Introduction to medfate modelling framework

Miquel De Cáceres, Rodrigo Balaguer

Ecosystem Modelling Facility, CREAF

Outline

  1. Purpose and development context
  2. Set of R packages
  3. Package installation and documentation
  4. Overview of medfate package functions
  5. Overview of medfateland package functions

M.C. Escher - Reptiles, 1943

1. Purpose and development context

Model scope

  • Being able to anticipate the impact of global change on forest ecosystems is one of the major environmental challenges in contemporary societies.

  • The set of R packages conforming the medfate modelling framework have been designed to study the characteristics and simulate the functioning and dynamics of forest ecosystems.

  • Climatic conditions are the main environmental drivers, with a particular focus on drought impacts under Mediterranean conditions.

  • Representation of vegetation accounts for structural and compositional variation but is not spatially-explicit (i.e. trees or shrubs do not have explicit coordinates within forest stands).

  1. Purpose and development context

Development context

Funding

I have been intensively working on medfate since 2013, when I obtained a Ramon y Cajal research fellowship from the Spanish government. Four other research projects (FORESTCAST, DRESS, BOMFORES, IMPROMED) have funded further developments.

Developments were also supported by CTFC (until March 2021) and CREAF, where I currently coordinate its Ecosystem Modelling Facility.

Contribution

A large number of people has contributed with ideas, data or code to the project:

Jordi Martínez-Vilalta (CREAF-UAB, Spain), Maurizio Mencuccini (ICREA, Spain), Juli G. Pausas (CIDE-CSIC, Spain), Pilar Llorens (CSIC, Spain), Rafa Poyatos (CREAF, Spain), Lluís Brotons (CREAF-CSIC, Spain), Antoine Cabon (WSL, Switzerland), Roberto Molowny (EMF-CREAF, Spain), Victor Granda (EMF-CREAF, Spain), Adriana Tovar (EMF-CREAF, Spain) Alicia Forner (MNCN-CSIC, Spain), Lluís Coll (UdL, Spain), Pere Casals (CTFC, Spain), Mario Beltrán (CTFC, Spain), Aitor Améztegui (UdL, Spain), Nicolas Martin-StPaul (INRA, France), Shengli Huang (USDA, USA), Enric Batllori (UB-CREAF, Spain), Santi Sabaté (UB-CREAF, Spain), Daniel Nadal-Sala (UB, Spain), María González (UPV, Spain), Núria Aquilué (CTFC, Spain), Mario Morales (UniZar, Spain), Léa Veuillen (INRAE, France), …

Institutions

The framework is the joint effort of different institutions:

  1. Purpose and development context

2. Set of R packages

Package suite

During the development of medfate ancillary functions were originally included in the package itself…

… but many of them were later moved into more specialized packages:

  1. Set of R packages

3. Package installation and documentation

Installation

In this course, we will use packages meteoland, medfate, medfateland, which are installed from CRAN (stable versions):

install.packages("meteoland")
install.packages("medfate")
install.packages("medfateland")

More frequent updates can be obtained if installing from GitHub:

remotes::install_github("emf-creaf/meteoland")
remotes::install_github("emf-creaf/medfate")
remotes::install_github("emf-creaf/medfateland")
  1. Package installation and documentation

Documentation

A reference book is available for detailed formulation of processes.

Several vignettes are available at the web pages of medfate and medfateland, including:

  1. How to create model inputs
  2. How to perform simulations
  3. Parameter estimation procedures
  4. Evaluation benchmarks
  5. Estimated computational times

This course can be found at the training section of the Ecosystem Modelling Facility website.

  1. Package installation and documentation

4. Overview of medfate package functions

Simulation functions

Three main simulation models can be executed in medfate:

Function Description
spwb() Water and energy balance
growth() Carbon balance, growth and mortality
fordyn() Forest dynamics, including recruitment and forest management

  1. Overview of medfate package functions

Plot/summary functions

Functions are included to extract, summarise and display the time series included in the output of each simulation function:

Function Description
extract() Reshapes daily or subdaily output into data frames.
summary() Temporal summaries of results.
plot() Display time series of the results.
shinyplot() Interactive exploration of results.
  1. Overview of medfate package functions

Sub-model functions

A large number of functions implement sub-models (i.e. modules) on which the simulation functions are built.

They are included in the package, as internal (they are not visible in function index).

Sub-model functions are grouped by subject:

Group Description
biophysics_* Physics and biophysics
carbon_* Carbon balance
fuel_* Fuel properties
fire_* Fire behaviour
hydraulics_* Plant hydraulics
hydrology_* Canopy and soil hydrology
light_* Light extinction and absortion
moisture_* Live tissue moisture
Group Description
pheno_* Leaf phenology
photo_* Leaf photosynthesis
root_* Root distribution and conductance calculations
soil_* Soil hydraulics and thermodynamics
transp_* Stomatal regulation, transpiration and photosynthesis
wind_* Canopy turbulence
  1. Overview of medfate package functions

5. Overview of medfateland package functions

Simulation functions

Package medfateland allows simulating forest functioning and dynamics on sets forests stands distributed across space, with or without spatial processes:

Function Description
*_spatial() Simulation on multiple forest stands
*_land() Simulations including spatial processes
fordyn_scenario() Regional simulations of climate/management scenarios

  1. Overview of medfateland package functions

M.C. Escher - Reptiles, 1943