South-East Regrowth Forest Growth & Yield Modelling - NSW

This document has been scanned from hard-copy archives for research and study purposes. Please note not all information may be current. We have tried, in preparing this copy, to make the content accessible to the widest possible audience but in some cases we recognise that the automatic text recognition maybe inadequate and we apologise in advance for any inconvenience this may cause.

· RES E."IS. R ( H _ P. 11 .p E R _ No. 2 4

e· SOUTH-EAST REGROWTH FOREST GROwrH AND YIELD MODELLING: DESIGN, METHODS AND PROGRESS

By Huiquan Bi

S TAT E

FORESTS RESEARCH DIVISION

SOUTH-EAST REGROWfH FOREST

GROWfH AND YIELD MODELLING:

DESIGN, METHODS AND PROGRESS

by

HUIQUANBI

RESEARCH DIVISION STATE FORESTS OF NEW SOlmI WALES

SYDNEY 1994

Research }>aper No. 24 September, 1994

The Author:

Huiquan Bi, Research Officer, Silviculture Section, Research Division, State Forests of New South Wales

Published by:

Research Division,

State Forests of New South Wales,

27 Oratava Avenue, West Pennant Hills, 2125

P.D. Box 100, Beecroft 2119

Australia.

Copyright ©' 1994 by State Forests of New South Wales

DDC 634.956011 ISSN 0729-5340 ISBN 0 7310 2200 9

le

'-

CONTENTS

SUMMARY

INI'RODUCI'ION

THE STUDY AREA

STAND CHARACTERISTICS OF THE IRREGULAR REGROWTH FORESTS

EXAMINATION OF THE EXISTING FOREST GROWTH DATA

DATA REQUIREMENI'S

THE FRAMEWORK

FOREST RESOURCES INFORMATION DATABASE

FOREST MANAGEMENI' DATABASE

SAMPliNG DESIGN AND FlEW PROCEDURES

1. THE INI'EGRATED PWT LAYOUT

2. PIN-POINI' LOCATIONS IN THE FlEW

3. THE SURVEY CRUISE

4. PIN-POINT PWT CE;NI'RE AT REGROWTH SITES

5. MEASURING AND SAMPliNG WITHIN THE INI'EGRATED PWT

VOLUME AND TAPER FUNCTIONS

1. INI'RODUCI'ION

2. RESULTS

TREE RING ANALYSIS

1. METHODS

2. TREE RING MEASUREMENI'S

FOREST GROWTH AND STAND DYNAMICS

1. PREliMINARY RESULTS

FOREST GROWTH MODE LUNG

FUTURE DIRECI'IONS FOR THE PROJECT

ACKNOWLEDGEMENI'S

REFERENCES

STA TB FORESTS OF NEW SOUTII WALES RESEARCH PAPER NO. 24

SOUTH-EAST REGRowrn FOREST GROwrn AND YIELD MODEIl.ING: DESIGN. METIIODS AND PROGRESS

ill

1

2

3

5

6

7

9

9

14

14

16

16

16

17

21

21

21

23

23

25

26

27

30

32

33

34

TABLES

Table 1. Areas of each forest type under the five rainfall classes within G1enbog 10 State Forest as extracted from the forest resources information database

Table 2. Logged, unlogged and never to be logged areas of each forest type within 12 G1enbog State Forest as extracted from the forest management database

Table 3. Numbe~ of plots allocated to each forest type within G1enbog State Forest 14 dUring the first stage 'of sampling

Table 4. Mean DBH, mean height and site characteristics of the forty 18 regrowth stands sampled

Table 5. Regression coefficients of combined variable underbark stem volume 22 equations for the six species and tlle validation statistics for the equations

Table 6. The estimated stand age, stand basal area, stand volume and stand 28 density of the 25 sample stands in regrowth brown barrel forests

FIGURES

Figure 1. The framework of the south-east regrowth forest growth and yield 8 modelling project

Figure 2. Distribution of forest type 154 and 152 in G1enbog State Forest 11 as extracted from the forest resources information database

Figure 3. Distribution of logged, unlogged and never 1l? be logged brown barrel 13 forests in G1enbog State Forest as extracted from the forest management database

Figure 4. A diagram of the integrated plot layout in a regrowth forest with 15 more regrowth trees than advanced· and old growth trees

Figure 5. The distribution of sample stands within G1enbog State Forest 20

Figure 6. Stand age of ten stands estimated through known logging and fire history 24 versus that estimated from tree rings

Figure 7. Mean annual increment (MAl) in regrowth stand volume in relation to 29 density of old-growth

Figure 8. Regrowth stand volume in relation to stand age and Old-growth density 29 (trees/ha)

PLATES

Plate 1. An irregular mixed-species regrowth stand in Glenbog State forest 4

Plate 2. An example of a disk synchronised for tree ring measurements 25

ii SOUTII-EAST REGROWTH FOREST GROWTH AND YIELD MODEI1.JNG: DESIGN, METIlODS AND PROGRESS

STATE FORESTS OF NEW SOUTII WALES RESEARCH PAPER NO. 24

e

SUMMARY

Regrowth forests constitute a substantial proportion of the 230,000 ha forest estate under the management . of State Forests in south-east New South Wales. They are managed for multiple use, with wood production and biological conservation being the most important management objectives. Managing these forests for multiple use has greatly increased the demand for accurate growth and yield information on forest management. In order to provide forest management the ability to determine the growth and yield of these regrowth forests, the South-East Regrowth Forest Growth and Yield Modelling Project was initiated in 1991 with an overall objective of developing growth and yield models for the regrowth forests in southeast New South Wales.

This report describes the design, methods and progress of this project. It starts with an account of current status of growth and yield modelling of mixed-species and mixed-age forests in Australia, then gives a brief description of the study area and the characteristics of the regrowtb stands. The limitations of the existing growth data and the data requirements of the project are recognised. The overall framework of the project .is outlined, within which areas of research and development work are clearly defined. These areas include developing forest resources information and management database, field sampling, tree ring analysis, developing volume and taper equations, forest growth and stand dynamics, forest growth modelling, merchantability modelling and harvesting modelling. The progress in each of these areas is. reported. Finally, future directions of the project are indicated.

STATE FORESTS OF NEW SOUIH WALES RESEARCH PAPER NO. 24

SOUIH-EAST REGROwm FOREST GROwm AND YIELD MODEUlNG: DESIGN. METIIODS AND PROGRESS ill

INTRODUCTION

'Essential to forest planning and management is the capacity to predict the growth and yield of forests both spatially and temporally. Traditionally developing such a capacity has been time consuming with

. considerable concerns in accuracy and timeliness of introduction. Recent advances in computer and infol1D.ation technology have led many forestry organizations to progress towards an integrated system which combines a forest resources da~base and growth models with a geographical infol1D.ation system for the detel1D.ination of future growth and yield (Deadman 1992, Gagnon et al. 1992). These integrated systeIns also facilitate the examination of alternative management strategies and harvesting options. Forest management will be increasingly dependent on the availability and reliability of such systems in which growth models play a pivotal role. .

e Through much of Australia's history, forest utilization was characterized predominantly by selective logging in old growth forests (phillis 1971, Carron 1985). The forests of south-eastNew South Wales have provided pulp wood and sawlogs to the local timber industry for many years (Anon. 1982). Regrowth forests now constitute a substantial proportion of the south-eastforests, and thus have become increasingly

. important for wood production. Land use management options forthese forests have been subject to much public debate since the beginning of woodcbipping at Eden (Carron 1985). Biological conservation has been recognized as an objective in multiple use forest management of the area (Richards et al. 1990). To manage the forests for multiple use more intensive management of the regrowth forests seems inevitable (Florence 1987). To achieve a balance between different management objectives more and realistic land use management options must be presented for public debate. Both require accurate growth and yield infol1D.ation of the regrowth forests.

The regrowth forests regenerated from selective logging usually develop into irregular, mixed-species and mixed-age forests. For a long time, in most forest services growth estimation oflogged-over forests was not a major priority (Rayner and Turner 1990a). It was not until 1966 that the first growth model for irregular forests was developed Using data from 62 continuous forest inventory plots in blackbutt (Eucalyptus piluIaris) dominated in Coopernook State Forest on the north coast of New South Wales, Turner (1966) derived a series of multiple linear regression equations to predict stand basal area and stand volume growth from only three predictor variables - stand density, species composition and site quality. More efforts on growth estimation of logged-over forests soon followed. Also using continuous forest inventory data Curtin (1970a) provided crude models for the estimation of stand volume and basal area increment from stand basal area and mean diameter for Yarratt State Forest on thenortb coast of New South Wales. Phillis (1971) proposed a simple arithmetical procedure to derive short-tel1D. yield prediction of irregular stands. under varying intensities of cut In addition to these stand level growth models, an individual tree growth model was developed by Kilgour (1982) for mixed-species and mixed-age messmate (E. obliqua) forests of southwest Victoria. In this model DBHOB increment of individual trees was a function of DB HOB, basal area, density, site index, and a dummy variable indicating the.status of past logging history. Simulations of some alternative management strategies for the messmate forests using this model were demonstrated by Kilgour and Rudra (1988).

