In ArthurPERE/biomass: Estimating Aboveground Biomass and Its Uncertainty in Tropical Forests

knitr::opts_chunk$set(echo = TRUE, cache = TRUE)
test <- TRUE
CACHE <- TRUE
require(knitr)
require(BIOMASS)

Load BIOMASS and datasets

Install BIOMASS (to be done once)

install.packages("BIOMASS")

Load the package

require(BIOMASS)
require(knitr) # To build tables in this document

Load the two datasets stored in the package

data(KarnatakaForest)
str(KarnatakaForest)
#
data(NouraguesHD)
str(NouraguesHD)

Select 10 plots for illustrative purpose

selecPlot <- KarnatakaForest$plotId %in% c("BSP2", "BSP12", "BSP14", "BSP26", "BSP28", "BSP30", "BSP34", "BSP44", "BSP63", "BSP65")
KarnatakaForestsub <- droplevels(KarnatakaForest[selecPlot, ])

Retrieve wood density

Check and retrieve taxonomy

First, check for any typo in the taxonomy

Taxo <- correctTaxo(genus = KarnatakaForestsub$genus, species = KarnatakaForestsub$species, useCache = T, verbose = F)
KarnatakaForestsub$genusCorr <- Taxo$genusCorrected
KarnatakaForestsub$speciesCorr <- Taxo$speciesCorrected

If needed, retrieve APG III families and orders from genus names

APG <- getTaxonomy(KarnatakaForestsub$genusCorr, findOrder = T)
KarnatakaForestsub$familyAPG <- APG$family
KarnatakaForestsub$orderAPG <- APG$order

Wood density

Retrieve wood density using the plot level average if no genus level information is available

dataWD <- getWoodDensity(
  genus = KarnatakaForestsub$genusCorr,
  species = KarnatakaForestsub$speciesCorr,
  stand = KarnatakaForestsub$plotId
)

The same but using the family average and adding other wood density values as references (here invented for the example)

LocalWoodDensity <- data.frame(
  genus = c("Ziziphus", "Terminalia", "Garcinia"),
  species = c("oenopolia", "bellirica", "indica"),
  wd = c(0.65, 0.72, 0.65)
)

dataWD <- getWoodDensity(
  genus = KarnatakaForestsub$genusCorr,
  species = KarnatakaForestsub$speciesCorr,
  family = KarnatakaForestsub$familyAPG,
  stand = KarnatakaForestsub$plotID,
  addWoodDensityData = LocalWoodDensity
)

Below the number of wood density value estimated at the species, genus and plot level:

# At species level
sum(dataWD$levelWD == "species")
# At genus level
sum(dataWD$levelWD == "genus")
# At plot level
sum(!dataWD$levelWD %in% c("genus", "species"))

Build height-diameter models

You may compare different models at once

result <- modelHD(
  D = NouraguesHD$D,
  H = NouraguesHD$H,
  useWeight = TRUE
)
kable(result)

Compute the local H-D model with the lowest RSE

HDmodel <- modelHD(
  D = NouraguesHD$D,
  H = NouraguesHD$H,
  method = "log2",
  useWeight = TRUE
)

Compute models specific to given stands

HDmodelPerPlot <- modelHD(
  D = NouraguesHD$D, H = NouraguesHD$H, method = "weibull",
  useWeight = T, plot = NouraguesHD$plotId
)
ResHD <- t(sapply(HDmodelPerPlot, function(x) c(coef(x$model), RSE = x$RSE)))
kable(ResHD, row.names = T, digits = 3)

Retrieve height data

Retrieve height data from a local Height-diameter model (Note that using a HD model built on French guianan trees for Indian trees is only for illustrative purpose here)

dataHlocal <- retrieveH(
  D = KarnatakaForestsub$D,
  model = HDmodel
)

Retrieve height data from a Feldpaush et al. (2012) averaged model

dataHfeld <- retrieveH(
  D = KarnatakaForestsub$D,
  region = "SEAsia"
)

Retrieve height data from Chave et al. (2012) equation 6

dataHchave <- retrieveH(
  D = KarnatakaForestsub$D,
  coord = KarnatakaForestsub[, c("long", "lat")]
)

Estimate AGB

Organize data

KarnatakaForestsub$WD <- dataWD$meanWD
KarnatakaForestsub$H <- dataHlocal$H
KarnatakaForestsub$Hfeld <- dataHfeld$H

Compute AGB(Mg) per tree

AGBtree <- computeAGB(
  D = KarnatakaForestsub$D,
  WD = KarnatakaForestsub$WD,
  H = KarnatakaForestsub$H
)

Compute AGB(Mg) per plot

AGBplot <- summaryByPlot(AGBtree, KarnatakaForestsub$plotId)

Compute AGB(Mg) per tree without height information (Eq. 7 from Chave et al. (2014))

AGBplotChave <- summaryByPlot(
  computeAGB(
    D = KarnatakaForestsub$D, WD = KarnatakaForestsub$WD,
    coord = KarnatakaForestsub[, c("long", "lat")]
  ),
  KarnatakaForestsub$plotId
)

