Modelling Natural Regeneration in Mountain Pine Beetle Affected Stands A Hybrid Model Approach Derek Sattler, M.Sc. Candidate Faculty of Forestry. University.

Modelling Natural Regeneration in Mountain Pine Beetle Affected Stands

A Hybrid Model Approach

Derek Sattler,M.Sc. Candidate

Faculty of Forestry. University of British Columbia, Vancouver, Canada.

Mountain Pine Beetle (MPB) Epidemic Lodgepole Pine (Pinus contorta var. latifolia)

Cumulative Volume Killed in All 'Pine' Units

0

250

500

750

1000

2000 2005 2010 2015 2020 2025year

tim

ber

vo

lum

e (1

0000

00's

of

m^

3)

Projected Kill

Observed Kill

Mil

lio

ns

of

m3

~ 80%

Dendroctonus ponderosae

Cumulative Volume Killed on the Timber Harvesting Landbase

Source: BC MoF, 2005

Stand Dynamics Post-MPB Attack

Highly variable snag fall rates (5 – 15 years)

Expect to see small tree release

Changing light dynamics

15-20 year regeneration delay

Challenge to model regeneration Uncertainty in Yield Projections

Candidate Growth Models:

1) SORTIE-ND Forest Ecology Model

2) PrognosisBC

Forest Management Tool

Input data: tree list, site info

Small treesHt then DBH growth

Large TreesDBH then Ht growth

MortalityCompetition, dbh, etc

Change in Crown

Regeneration

results

Thinning

smoothing

PROGNOSISBC Model Flow

Project Specific PrognosisBC Advantages

1) Calibrated using local data

2) Designed for complex, mixed stands

3) Includes Site factors – transportable

4) Government supported model

PrognosisBC Project Disadvantages

1) Poor results with Regeneration Submodel

2) No Post-MPB specific Mortality Submodel

3) Not Spatially Explicit

(i.e., Clumped vs. Even distribution)

SORTIEND Model Flow

Input data: tree list, location

Seedling/SaplingsDiameter then Ht

Large TreesDBH then Ht growth

Change crown size

Mortality

Regeneration

results

Light

Stem Map

Thinning

SORTIE Project Specific Advantages:

1. Episodic Mortality Behaviour

2. One year cycles for simulated runs

3. Post-MPB specific snag fall down function

4. Light mediated model

Project Specific Disadvantages:

1. Has not been calibrated for study area

2. Less precision in G & Y estimates

3. Over-simplified crown allometry

4. Used Less (?)

adbhCsCrownRadiu 1

bHeightCtCrownHeigh 2

Hybrid Model (SORTIE + PrognosisBC)

Advantages of Hybrid Approach:

1) Natural Regen Following MPB – Dynamic- Process-based Model

2) Tree Growth through Empirical Model

3) Uses Existing Models

Hybrid Model FlowSortie-ND

O/S + U/S tree list (from field data)

-

Time 1 (MPB attack)

Defined by ?

PrognosisBC

O/S + U/S tree list (from field data)

Sortie-ND

New O/S + U/S tree list following simulation

New Seedlings

PrognosisBC

New O/S+ U/S tree list following projection

Imputation from SORTIE

Time 2 (Post MPB attack)

PrognosisBC

O/S + U/S + New Seedlings projected in Prognosis

Regeneration submodel ‘off’

Time 3

Preliminary Results

Tested SORTIE-ND using CFS data (R. Scott)

(1987, 2001) SORTIE behaviour selection:

O/S + U/S + Initial Mortality + Subsequent Mortality– Non-spatial Seed dispersal– Number of Seeds = f (Basal Area parent trees) – Proportional Seedling Establishment– Light dependent mortality

Ht Class = 0.1-0.5cm

0

2000

4000

6000

8000

0 2000 4000 6000 8000

Observed Stems Per Hectare

Pre

dic

ted

SP

H

Lodgepole PineOther ConifersDeciduous trees.

a)

Ht Class = 1.0-1.5

0

500

1000

1500

0 500 1000 1500

Observed Stems Per Hectare

Pre

dic

ted

SP

H

c)

Species 0.1-0.5 1.0-1.5 0.1-0.5 1.0-1.5Pine -3159 -154 3999 391Conifers -75 -7 270 16Deciduous 94 180 204 383

BIAS RMSE

n = 9 stands

Lodgepole PineOther ConifersDeciduous trees.

Modifications to SORTIE-ND

1. Bath seed rain function

2. Height/DBH allometry

3. Light-dependent mortality

4. Crown allometry

Crown allometry

Crown Ratio (CR):

Xe

aRC

1ˆ

Crown Allometry Results

Pseudo – Rsquare

Model lnCCF H/D TPH H Slope Elevation

0.34 0.07 0.12 0.03 0.25 0.01 0.01

Standard Error of Estimate (SEE)

Model lnCCF H/D TPH H Slope Elevation

0.17 0.20 0.20 0.21 0.18 0.21 0.21

ElevationgSlopefSPHeCCFd

HtcDHbCR

ln

/0

11ln

Next Steps for the Hybrid Model

1. Crown Width Model

2. Other SORTIE-ND parameter adjustments• Using new dataset

3. Identification of ‘Hand-off’ point

4. Efficient Linkage (SORTIE to Prognosis)

Outstanding Questions

1. How to determine hand-off point between SORTIE-ND and PrognosisBC?

2. Does the Hybrid Model improve upon MSN results?

3. Does the Hybrid Model improve upon SORTIE alone, Prognosis alone?

• How to test this?

Acknowledgments

Data For Preliminary Analyses:

Natural Resource Canada (Brad Hawkes) - MBPI

Funding:

British Columbia Forest Science Program

Supervisor:

Dr. Valerie LeMay

Committee Members:

Peter Marshall, Bruce Larson, Dave Coates

Preliminary Analysis: Prognosis Technical Support:

Robyn Scott Donald Robinson, ESSA