These existing models represent only a limited development of growth models for the vast extent of irregular forests in Australia; for many irregular forests growth models have not been developed (Rayner and Turner 1990a). Furthel1D.ore, since forest management has changed significantly over the past twenty years, the existing models may be inappropriate for growth and yield prediction of irregular regrowth forests generated by current management practices. Without appropriate growth and yield models forest management is restricted to very general silvicu1tural strategies and management options over broad


SOUTII-EAST REGROWTH FOREST GROWTH AND YIELD MODEU..ING: DESIGN. METIIODS AND PROGRESS 1

geographical areas, a: wide range offorest types and stand conditions. Such generalizations can potentially distort the true productive potential and optimal forest level management strategies (Rayner and Turner 1990a).

The requirements for growth. and yield information for decision making were the circumstances under which the South-East Regrowth. Forest Growth. and Yield Modelling Project was initiated in 1991. The Commonwealth Department of Primary Industry and Energy provided funding in addition to support from State Forests of New South Wales. The overall objective of the project is to develop growth. and yield models for the regrowth. forests in the south-east of New South Waies. This report outJ#ies the design and progress of the project.

THE STUDY AREA

There are about 230,000 ha of forests under the management of State Forests in south-east New South Wales. The region lies within latitude 36~ I' to 37~8' south and longitude 149'lO1' to 1S00S0' east. It can be broadly divided into a tableland area and a coastal area, each with a diversity of forest types. The tableland forests consist of mostly wet sclerophyll forests covering approximately 8S,000 ha with elevation ranging from 800 to more than 11oom. HereEucalyptusjastigataisthemostimportantcommercial species. It can occur in pure stands, but often grows in association with other eucalypts such as E. obliqua, E. nitens, E. viminalis, E. elata, E. cypeUocarpa and E. radiata. The mean annual rainfall in these areas ranges from 700 mm to 11oomm. The mean annual temperature is inthe range of8.so-13.SoC and the mean minimum for the coldest month is -40 to 3°C. The coastal forests consist of mostly dry sclerophyll forests where E. sieberi is the predominant commercial species. Other species often associated with E. sieberi include E. obliqua, E. cypellocarpa, E.fraxinoides, E. dalrympleana, E. muellerana, E. radiata, E. dives, E.longifolia, E. globoidea and E. viminalis. The coastal area is warmer and receives less rainfall than the tableland area. The mean annual rainfall ranges from 700 mm to 1000 mm. The mean annual temperature is between 13° and 1SoC and the mean minimum temperature in the coldest month is in the range of SO-8°C.

Wildfire has been a frequent event in the region, with those in 1939 and 19S2 being the most destructive recorded. It has had a pronounced effect on the regeneration and age structure of the forests (Bridges 1983). High intensity fire can destroy all trees, while low intensity fire damages some trees more than others in the stand. Relatively even-aged regrowth. stands usually develop after high intensity fires; mixed-aged stands often emerge after low intensity ones. In general the regrowth. forests after fire are a mosaic of relatively even-aged and mixed-aged stands.

The relatively accessible parts of this area were selectively logged primarily for sawmilling for many years under very limited supervision (Carron 1985). After the State Government and Harris-Daishowa reached an agreement to supply woodchips for export to Japan in 1967, pulp wood was also harvested in the area. Integrated logging initially concentrated on the coastal area and moved to the tableland area in the late 1970s. Resource assessment was carried out through air photo interpretation in late 1960s and early 19708 (Anon. 1982). Regrowth. forests have become increasingly important in wood production as large areas

2 sourn-EAST REGROwrH FOREST GROwrH AND YIELD MODElLING: DESIGN, METIlODS AND PROGRESS

STATE FORESTS OF NEW sourn WALES RESEARCH PAPER NO. 24

of old growth forests in the region, previously available for logging. have become unavailable to forestry. However, the distribution and extent of regrowth forests are not known. Inventory of regrowth forests in the area is being carried out.

TIle management system has undergone a considerable change since the inception of integrated harvesting in 1969 (Bridges 1983). TIle changes in harvesting unit, burning regimes. erosion control, tree retention and wildlife management were described in detail by Bridges and Dobbyns (1991). Under the cunent alternate coupe system the average size of a coupe for harvesting is 60 ha. Within the harvested coupe, wildlife habitat trees. seed trees, potential sawlogs and advanced growth are retained according to management specifications. This management system generates irregular mixed-species and mixed-age regrowth forests with a wide range of age and size structures. TIle residual trees after logging in these forests have a diversity of spatial distribution patterns.

STAND CHARACTERISTICS OF THE IRREGULAR

REGROWTH FORESTS

Because regrowth forests in the region resulted mostly from logging andIor fire, their stand structures reflect to a large extent these past disturbances. 1bere may be five or more age and crown strata in the irregular mixed-species regrowth forests (Plate 1): (1) old-growth stra1UIll, (2) advanced growth stra1l1m, (3) major regeneration stra1UIll, (4) minor regeneration stra1UIll and (5) shrub stratum. Trees within each stratum mayor may not be of the same age. The age range of trees within the same stra1l1m may very from one to many years. 1be old-growth stra1UIll is characterised by large living old trees and also, in some cases, standing dead trees. These senescent trees usually have many dead limbs, broken branches and epiconnic shoots along the stem. 1be branches tend to be thicker and branching angles relatively greater. 1be crown volume is much reduced and relatively small for the diameter of the stem. TIle foliage is more or less vertically distributed along the stem. Trees in the advanced growth stra1UIll generated before, and survived after, the last major disturbance in the stand. Still vigorously growing. these trees may be the tallest in the stand with large and dense crowns. Rarely, epiconnic shoots can be found along the stem. TIle major regeneration stra1UIll may have a greater number of trees in the stand than any other stra1UIll. Growing up shortly after the last major disturbance trees in this stratum have a narrower age range and fonn the most important part of stand growth. Competition strongly affects the growth of these trees. Sometimes another stratum of much younger regeneration may be identifiable. Trees in these strata may belong to the second wave of regeneration after the last major disturbance. They often fill in the gaps in the stand fonning a minor component of the regrowth. Wildfire has also resulted in a substantial number of coppice trees in the forests. These trees grow mainly from the base of old stumps after fire, and often have double or multiple leaders with a thicker base than trees originated from seedlings.

STA TB FORESTS OF NEW SOU1lI WALES RESEARCH PAPER NO. 24

SOU1lI-HAST REGROwrn FOREST GRowrn AND YIBLD MODBU.ING: DESIGN. MB1lIODS AND PROGRBSS 3

Plate 1. An irregular mixed-species regrowth stand in Glenbog State Forest

4 SOU1H-EAST REGROwrn FOREST GROwrn AND YIELD MODElLING: DESIGN, METIIODS AND PROGRESS

STA TB FORESTS OF NEW SOU1H WALES RESHAR.CH PAPER NO. 24

EXAMINATION OF THE EXISTING FOREST GROWTH

DATA

In the coastal area. about 250 growth and thinning plots have been established since 1968, mainly in relatively even-aged fire or logging regrowth forests of E. sk~ri. Many growth plots were established to cover four logging years (1970n1, 1975n6, 1980/81 and 1985/86) which were thought to provide a range of site conditions. Forty nested plots are in four intensive study areas which were initially set up for monitoring the regeneration and early growth of the regrowth stands. The thinning plots are located in four thinning trials in 1939, 1952 and 1964 fire regrowth stands. These growth and thinning plots are distributed mainly in four State Forests: Nadgee, East Boyd, T'unbillica and Yambulla. 1be areas where plots were established receive annual rainfall ranging from 800 mm to more than 1100 mm. 1be elevation of these areas varies from 20 m to 650 m above sea level. 1be parent material in many of these areas is either Devonian or Ordovician sediments (Kelly and Turner 1978).

Measurements of these plots include DBH of all trees and the height of a few dominant trees in each plot. Measurements have been carried out at irregular time intervals. Many plots were measured annually in the 70s. In more recent years most plots have been measured every three to five years. In the intensive study areas only one plot out of ten was measured in each area during the last measurement period. There is no comprehensive document about these growth plots and thinning experiments. Records of establishment, measurement and fire history are available for some plots, rut not complete for others. No growth and thinning plots were established after 1983.

Since fire is a frequent event in this region many plots have been affected by fire to some extent. In total about one-third of the plots have been affected by fire. Fire disturbances have brought a great deal of 'noise ' into the growth data making data analysis very difficult. Bruskin and Home (1990) found a number of shortcomings with the growth data from Eden specifically: (1) data from a large number of plots were not useful because of damage caused by the 1980 fire; (2) a number of plots did not have all trees counted and measured, whilst other plots had all of the trees measured in some years but not in other years; and (3) the growth plots are not representative of current management practices for overstorey retention. Their findings reflected more or less the quality of the growth data from these plots. Apart from these problems, many plots have only been measured three or four times and the data have not yet covered a big enough age range for modelling the long term growth of the regrowth forests. Furthermore, because only few dominant trees were measured for height it is impossible to estimate stem volume of trees using the current volume equations for E. skberi which require both DBH and height data

In the tableland area no growth plots have been established, so there is a complete lack of growth data for the forests located there.

