clAggregatedMortality Class Reference

Aggregated mortality version 1.0. More...

#include <AggregatedMortality.h>

Inheritance diagram for clAggregatedMortality:

clMortalityBase clBehaviorBase clWorkerBase List of all members.

Public Member Functions

 clAggregatedMortality (clSimManager *p_oSimManager)
 Constructor.
 ~clAggregatedMortality ()
 Destructor.
void DoShellSetup (xercesc::DOMDocument *p_oDoc)
 Does behavior setup.
void PreMortCalcs (clTreePopulation *p_oPop)
 Performs mortality.
float GetMortalityEpisodeProbability ()
 For testing purposes.
float GetTreeMortalityProb ()
 For testing purposes.
bool DoMort (clTree *p_oTree, const float &fDbh, const short int &iSpecies)
 Determines mortality for a tree.

Protected Member Functions

void ReadParFile (xercesc::DOMDocument *p_oDoc)
 Reads in parameters from the parameter file.
void GetDeadCodes (clTreePopulation *p_oPop)
 Gets codes for the "dead" data member for each tree type to which this behavior applies.
void SetUpAppliesTo (clTreePopulation *p_oPop)
 Sets up the mp_bAppliesTo matrix.
bool IsValid (clTree *p_oTree)
 Tests a tree to see whether it is eligible for killing.
void KillClump (clTreePopulation *p_oPop, clTree *p_oTree, int iNumToKill)
 Kills a single clump of trees.

Protected Attributes

int ** mp_iDeadCodes
 Return codes for the "dead" bool tree data member.
bool ** mp_bAppliesTo
 Matrix for which tree types and species this behavior is applied.
float m_fMortalityEpisodeProbability
 Per-timestep probability of a mortality episode occuring, 0 - 1.
float m_fMortalityProb
 Per-timestep proportion of trees to kill, 0 - 1.
int m_iNumTreesToClump
 Number of trees to kill together in a clump.
float m_fClumpingParameter
 Clumping parameter for the negative binomial distribution, if applicable.
int m_iTotalNumSpecies
 Total number of species.
bool m_bClumpSizeDeterministic
 If true, the size of a single clump of trees to kill is deterministic.

Detailed Description

Aggregated mortality version 1.0.

This behavior kills trees in a way that is clumped in both time and space.

Mortality occurs in discrete episodes, governed by a return interval specified by the user. Each timestep, the probability of a mortality episode is T/RI, where T is the number of years per timestep and RI is the return interval in years. At each timestep, a uniform random number is compared to this probability to decide whether or not an episode of mortality occurs. If not, no trees are killed by this behavior.

When a mortality episode occurs, trees are killed in clumps of a size set by the user. Each tree has an equal probability of being the first tree to die in a clump, and is evaluated individually. The probability of mortality is the total mortality probability given by the user divided by the size of the clump. This probability applies only to trees to which this behavior is applied. Other trees are not killed by this behavior, either directly or as part of a clump.

Trees are killed in clumps. The number of trees in a clump can either be deterministic, meaning the clumps are always the same size, or it can be drawn from a negative binomial probability distribution.

For each tree in the pool, the behavior uses a random number to see if it will start a clump. If it will, the behavior kills it, then its closest neighbors to which the behavior applies, until it has killed the number of trees in the clump. If the clump size is 1, then this behavior replicates the clStochasticMort method.

The mortality rate is given as per year. The per timestep amount to kill is 1-(1-AD)^X, where AD is the amount of damage and X is the number of years per timestep.

Because of the clumping thing, this chooses all its trees to kill in the PreMortCalcs function.

There could be problems if there are other mortality behaviors that come before this one. Not in a code way, but they might kill again trees this behavior has chosen to kill; so fewer than expected trees might die.

The namestring for this class is "aggregatedmortshell". The parameter file call string is "Aggregated Mortality".

This behavior is only descended from clMortalityBase in order to take advantage of the registration of the "dead" tree data member.

Copyright 2006 Charles D. Canham

Author:
Lora E. Murphy
Edit history:
-----------------
July 19, 2006 - Created (LEM)


Constructor & Destructor Documentation

clAggregatedMortality::clAggregatedMortality ( clSimManager p_oSimManager  ) 

Constructor.

Parameters:
p_oSimManager Sim Manager object.

clAggregatedMortality::~clAggregatedMortality (  ) 

Destructor.

Frees memory.


Member Function Documentation

bool clAggregatedMortality::DoMort ( clTree p_oTree,
const float &  fDbh,
const short int &  iSpecies 
) [virtual]

Determines mortality for a tree.

This just returns the value in the "dead" bool tree data member.

Parameters:
p_oTree Tree being evaluated.
fDbh DBH of tree being evaluated.
iSpecies Species of tree being evaluated.

