ERS186: Environmental Remote Sensing

1
ERS186Environmental Remote Sensing

• Lecture 9
• Soils

2
Overview

• Applications
• Soil Science
• Physical Principles
• Reflectance (specular and diffuse scattering)
• Absorption bands
• Dielectric constants
• Sensors
• RADAR
• Thermal
• Hyperspectral

3
Definitions

• Soil the weathered material between the surface
  of the Earth and the bedrock.
• Soils are composed of different composition and
  sizes of particles of inorganic mineral and
  organic matter
• Particles are about 50 of the soil volume, pores
  occupy the rest of the space. Pores can contain
  air or water (or ice!)
• Soils have vertical zonation (soil horizons)
  created by biological, chemical and physical
  processes

4
Soil Horizons

• O horizon gt 20 partially decayed organic matter
  (humus)
• A horizon zone of eluviation/leaching water
  leaches many minerals often pale and sandy
• E horizon mineral layer with loss of some
  combination of silicate clay, iron, aluminum
• B horizon zone of illuviation materials leached
  from other zones end up here often lots of clay
  and iron oxides
• C horizon weathered parent material mostly
  mineral
• W horizon water layer Wf if permenantly frozen
• R horizon bedrock

5
Soil Grain Size
6
Soil Grain Size

• Different size particles play different roles in
  soil
• Sand (0.05 to 2.0 mm) large air spaces, rapid
  drainage of water
• Silt (0.002 to 0.05 mm) enhance movement and
  retention of soil capillary water
• Clay (lt 0.002 mm) enhance movement and retention
  of soil capillary water carry electrical charges
  which hold ions of dissolved minerals (e.g.
  potassium and calcium)

7
Soil Texture

• Proportion of sand, silt and clay in a soil (or
  horizon), usually calculated as weight for each
  type of particle.
• These s can be broken up into different
  soil-texture classes.

8
Soil Taxonomy

• Similar to biological taxonomy -- dichotomous
  keys based on soil profiles, soil color,
  soil-texture class, moisture content, bulk
  density, porosity, and chemistry are used to ID
  different types of soils.

9
The Question

• What are the important properties of a soil in an
  RS image?
• Soil texture (proportion of sand/silt/clay)
• Soil moisture content
• Organic matter content
• Mineral contents, including iron-oxide and
  carbonates
• Surface roughness

10
Radiance of Exposed Soil

• Lt Lp Ls Lv
• Lt at-sensor radiance of a pixel of exposed
  soil
• Lp atmospheric path radiance, usually needs to
  be removed through atmospheric correction
• Ls radiance reflected off the air-soil
  interface (boundary layer)
• Soil organic matter and soil moisture content
  significantly impact Ls typically characterize
  the O horizon, the A horizon (if no O), or lower
  levels if A and O are nonexistant.
• Lv volume scattering, EMR which penetrates a
  few mm to cm.
• penetrates approximate 1/2 the wavelength
• Function of the wavelength (so RADAR may
  penetrate farther), type and amount of
  organic/inorganic constituents, shape and density
  of minerals, degree of mineral compaction, and
  the amount of soil moisture present.

11
Exposed Soil Radiance
12
Exposed Soil
(Wavelength of C-band is approximately 5 cm
L-band, 30 cm.)
13
Interpretation of last slide

• There is water in the rivers (reason glitter in
  color IR)
• Moderate wind possibly from upper left is
  ruffling water producing small capillary waves
  (reason thats the direction of blown sand from
  dunes in image, glitter is spread out indicating
  ruffled surface, rivers yellow in radar
  composite indicating high return in C band and L
  band HV but low return in L band HH)
• Pattern of water channels apparent beneath sand
  dunes.
• Dark red soil surface on left side of composite
  radar image might have low density of scattered,
  possibly loose, surface pebbles (reason dark red
  indicates moderate C band return, little L band
  return so the surface is smooth on a scale of 30
  cm but a little rough on a scale of 5 cm And the
  slight roughness in C band is depolarizing the
  signal, converting H polarization to V much like
  what a small pebble might do.)
• The blue areas of soil are probably hard packed
  fairly smooth clay. (reason The lack of red
  suggests a smooth specularly reflecting surface
  at a scale of 5 cm, one sufficiently smooth that
  it tends not to depolarize the incident signal
  the L band HH signal is returned by the very
  minor roughness very minor undulations not
  rocks - on the soil surface if the surface
  included rocks, presumably depolarization would
  occur in both C and L bands and we would observe
  more red and green mixed together with the blue.)
• I do not have an explanation for the
  predominately green areas in the center of the
  image.

