vignettes/parametrization/SpeciesParameterization.Rmd
SpeciesParameterization.RmdThis vignette describes the procedures used to obtain the species
parameter data tables (SpParamsES, SpParamsFR
and SpParamsUS) included in package
medfate. Each species parameter table is meant to be
used on forest plot locations belonging to a given National Forest
Inventory (NFI). Hence, the taxon entities chosen (rows in
SpParams) are intended to represent plant taxa reported in
the NFI.
IMPORTANT: The vignette is not self-contained, in the sense that it cannot be reproduced without access to data sets that are not included. Nevertheless, it is intended to serve as example of species parameterization for other regions.
Installed version of medfate should be ver 2.9 or higher.
install.packages("medfate")If the version at CRAN is older, the package should be installed from
GitHub via remotes::install_github():
remotes::install_github("emf-creaf/medfate")Package medfateutils is required for the species parameterization process, as it includes several functions that facilitate parsing data sources into parameter values. Package medfateutils is installed only from GitHub:
remotes::install_github("emf-creaf/medfateutils")Package medfuels is also needed because it includes allometric coefficients for Mediterranean shrub species. As before, package medfuels is installed only from GitHub:
remotes::install_github("spif-ctfc/medfuels")Once we have medfate, medfateutils and medfuels, we load them and other common packages that we will employ in this vignette:
## Warning: replacing previous import 'dplyr::filter' by 'stats::filter' when
## loading 'medfateutils'The set of species parameters needed to run the models included in
medfate are described in a dataframe called
SpParamsDefinition:
data("SpParamsDefinition")
SpParamsDefinition$ParameterName## [1] "Name" "SpIndex" "AcceptedName"
## [4] "Species" "Genus" "Family"
## [7] "Order" "Group" "GrowthForm"
## [10] "LifeForm" "LeafShape" "LeafSize"
## [13] "PhenologyType" "DispersalType" "Hmed"
## [16] "Hmax" "Z50" "Z95"
## [19] "fHDmin" "fHDmax" "a_ash"
## [22] "b_ash" "a_bsh" "b_bsh"
## [25] "a_btsh" "b_btsh" "cr"
## [28] "BTsh" "a_fbt" "b_fbt"
## [31] "c_fbt" "a_cr" "b_1cr"
## [34] "b_2cr" "b_3cr" "c_1cr"
## [37] "c_2cr" "a_cw" "b_cw"
## [40] "a_bt" "b_bt" "LeafDuration"
## [43] "t0gdd" "Sgdd" "Tbgdd"
## [46] "Ssen" "Phsen" "Tbsen"
## [49] "xsen" "ysen" "SLA"
## [52] "LeafDensity" "WoodDensity" "FineRootDensity"
## [55] "conduit2sapwood" "r635" "pDead"
## [58] "Al2As" "Ar2Al" "LeafWidth"
## [61] "SRL" "RLD" "maxFMC"
## [64] "minFMC" "LeafPI0" "LeafEPS"
## [67] "LeafAF" "StemPI0" "StemEPS"
## [70] "StemAF" "SAV" "HeatContent"
## [73] "LigninPercent" "LeafAngle" "LeafAngleSD"
## [76] "ClumpingIndex" "gammaSWR" "alphaSWR"
## [79] "kPAR" "g" "Tmax_LAI"
## [82] "Tmax_LAIsq" "Psi_Extract" "Exp_Extract"
## [85] "WUE" "WUE_par" "WUE_co2"
## [88] "WUE_vpd" "Gswmin" "Gswmax"
## [91] "Gs_Toptim" "Gs_Tsens" "Gs_P50"
## [94] "Gs_slope" "VCleaf_kmax" "VCleaf_P12"
## [97] "VCleaf_P50" "VCleaf_P88" "VCleaf_slope"
## [100] "Kmax_stemxylem" "VCstem_P12" "VCstem_P50"
## [103] "VCstem_P88" "VCstem_slope" "Kmax_rootxylem"
## [106] "VCroot_P12" "VCroot_P50" "VCroot_P88"
## [109] "VCroot_slope" "Vmax298" "Jmax298"
## [112] "Nleaf" "Nsapwood" "Nfineroot"
## [115] "WoodC" "RERleaf" "RERsapwood"
## [118] "RERfineroot" "CCleaf" "CCsapwood"
## [121] "CCfineroot" "RGRleafmax" "RGRsapwoodmax"
## [124] "RGRcambiummax" "RGRfinerootmax" "SRsapwood"
## [127] "SRfineroot" "RSSG" "MortalityBaselineRate"
## [130] "SurvivalModelStep" "SurvivalB0" "SurvivalB1"
## [133] "SeedProductionHeight" "SeedMass" "SeedLongevity"
## [136] "DispersalDistance" "DispersalShape" "ProbRecr"
## [139] "MinTempRecr" "MinMoistureRecr" "MinFPARRecr"
## [142] "RecrTreeDBH" "RecrTreeHeight" "RecrShrubHeight"
## [145] "RecrTreeDensity" "RecrShrubCover" "RecrZ50"
## [148] "RecrZ95" "RespFire" "RespDist"
## [151] "RespClip" "IngrowthTreeDensity" "IngrowthTreeDBH"It is obvious that such a sheer number of parameters cannot be searched for manually, even for a small number of target species. Hence, it is important to draw values from allometric and plant trait data bases. Not all model parameters have their counterpart in publicly-available databases, and even when using data bases it is unlikely that one will find appropriate parameter values for all species. To help with this situation, medfate has inbuilt imputation procedures to estimate missing values for many parameters before starting model simulations. However, imputed parameter estimates will normally be worse than estimates coming from data bases. Hence, we should put a strong effort in finding source parameter data before relying on inbuilt imputation.
The following sections describe different steps that we used to obtain the species parameter table, which can be grouped into five main tasks: