Awesome

PROSAIL Python Bindings

J Gomez-Dans (NCEO & UCL) j.gomez-dans@ucl.ac.uk

Install using Anaconda

You should be able to easily install this using Anaconda (only tested on Linux!) with

conda install -c jgomezdans prosail

I think it might work on both Python 2.7 and 3.6. But I'm only a scientist, so expect car crashes!

Description

This repository contains the Python bindings to the PROSPECT and SAIL leaf and canopy reflectance models, respectively. Both models have been rewritten and coupled in Python, with some changes to improve on efficiency. The bindings implement the following models:

• PROSPECT: versions 5 and D. Flexibility to add/modify leaf absorption profiles.
• SAIL: FourSAIL version. The thermal extension of the model is also implemented, although this hasn't been widely tested.
• Simple Lambertian soil reflectance model

I have used as a benchmark the codes available from Jussieu.

A recent(ish) review on the use of both RT models is availabe in this paper_.

Installing the bindings

The installation of the bindings is quite straightforward: unpack the distribution and run the following command

python setup.py install

This assumes that you have the following things installed:

• Numpy
• Scipy
• The LRU_Cache library for Python 2.7
• Numba

Most of these things can be installed quite easily using Anaconda Python. In this case, you can probably just install everything you need with

  conda install python=2.7 numpy numba scipy
  pip install -U backports.functools_lru_cache
  

The bindings should then install without any issue.

Using the bindings

Once you import the bindings into the namespace with

import prosail

you can then run SAIL (using prescribed leaf reflectance and transmittance spectra, as well as canopy structure/soil parameters), PROSPECT and both (e.g. use PROSPECT to provide the spectral leaf optical properties).

run_sail

To run SAIl with two element arrays of leaf reflectance and transmittance sampled at 1nm between 400 and 2500 nm rho and tau, using a black soil (e.g. zero reflectance), you can just do

rho_canopy = prosail.run_sail(rho, tau, lai, lidfa, hspot, sza, vza, raa, rsoil0=np.zeros(2101))

Here, lai is the LAI, lidfa is the mean leaf angle in degrees, hspot is the hotspot parameter, sza, vza and raa are the solar zenith, sensor zenith and relative azimuth angles, and rsoil0 is set to an array of 0s to define the soil reflectance.

You have quite a few other options:

• You can use a different way of specifying the leaf angle distribution (by default we use a Campbell distribution with one single parameter, but you might want to use the Verhoef distribution). The Verhoef distribution is selected by adding the extra keyword typelidf=1 and the two parameters are given by lidfa and the additional optional parameter lidfb.
• You can use the internal soil spectrum model. This model is basically rho_soil = rsoil*(psoil*soil_spectrum1+(1-psoil)*soil_spectrum2). The first spectrum is a dry soil, the second one a wet one. You can also set the spectra using the soil_spectrum1 and soil_spectrum2 keywords.
• By default, we return the surface directional reflectance, but you can choose other reflectance factors (e.g. BHR, DHR, HDR).

run_prospect

To calculate leaf reflectance and transmittance using the PROSPECT model, you can use the run_prospect function. You can select either the PROSPECT-5 or PROSPECT-D versions (by default, version 'D' is used). A call to this would look like:

lam, rho, tau = prosail.run_prospect(n, cab, car, cbrown, cw, cm, ant=8.0)

Where the parameters are all scalars, and have their usual PROSPECT meanings (see table below). ant stands for anthocyannins, which isn't present in PROSPECT-5.

To do the same for PROSPECT-5...

lam, rho, tau = prosail.run_prospect(n, cab, car, cbrown, cw, cm, prospect_version='5')

You can change a number of things when calling PROSPECT, but I can't be arsed documenting it now.

run_prosail

The marriage of heaven and hell, PROSPECT being fed into SAIL in one go! Same options as the two other functions put together:

rho_canopy = prosail.run_prosail(n, cab, car, cbrown, cw, cm, lai, lidfa, hspot, tts, tto, psi, \
                    ant=0.0, alpha=40.0, prospect_version='5', typelidf=2, lidfb=0.0, \
                    factor='SDR', rsoil0=None, rsoil=None, psoil=None, \
                    soil_spectrum1=None, soil_spectrum2=None)

The parameters

The parameters used by the models and their units are introduced below:

Specifying the leaf angle distribution

The parameter typelidf regulates the leaf angle distribution family being used. The following options are understood:

• typelidf = 1: use the two parameter LAD parameterisation, where a and b control the average leaf slope and the distribution bimodality, respectively. Typical distributions are given by the following parameter choices:

• typelidf = 2 Ellipsoidal distribution, where LIDFa parameter stands for mean leaf angle (0 degrees is planophile, 90 degrees is erectophile). LIDFb parameter is ignored.

The soil model

The soil model is a fairly simple linear mixture model, where two spectra are mixed and then a brightness term added:

rho_soil = rsoil*(psoil*soil_spectrum1+(1-psoil)*soil_spectrum2)

The idea is that one of the spectra is a dry soil and the other a wet soil, so soil moisture is then contorlled by psoil. rsoil is just a brightness scaling term.