eden forest resource and management system report · eden forest resource and management system...

Eden Forest Resource And ManagementSystem Report

A report undertaken for the NSW CRA/RFA Steering Committee12/5/98

EDEN FOREST RESOURCEAND MANAGEMENTSYSTEM REPORT

STATE FORESTS OF NSW &BUREAU OF RESOURCE SCIENCES

A report undertaken for the NSW CRA/RFA SteeringCommittee

Project numbers: NE 09/FRA, NE 10/FRA, NE 11/FRA,NE 12/FRA, NE 13/FRA, NE 14/FRA, NE 15/FRA, NE 20/FRA,

NE 22/FRA

12/5/98

Report Status

This report has been prepared as a working paper for the NSW CRA/RFA Steering Committee under thedirection of the FRAMES Technical Committee. It is recognised that it may contain errors that require

correction but it is released to be consistent with the principle that information related to the comprehensiveregional assessment process in New South Wales will be made publicly available.

For more information and forinformation on access to data contactthe:Resource and Conservation Division, Departmentof Urban Affairs and Planning

GPO Box 3927SYDNEY NSW 2001

Phone: (02) 9228 3166Fax: (02) 9228 4967

Forests Taskforce, Department of Prime Ministerand Cabinet

3-5 National CircuitBARTON ACT 2600

Phone: 1800 650 983Fax: (02) 6271 5511

© Crown copyright May 1998

This project has been jointly funded by the New SouthWales and Commonwealth Governments. The workundertaken within this project has been managed bythe joint NSW / Commonwealth CRA/RFA SteeringCommittee which includes representatives from theNSW and Commonwealth Governments andstakeholder groups.

The project has been overseen and the methodologyhas been developed through the FRAMES TechnicalCommittee which includes representatives from theNSW and Commonwealth Governments andstakeholder groups.

DisclaimerWhile every reasonable effort has been made toensure that this document is correct at the time ofprinting, the State of New South Wales, its agents andemployees, and the Commonwealth of Australia, itsagents and employees, do not assume anyresponsibility and shall have no liability, consequentialor otherwise, of any kind, arising from the use of orreliance on any of the information contained in thisdocument.

CONTENTS

Executive summary

1. INTRODUCTION 11.1 Purpose 1

1.2 FRAMES and the Eden CRA 1

1.3 Objectives 1

1.4 Scope 2

1.5 Description of the Eden CRA Region 21.5.1 Native forests 21.5.2 Plantations 3

1.6 Methods 3

2. PRE CRA YIELD ASSESSMENTPROCEDURES 5

2.1 Existing data 5

2.2 Previous studies 52.2.1 Summary of EIS outcomes 52.2.2 Summary of DFA/IFA outcomes 52.3 Conclusion 6

3. DATA COLLECTION 73.1 Review of resource data 73.1.1 Project objectives 73.1.2 Methods 73.1.3 Results 8

3.2 Retained tree inventory 11

3.2.1 Project objectives 113.2.2 Methods 113.2.3 Results 11

3.3 Development of net harvestable area 123.3.1 Project objectives 123.3.2 Methods 123.3.3 Identification of exclusions from harvesting 123.3.4 Derivation of erosion hazard categoriesand stream buffers 133.3.5 Results 14

3.4 Site productivity index 143.4.1 Project objectives 143.4.2 Methods 14

3.4.3 Results 15

3.5 Sawlog species mix 153.5.1 Project objectives 153.5.2 Methods 153.5.3 Results 15

4. MODEL DEVELOPMENT ANDAPPLICATION 17

4.1 Growth and yield model 174.1.1 Project objectives 174.1.2 Methods and results 174.1.3 Review of the STANDSIM model 174.1.4 Conclusion 21