STA'rn FORESTS OF NEW SOUlH W ALBS RESEARCH PAPER NO. 24

SOUlH-EAST REGROwrn FOREST GROwrn AND YIELD MODBLLING: DESIGN. MElHODS AND PROGRESS S

DATA REQUIREMENTS

Ideally, data for growth and yield modelling in general should come from randomly located permanentandl or temporary growth plots which cover the geographic range of the forest and encompass a broad range of forest types (Vanclay 1991). Growth data for each forest type should cover the full range of environmental gradients within which the forest type occurs and a wide range of stand conditions under which the particular forest type exists. The number of plots needed may vary with forest type. Forexample. in tropical rain forests, 50 to 100 randomly located plots are recommended for each forest type (Synnott 1979). Also. growth data should be of high quality because modelling approaches and techniques depend on the quality of the available data. Generally, individual-tree level data are required to ensure flexibility in modelling approaches (Rayner and Turner 1990b). Based on the experiences in New Zealand, Tennent (1988) suggested that it is better to have a small quantity of high quality data than to have a large quantity oflow quality data. Furthermore, growth data should come from plots large enough to be representative of the forest stand so that stand attributes can be accurately estimated.

Existing growth data are very limited for constructing growth and yield models for the south-east regrowtb forests. This limitation necessitates that the present project starts with the collection of growth data. Because of the diversity of forest types and the irregular stand characteristics of the regrowth forests, growth data must be collected from a large number of plots covering all commercially important forest types and a wide range of stand conditions. Since there is a complete lack of growth data for the tableland forests, which represent a significant area for wood production in the region, sampling temporary plots in tableland forests to obtain growth data quickly is the priority of the project However, the scope of field sampling is limited by the resources currently available. Within the constraints of the available resoun:e, a total of 40 plots was targeted in Glenbog State Forest during the first stage of sampling. Other tableland forests will be sampled at a later stage.

6 sourn-8AST RBGROWTH FOREST GROWTH AND YIELD MODmLING: DBSIGN, MB1HODS AND PROGRESS

STA TB PORBSTS OF NEW SOUlll WALES RESEARCH PAPBRNO. 24

THE FRAMEWORK

TIle overall objective of the project is to develop growth and yield models for the south-east regrowth forests which cover a large geographical area and encompass a diversity of forest types. Because of the lack of growth data. the project has to start with a very basic step, i.e. collection of growth data A research project of these dimensions requires its areas of work be identified and components clearly defined so that the research work can be carried out most efficiently. For growth modelling the following areas of work were identified:

• Forest resources database development: to provide a basis for sampling design and growth prediction.

• Field sampling: to obtain growth and taper data.

• Tree ring analysis: to obtain growth data of individual trees.

• Developing volume and taper equations: to predict stem volume and taper from DBH and height.

• Site productivity index: to derive an indicator of site productivity for irregular forests.

• Forest stand dynamics: to examine the growth of, and interaction among, trees in the irregular regrowth stands.

• Growth modelling: to construct growth models to predict forest growth.

For yield modelling some extra work will have to be carried out in the following areas:

• Forest management database: to establish the application database for yield prediction.

• Defects modelling: to determine the extent of defects and merchantability of trees.

• Harvest modellings: to determine merchantable yield and stem damage caused by harvesting.

1bese component areas together form a framework which maps out the strategy of the south-east regrowth forest growth and yield modelling project (Fig. 1). Within each component area, one or more smaller projects were formulated. Projects related to growth modelling are being carried out concurrently. The design and progress of these projects are described in the remainder of this report.

STA TB FORESTS OF NEW SOU1ll WALES RESEARCH PAPER NO. 24

SOU1ll-EAST REGROwm FOREST GRowm AND YIELD MODBLLING: DESIGN. METHODS AND PROGRESS 7

South-East Regrowth Forest Growth and Yield Forecasting Project

Forest resources and

management information

database development

Forest f----f resources

database

Forest I---~ managemen

database

Forest growth

and yield models

development

Volume and taper models

Tree ring analysis

Forest growth and stand dynamics

Forest growth modelling

Harvest modelling

South-East Regrowth Forest Growth and Yield Forecasting System

Figure 1. The framework of the south-east regrowth forest growth and yield modelling project.

8 SOunI-BAST REGROwm FOREST GROwm AND YIELD MODElLING: DESIGN. ME11IODS AND PROGRESS

STA TB FORESTS OF NEW sourn WALES RESEARCH PAPER NO. 24

FOREST RESOURCES INFORMATION DATABASE

Objective: to establish a comprehensive forest resources database for the Eden region which contains largely static attributes of the forests such as forest type, geographical location, rainfall, parent material, elevation and aspect.

A forest resources database is a prerequisite for sampling design at present and predicting forest growth and yield in the region at a later stage when growth and yield models are developed. Due to the lack of a geographical information system, data collection has been done manually by a cartographer through dotgrid counting. For each compartment the total area is being split into a number of smaller patches, each with a unique set of attributes including forest type, rainfall, parent rock type, elevation, aspect and management category. Such a database is essential for forest management to take the advantage of current development in information technology and will eventually enhance the geographical information system the State Forests is developing.

This work started in April, 1991 and so far has covered the entire tableland area. Except for a part of Yambulla State Forest, the coastal forests have not been covered since many of them have not been typed. The database is now available to State Forests' staff for both management and research purposes. Examples of what the database can provide at this stage are shown in Table 1 and Figure 2.

FOREST MANAGEMENT DATABASE

Objective: to build a secondary database (secondary to the forest resources database) which contains largely the dynamic attributes of the forests (Le. how the forests are being manipulated by management and distwbed by fire) such as areas oflogged, unlogged and never to be logged forests, thinning, harvesting and fire dates, harvesting rotation, yield of pulpwood and saw logs per hectare in each logging cycle.

This database is in part an adjunct to the forest resources information database and records the on-going changes in the forests. It is an important part of the forest growth and yield forecasting system being developed as it is the working level to which the growth and yield models will be applied. Apart from its importance in growth and yield prediction, the database will also support the day to day forest management as well as forest planning and decision making.

Work on this database so fachas enabled some useful information to be extracted from the tableland forests. For example, the areas of each forest type in Glenbog State Forest under the three logging categories can be extracted from the database in tabular form (Table 2). A map of forest type 154 under three logging categories (logged, unlogged and never to be logged) in Glenbog State Forest derived from the database is shown in Figure 3. Information on the recorded sawlog and pulp wood yield from the logged compartments can also be obtained. Fire history, an important attribute for growth and yield modelling, has yet to be included in the database.

STA TB PORESTS OF NEW SOUTH W ALBS RESEARCH PAPER NO. 24

SOUTH-EAST RBGROwm FOREST GRowm AND YIBLD MODELLING: DESIGN. ME11JODS AND PROGRESS 9

-o

~~ h ~~ ~!

h ~~ ;a~

5~ ~

~ ~

El ~i i~ zi~ p ~

Table 1. Areas of each forest type under the five rainfall classes within Glenbog State Forest as extracted from the forest resources infonnation database.

e e

.-c :J C. cc E c cc cc ~ C/I :J ~

Cl .~ ..r:::: .-... 0 z

5960000

5955000

5950000

5945000

5940000

5935000

5930000

5925000

5920000

707000 712000 717000 722000 727000 732000

Easting (Australian map unit)

Figure 2. Distribution of forest type 154 (closed symbols) and 152 (open symbols) in Glenbog State Forest as extracted from the forest resources information database.

STA TB FORESTS OF NEW SOU1H WALES RESEARCH PAPER NO. 24

SOU1H-BASI'REGROWTH FOREST GROWTH AND YIBLD MODELLING: DESIGN. MEllIODS AND PROGRESS 11

Table 2. Logged, unlogged and never to be logged areas of each forest type within Glenbog State Forest as extracted from the forest management database.

12 SOUTIl-EAST REGROwrn FOREST GROwrn AND YIELD MODEUJNG: DESIGN. METIlODS AND PROGRESS


... '2 ::l a. «I

E c .~ "ii .... ... III ::l

~ Cl C :.c ... .... 0 z

5960000

5955000

5950000

5945000

5940000

5935000

5930000

5925000

5920000

707000 712000 717000 722000 727000 732000


Figure 3. Distribution of logged (open symbols), unlogged (closed symbols) and never to be logged (crosses) brown barrel forests (forest type 154) in Glenbog State Forest as extracted from the forest management database.

STA TB FORBSTS OF NBW SOUnI WALES RESEARCH PAPER NO. 2.4

SOUnI-EAST REGROwn! FOREST GROWTH AND YlELD MODELUNG: DESIGN, ME1HODS AND PROGRESS 13

SAMPLING DESIGN AND FIELD PROCEDURES

Based on the forest resources infonnation database, a stratified random sampling scheme was designed for Glenbog State Forest using forest types as defined in Anon. (1989) as strata. The total number of sample plots is forty at this stage. TIle number of sampling plots in each stratum is proportional to the total area of the stratum (Le. the total area of each forest type in the State Forest). No plots were allocated to commeICially insignificant forest types (Table 3).

