R2BayesX (version 1.1-5)

ForestHealth: Forest Health Data

Description

The data set consists of 16 variables with 1796 observations on forest health to identify potential factors influencing the health status of trees and therefore the vital status of the forest. In addition to covariates characterizing a tree and its stand, the exact locations of the trees are known. The interest is on detecting temporal and spatial trends while accounting for further covariate effects in a flexible manner.

Usage

data("ForestHealth")

Format

A data frame containing 1793 observations on 16 variables.

id:

tree location identification number.

year:

year of census.

defoliation:

percentage of tree defoliation in three ordinal categories, `defoliation < 12.5%', `12.5% <= defoliation < 50%' and `defoliation >= 50%'

x:

x-coordinate of the tree location.

y:

y-coordinate of the tree location.

age:

age of stands in years.

canopy:

forest canopy density in percent.

inclination:

slope inclination in percent.

elevation:

elevation (meters above sea level).

soil:

soil layer depth in cm.

ph:

soil pH at 0-2cm depth.

moisture:

soil moisture level with categories `moderately dry', `moderately moist' and `moist or temporarily wet'.

alkali:

proportion of base alkali-ions with categories `very low', `low', `high' and `very high'.

humus:

humus layer thickness in cm, categorical coded.

stand:

stand type with categories `deciduous' and `mixed'.

fertilized:

fertilization applied with categories `yes' and `no'.

References

Kneib, T. & Fahrmeir, L. (2010): A Space-Time Study on Forest Health. In: Chandler, R. E. & Scott, M. (eds.): Statistical Methods for Trend Detection and Analysis in the Environmental Sciences, Wiley.

G\"ottlein A, Pruscha H (1996). Der Einfuss von Bestandskenngr\"ossen, Topographie, Standord und Witterung auf die Entwicklung des Kronenzustandes im Bereich des Forstamtes Rothenbuch. Forstwissens. Zent., 114, 146--162.

See Also

bayesx

Examples

if (FALSE) {
## load zambia data and map
data("ForestHealth")
data("BeechBnd")

fm <- bayesx(defoliation ~  stand + fertilized + 
  humus + moisture + alkali + ph + soil + 
  sx(age) + sx(inclination) + sx(canopy) +
  sx(year) + sx(elevation),
  family = "cumlogit", method = "REML", data = ForestHealth)

summary(fm)
plot(fm, term = c("sx(age)", "sx(inclination)", 
  "sx(canopy)", "sx(year)", "sx(elevation)"))
}