Transition Probabilities

From efdm
Jump to: navigation, search

The transition probabilities for forest areas subject to different activities such as “no-management”, thinning and final felling can be either read from files or estimated (automatically) at the start of every simulation. The choice of whether or not to read or estimate the probabilities should be indicated in the Activities file. Note, that in the current version, estimation “on the fly” is only possible for the “no-management” activity.

"No management"

Using plot data at two different time steps

To estimate the no management activity, a data set giving the state of individual plots (or other units) at two different points of time, where the “state” is expressed in terms of classes according to definitions in “factors.txt” is needed. In the toy dataset, the “nomgmtdata.txt” contains the input data which is necessary for the estimation procedure. This data is produced by classifying raw data and looks like this:

 species siteclass vol0 vol1 age0 age1
 BT SE11 10 10 8 9
 BT SE11 1 3 1 2
 BT SE11 11 12 10 11
 BT SE11 12 13 11 12
 BT SE11 13 14 12 13
 BT SE11 14 14 13 14
     ...

Volume and age must be given for two different Times Steps. The pre.estimate function allows the user to estimate the transition matrix for a no-management situation using the hackfunctions.r

Sometime the raw data does not include volume at two points in time but rather volume and the five-year increment. The volume from five years before can be derived by subtracting the increment from the present volume. It is important to consider the level of activity within the measurement plots. All plots with activities should probably be excluded (since we are estimating “no management” probabilities). To produce a file with no-management transition probabilities the function “pre.estimate (…)” (in hackfunctions.r [1]) can be used. A file with no-management probabilities is needed, for example, when you use the makeathinP function (also in hackfunctions.r) to compute the transition probabilities for thinning areas.

Using growth and yield data

Personal tools
Namespaces

Variants
Actions
Navigation
Toolbox