Package overview

Introduction

Being able to anticipate the impact of global change on forest ecosystems is one of the major environmental challenges in contemporary societies. However, uncertainties in how forests function and practical constraints in how to integrate available information prevent the development of robust and reliable predictive models. Despite the amount of knowledge accumulated about the functioning and dynamics of Mediterranean forests, scientists should make coordinate their efforts to address the challenge of integrating the different global change drivers in a modelling framework useful for research and applications.

The R package medfate has been designed to study the characteristics and simulate the functioning and structural dynamics of forest ecosystems. Climatic conditions are the main environmental drivers, with a particular focus on drought and fire 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). This representation is chosen so that package functions can be easily applied to forest plot data from national forest inventories. Since the package intends to facilitate predictions of not only forest functioning but also forest structural and compositional dynamics, the taxonomic identity of plants is stored, and parameter values need to be provided for each taxonomic entity (but the package could be used with functional groups).

Currently, the distributed R package does not include any vignette, but the package website includes articles covering model simulation examples, sensitivity analysis, parameter specification, model evaluation and applications. In addition, complete documentation on the design and formulation of the simulation models can be found at the medfate reference book.

Dynamic simulation functions

Three main kinds of simulations can be done in medfate, each model building on the previous ones.

Water/energy balance

Eco-hydrological processes are fundamental for the simulation models included in the medfate package. In particular, the package allows the simulation of water balance of soils and plants within forest stands. Processes affecting soil water content include rainfall, canopy interception, infiltration and runoff, percolation and deep drainage, soil evaporation and plant transpiration. In medfate, the soil water balance of a forest is primarily used to predict drought stress for living plants in it. Soil/plant water balance can be studied for a given forest stand using function spwb(). Function spwb() can be run using a different level of complexity. The basic approach focuses on soil water balance and strongly simplifies processes underlying plant transpiration. In contrast, the advanced approach is computationally more demanding but provides an explicit simulation of processes regulating stomatal behaviour and water transport through the plant, which also requires an explicit energy balance.

Examples of simulation with spwb() under the two approaches are provided in articles Basic water balance and Advanced water and energy balance, respectively.

Carbon balance, growth and mortality

Changes in leaf area and plant growth are key to evaluate the influence of climatic conditions on forest structure and function. Processes affecting changes leaf area and plant size are those involved in water, energy and carbon balances, as well as those directly affecting meristematic activity (e.g. phenology or other sink limitations). Carbon balance arises from the interplay between carbon assimilation via photosynthesis and the respiration costs required for the maintenance of existing cells and the formation of new tissue. Water and carbon balances are coupled through the regulation of gas exchange done by leaf stomata. Plant growth is affected by the availability of carbon (source limitation), but also by temperature and water status (sink limitation). In addition, water and carbon status of cohort plants can increase the likelihood of mortality, resulting in a decrease of the number of individuals in the cohort.

Package medfate allows simulating daily water/carbon balances, growth and mortality of a set of cohorts (competing for light and water) in a single forest stand using function growth(), which adds carbon balance, growth and mortality processes to those simulated by function spwb(). As before, function growth() can be run using two levels of complexity which match the two transpiration modes of function spwb().

An example of simulation with growth() is provided in article Forest growth.

Forest dynamics

Changes in forest structure and composition result from the interplay of demographic processes (growth, mortality and recruitment) and may include disturbances such as forest management. The package includes function fordyn(), which allows simulating these processes at yearly time steps on a given forest stand. Function fordyn() builds on the previous two simulation functions and incorporates recruitment and forest management to the set of simulated processes.

An example of simulation with fordyn() is provided in article Forest dynamics.

Plots, summaries and post-processing

Simulation models produce multiple outputs and it is important to learn how to visualize them and extract information for further analysis. Each simulation function returns an output object whose S3 class has the same name as the simulation function (e.g. spwb() returns an object of class spwb).

