HeightPredict: Predicted Height Values
In ryanmismith/inventoryfunctions: Forestry Inventory Functions

HeightPredict

R Documentation

Predicted Height Values

Description

This function predicts the heights of any trees that have missing height values. If no height values are provided, heights will be predicted using the FVS acadian growth model (formula citation???). If height values are provided, this function will leverage the provided height, SPP, Stand, and Plot data by running the predicted heights and provided heights through the following equation (HT ~ HTPred + (1|SPP) + (1|Stand/Plot)).

Usage

HeightPredict(Stand, Plot, SPP, DBH, CSI, CCF, BAL, HT = NULL)

Arguments

`Stand`	Stand ID for Plot where the nth tree is located.
`Plot`	Plot ID for Plot where the nth tree is located.
`SPP`	Species observation for every tree (FVS species code)
`DBH`	Diameter at breast height in cm.
`CSI`	Climate site index for each tree.
`CCF`	Plot based Crown Competition Factor for each tree.
`BAL`	Basal area of larger trees within the plot
`HT`	Measured HT values for all trees with measured heights (trees with no heights should be entered as 0)

Details

Stand, Plot, and SPP will only be included as random variables when either at least 5 Unique Stands, Plots and Spp (HT ~ HTPred + (1|SPP) + (1|Stand/Plot)), 5 Unique Plots and Spp (HT ~ HTPred + (1|SPP) + (1|Plot)), or 5 Unique SPP are entered (HT ~ HTPred + (1|SPP)). HTPred is the predicted heights using the acadian growth model formulas. If there are not enough categorical variables for a LMM a simple lm will be run as (HT ~ HTpred) from the acadian growth model.

All measured height values will override predicted values in the output.

Each Plot and Stand needs to have its own unique ID! If plot IDs are recycled they will be lumped together as a single plot in this function.

A simple solution if plot IDs are recycled in each stand (ex Stand 1 Plots 1,2,3,4 - Stand 2 Plots 1,2,3,4) is to create a new variable PlotID. df$PlotID <- paste(Stand, Plot, sep = "-"). This will give you unique plot IDs for every plot. ex: Stand 1 Plots 1_1, 1_2, 1_3, 1_4 - Stand 2 Plots 2_1, 2_2, 2_3, 2_4, etc.

This function requires that all data be entered as a vector of length n. See example.

Value

This function returns a numeric vector of length n with values for all trees with missing heights.

This function returns the random variables used (if any) and a plot of the model's residuals.

Author(s)

Ryan Smith

References

NEED TO WRITE REFERENCES

Examples

## Not run: 

 ###### RUNNING THE SCRIPT ######
 ###### REQUIRES FULL VECTORS #####
 ###### AS SEEN HERE ######

 trees$HT <-  HeightPredict(trees$Stand, trees$Plot, trees$SPP, trees$DBH,
                            trees$CSI, trees$CCF, trees$BAL, trees$HT)

 ##### RUN WITH FULL VECTORS AND NOT WITH MAPPLY #####

data(tree_data)
trees <- tree_data

cord <- data.frame(trees$X, trees$Y)
sp::coordinates(cord) <- cord
sp::proj4string(cord) <- sp::CRS("+proj=longlat +datum=WGS84")
CSI <- raster::raster("EastSI_ENSEMBLE_rcp60_2030.tif")
CSIdata   <- raster::extract(CSI, cord, method = 'simple', df = TRUE)
trees$CSI <- CSIdata$EastSI_ENSEMBLE_rcp60_2030

trees <- trees %>%
dplyr::mutate(
 EXPF = EXP.F(DBH, BAF),
 SDIPlot = SDI.Plot(Stand, Plot, Tree, DBH, EXPF),
 SDIMax = SDI.Max(Stand, Plot, Tree, SPP, DBH = DBH, EXPF = EXPF, CSI = CSI, X_Coord = X, Y_Coord = Y),
 BA = BA(DBH),
 CCF = CrownCompF(Stand, Plot, Tree, SPP, DBH, EXPF),
 ID = Unique.ID(Stand, Plot),
 RD = RD(SDIPlot, SDIMax),
 BAPH = BAPH(Stand, Plot, BA, EXPF),
 TPH = TPH(Stand, Plot, DBH, EXPF)
)
trees <- trees %>%
 dplyr::group_by(ID) %>%
 dplyr::arrange(desc(DBH), .by_group = TRUE)
trees <- trees %>%
 dplyr::mutate(
   BAL = BA.Larger.Trees(ID, DBH, BA)
 )

 trees$HT <-  HeightPredict(trees$Stand, trees$Plot, trees$SPP, trees$DBH,
                            trees$CSI, trees$CCF, trees$BAL, trees$HT)


## End(Not run)

ryanmismith/inventoryfunctions documentation built on Aug. 5, 2022, 2:22 a.m.