14
Basic Dry Soil Spectrum
gtgtgtKey characteristic of soil spectrum
increasing reflectance with increasing wavelength
through the visible, near and mid infrared
portions of the spectrum
15
Soil Moisture

• Water coats particles, filling air spaces and
  reducing the amount of multiply scattered light,
  so soils with more moisture will be darker in the
  VNIR and SWIR than drier soils.
• Moist soils will also be darker in the SWIR
  region where water absorption increases
  significantly with increasing wavelength.
• The depths of the water absorption bands at 1.4,
  1.9 and 2.7 ?m can be used to determine soil
  moisture.

16
Soil Moisture and Texture

• Clays hold more water more tightly than sand.
  
• Thus, clay spectra display more prominent water
  absorption bands than sand spectra.
• AVIRIS can be useful for quantifying these
  absorption features.

17
Soil Moisture from RADAR

• Dielectric constant of water is about 80. (sq.
  root of dielectric constant index of refraction
  9 for water at longer radar wavelengths)
• Dielectric constants of anything dry dry soil,
  for example are much, much smaller, generally
  less than 5.
• Thus, adding water (80) to anything dry (lt5)
  dramatically increases the dielectric constant of
  the mixture of the two when measured at radar
  wavelengths.
• The bigger the difference in the speeds of light
  in air and soil, the bigger the reflection at
  their interface Thus,
• Higher dielectric constants (more moisture) yield
  dramatically higher RADAR backscatter.
• (Also remember that radar images are usually
  display using a log rather than linear scale. So
  the differences displayed in this image are huge.)

Melfort, Saskatchewan, Canada, ERS-1 Rainfall
was incident on the lower half of the image but
not on the upper half.
18
Soil Moisture from Thermal Sensors

• Water has a higher thermal capacity than soil and
  rock.
• Moist soils will change in temperature more
  slowly than dry soils.

19
Soil Moisture from Thermal Sensors

• Daedalus thermal image (night time). If we had a
  daytime image to compare it to, we could see the
  amount of change in temperature and make
  inferences on the soil moisture content (less
  change more moisture).

20
Identifying Clayey Soils
Soils with a large amount of clay exhibit
hydroxyl absorption bands at 1.4 and 2.2 ?m. 2.2
?m is more useful since it doesnt overlap the
water absorption feature.
21
Soil Organic Matter

• Organic matter is a strong absorber of EMR, so
  more organic matter leads to darker soils (lower
  reflectance curves).

22
Soil Organic Matter

• Organic matter content in the Santa Monica
  mountains mapped using AVIRIS (Palacios-Orueta et
  al. 1999).

23
Iron Oxide

• Recall that iron oxide causes a charge transfer
  absorption in the UV, blue and green wavelengths,
  and a crystal field absorption in the NIR (850 to
  900 nm). Also, scattering in the red is higher
  than soils without iron oxide, leading to a red
  color.

24
Iron Oxide

• Iron content in the Santa Monica mountains mapped
  using AVIRIS (Palacios-Orueta et al. 1999).

25
Surface Roughness

• If a surface is smooth (particles smaller than
  wavelength), specular reflection is important.
• No return surface dark unless sensor
  correctly positioned and pointed in specular
  direction.
• Smooth soil surfaces tend to be clayey or silty,
  often are moist and may contain strong absorbers
  such as organic content and iron oxide.
• Conversely, a rough surface scatters EMR and thus
  appears bright.
• But paradoxically, microwave data of well drained
  sands are often very bright, regardless of angle.
  Why?

26
Surface Roughness

• C/X-SAR (l5/1 cm) image of Oxford County,
  Ontario, Canada Crop residue on the soil
  surface can diminish soil erosion.
• Conventional tillage produces a much rougher
  surface than no-till, and therefore brighter
  backscatter.
• The goal was to determine if tillage practices
  could be identified using SAR imagery.
• A Philosophical Comment
• Do you think the various tillage practices could
  be identified? At regional to global scales?
  With certainty?
• Except in geology, remote sensing generally
  involves a mapping of one observation to many
  possible causes. So reduced backscatter might
  indicate no-till in Oxford County on one date
  but not necessarily no-till in ag fields
  anywhere else - around Davis, for example.
• However, by collecting observations at multiple
  wavelengths, times, dates, view angles, sun
  angles, polarizations, etc. and taking account of
  scene contextual information, we usually are able
  to narrow the range of possibilities and
  ultimately extract good quality information about
  the scene.