Radiative transfer of leaves
¶
J Gómez-Dans (UCL & NCEO)
¶
j.gomez-dans@ucl.ac.uk
What is radiative transfer theory?
¶
The properties of the leaves and the canopy are
combined
For analysis, untangle response of the leaves and canopy.
We want to understand the
processes
and
properties
that
control
sunlight being reflected by an e.g. canopy.
RT tries to derive a physical description of the
fate of photons
as they are
scattered
and
absorbed
within a canopy.
Let's start with
leaves
... then we'll embed them in a canopy.
Modelling leaf optical properties
¶
Photon hits a leaf, it can be
absorbed
($A$),
scattered
($R$), or
transmitted
($T$).
From energy conservation $A+R+T=1$
Implicit dependence on wavelength ($\lambda$)
How does a leaf look like?
¶
Monocot
¶
Dicot
¶
Leaf reflectance: directional aspects
¶
The wax cuticle is bright, and scatters incoming radiation with a strong directional signal in the
specular
direction, $R_s$.
Specular response broadened by the leaf's surface
microtopography
.
At interfaces between different materials, photons are scattered.
Internal contribution $R_d$ is
Lambertian
(isotropic), the result of photons interacting with internal leaf structure and pigments.
Leaf reflectance: spectral aspects
¶
Main spectral regions:
VIS $[400-700nm]$: absorption of radiation in green leaves. Chlorophyll, carotenoids, ..
NIR $[700-1100nm]$: multiple scattering due to air-cell wall interfaces,
SWIR $[>1100nm]$: mostly water absorption, note peaks at 1450, 1950 and 2500 nm.
We will focus on the Lambertian effect
This is a consequence of internal photon scattering
therefore
, it should tell us something about the internal structure and composition of the leaf.
Although specular component is important, it is often neglected!
A simple leaf model
¶
A monocot leaf looks like a
slab
What's weird about this photo? ;-)
We can model it by considering the internal reflections of a bean inside a plate:
Solution to plate model given Airy in 1833, applied to leaves by Allen and others in 1970s
We can write a fairly compact expression for the reflectance and transmittance of the plate that only depends on
The index of reflection of the plate, $n$, and
The absorption coefficient $k$.
The model works surprisingly well for monocots, but fails for more complex leaves or senescent leaves.
A multilayer leaf model
¶
Numerous photon scatter events take place at the
air-cell boundaries
within the leaf mesophyll
So extend model to have a
stack of plates
separated by
air interfaces
to account for these interactions
Problem solved by Stokes in 1862.
Can calculate leaf reflectance and transmittance of $N$ layers as a function of a single layer ($R(1)$ and $T(1)$):
$$ \frac{R(N)}{b^{N}-b^{-N}}=\frac{T(N)}{a-a^{-1}}=\frac{1}{ab^{N}-a^{-1}b^{-N}}, $$
where
$$ \begin{align} a &=\frac{1+R^{2}(1)-T^{2}(1) + \Delta}{2R(1)}\\ b &=\frac{1-R^{2}(1)+T^{2}(1) + \Delta}{2T(1)}\\ \Delta &= \sqrt { (T^{2}(1)-R^{2}(1)-1)-4R^{2}(1)}. \end{align} $$
Can make number of layers a
real number
not an integer (e.g. 1.3 layers)
$\Rightarrow$
PROSPECT
PROSPECT
predicts leaf reflectance and transmittance as a function of
$N$ (the number of layers $\sim \left[1.5-2.5\right]$, a sort of internal complexity of the leaf parameter)
Calculates $k$ from the concentrations of
Chlorophyll a+b
Leaf water
"Brown pigment"
"Dry matter"
Carotenoids
Chlorophyll a+b specific absorption
¶
Carotenoids specfic absorption
¶
Dry matter specific absorption
¶
Water specific absorption
¶
Cellulose and lignine as well as proteins have particular absorption fingerprints
However, absorption coincides with water absorption
Hence these responses lumped into "Dry matter".
Recap
¶
Leaf BRDF
Directional component due to surface specular reflection
Lambertian component due to internal interactions
Leaves can be modelled as stacks of plates
Parsimonious model with a handful of parameters
PROSPECT
:
N layers ($N\in R$)
Chlorophyll concentration $(\mu gcm^{-2})$
Carotenoids $(\mu gcm^{-2})$
Equivalent water thickness $(cm)$)
Dry matter $(gcm^{-2})$