Evaluating and modelling radiata pine wood quality in the · PDF fileForest Quality more...

Forest Qualitymore valuable trees


FWPA Project PNC325-1314: Evaluating and modelling radiata pine wood quality in the

Murray valley region

Geoff DownesForest Quality Pty. Ltd.

Dave DrewForest Forecasting

Validation and development of eCambium within an evaluation of Pinus

radiata wood variability in the Murray Valley Basin

Jul 2014 – Aug 2016


Wealth from forestry

• The ability to turn sunlight into product and add value to it.

y = 93.61x + 55.20R² = 0.57

$300

$325

$350

$375

$400

$425

$450

2.9 3 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9

Site

ave

rage

lum

ber

valu

e ($

.m-3

)

Site-average log Acoustic wave velocity (km.sec-1)


eCambium’s prediction of site value

• eCambium predictions made 4 months before logs were harvested

• Partly dependent on grade thresholds for converting MOE into grade

y = 0.41x + 239.37R² = 0.62

$225

$250

$275

$300

$325

$350

$375

$400

$425

$225 $250 $275 $300 $325 $350 $375 $400 $425

Act

ual c

alcu

late

d va

lue

$/m

3

eCambium predicted $/m3

VicNSW

Bago cpt 66 high elevation


eCambium’s prediction of log MOE


y = 1.01xR² = 0.50

6

7

8

9

10

11

12

13

14

6 7 8 9 10 11 12 13 14

Mea

n lo

g M

OE

(GPa

)

eCambium predicted BH MOE (GPa)

Vic

NSW

Bg066

Me111

Hv013a


Overview of eCambium (Version 2)

• eCambium predictions of wood variation link effects of growth on board volume and quality


Drilling down into the lower level values

• Daily predictions of

• Wood density• MFA• Fibre diameter• Fibre wall

thickness


Tree growth predictions

Predictions of• Height• DBH• Stand Volume• Soil water

availability• Predawn water

potential


eCambium’s main user interface


Defining site


Defining regime


Defining weather


Building a “scenario”


Murray Valley Basin ValidationProject Objectives•Regional assessment of wood property variation from P.radiata sites in the Murray Valley Basin

•close to harvest•represent wide range of resource variability

•Obtain a representative data set describing site average andvariance data of tree and wood properties as a basis forevaluating the performance of the eCambium modelling tool

•Produce a “commercially-ready” version of eCambium


Site SelectionBuccluegh

Greenhills

BagoCarabost

Maragle

Shelley

Ovens Valley

Benalla


Stage 1: NDE Resource Evaluation

• 53 sites• 30 trees per site

• DBH• Outerwood density cores • Standing tree acoustic velocity

(ST300)• 6 trees per site

• Tree height• Branch diameter• Bark thickness


eCambium prediction of outerwood density

50% of the variation in site mean OWD was

predicted by eCambium


Empirical prediction of outerwood density

OWD = 396.6–1.93(Branch Diameter) + 0.174(Spring Rainfall) + 2.43(Age)

• 38% variance explained

UT_OWD = 392–2.36(Branch Diameter) + 0.63(Autumn Rainfall) + 4.64(Max. Autumn T)+2.45(Age)

• 55% variance explainedT1_OWD = 1187–2.26(Branch Diameter) – 0.476(DBHOB) – 0.066(Annual

Rainfall) – 21.55(Ave Temp) – 0.07(Annual Radiation) + 10.1(Age)

• 71% variance explained


Problem of 50mm outerwood cores

0

200

400

600

800

1000

1200

1400

0 25 50 75 100 125 150 175 200

a. CB001 250-100

Mean wood density (kg/m³)

5 yr50 mm

0

200

400

600

800

1000

1200

1400

1600

0 25 50 75 100 125 150 175 200

b. CB001 1000-1000

Mean wood density (kg/m³)

50 mm5 yr

50 mm

5 yr

50 mm5 yr


Comparison with actual Silviscan data allowed

eCambium to be evaluated at the annual

ring level to assess whether the model was predicting radial trends

accurately.