Table 3. Number of plots allocated to each forest type within Glenbog State Forest during the first stage of sampling.

1. THE INTEGRATED PWT LAYOUT

Because the growth of regrowth trees can be strongly affected by the size, density and spatial location of old growth and advanced growth trees, the stand condition characterised by old-growth and advanced growth must be taken into account when making growth predictions for regrowth stands. Whenmeasuring regrowth trees in the sampling plots, the old-growth and advanced growth trees surrounding the plot should also be assessed. Since the density of old and advanced growth trees is usually low, a two tier system combining both fixed and variable plots techniques is adopted (Fig. 4). The fixed plot should have a radius of atleast 15 m and contain at least 50 regrowth trees. The variable plot should contain 15 to 20 old-growth and advanced growth trees.

SOUIH-liAST REGROwm FOREST GROwm AND YIELD 14 MODELLING: DESIGN. METHODS AND PROGRESS

STA TB FORESTS OF NEW SOUIH w AI..BS RESIiARCH PAPER NO. 24

.. ' ,. o· •

it • •• . :' o

• • • , .' ',. ... .,.

•

•

" ..

'. ... . "~ ..

......

•

•

•

•• ".

"

.. '

o •

• •

• ". . .....

._ ..• _. __ ._ ... _ .0- _. __ ._ ... _' __ .

--.--.- . . -'-. ". ." .. -.- .. , ................ -.-.-.-..

-. .. ..... ... .... . . ......

. ' . •

• • •

• • • • • •

•

• • •

o • • •

• • •

•

'-. c5-........ -... -~ ... _._._. o

......... ... ... . . -', "" . . .... , ........ . -.... _- . .

"'0 ,.. .... / ',

•

••

•

•

.. '

.. , .......... ~·'/9···· • .... / ";< -,\

. .. 0 .-- .. '

' . ./ '., . ...

.' .'

• . ' .

• • • . .

•• •

. :'

•

,,/ . .. / ~; ... / .~ ... // ..

• . .. .... • . ...... . . . . . ,. , .. ,.~;.... . ..... .'

.... ... ........... . - ' ................. -..... -... --'

" . ,.-

Figure 4. A diagram of the integrated plot layout in a regrowth forest with more regrowth trees (closed symbols) than advanced and old growth trees (open symbols). The inner circle represents the fixed plot and the dotted circles represent the variable ploL Two different critical angles (basal area facton of the wedge prism) were drawn to show the effect of critical angle on the nwnber and size of advanced and old growth trees being sampled.

STATB PORBSTS OF NBW SOlTIH W ALBS RBSBAllCH PAPER NO. 24

SOU1H-BAST RBGROW'llI PORBST GROWTH AND YIBLD MODBUJNO: DBSIGN, MB1HODS AND PROGRESS

2. PIN-POlNI' WCATIONS IN THE FlEW

The sampling sites on the map can not always be easily located in the field. The first step to pin-point a location in the field is to select the most reliable and the easiest reference point which can be pin-pointed on the map as well as in the field. 1ben a protractor and a ruler are used to find out the bearing and the distance from the reference point to the sample site. The difference between map north and the magnetic north should be adjusted in this process. Finally, one person sets off with a compass and a measuring chain to locate the site in the field, whilst the other person remains at the reference point holding the other end of the chain.

3. THE SURVEY CRUISE

Because there is little infonnation available for locating the regrowth forests on a map, a survey cruise in the field of a large number of randomly selected sites in each stratum has to be carried out in order to find the regrowth sites suitable for sampling. The results of the survey cruise can provide an estimate of the proportion of regrowth forests in each stratum.

Ifasitehappens to be an old-growth site, a quick measurement of stand basal area should be made. Because of the size structure of Old-growth forests two wedge prisms of different basal area factors should be used, with one for measuring old-growth and advanced growth trees and the other for smaller regrowth trees. An estimate of stand density will be made by counting the number of trees in four 10 x 10 m visual squares. A visual square is formed by a person standing in the field as the corner of the square with his two arms stretched out at a right angle to each other to define the sight borderlines of the square. After counting trees in one square the person rotates 90 degrees to count the number of trees in the adjacent square until all four squares are counted. An estimate of the average height of the tallest five trees should be recorded. The assessment does not have to be carried out in stands with understorey too dense to use the wedge.

4. PIN-POINT PWT CENI'RE AT REGROWIH SITES

Once the centre of a sample site is located, a bearing and a distance (within the range of 0 to 50 meters away from the centre of the site) are taken randomly to detennine the centre of the sample plot.

SOUI'H-BAST REGROwm FOREST GROwm AND YIELD 16 MODHLLING: DESIGN. MH11:IODS AND PROGRESS

STAlE FORESTS OF NBW SOUI'H WALES RESHARCH PAPER. NO. 24

5. MEASURING AND SAMPUNG WITHIN THE INTEGRATED PLOT

1. Locate and maIk. the centre of the plot with a peg.

2. Set up the theodolite and decide the radius of the plot.

3. Measure DBH and balk thickness at breast height, mark. breast height and code trees with paint spray.

4. Trees smaller than 10 cm in diameter will not be measured, but the species and the number of stems within the plot should be recorded.

5. Map trees with theodolite and a tape.

6. Check the measurements before destructive sampling.

7. Cut down regrowth and advanced growth trees with a chainsaw (start with small trees).

8. Measure tree height, bole height and merchantable bole height

9. Measure diameters and bark. thickness along tree bole at 1.5 m intervals until after a diameter smaller than 5 cm is recorded (trees with DBH greater than 10 cm will be measured for taper line).

10. Cut a disk. at the maIk.ed point of each tree and code the disk.

11. Cut taper line disks from five dominant trees (five largest regrowth or advanced growth trees in DBH) and code the disks for stem analysis.

12. Take a block sample from each of the old growth trees within the fixed plot

13. Measure the DBH and height of old growth trees immediately outside the fixed plot

14. Measure the basal area of old growth trees in the variable plot (including trees both inside and outside the fixed plot) at the centre of the plot using a wedge.

15. Measure dead trees and take disk samples from them in the same was as for live trees.

16. Load the disks onto a vehicle.

Progress

Over the past two years more than forty plots comprising more than 2,000 trees, mostly in Olenbog State Forest, have been sampled (Fig. 5). These plots represented a range of growth and stand conditions on the tablelands (Table 4). The first eight plots were sampled to verify annual ring production of trees in the tableland forests. They are not included the forty sample plots shown in Table 3.

STA TB FORESTS OF NEW SOU1H W ALBS RESEARCH PAPER NO. 24

SOU1H-BAST RBGROWTH FOREST GROWTH AND YIBLD MODELLING: DESIGN, MElHODS AND PROGRESS 17

-00

~~ ~~ ~§ ~~ S'" .z ~

h ~~ ~~ m~ ~

~ ;1

;1 ~~ i~ ~~

Table 4. Mean DBH, mean height and site characteristics of the forty regrowth stands sampled.

e e

e e

n Table 4. (COOL)

Forest ~~ z"ll

~~ 21 GlqtbO$ ::: 154 :: zt OJellbo 154

~ .:.: g

:23... GlCrtbog ', 154 ~ ::.' !M':" dkh@g, '154 , i :.: ,", :.:.

' 2$ Gtcitb6g : 154: 26 GJenbog 155

'l:J Glen:bog 152

28 Gtenbog 152

~ 29 G1enbog 152

"30 :: 0Jenbo~' ,1S1-~~ 8§ ~~ .. ~ 34 G1enbog 154 ~~ :::3$

O~.,: 159 .~ i :3<; Gleribog: 154

~~ ~~ ~~ ~~ i~ -10

5960000

5955000

5950000

--·c :J

g. 5945000 E c .!! "ii ... -III :J ::;. ~ 5940000 :c 1:: 0 Z

5935000

Forest Type

5930000 +

5925000

707000

150 0

151 6-

152 0 154 • 155 :&

159 •

712000

~/ l{

• •

0 • •• • • • 0 • • • o o.

A ·Cl • Q) •

~ r ~ L

~ I

717000 722000 727000 732000


Figure 5. 1be distribution of sample stands within Glenbog State Forest. TIle forest types of these stands are messmate (150), brown barrel-messmate (151), messmate-gum (152), brown barrel (154), brown barrel-gum (155) and mountain/manna gum (159). Sample stands in other tableland forests are listed in Table 4.

SOUTIl-EAST REGROwm FOREST GRowm AND YIELD 20 MODElLING: DESIGN. MBTIlODS AND PROGRESS

ST A 'rn FORESTS OF NEW SOUTIl W AIJlS RESEARCH PAPER NO. 24

e

e

VOLUME AND TAPER FUNCTIONS

1. INI'RODUCI'ION

Accurate stem volume prediction is a prerequisite for intensive forest management It is also an essential component of forest growth and yield models. However, volume equations have not previously been available for any species in the tableland area to aid forest management The objectives of research in this area are to develop volume equations and taper functions for major commercial species in both the tableland and coastal forests.

2. RESULTS

Since the work of developing volume equations was described in detail by Bi (1994a, 1994b), only a summary of the results is provided here.