Compute AGB(Mg) per tree with Feldpausch et al. (2012) regional H-D model

AGBplotFeld <- summaryByPlot(
  computeAGB(
    D = KarnatakaForestsub$D, WD = KarnatakaForestsub$WD,
    H = KarnatakaForestsub$Hfeld
  ),
  plot = KarnatakaForestsub$plotId
)

Propagate AGB errors

Organize data

KarnatakaForestsub$sdWD <- dataWD$sdWD
KarnatakaForestsub$HfeldRSE <- dataHfeld$RSE

Propagate error for all tree at once using the local HD model constructed above (modelHD), i.e. non-independent allometric errors will be assigned to all trees at each iteration, independently of plots.

resultMC <- AGBmonteCarlo(D = KarnatakaForestsub$D, WD = KarnatakaForestsub$WD, errWD = KarnatakaForestsub$sdWD, HDmodel = HDmodel, Dpropag = "chave2004")
Res <- summaryByPlot(resultMC$AGB_simu, KarnatakaForestsub$plotId)
Res <- Res[order(Res$AGB), ]
plot(Res$AGB, pch = 20, xlab = "Plots", ylab = "AGB (Mg/ha)", ylim = c(0, max(Res$Cred_97.5)), las = 1, cex.lab = 1.3)
segments(seq(nrow(Res)), Res$Cred_2.5, seq(nrow(Res)), Res$Cred_97.5, col = "red")

Using the Feldpaush regional HD averaged model (code only given)

resultMC <- AGBmonteCarlo(
  D = KarnatakaForestsub$D,
  WD = KarnatakaForestsub$WD,
  errWD = KarnatakaForestsub$sdWD,
  H = KarnatakaForestsub$Hfeld,
  errH = KarnatakaForestsub$HfeldRSE,
  Dpropag = "chave2004"
)

Res <- summaryByPlot(resultMC$AGB_simu, KarnatakaForestsub$plotId)
Res <- Res[order(Res$AGB), ]
plot(Res$AGB, pch = 20, xlab = "Plots", ylab = "AGB (Mg/ha)", ylim = c(0, max(Res$Cred_97.5)), las = 1, cex.lab = 1.3)
segments(seq(nrow(Res)), Res$Cred_2.5, seq(nrow(Res)), Res$Cred_97.5, col = "red")

Per plot using the Chave et al. (2014) Equation 7 (code only given)

resultMC <- AGBmonteCarlo(
  D = KarnatakaForestsub$D,
  WD = KarnatakaForestsub$WD,
  errWD = KarnatakaForestsub$sdWD,
  coord = KarnatakaForestsub[, c("long", "lat")],
  Dpropag = "chave2004"
)
Res <- summaryByPlot(resultMC$AGB_simu, KarnatakaForestsub$plotId)
Res <- Res[order(Res$AGB), ]
plot(Res$AGB, pch = 20, xlab = "Plots", ylab = "AGB (Mg/ha)", ylim = c(0, max(Res$Cred_97.5)), las = 1, cex.lab = 1.3)
segments(seq(nrow(Res)), Res$Cred_2.5, seq(nrow(Res)), Res$Cred_97.5, col = "red")

Some tricks

Mixing measured and estimated height values

If you want to use a mix of directly-measured height and of estimated ones, you may do the following steps.

(@) Build a vector of H and RSE where we assume an error of 0.5 m on directly measured trees

NouraguesHD$Hmix <- NouraguesHD$H
NouraguesHD$RSEmix <- 0.5
filt <- is.na(NouraguesHD$Hmix)
NouraguesHD$Hmix[filt] <- retrieveH(NouraguesHD$D, model = HDmodel)$H[filt]
NouraguesHD$RSEmix[filt] <- HDmodel$RSE

(@) Apply the AGBmonteCarlo by setting the height values and their errors (which depend on wether the tree was directly measured or estimated)

wd <- getWoodDensity(NouraguesHD$genus, NouraguesHD$species)
resultMC <- AGBmonteCarlo(
  D = NouraguesHD$D, WD = wd$meanWD, errWD = wd$sdWD,
  H = NouraguesHD$Hmix, errH = NouraguesHD$RSEmix,
  Dpropag = "chave2004"
)
Res <- summaryByPlot(resultMC$AGB_simu, NouraguesHD$plotId)
Res <- Res[order(Res$AGB), ]
plot(Res$AGB, pch = 20, xlab = "Plots", ylab = "AGB (Mg/ha)", ylim = c(0, max(Res$Cred_97.5)), las = 1, cex.lab = 1.3)
segments(1:nrow(Res), Res$Cred_2.5, 1:nrow(Res), Res$Cred_97.5, col = "red")

Add your tricks

Please contact Maxime ([e-mail]) if you would like to add here a code that may be useful for users (code authorship will be respected)

ArthurPERE/biomass documentation built on May 18, 2019, 2:33 a.m.