Implementations of summary() and plot() are available for simulation output objects, which facilitates displaying and summarizing information.
A generic function shinyplot() (as well as its implementation for different output objects) allows an interactive exploration of simulation results.
Model outputs can be extracted to simple data.frame objects through function extract().

Additional package function are meant to be used on simulation results and produce time series of additional (derived) properties:

droughtStress() : Plant/stand drought stress indices
waterUseEfficiency() : Water use efficiency metrics
resistances() : Hydraulic resistances to water transport
fireHazard() : Potential fire behaviour (see below)

Fuel properties and fire hazard

Vegetation functioning and dynamics have strong, but complex, effects on fire hazard. On one hand, growth and death of organs and individuals changes the amount of standing live and dead fuels, as well as downed dead fuels. On the other, day-to-day changes in soil and plant water content changes the physical properties of fuel, notably fuel moisture content.

Package medfate provides functions to estimate fuel properties and potential fire behaviour in forest inventory plots. Specifically, function fuel_stratification() estimates the division of live fuels in the stand between understory and canopy strata; and fuel_FCCS() calculates fuel characteristics from a forest object following an adaptation of the protocols described for the Fuel Characteristics Classification System (Prichard et al. 2013). In FCCS, fuelbed is divided into six strata, including canopy, shrub, herbaceous vegetation, dead woody materials, leaf litter and ground fuels. All except ground fuels are considered here. The intensity of burning depends on several factors, including topography, wind conditions, fuel structure and its moisture content, which is determined from antecedent and current meteorological conditions. A modification of the Rothermel’s (1972) model is used in function fire_FCCS() to calculate the intensity of surface fire reaction and the rate of fire spread of surface fires assuming a steady-state fire. Both quantities are dependent on fuel characteristics, windspeed and direction, and topographic slope and aspect.

Internal package functions

Package medfate contains many more functions than those visible in reference lists. Since v.4.3.2 multiple functions are labelled as internal, meaning that they are accessible and documented, but they are not visible in reference list, so that the sheer number of functions does not overwhelm users.

Plant, species and stand attributes

The package includes a number of functions to examine properties of the plants conforming the forest object, summary functions at the stand level or vertical profiles of several physical properties:

plant_*: Cohort-level information (species name, id, leaf area index, height…).
species_*: Species-level attributes (e.g. basal area, leaf area index).
stand_*: Stand-level attributes (e.g. basal area).
vprofile_*: Vertical profiles (light, wind, fuel density, leaf area density).

Sub-model functions

Many of the functions included in medfate are internally called by simulation functions. Some of them are made available to the user to facilitate a deeper understanding the different sub-models and a more creative use of the package, but most users can ignore them.

Sub-model functions are grouped by subject, which is included in the name of the function. The different sub-model functions are (by subject):

biophysics_*: Physical and biophysical utility functions.
carbon_*: Carbon balance.
fire_*: Fire severity.
fuel_*: Fuel properties.
hydraulics_*: Plant hydraulics.
hydrology_*: Canopy and soil hydrology (rainfall interception, soil evaporation, soil infiltration).
light_*: Radiation extinction and absorption.
moisture_*: Live tissue moisture.
pheno_*: Leaf phenology.
photo_*: Leaf photosynthesis.
root_*: Root distribution and conductance calculations.
soil_*: Soil hydraulics and thermodynamics.
transp_*: Stomatal regulation and resulting transpiration/photosynthesis.
wind_*: Canopy turbulence.

Companion packages

During the development of medfate some functions have been originally placed there and then moved to more specialized packages which evolve together with medfate:

Package meteoland allows generating daily weather input for simulation models in medfate.
Package medfateland extends medfate by allowing simulations to be performed in a spatially explicit context.
Package traits4models provides functions to help creating species parameter inputs for medfate and medfateland simulation functions.

The set of R packages are developed and maintained by the Ecosystem Modelling Facility at CREAF (Spain).

Miquel De Caceres

2024-11-05