4.2 Modifier models 224.2.1 Project objectives 224.2.2 Methods 224.2.3 Results 22

5. DATA COMPILATION ANDANALYSIS 23

5.1 Compile and document database 235.1.1 Project objectives 235.1.2 Methods 235.1.3 Results 23

5.2 Yield estimator and scheduling toolbox 245.2.1 Objectives 245.2.2 Methods 245.2.3 Results 24

6. OUTPUTS 276.1 Key outputs of Eden FRAMES 276.1.1 Projected volumes 276.1.2 Integration and development of scenarios for Eden 276.2 After the Regional Forest Agreement 28

FIGURES

1a Eden FRAMES Methodology 4

TABLES

1a Softwood Resource Volumes from Plantationsin the Eden CRA Region 3

3a Exclusions in the Eden CRA Region 13

4a Stratification of Regrowth Areas 21

4b STANDSIM Predicted Yield Tables 21

6a Annual Wood Resource Volumes from NativeState Forests in the Eden CRA Region 28

6b Total Predicted Harvestable Volume (m3) fromthe Multi-aged Forest by Species and DiameterClass 28

REFERENCES 29

APPENDICES

1 IFA Wood Resource Study Summary

2 Validation of Eden Wood Resources Data

3 IFA Retained Tree Report

4 Growth of Sawlogs in Multi-aged Forest atEden

5 Modification of the STANDSIM Model for theEden Management Area

6 Yield Equations for Regrowth ForestsRegenerated from Fire on the Southeast Coastof NSW

7 Eden FRAMES User Guide

8 FRAMES Submission to CRA/RFA SteeringCommittee

EXECUTIVE SUMMARY

This report has been prepared for the jointCommonwealth/State Steering Committee whichoversees the comprehensive regional assessmentsof forests in New South Wales.

The comprehensive regional assessments (CRAs)provide the scientific basis on which the State andCommonwealth governments will sign regionalforest agreements (RFAs) for the major forests ofNew South Wales. These agreements willdetermine the future of the State’s forests,providing a balance between conservation andecologically sustainable use of forest resources.

This report details the objectives, methods andoutcomes of the Forest Resource andManagement System (FRAMES) projects thatwere undertaken as part of the Eden CRA.

Various projects were carried out to determine thecurrently available wood resources in the forestsof the Eden CRA Region and the expected futureyields which can be sustained over time.

The timber resource in the Eden CRA Region canbe categorised into two main components, themulti-aged forest, and regrowth forest.

The multi-aged forest resource is expected toprovide the majority of sawlog supply in the EdenCRA Region over the next 20 years until aboutthe year 2015. From around 2016 onwards, it isanticipated that sufficient area of the regrowthresource will have reached a size where it alonecan supply the pulpwood and sawlogcommitments previously obtained from the multi-aged forest.

The timber currently available for harvesting isshown in Table Ea. The volumes have beencalculated using the current land tenure andapplying current management prescriptions but donot take account of fauna moratorium areas whichwill be considered as part of the RFA.

There is potential impact from wildfire which wasnot taken into account in tables Ea and Eb andapplying current management prescriptions. TableEb shows total predicted harvestable volume forthe multi-aged forest by species group anddiameter class. This data has been extracted fromthe Eden FRAMES database.

The FRAMES Technical Committee recognisedthat the certainty of these estimates varies withtime period. The relevant time periods are:

n from 1997 to 2019 estimated yields are drawnprimarily from the multi-aged forest;

n from 2020 to 2040 estimated yields are largelydrawn from fire regrowth and thinning oflogging regrowth.

There are many uncertainties in estimating timberyields from forests with high inherent variability,such as the Eden forests. In considering thescenarios, FRAMES estimates of mean sawlogyields per hectare from the multi-aged forest forthe period 1997-2020 have confidence limits of+/- 30%. Although some of this variabilitycancels out as compartments are aggregated toproduce an annual yield, that annual yield maystill have reasonably wide confidence limits.There is some uncertainty regarding areas andgrowth rates for regrowth stands harvested up to2040 and no confidence limits can be assigned.

Beyond 2040 estimated yields are drawnprimarily from logging regrowth which has notbeen well sampled. Estimates are indicative only.

