EUR 14850 EN .
European Commission, Science Research & Development, pp 1-21.
Pickup, G., and Chewings, V.H., 1996, Identifying and measuring land degradation processes using remote sensing. In The
use of remote sensing for land degradation and desertification monitoring in the Mediterranean basin (Edited by Hill, J., and
Peter, D.,). Proceedings of an experts workshop , Valencia, Spain, 13-15 June 1994.
Shoshany, M., Kutiel, P. and Lavee, H. (1995). Seasonal vegetation cover changes as indicators of soil types along a
climatological gradient: a mutual study of environmental patterns and controls using remote sensing. International Journal
of Remote sensing 16: 2137 - 2151.
Shoshany, M., and Lavee, H., 1998, Soil erodibility along the climatic gradient of the Judean Desert. Judea and Samaria
Research. 62-69. (in Hebrew)
Stoner, E., R., and Baumgardner, M., F., 1981, Characteristic variations of reflectance of surface soils. Soil Sci. Soc. Am. J.,
45, 1161-1165.
Wischmeier, W.H., and Smith, D., D., 1978, Predicting rainfall erosion losses, a guide to conservation planning. Agr.
Handbook No. 537, USDA, Washington DC, 58 pp.
Yassoglou , N., 1996, Land and desertification. Desertification in a European context, Physical and socio-economic aspects.
Report EUR 15415EN . European Commission, Science Research & Development, pp 35-55.
1388
International Archives of Photogrammetry and Remote Sensing. Vol. XXXIII, Part B7. Amsterdam 2000.