Taper data from more than 1,900 trees were used to develop and validate combined variable (I)2H) underbaIX volume equations for six species, E. fastigata, E. ob/iqua, E. radiata, E. nitens, E. viminalis and E. cypellocarpa (Table 5). Average bias of volume prediction varied from -0.009335 m' for E. cypellocarpa to 0.013521 m] for E.fastigata, with -0.001404 m] for E. viminalis being the closest to zero. Mean absolute deviation of volume prediction ranged from 0.012126 m3 for E. viminalis to 0.043568 m3 for E.fastigata. Root mean squared deviation of volume prediction was generally greater than 0.10 for the equations, with the exception of the equation for E. viminalis. A further evaluation along DBHOB classes for E.fastigata showed that percentage bias was not greater than five percent for all DBHOB classes and percentage precision was mostly smaller than 15 percent Volume equations for species other than E.fastigata are provisional at this stage. They will be updated when more taper data become available within the next two years.

A lower stem form quotient, calculated as the ratio of diameter at 4.5 m above ground level to DBH, was incorporated into volume equations as a measure of stem form for E.fastigata. This incorporation resulted in a substantial improvement in precision of stem volume prediction. The added precision obtained by including the lower stem form quotient into volume equations should prove useful in thinning and fertilization studies where more accurate volume estimation is required to detect tree responses in volume growth to specific treatments.

ST A TB FORESTS OF NEW SOUTII WALES RESEARCH PAPER NO. 24

SOUTIl-EAST REGROwrn FOREST GRowrn AND YIELD MODELUNG: DESIGN. METHODS AND PROGRESS 21

~

~~

~~ 9§ ~~ s'" .Z ~

h ~~ ;8~

~~ ~

~ ~

;1 ~~ o~ :;~

i~ ~

~~

Table S. Regression coefficients of combined variable underbark. stem volume equations (V = a + bIYH) for the six species and the validation statistics for the equations. MADVp denotes mean absolute deviation in volume prediction; RMSDVp denotes root mean squared deviation in volume prediction. Standard errors of the regression coefficients are in brackets.

e e

TREE RING ANALYSIS

Tree ring analysis has been a common method of reconstructing the growth history of temperate forest trees in the Northern Hemisphere, where deciduous species and conifers generally grow in a more predictable pattern in response to regular changes of season.. In Australia, the lack of contrast in seasons in general and the opportunistic growth habit of eucalypts (Jacobs 1955) in particular can make tree rings less clear and regular than many of their Northern Hemisphere counterparts. However, trees growing in places where there is a relatively strong seasonal contrast can produce more reliable rings than trees growing in areas with less obvious seasonal patterns. The potential of tree ring analysis in obtaining data on growth and history of past disturbances in Australia has been demonstrated for both eucalypts (Mazanec 1968, Readshaw and Mazanec 1969, Banks 1982, Keith 1982, Wllliams 1989, Home and Robinson 1990, Strasser 1992, Rayner 1992) and planted conifers (Nicholls and Wright 1976, Booth et al. 1988).

In the study area of this project. tree ring analysis has been used in two studies of past disturbances in the tableland forests, one in Glenbog State Forest (J.C.G. Banks pers. comm.) and the other in Coolangubra State Forest (Rymer 1991). The species involved in these two studies included E.fastigata,E. cype/locarpa, E. radiata, E. viminalis, E. sieberi and Acaciafloribunda. These studies indicated that tree ring analysis could be used to obtain growth data from the tableland forests where there is a strong seasonal contrast in climate. However, the assumption of annual ring production has to be verified before tree ring measurements can be formally adopted for obtaining growth data.

Because of the lack of growth data for the tableland forests, an efficient and reliable method of obtaining data covering the last 30-50 years' growth has to be found for the regrowth forests. Research in this area aims to - (1) examine tree ring formation of eucalypt species in the tableland forests so that the information stored in the rings can be interpreted; and (2) establish a routine method to obtain data from tree rings for growth and stem analysis.

1. METHODS

If climate data were available, the assumption of annual ring production would be verified through examining the correlation between climate changes and ring width patterns. However, the lack. of climate data from meteorological stations close enough to the study area prevented this approach. Another way of verifying the assumption is to examine the ring counts from sample trees in stands with known establishment, logging or fire history to see if the number of rings correlates closely to the estimated age of the stand. The latter approach was used to carry out a preliminary study to verify the assumption of annual ring production.

Eight regrowth stands with known logging history (plot 1-8 in Table 4) and fifteen trees from two staOOs in an E. mtens plantation in the tableland area were sampled. Disks were taken at breast height from trees in the eight regrowth stands. Two cores were taken from each tree at breast height in the two plantation plots. The disk and core samples were sanded. Then ring counts were performed by a person uninformed about the logging and establishment history of the stands.

For the eight sample stands the average number of rings from ten regrowth trees largest in DBH was used as the estimated stand age. For the two sample stands in the plantation. the average number of rings of all the sample trees was taken as the estimated stand age.

ST A TB FORESTS OF NEW SOUTIl WALES RESEARCH PAPER NO. 24

SOUTIl-EAST REGROwrn FOREST GRowrn AND YIBLD MODELLING: DESIGN. MBTIIODS AND PROGRESS 23

The estimated stand age was plotted against the known age of the stand. TIle results showed that the estimated stand age by ring counts agreed reasonably well with the stand age estimated from logging and establishment history (Fig. 6). 'The preliminary results suggested that tree ring analysis can be used to obtain growth data for tableland forests.

45

~ 40

£ C 35 £ '" :.a 30 co c:: ·Ch ~ 25 E o

..= 20 "0 ~ (;j

.§ 15 '" ~ ~

~ 10 "0 c::

'" ci5 5

o o 5 10 15 20 25 30 35 40 45

Stand age estimated from tree rings (years)

Figure 6. Stand age of ten stands estimated through known logging and fire history versus that estimated from tree rings. TIle lines above and below the central line indicate a difference of four years between the two stand age estimates.

SOUTII-EAST REGROwrn FOREST GROwrn AND YIELD 24 MODHU.ING: DESIGN, MEllIODS AND PROGRESS

STATB FORESTS OF NEW sourn WALES RESEARCH PAPER NO. 24

2. TREE RING MEASUREMENTS

Obtaining growth data from tree rings is much faster than the conventional methods of establishing permanent growth plots. However, reading tree rings is still very time consuming and resource demanding. Disk samples taken from the field are air dried in a shed. After drying, one side of each disk is sanded using a large belt sander, first with a rough belt to sand down the surface, and then a fine belt to achieve the required clarity for tree ring measurements. Since tree rings are often asymmetrical, to obtain an accurate estimate of ring width four measurements of ring width are taken from each ring at different points. For most disks four radii are drawn perpendicular to each other from pith to the cambium on the sanded surface of the disks. For disks with cracks or defects the radii may not be perpendicular to each other. Tree rings vary not only in clarity but also in the way they fonn. For measurement pwposes two categories of rings are defined - primary and secondary rings. Primary rings are complete rings with well defined latewood bands of a distinct colour that is daIXerthan the earlywood bands. Secondary rings are incomplete, tapering towards the inside or outside ring, often with diffused boundary between earlywood and latewood because latewood band is not well defined. Most rings are primary rings. However, since the width of each ring are to be measured on four radii, it must be absolutely certain that the four measurements are of the same ring. To acquire this certainty four measuring points are synchronised before tree ring measurement. Each primary ring along a chosen radius is marlced and numbered. Using this chosen radius as a principle radius, each ring is traced to the other measurement points so that they could be coded with the same ring number to that of the principle radius (plate 2). Finally tree ring measurements are taken undera stereo-microscope using a Henson Incremental Measuring Machine linked to a computer with the software 1RIMS which takes the digital signal and stores the data. In addition to ring width, sapwood thickness and the occurrence of gum veins are also recorded so that fire history and past variations in climate can be reconstructed by identifying and interpreting the marlcs left by wildfires and adverse growth conditions on the discs.

Plate 2. An example of a disk synchronised for tree ring measurements.

STATB FORESTS OFNIiW SOU1lI WALES RESEARCH PAPER NO. 24

SOU1lI-EAST REGROwm FOREST GRowm AND YIElD MODELLING: DESIGN, METHODS AND PROGRESS 25

- - - - - -----------------------

FOREST GROWTH AND STAND DYNAMICS

Sound growth models can only be constructed with a good understanding of the dynamics of forest stands to which the models are to be applied. Stand dynamics are concerned with the growth and death of individual trees in a stand through time. As described before, the irregular mixed-species regrowth stands consist of trees of different sizes and ages which may have originated from seeds, coppice or lingnotubers. The stand dynamics of such forests have been recognised as being complex (e.g. Florence 1970, Opie et al. 1978), but have remained unclear to a large extent This lack of understanding of stand dynamics constrains future approaches to growth and yield modelling of irregular mixed-species forests (Rayner and Turner 199Ob).