TABLE Ea ANNUAL WOOD RESOURCEVOLUMES FROM NATIVE STATE FORESTS IN

THE EDEN CRA REGION

Period Current land tenure includingany areas from which harvestinghas been deferred

Sawlog1 (m3)per annum

Pulp1 (t) perannum

1997-2019 28 300 403 000 2020-2040 28 300 444 000 2040+ >30 000 ~500 000 1 From the multi-aged forest and regrowth, includingthinning

TABLE Eb: TOTAL PREDICTEDHARVESTABLE VOLUME (m3 ) FROM THEMULTI-AGED FOREST BY SPECIES AND

DIAMETER CLASS

Sawlog diameter class (centrediameter under bark)

Species Group <40 cm (m3) ≥40 cm (m3) Silvertop Ash 10 000 134 000 Stringybarks 15 000 149 000 Messmate 3 000 56 000 Spotted Gum 3 000 12 000 Monkey Gum 5 000 68 000 Tablelands spp1 3 000 64 000 Specials2 1 000 3 000 Other 3 000 26 000 1 Mountain gum, brown barrel, shining gum, mannagum 2 Grey box, grey gum, ironbark, white box

Based on this information, allowable cutcalculations can be estimated for the currentplanning horizon (Forest Essentials Pty Ltd,1997b).

Eden FRAMES identified a need for moreinventory data for future evaluations of theregrowth resource.

1

1. INTRODUCTION

1.1 PURPOSE

To develop an RFA for Eden it is necessary tounderstand the currently available wood resourcesin the Eden CRA Region’s forests, and theexpected future yields which can be sustainedover time.

To this end, a Technical Committee wasestablished, with representatives from State andCommonwealth Governments and variousstakeholder groups, to develop methods andoversee assessment work. The Forest Resourceand Management System (FRAMES) TechnicalCommittee in its Technical Framework identifiedits aims as:

n determining ecologically sustainable woodflows and expected sawlog sizes and qualitiesfor a range of management options and avarying resource base for use in RFAintegration;

n providing reliable ecologically sustainableyield figures as a basis for an RFA between theCommonwealth and NSW;

n providing a basis for ongoing ecologicallysustainable management of wood flows byState Forests of NSW;

n providing information on resourcecharacteristics for use in long term planning bywood-based industries in NSW; and

n providing information on and validation of themodelled effects of environmental andsilvicultural options for use in developingproposals for ecologically sustainable forestmanagement.

The ecological sustainability of wood flows wasprimarily addressed by the ESFM (EcologicallySustainable Forest Management) TechnicalCommittee.

1.2 FRAMES AND THE EDEN CRA

A detailed Forest Resource and ManagementSystem (including detailed inventory) is currentlybeing applied to State forest areas which are

included in the other New South WalesComprehensive Regional Assessments (CRAs).

However, in setting out to establish wood flowsover time for the Eden CRA Region, theFRAMES Technical Committee decided inJanuary 1997 that there would not be sufficienttime to complete a full strategic inventory forEden, as is being carried out for the rest of theState. Separate Eden-specific projects weretherefore developed, with the expectation that arevised and externally validated wood resourcesdatabase, plus improved associated models, wouldprovide a reasonable foundation for estimates ofharvestable log volumes and sustained yield. Thiscombined assessment is referred to as EdenFRAMES.

The work carried out for Eden has included:

n refinement and independent review of existinginformation on wood resources;

n revised estimates of forest growth rates toenable the prediction of future wood resources;

n revised calculations of potential future sawlogand pulpwood yields from the Eden region.

This report details the methods used and theresults of the Eden FRAMES projects.