Implements clMortalityBase.

void clAggregatedMortality::DoShellSetup ( xercesc::DOMDocument *  p_oDoc  )  [virtual]

Does behavior setup.

Calls the following functions:

  1. ReadParFile()
  2. GetDeadCodes()
  3. SetUpAppliesTo()
Then final calculations for mortality episode probability and amount of damage are made.
Parameters:
p_oDoc DOM tree of parsed input file.

Reimplemented from clMortalityBase.

void clAggregatedMortality::GetDeadCodes ( clTreePopulation p_oPop  )  [protected]

Gets codes for the "dead" data member for each tree type to which this behavior applies.

Parameters:
p_oPop Tree population object.
Exceptions:
modelErr if the codes are not available for every tree type to which this behavior is applied, or if this behavior is applied to seedlings.

float clAggregatedMortality::GetMortalityEpisodeProbability (  )  [inline]

For testing purposes.

Returns:
Probability of mortality

float clAggregatedMortality::GetTreeMortalityProb (  )  [inline]

For testing purposes.

Returns:
Mortality probability

bool clAggregatedMortality::IsValid ( clTree p_oTree  )  [protected]

Tests a tree to see whether it is eligible for killing.

It must be of a species and tree type to which this behavior is applied, and not dead.

Parameters:
p_oTree Tree to test.
Returns:
true if the tree is eligible; false if not.

void clAggregatedMortality::KillClump ( clTreePopulation p_oPop,
clTree p_oTree,
int  iNumToKill 
) [protected]

Kills a single clump of trees.

This function gets the coordinate of the first tree in the clump. It searches a circle of an ever-widening radius until it finds enough trees to make a complete clump. It then sorts the trees it finds in distance order so that the trees closest to the target will be killed to make the clump.

Parameters:
p_oPop Tree population object.
p_oTree First tree in the clump.
iNumToKill Number of trees to kill.

void clAggregatedMortality::PreMortCalcs ( clTreePopulation p_oPop  )  [virtual]

Performs mortality.

A random number compared to the probability of a mortality episode decides if an episode occurs. If not, this exits. If so, this goes through all the trees to which it applies. For each, if the tree is not already "dead", then a random number decides whether it begins a dead clump. To find the trees in the clump, it searches in a circle of radius 5 meters for neighbors. The trees eligible to die are killed in order of distance until the number in the clump has been reached. If not enough trees are found in 5 meters, the circle is widened in 5 meter increments until enough trees are found.

Parameters:
p_oPop Tree population.

Reimplemented from clMortalityBase.

void clAggregatedMortality::ReadParFile ( xercesc::DOMDocument *  p_oDoc  )  [protected]

Reads in parameters from the parameter file.

Parameters:
p_oDoc DOM tree of parsed input file.
Exceptions:
modelErr if:
  • The value for amount to cut is not between 0 and 1
  • The return interval is less than 0

void clAggregatedMortality::SetUpAppliesTo ( clTreePopulation p_oPop  )  [protected]

Sets up the mp_bAppliesTo matrix.

Parameters:
p_oPop Tree population object.


Member Data Documentation

bool clAggregatedMortality::m_bClumpSizeDeterministic [protected]

If true, the size of a single clump of trees to kill is deterministic.

If false, the size is chosen from a negative binomial probability distribution.

float clAggregatedMortality::m_fMortalityEpisodeProbability [protected]

Per-timestep probability of a mortality episode occuring, 0 - 1.

Read from the parameter file as a return interval, then this value is calculated as T/RI, where T is the number of years per timestep and RI is the return interval in years.

float clAggregatedMortality::m_fMortalityProb [protected]

Per-timestep proportion of trees to kill, 0 - 1.

Read from the parameter file as an annual amount, then this value is calculated as (1-(1-AD)^X)/C, where AD is the amount of damage, X is the number of years per timestep, and C is the clump size.

int clAggregatedMortality::m_iNumTreesToClump [protected]

Number of trees to kill together in a clump.

This is either a deterministic value or the mean of the negative binomial probability distribution from which to choose a value.

int clAggregatedMortality::m_iTotalNumSpecies [protected]

Total number of species.

Primarily for the destructor.

bool** clAggregatedMortality::mp_bAppliesTo [protected]

Matrix for which tree types and species this behavior is applied.

Array index is sized total number of species by total number of life history stages.

int** clAggregatedMortality::mp_iDeadCodes [protected]

Return codes for the "dead" bool tree data member.

Array index one is sized total number of species by total number of life history stages.


The documentation for this class was generated from the following file:
Generated on Wed Nov 29 08:42:20 2006 for SORTIE Core C++ Documentation by  doxygen 1.4.7