Competition between individual trees is one of the major factors influencing stand dynamics. The understanding ofhow competition processes operate have important ramifications for constructing growth models (West 1988a). A number of studies in even-aged eucalyptus stands have provided a clearer understanding of the competition processes and a better interpretation of the stand dynamics of these evenaged regrowth stands (Curtin 1964, Opie 1968, Curtin 197~, Ashton 1975, Ashton, 1976, Incolll979b, West 1981a, West 1981b, West 1984, West 1988b, Hamilton 1988, Home and Robinson 1990). Because trees essentially use the same resources, competition processes in even-aged monocultures and irregular mixed-species forests may be more or less the same in mechanism. TIle knowledge of stand dynamics of even-aged eucalypt forests can provide a starting basis for our understanding of the stand dynamics of irregular forests.

One important aspect of stand dynamics of irregular forests about which information is most urgently required by forest management is the effects of old-growth retention on the growth of regrowth trees. Neighbourhood analysis of growth and competition of individual trees have shown that old growth has a suppressive effect on regrowth in river red gum (Opie 1968), silver-top ash (Incoll 1979a), messmate (Kellas et al. 1982), karri (Rotherram 1983), alpine ash (Bowman and Kirkpatrick 1986a, 1986b, Ellis et al. 1987) and box-ironbark. forests (Kellas et al. 1987). Incoll (1979a) extrapolated his results to stand level by constructing hypothetical stands of 30-year-old regrowth E. sieberi with different densities of old growth under the assumption of the suppressive effects of old growth on regrowth being additive. He showed that the reduction in regrowth stand volume was more than 75 percent when the density of old growth, with an average crown diameter of 10.6 m, reached 20 trees/ha. Apart from his findings, little published data are available on the effects at stand level that old growth have on regrowth. Stand level characteristics of regrowth forests such as stand growth and stand structure as affected by old growth retention are largely unknown Because the suppressive effects of old growth on regrowth can vary with forest type and site quality (Opie 1968, Incolll979a, Rotherram 1983) and are likely to be inconsistent through time, quantitative relationships between old growth retention and the volume of regrowth stands over a wide range of sites and ages are urgently needed for management prescriptions of old growth retention.

One of the objectives of research work in the area of forest growth and stand dynamics is to quantify the effects of old growth retention on the growth of regrowth stands in the south-east tableland forests.

SOtrrH-HAST REGROwrn: FOREST GROwrn: AND YIELD 26 MODIiLLING: DESIGN. MB1HODS AND PROGRESS

STATE FORESTS OF NEW sourn W ALBS RESBARCH PAPHR NO. lA

1. PREliMINARY RESULTS

Up till now tree ring data have been obtained from 25 plots in regrowth forests of brown barrel (Forest Type 154). The estimated stand age of these stands ranged from 20 to 63 years (Table 6). Regrowth stand basal area of these stands varied between 14.35 and 54.29 ml. Regrowth stand volume was between 63.79 and 442.91 m'. The density of regrowth trees varied between 276 to 1847 trees/ha. whilst the density of oldgrowth in these stands was between 0 and 102 trees/ha.

The mean annual increment in regrowth stand volume decreased rapidly with the increase in old-growth density (Fig. 7) :

10 MAl = 3.391 - 0.014 N - 0.349ln age R2:0.79

where MAl is the mean annual increment in regrowth stand volume in m'ha-1yrl. N is the density of oldgrowth in treesIha and age is the estimated regrowth stand age. As expected regrowth stand volume also decreased with increasing old-growth density (Fig. 8):

In V = 3.391 -0.014 N + 0.651 In age R2:0.78

where V is regrowth stand volume in m'. Although these preliminary results provide the best available information at present for evaluating the potential loss in productivity of regrowth forests at varying intensities of old-growth retention. the confounding effect of old-growth density and site quality on regrowth productivity must be kept in mind when interpreting the results since field observations suggest that old-growth density is conelated with site quality. A further study will incorporate site quality into the equation so that stand yield tables can be developed for sites with different qualities and old-growth retention.

STA TB FORESTS OF NEW SOtml WALES RESEARCH PAPER NO. 24

SOtml-EAST REGROwm FOREST GRowm AND YIBLD MODEU.lNG: DESIGN. ME1ll0DS AND PROGRESS 27

~

~!S o~

h ~~ C5~ .z ~

~ ~I ~~ ~~ i~ ~

~ ;J

I ;~

§~ ~~ ~~

Table 6. The estimated stand age, stand basal area, stand volume and stand density of the 25 sample stands in regrowth brown barrel forests.

e e

12

10

i 8 g ~

6

~ 'S ..... 4 < ::g

2

o +-----~----~~----~----_+----~ o 20 40 60 80 100

Stand density of old-growth (treeslha)

Figure 7. Mean annual increment (MAl) in regrowth stand volume in relation to density of old-growth. The number nex~ to each line indicates stand age.

500

0

400

........ '" r::: 20 S ';; 300 a ::s

"0 40 ;;.-

"0 r:: .5 '"

..t: 200 ~. e bO Cl)

~

100 0-

o 20 25 30 35 40 45 50 55 60

Stand age (years)

Figure 8. Regrowth stand volume in relation to stand age and old-growth density (trees/ha) as indicated by the number on top of each line.


SOUTIl-EASTREGROWTH FOREST GROWTH AND YIELD MODEllING: DESIGN, ME1HODS AND PROGRESS 29

FOREST GROWTH MODELLING

This section aims to - (1) provide an outline of the approach being adopted for modelling the irregular mixed-species regrowth forests of south-east New South Wales; (2) give an account of the constraints of available information for modelling the growth of the irregular forests; (3) illustrate some desired"features of the growth models to be constructed; and (4) identify the component models which need to be constructed for growth prediction. Literature on modelling techniques are extensive and it is beyond tbe scope of this report to provide a review. However, it is expected that modelling techniques will have to be developed and tested when work on growth modelling starts after enough data are obtained.

The majority of growth and yield models used in forestry are empirically constructed and fall into three broad categories: (1) whole stand models, (2) diameter class models and (3) individual tree models (Davis and Johnson 1986, Daniels and Burkhart 1988). Whole stand models usually use stand level variables such as stand basal area, stand density, site index, stand age to predict future stand growth and diameter distribution. Diameter class models divide a stand into a number of diameter classes (cohorts) and use the

" classes as the basic units for modelling. Individual tree models can be either distance-indePendent or distance-dependent (Munro 1974). Distance-independent individual tree models do not use the spatial distribution of individual trees iil a stand, and therefore do not take competition between individual trees into account. Distance-dependentmodeis use the spatial distribution of trees and take the distance and size of neighbours into account. The three categories represent modelling approaches which vary in structural complexity and output detail.

Stand level approaches have seldom been adopted in modelling uneven-aged mixed species forests (e.g. Turner 1966, Moser and Hall 1969, Mendoza and GumpalI987). More often diameter class models and individual tree model are employed (e.g. Stage 1973, Ek. 1974, Leak and Graber 1976, Monserud and Ek. 1977, Leary 1979, Hann 1980, Kilgour 1982, Goodwin 1988, Vanc1ay 1989). Because individual tree models provide the greatest flexibility in predicting growth under a wide range of stand conditions and management alternatives, an individual tree approach is the most desirable for modelling the irregular mixed-species regrowth eucalypt forests of south-east New South Wales.

In choosing this approach, the constraints of available information for growth modelling of the irregular forests and some desired features of the growth models to be constructed were also considered. The constraints are:

1. There will be only a small database for a wide range of forest types and stand conditions.

2. Mortality data will not be available in the near future.

3. There is a very limited understanding of stand dynamics of irregular mixed-species forests generally.

4. Data obtained from current sampling may not reflect future management practice.

5. There is limited experience in modelling irregular eucalypt forests in Australia.

SOUTH-EAST REGROWTH FOREST GROWTH AND YIELD 30 MODELLING: DESIGN. METIlODS AND PROGRESS

STATE FORESTS OF NEW SOUTH WALES RESEARCH PAPER NO. 24

",

---------------------------------------------------------------- -

However, the growth models to be ~nstructed under these constraints should have some desired features, which include: .: .

1. Efficiency in ~ing the limited growth information.

2. Simplicity in structure so that models can be easily updated when new data are available.

3. Incorporation of site information (such as what is provided by the forest resources information database) to make future links with forestry GIS possible.

4. Capacity to predict forest growth under varying stand conditions in sufficient detail for management decisions.

5. Flexibility in simulating tree and stand growthundersilvicultural and management alternatives, even when data for which do not presently exist

6. Ooseness to biological growth processes of trees in the irregular forests.

To have such desired features, the growth models must consist of some essential component models:

(a) Diameter growth model

(b) Height growth model

(c) Mortality model

(d) Ingrowth model

(e) Site index model

(t) Bark thickness model

These component models will be developed individually and incorporated into a system with input and . output specifications. The design of the system needs to corisideruserrequirements and system efficiency.

STATE FORESTS OF NEW SOUTH WALES RESEARCH PAPER NO. 24

SOUTH-EASTREGROWTH FOREST GROWTH AND YIELD MODElLING: DESIGN, MEmODS AND PROGRESS 31

FUTURE.DIRECTIONS FOR THE PROJECT

The south-east regrowth forest growth and yield modelling project requires an appreciable level of mathematical, statistical skills as well as forest silvicultural expertise. Projects of this dimension and complexity usually involve long-term research efforts from well co-ordinated team w~rk. However, all interim results are of immediate use to forest management

The rate of output from the project depends largely on maintained level of resources, but planned progress can be expected in the follOwing areas:

1. A technical document and a simple application software will be written to assist management staff to extract required information from the forest resources and management database within a year's time. Also procedures to update the database will be established.