1.3 OBJECTIVES

FRAMES for Eden was designed to:

n produce estimates of harvestable volumes ofthe older multi-aged forest by quality, size andspecies classes as well as overall value ($);

n provide models for growth and yield of theregrowth resource;

n combine multi-aged forest volumes andregrowth volumes to estimate sustained yieldsof timber;

n provide data and methods for calculatingsustained yield for Commonwealthaccreditation;

n identify wood supply potential by area;

n provide information (harvestable woodvolume, sustained yield, and information on

Eden FRAMES report 12 May1998

2

size, species and quality mix) for assessmentof industry development potential;

n provide capacity to simulate some of theeffects on wood flows of harvesting intensity,so that this information may be used withother models to assess the consequent impacton other environmental values;

n derive an estimate of gross and net harvestablearea for each management unit;

n allow resource information to be displayedspatially;

n provide a scheduling tool for combining andaveraging the flow of wood from specificharvesting operations; and

n provide resource information as inputs into theForest Resource Use Model (FORUM) toenable modelling of economic and socialimpacts.

1.4 SCOPE

Due to time limitations and the features ofexisting data, FRAMES for Eden was designed tobuild upon the forest resource information systemin use in the Eden CRA Region.

Existing data were used to the fullest extentpossible. A suite of projects was developedspecifically to match the data and timeavailability, and requirements for the region.These projects were designed to provide a revisedand externally validated database and a predictivemodel for use in estimating potential yields underalternative land use scenarios.

1.5 DESCRIPTION OF THE EDENCRA REGION

The Eden CRA Region adopts the same boundaryas State Forests of NSW (SFNSW) EdenManagement Area. Its area totals about 800 000hectares and extends, in broad terms, fromBermagui and Nimmitabel in the north toDelegate and Cape Howe in the south.

1.5.1 Native forests

The timber resource in the Eden CRA Region canbe categorised into two main components:

n the multi-aged forest; and

n the essentially even-aged regrowth forest.

The multi-aged forest resource includes unloggedforests and forest areas which have beenselectively harvested in the past and represent acurrently available bank of wood resource.

The regrowth forest in the Eden CRA Regionincludes regrowth originating from fires (wherethe fires have been of sufficient intensity to killmost overstorey trees) as well as from integratedharvesting. The regrowth resource will be animportant future source of wood for the region’sindustries. There have been four major fireseasons affecting the resource, 1952, 1968, 1972and 1980. Since 1967, approximately 170 000hectares have been burnt.

Even where trees have not been killed, the longhistory of wildfire in the region has affected logquality in the multi-aged forests. Due to theyounger average age and reduced exposure todamaging wildfire, the regrowth resource isanticipated to be of considerably higher qualitythan the multi-aged forest sawlogs. TheSTANDSIM growth model (refer Section 4.1)predicts that the regrowth resource will producehigher volumes per hectare of sawlogs than themulti-aged forest.

All data was collected for area units known ascoupes which are the area basis for planningharvesting operations. Each coupe has an area ofapproximately 50 hectares.

The alternate coupe harvesting system is appliedin Eden. This is a planning strategy to spread theenvironmental impacts of logging over space andtime by dividing a compartment into smaller areas(coupes) and logging the coupes in an alternatingpattern (SFNSW, 1994). Logging of alternatecoupes is intended to be separated in time as faras practicable and is generally around 15 years.

Younger regrowth areas result from harvestingoperations conducted from 1970 onwards. About120 000 hectares have been harvested byintegrated harvesting operations since 1970.

The multi-aged forest resource is expected toprovide the majority of sawlog supply in the EdenCRA Region over the next 20 years until aboutthe year 2015. From around 2016 onwards, it isanticipated that sufficient area of the regrowthresource will have reached a size where it alonecan supply the pulpwood and sawlogcommitments previously obtained from the multi-aged forest.

12 May 1998 Eden FRAMES report

3

There is also estimated to be around 90 000hectares of private forest in the CRA Region.Management of this resource and its potentialyields are not clear.

1.5.2 Plantations

SFNSW plantations of Pinus radiata in the EdenCRA Region have a net area of 31 000 hectares.Most are concentrated in the Bombala area,between Craigie, Rockton, Towamba andCathcart, with smaller plantations in the Glenbogand Glen Allen areas. Most of the resource is lessthan 25 years old, with some areas reachingsecond and third thinning age. Significantincreases in harvest volume are planned over thenext decade (see Table 1a).