Radial trends


IML Resistograph

• 30 secs per tree• Provides radial variation• No Laboratory work• Easy –to-use

y = 0.15x + 17.06R² = 0.88

300

350

400

450

500

550

2000 2500 3000 3500 4000

Oute

rWoo

d Den

sity (

kg/m

3)

Outerwood Resi (resistance units)

a.


eCambium site value predictions

1:HV013

8264

$329

Count =DBHUB =

0.250.500.75

2:HV013a

10268

$347

2:ST048

7242

$252

3:BR019

12278

$265

5:MG001

7242

$310

6:MG001

13316

$314

7:MG001

21390

$340

8:ME111

15330

$315

9:NN001

21386

$355

10:HC427

13316

$320

11:MH00

13316

$219

Count =DBHUB =

0.250.500.75

12:WT00

13302

$318

13:HL224

13304

$315

14:WC15

21372

$314

15:EV002

13312

$276

16:JN058

21378

$342

17:KO057

18350

$345

18:LV015

16334

$348

19:LV018

10274

$305

20:LV018

13314

$337

21:GO024

20364

$247

Count =DBHUB =

0.250.500.75

22:BU019

13308

$337

23:BU019

8262

$317

24:TR016

12284

$307

25:TR014

15332

$344

26:GA003

21374

$314

27:WB026

38506

$369

27:WB02

16340

$282

28:AR119

20354

$352

29:GC80

16334

$348

30:OC119

13320

$330

Count =DBHUB =

0.250.500.75

31:OC119

12290

$295

32:OC101

12282

$321

33:OC101

21392

$369

34:WJ115

18352

$357

35:SP118

13298

$223

36:SP118

20370

$217

37:BI104T

13312

$386

38:BI133

20358

$363

39:MA04

12286

$302

40:MA05

13308

$309

Count =DBHUB =

0.250.500.75

41:BG582

12288

$303

42:BG587

13308

$321

43:BG583

21380

$362

44:BG066

12280

$318

45:GH849

13322

$236

46:GH849

20358

$277

47:GH845

13324

$331

48:GH845

20368

$268

49:GH828

18352

$232

49:GH828

13312

$229

Count =DBHUB =

0.250.500.75

50:CB001

13302

$323 51:MU20

7250

$354 52:CB011

12280

$368 53:CB018

12284

$328

utilityF5MGP10MGP12MGP15

Boar

d Pr

opor

tions


Sawmill study sites value predictions

2:HV013avalue = $ 344 m^3

10268

utilityF5MGP10MGP12MGP15

Board Count =DBHUB =

0.0000

0.0625

0.1250

0.1875

0.2500

8:ME111value = $ 320 m^3

15330

11:MH001value = $ 250 m^3

13316

18:LV015value = $ 344 m^3

16334

19:LV018bvalue = $ 310 m^3

10274


0.0000

0.0625

0.1250

0.1875

0.2500

20:LV018avalue = $ 334 m^3

13314

37:BI104T1value = $ 375 m^3

13312

40:MA044UTvalue = $ 316 m^3

12286

42:BG587T1value = $ 325 m^3

13308


0.0000

0.0625

0.1250

0.1875

0.2500

44:BG066UTvalue = $ 328 m^3

12280

47:GH845T2value = $ 331 m^3

13324

52:CB011T1value = $ 360 m^3

12280

Boa

rd V

olum

e (m

^3)


Log Preparation


Comparing actual and eCambium log end images


Barcode application to log ends


Logs weighed for green density calculation


Logs loaded into mill infeed


Resultant boards imaged


eCambium’s prediction of site value


• Partly dependent on grade thresholds for converting MOE into grade

y = 0.41x + 239.37R² = 0.62

$225

$250

$275

$300

$325

$350

$375

$400

$425

$225 $250 $275 $300 $325 $350 $375 $400 $425

Act

ual c

alcu

late

d va

lue

$/m

3

eCambium predicted $/m3

VicNSW

Bago cpt 66 high elevation


Conclusions: 1 of 2• eCambium explained similar OWD variance to empirical models,

• These were fitted regressions and not applied in a predictively • What is a commercially acceptable?

• Discussions with industry: 60% or better is what most would be comfortable with.

• eCambium predictions in range using general inputs for site descriptions• publically available database values• FR constant at 0.3, • rock content constant at zero

• Parameter optimization, across Tasmania, Green Triangle, Victoria and NSW yielded calibration r2 > 60%.

• Comprehensive optimization and further development should improve predictions.


Conclusions: 2 of 2

• eCambium explained 50-60% of the site average sawlog value and log MOE

• Cheap and simple way to obtain broad-scale insights into wood property and value variation across an estate

• Used to target field sampling programs• combine with Resistograph as a cost-effective approach to sampling for density

and AWV.

• Model setup and operation has been simplified • ready for commercial application


Industry workshops

• Two workshops will provide an opportunity for FWPA levy payers to try the model

• Wed 6th July 2016• FCNSW offices, Tumut

• Friday 8th July 2016• HVP offices, Melbourne


Questions:

Date post:	17-Mar-2018
Category:	Documents
Upload:	doanh
View:	221 times
Download:	2 times

Evaluating and modelling radiata pine wood quality in the · PDF fileForest Quality more...

Documents