2. Field sampling of atleast 60 more temporary plots in the tableland area will be completed over the three years.

3. Between 60 and 100 permanent growthplots alongside the temporary plots will be established to obtain future growthandmortality dataforbuildingmortalityfunctions andmodel validation. This work. will require a year of field work..

4. A plan for establishing permanent growth plots in the coastal area will be prepared and implemented within a year.

5. Provisional volume equations developed for the important commercial' species in the tableland forests will be updated and finalised in two to three years' time. During this time volume equations to variable top limit diameters and taper equations will also be developed.

6. Tree ring measurements will continue even after field sampling stops. In two years' time the measurements will provide diameter growth data from at least 80 plots. Efforts will be made to establish a standard tree ring chronology and a method of crossdating in the tableland area.

7. The effects of overstorey on the growth of regrowth trees will be examined at both stand and individual tree level. The study will be completed within two years.

8. A system for growth prediction will be designed for the tableland regrowth forests within two years. The core of the growth models will be developed and validated within three years. Other component models will have to be developed until data are available.

9. A collaborative research on defects with Biology Section will be established and initial results will be obtained within a year. These results will provide a basis for merchantability modelling.

10. Collaborative research work. with researchers in other research organisations will be set up to investigate the possibility of incorporating mechanistic modelling approach in developing di.stan~-dependentindividualtreemodeltosimulatetreeandstandgrowthundersilvicultural and management alternatives for which data do not presently exist

SOUTIl-EAST REGROwn! FOREST GROwn! AND YIELD 32 MODEILlNG: DESIGN. METIlODS AND PROGRESS

STA TB FORESTS OF NEW SOUTIl WALES RESEARCH PAPER NO. 24

The progress in these areas will provide a good basis for yield modelling wheremerchantability and harvest models are the two essential components. , They will be developed after work on growth modelling is completed. It can be expected that, with appropriate inventory'data on residual stands after logging, the final product of this project will provide forest management the ability to determine the growth and yield and examine alternative management strategies of regrowth forests in south-east New South Wales.

ACKNOWLEDGEMENTS

, The study was funded in part by a grant from the Commonwealth Department of Primary Industry and Energy. I wish to thank Dr. J. Turner, Mrs. Marcia Lambert, Dr. Tun Shields, Mr. Vic Jurskis and other State Forests staffboth in Sydney and Eden for their support and encouragement. The field work is being co-ordinated by Mr. Vic Jurskis in Eden and I thank him for his continual support. A number of people have been involved in field work, mapping, s,ample processing, tree ring measurements, data entry, data maintenance and various other aspects of the project. For their contributio~ I thank Richard Colly, lan Ralph, Dalibor Hatich, Samantha ~andanayake, Ying Yan and Simon van HoIst. I also thank Joy Gardner for presenting the manuscript for production.

STATE FORFSTS OF NEW SOUTII WALES RESEARCH PAPER NO. 24

SOUTIl-EAST REGROWTH FORFST GROWTH AND YIELD MODEUJNG: DESIGN, METIlODS AND PROGRESS 33

REFERENCES

Anon. (1982). Eden native forest management plan. For. Comm. N.S.W.

Anon. (1989). Forest types in New South Wales. For. Comm. N.S.W.

Ashton, D. H. (1975). The root and shoot development of Eucalyptus regnans F. Muell. Aust. J. Bot. 23: 867-887.

Ashton, D. H. (1976). The development of even-aged stands of Eucalyptus regnans F. Muell. in central Victoria. Aust. J. Bot. 24: 397-414.

, Banks, J. C. o. (1982). The use of dendrochronology in the inteIpretation of the dynamics of the snow gum forest Ph.D Thesis. Australian National University, Canberra.

Bi, H. (1994a): Improving stem volume estimation of regrowth Eucalyptus /astigata with a lower stem fOIm quotient Aust. For. (In press). '

Bi, H. (1994b). Volume equations for six Eucalyptus species on the south-east tablelands of New South Wales. State Forests o/N.S.W. Res. Pap. No. 23. 11 pp.

Booth, T. H., Serra, R O. and Wells, K. F. (1988). Tree ring measurement and growth modelling. In Modelling Trees, Stands and Forests. (Eds. J. W.Leech, RE.McMurtrie, P. W. West, RD. Spencer andBM.Spencer). Proc. WOIXshop, University of Melbourne, August26-28, 1985. BulletinNo. 5, School of Forestry, University of Melbourne.

Bowman, D. M. J. S. and Kirkpatrick, J. B. (1986a). Establishment, suppression and growth of Eucalyptus delegatensis R T. Baker in multi aged forests. 11. Sampling growth and its environmental correlates. Aust. J. Bot. 34: 73-80.

Bowman, D. M. J. S. and Kirkpatrick, J. B. (1986b). Establishment, suppression and growth of Eucalyptus delegatenSis R T. Baker in multiaged forests. m. 'lntraspecific allelopathy, competition between adult and Juvenile for moisture and nutrients, and frost damage to seedlings. Aust. J. Bot. 34: 81-94.

Bridges, R O. (1983). Integrated logging and regeneration in the silvertop ash-stringybark forests of the Eden Region. For. Comm. N .s.W. Res. Paper No. 2. 27 pp.

Bridges, R O. and Dobbyns, O. R (1991). The dry sclerophyll silvertop ash-stringybm forests of southeastern N.S.W. In Forest Management in Australia. (Eds. F.H.McKinnell, E.RHopkins and J.E.D.Fox). Surrey Beatty and Sons, Chipping Norton.

Bruskin, S. and Home, R (1990). An analysis of growth data from eucalypt stands in the coastal forests of the Eden region. For. Comm. N.s.W. Tech. Pap. No. 53. 20 pp.

Canon, L. T. (1985). A history offorestry in AustraJia. Australian National University Press, Canberra. 355 pp.

SOUTII-EASf REGROWTII FOREST GROWTII AND YIELD 34 MODElLING: DESIGN, METIIODS AND'PROGRESS


Cortin, R. A. (1964) .. Stand density and·the relationship of crown width to diameter and herght in Eucalyptus obliqua. Aust. For. 28: 91-105.

Cortin, R. A. (1970a). ADalysis of growth in a mixed eucalypt foreSt. For. Comm. N.S.W. Res. Note No. 24. 31 pp.

Cortin, R. A. (1970b). Dynamics of tree and crown structure in Eucalyptus ob/iqua. For. Sci. 16: 321-328.

Daniels, R. F. and Burldlart, H. E. (1988). An integrated system of forest stand models. For. Ecol. & Alanage.23: 159-177.

Davis, L. S. and Johnson, K. N. (1986). Forest management (3rd Edition). McGraw-Hill, New York. 790 pp.

Deadman, M. W. (1992). Integrated information systems: Do we really have them? In Integrating Forest Information over Space and Time (Eds. G.Wood and B.Tumer). Proc. IUFRO Conference, 13-17 January, 1992, Canberra, Australia.

Ek, A. R. (1974). Nonlinear models for stand table projection in northern hardwood stands. Can. J. For. Res. 4: 23-27.

Ellis, RH., Ratkowsky, D. A., Mattay, J. P. and Rout, A. F. (1987). Growth of Eucaiyptlf.S delegatensis following partial harvesting of multi aged stands. Aust. For. 50: 95-105.

Florence, R G. (1970). Conservative silvicultural practice in irregular blackbutt forest Comm. For. Rev. 49: 86-93.

Florence R. G. (1987) Directions for silviculture and multiple use. In Prospects for Australian Hardwood Forests. (Eds. J.Dargavel and G.Sheldon). CRES Alonograph 19, Canberra. pp. 334.

Gagnon, P., Bedard, Y. and Edwards, G. (1992). Fundamentals Qf space and time and their integration into forestry geographic databases. In Integrating Forest Information over Space and Time. (Eds. G.Wood and B.Turner).Proc.IUFRO Conference, 13-17 January, 1992, Canberra, Australia.

Goodwin, A. N. (1988). Adaption of the 'STEMS' growth model to Eucalyptus forest. In Forest Growth Modelling and Prediction Vol. 1. Proceedings oflUFRO Donf.· Aug. 1987, MN. USDA For. Ser., Gen. Tech. Rep. NC-120., pp~ 76-85.

HamUton, F. (1988). Spatial pattern and stand development in even-aged stands with contagious spatial . patterns. In Modelling Trees, Stands and Forests. (Eds. J.W.Leech, R.E.McMurtrie, P.W.West;

RD. Spencer and B.M.Spencer). Proc. Workshop, University of Melbourne, August 26-28, 1985. Bulletin No. 5, School of Forestry, University of Melbourne.

Hann, D. W. (1980). Development and evaluation of an even- and uneven-aged ponderosa pine/Arizona fescue stand simulator. USDA For. Ser. Res. Pap. INT-267. 95 pp.