The area of privately owned pine plantations inthe Eden CRA Region is around 2 000 to 3 000hectares.

About 1 100 hectares of eucalypt plantations havebeen established in the area. Most are less than 10years old and principally targeted at the pulpwoodmarket. These include 800 hectares ofpredominantly Eucalyptus nitens plantation andabout 300 hectares of smaller farm woodlotsrecently planted as joint ventures between farmersand SFNSW, comprising E. nitens, E. botryoides,E. saligna and E. agglomerata. Yields areunknown.

TABLE 1A SOFTWOOD RESOURCEVOLUMES FROM PLANTATIONS IN THE

EDEN CRA REGION

Period Sawlogs(m3) perannum

Pulpwoodroundwood

(t) perannum

Sawmillresidues(t) perannum

Current Harvest 50 000 35 000 -

Potential 2002 a, b, c 290 000 200 000 87 000

Potential 2010 a, b, c, 380 000 270 000 114 000

Potential 2020 a, b, c 550 000 270 000 165 000

a) Potential 2002 level of yield based on currentplantation resource; Potential 2010 and Potential 2020levels of yield assume an expansion of the plantationestate by 10,000 hectares and 20,000 hectaresrespectively. b) Sawlog availability predicated on the sale of allpulpwood. c) Potential yields represent sustained annual levelsassuming an even distribution of age classes, which isnot currently the case.

Source: SFNSW, BRS (Social and Economic TechnicalCommittee).

1.6 METHODS

The FRAMES Technical Committee developedmethods and oversaw projects. Figure 1a showsthe methodology used for Eden FRAMES. Theassessment consisted of the following phases:

n pre CRA yield assessment procedures;

n data collection;

n model development and application;

n data compilation and analysis; and

n outputs.

The following chapters outline what was involvedin each of these phases.

Eden FRAMES report 12 May1998

4

FIGURE 1a EDEN FRAMES METHODOLOGY

NE09/FR A - Review of resource data for

multi-aged forest and regrow th in the Eden

CRA Region

N E13/FRA - D evelopment of a site productivity index

N E11/FRA - Retained tree inventory

N E10/FRA -Compilation &

documentation ofthe database

N E14/FRA -D evelopment

of a growth andyield

model for regrowth

N E22/FRA -D evelopment of a yield estimating

andscheduling

toolbox

N E15/FRA -Development ofmodifier models

N E12/FRA - D evelopment of net harvestable area

Pre CRA yield assessmentprocedures:- IA P W ood Resources Studyfor Eden (and past studies),strengths and w eaknesses- M ethodology developmentfor Eden

Data collection

M odeldevelopment

& application

Pre CRA yieldassessmentprocedures O utputs

Run toolbox to determine:- varying resource supply areas;- varying area-based prescriptions;- varying multi-aged forest harvesting intensities

NE20/FRA - Sawlog species mix

Data compilation& analysis

5

2. PRE CRA YIELDASSESSMENTPROCEDURES

2.1 EXISTING DATA

In the course of its normal forest managementactivities, SFNSW has been carrying out projectsto improve its estimates of the timber volumes perhectare which can be expected from the multi-aged and regrowth components of the resource.

Existing available SFNSW data includes:

n inventory data (broad based and compartmentlevel); and

n actual historical yields for harvested coupesdating back to 1970.

A major inventory of the resource was undertakenin 1987. This 1987 timber inventory was builtupon the recognition that the managementpractice of alternate coupe harvesting in the EdenCRA Region allows for actual volumes per areaand species and size mix to be cross-referenced toadjoining unlogged coupes. As a result, futurewood volumes can be estimated in adjacentunlogged coupes, recognising that adjustmentsneed to be made for differences between currentand historical net harvestable area and treeretention prescriptions.