Home, R. and Robinson, G. (1990). Growth response following variable espacementof28-year-old alpine ash regeneration in New South Wales. Aust. For. 53: 280-289.

Incoll, W. D. (1979a). Effect of overwood trees on growth of young stands of Eucalyptus sieberi. Aust. For. 42: 110-116.


SOUTIl-EAST REGROWTII FOREST GROWTII AND YIELD MODELLING: DESIGN, METIlODS AND PROGRESS 35

Incoll, W. D. (1979b). Root system investigations in stands of Eucalyptus regnans. For. Comm. Vie. Forestry Technical Paper No. 23: 23-32.

Jacobs, M. R (1955). Growth habits of the eucalypts. Commonwealth Forestry & Timber Bureau, Canberra.

Keith, H. (1982). Investigation of the growth response of Eucalyptus paucijlora to clini.ate and fire, using dendrochronology, in the Brindabella range, A. C.T. B.Sc. (Hons.) Thesis, University of New South Wales, Sydney.

Kellas, 1. D., Owen, J. V. and Squire, R O. (1982). Response of Eucalyptus sideroxylon to release from competition in an irregular stand. For. Comm. Vic. Forestry Technical Paper No. 29: 33-36.

Kellas, J. D., Edgar, J. G. and Squire, R O. (1987). Response of messmate stringybarkregrowth to release in irregular stands of mixed eucalypts. Aust. For. 50: 253-259

Kelly, J. and Turner, J. (1978). Soil nutrients - vegetation relationships in the Eden area, N.S.W. I. Soil nutrient survey. Aust. For. 41: 127-134.

Kilgour, B. A. (1982). A yield model for uneven-aged messmate stands of South West,Victoria. M.Sc. (For.) Thesis, University of Melbourne. 127 pp.

Kilgour, B. A. and Rudra, A. B. (1988). The simulation of some alternative management strategies for the messmate forests of South West Victoria. I nModellingTrees, Stands and Forests. (Eds. J. W.Leech, RE.McMurtrie, P.W.West, RD. Spencer and B.M.Spencer). Proc. Workshop, University of Melbourne, August 26-28, 1985. Bulletin No. 5, School of Forestry, University of Melbourne.

Leak, W. B. and Graber, R. E. (1976). Seedling input, death and growth in uneven-aged northern hardwoods. Can. J. For. Res. 6: 368-374.

Leary, R A. (1979). Design. In A Generalized ' Forest Growth Projection System Applied to the Lake States Region. USDA For. Ser. Gen. Tech. Rep. NC-49. pp. 5-15.

Mazanec, Z. (1968). Influence of defoliation by the phasmatid Didymuria violescens on seasonal diameter growth and the pattern of growth rings in alpine ash. Aust. For. 32: 3-14. '

Mendoza, G. A. and Gumpal, E. C. (1987). Growth prediction of a selectively cut-over forest based on residual inventory. For. Ecol. & Manage. 20: 253-263.

Monserud, R. A. and Ek, A. R. (1977). Prediction ofunderstorey tree height growth in northern hardwood stands .. For. Sci. 23: 391-400.

Moser, J. W. and Hall, O. F. (1969). Deriving growth and yield functions for uneven-aged forest stands. For. Sci. 2: 183-188. '

Munro, D. D. (1974). Forest growth models - a prognosis. In Growth Models forTree and Stand Simulation (Ed. J.Fries). Proc.IUFRO Working Party S4.01-4 meetings, 1973. Dep. For. Yield Res. Royal College of Forestry (Stockholm). Res. Note No. 30.

Nicholls, J. W. P. and Wright, J. P. (1976). The effect of environmental factors on wood characteristics. 3. The influence of climate and site on young.Pinus radiata materiah Can. J. For. Res. 6: 113-121.

soum-EAST REGROWTII FOREST GROWTII AND YIELD 36 MODEILING: DESIGN, METIlODS AND PROGRESS

STATE FORESTS OF NEW soum WALES RESEARCH PAPER NO. 24

Opie, J. E. (1968). Predictability of individual tree growth using various definitions of competing basal area For. Sci. 14: 314-323.

Opie, J. E., Curtin, R. A and Incoll, W. D. (1978). Stand management In Eucalypts for Wood Production. (Eds. W.E.Hillis and AG.Brown). CSIRO, Australia. 434 pp.

Phillis, K. J. (1971). Management inventory of irregular broad-leaved forests of Australia. Ph.DThesis, Australian National University.

Rayner, M. E. (1992). Application of Dendrochronology, stem analysis and inventory data in the estimation of tree and stand ages inkarri forest Department of Conservation andLandManagement, W.A. Technical Report No. 27.

Rayner, M. E. and Turner, B. J. (1990a).·Growth and yield modelling of Australian eucalypt forests I. Historical development Aust. For. 53: 224-237.

Rayner, M. E. and Turner, B. J. (1990b). Growth and yield modelling of Australian eucalypt forests n. Future trends. Aust. For. 53: 224-237. .

Readshaw, J. L. and Mazanec, Z. (1969). Use of growth rings to determine past phasmatid defoliations of alpine ash forests. Aust. For. 33: 29-36.

Richards,B.N., Bridges,R G., Curtin,R A.,Nix,H.A,Shepherd,K.R and Turner, J.(l990). Biological conservation of the south-east forests. Report of the Joint Scientific Committee. Commonwealth of Australia.

Rotheram, I. (1983). Suppression of growth of surrounding regeneration by veteran trees of karri (Eucalyptus diversicolor). Aust. For. 46: 8-13.

Rymer, A (1991). Investigations into the environmental history of the Waratah Creek Catchment area, Coolangubra State Forest, South-Eastern New South Wales. B.Sc. (Hons) thesis. Australian National University, Canberra.

Stage, A R. (1973). Prognosis model for stand development USDA For. Ser. Res. Pap. INT-137.

Strasser, M. (1992). Computer modelling of montane eucalypt forests. Ph.D Thesis, Australian National University, Canberra.

Synnott, T. J. (1979) .. A manual of pelIDanent plot procedures for tropical rainforests. Tropical Forestry Papers 14. Commonwealth Forestry Institute, Oxford.

Tennent, R B. (1988). A New Zealand data collection procedure for growth modelling. In Modelling Trees, Stands and Forests. (Eds. J.W.Leech, RE.McMurtrie, P.W.West, RD.Spencer and BM.Spencer). Proc. Workshop, University of Melbourne, August 26-28, 1985. Bulletin No. 5, School of Forestry, University of Melbourne.

Turner, B. J. (1966). An investigation into the growth of irregular eucalypt stands in New South Wales. For. Comm. N.s.W. Res. Note No. lB. 55 pp.

Vanclay, J. K. (1989). A growth model for north Queensland rainforests. For. Ecol. & Manage. 27: 245-271.

STATE FORESTS OF NEW SOUTII WALES . RESEARCH PAPER NO. 24

SOUTII-EASf REGROWTH FOREST GROWTH AND YIELD MODEllING: DESIGN, METIlODS AND PROGRESS 37

Vanclay, J. K. (1991). Review: Data requirements for developing growth models for tropical forests. Comm. For. Rev. 70: 248-27l.

West, P. W. (1981a). Simulation of diameter growth and mortality in regrowth eucalypt forest of southern Tasmania For. Sci. 27: 603-616.

. .

West, P. W. (1981b). Comparative growth rates of several eucalypts in mixed-species stands in southern Tasmania N. Z. J. For. Sci. 11: 45-52.

West, P. W. (1984). Inter-tree competition and small scale pattern in monocu1ture of Eucalyptus obliqua L 'Herit Aust. J. Ecol. 9: 405-411.

West, P. W. (1988a). Competition processes and modelling growth of individual trees inmonoculture. In Modelling Trees, Stands and Forests. (Eds. J.W.Leech, RE.McMurtrie, P.W.West, RD. Spencer andBM.Spencer). Proc. Worlcshop, University of Melbourne, August 26-28, 1985. BulletinNo. 5, School of Forestry, University of Melboume.

West, P.W. (1988b). Modelling diameter growth of individual trees in thinned stands of Eucalyptus obliqua L'Herit In Modelling Trees, Stands and Forests. (Eds. J. W.Leech, R.E.McMurtrie, P.W.West, RD. Spencer and BM.Spencer). Proc. Worlcshop, University of Melboume, August 26-28,1985. Bulletin No. 5, School of Forestry, University ofMelboume.

Williams, J. E. (1989). Ecology of sub-alpine eucalypt forests in relation to environmental gradients. Ph.D Thesis, Australian National University, Canberra.

SOUTIl-EAST REGROWTH FOREST GROWTH AND YIELD 38 MODEUlNG: DESIGN, METHODS AND PROGRESS

STA TB FORESTS OF NEW SOUTII WALES RESEARCH PAPER NO. 24

The Research Division produces publications in Technical Paper, Research Paper and Forest Resources Series.

Copies and further information are available from:

Publications Officer Research Division State Forests of NSW PO Box lOO Beecroft NSW 2119

TeI. (02) 872·0111 Fax. (02) 871 6941

'-

South-East Regrowth Forest Growth & Yield Modelling - NSW

Documents

Transcript of South-East Regrowth Forest Growth & Yield Modelling - NSW