2.2 PREVIOUS STUDIES

The results of a 1987 timber inventory formed thebasis for the 1994 Eden Environmental ImpactStatement (EIS) predictions and the 1996 InterimForest Assessment (IFA) Wood Resources Study(WRS).

The following sections present the outcomes ofboth the EIS and IFA.

2.2.1 Summary of EIS Outcomes

The outcome of the Eden EIS was that woodresource volumes were set at 59,000m3/yr ofsawlog and 504,000t/yr of pulpwood. It was

determined that actual wood resource volumeswere to be reviewed in five years.

2.2.2 Summary of DFA/IFA Outcomes

The Deferred Forest Agreement (DFA) betweenthe Commonwealth and New South WalesGovernments (signed January 1996) provided thatcommercial timber harvesting operations werepermitted to take place only in compartmentsidentified in the DFA. Approximately 340compartments in the Eden CRA Region werelisted under the DFA.

The 1996 IFA was concluded subsequent to theDFA. Some of the key conservation initiatives ofthe IFA that relate to the Eden CRA Region were:

n the reservation of approximately 45,000hectares of new national park in the southeastforests;

n approximately 42,000 hectares were includedin the Interim Deferred Forest Area (IDFA).Of this, 36,000 hectares (approximately) ofState forests in the Eden section of the IDFAwas to be considered during the CRA foraddition to the South East Forests NationalPark. In addition, approximately 64,000hectares of State forests were included in aCommonwealth DFA, pending the CRA/RFAoutcome.

Some of the key industry initiatives in the IFArelating to the Eden CRA Region included:

n quota grade sawlog supply will be 26,000m3

until the completion of the CRA, and thenmaintained at a minimum of 20,000m3 per yearfor the length of the agreement negotiated forrevised sawlog arrangements for the EdenCRA Region;

n timber commitments to be supplied fromoutside IDFA and Commonwealth DFA

Eden FRAMES report 12 May 1998

6

compartments in the Eden CRA Region untilthe RFA is completed.

A discussion of the IFA WRS study is providedin Appendix 1. For further information about theIFA, reference should be made to the DraftInterim Forestry Assessment Report (RACAC,June 1996).

2.3 CONCLUSION

The pre-CRA methodology had a number oflimitations as identified in the following chapterand Appendix 1.

The CRA assessment of Eden timber resourcesdetailed in this report comprised a review ofexisting data and studies, significantly refiningthe methods used in the IFA Wood ResourcesStudy to essentially rebuild the multi-aged forestdatabase and construct a regrowth database andscheduling tool.

7

3. DATA COLLECTION

3.1 REVIEW OF RESOURCE DATA

Review of Resource Data for Multi-agedForest and Regrowth in the Eden CRARegion

Project No. NE/09 FRA

3.1.1 Project objectives

The objective of this project was to quantify thereliability and limitations of work which had beenundertaken to predict the resource available perhectare in the current multi-aged and regrowthforests in the Eden CRA Region. The projectexpanded to include improved methods fordefining base resource data for both multi-agedforest and regrowth.

3.1.2 Methods

The resource data for multi-aged and regrowthforests was independently reviewed by ForestEssentials Pty Ltd and their report Validation ofEden Wood Resource Data (1997a) is included inAppendix 2. References from that report areshown in italics in Sections 3.1.2 and 3.1.3.

Multi-aged forest

The project involved:

n documentation, review and refinement of themethodology for predicting harvestablevolumes; and

n a statistical analysis of the relationshipbetween predicted yields (from the 1987inventory) and actual yields.

INVENTORY OF THE MULTI-AGEDFOREST (MAF)

A significant problem with evaluating themethodology was the lack of documentation ontechniques at the time when the inventory wasplanned and executed. The following descriptionof methods was obtained by discussion with staffat Eden.

Compartment boundaries for the EdenManagement Area (EMA) were delineated in theearly 1970s, and coupe boundaries weredetermined for all productive forest in 1977 Theywere revised to comply with PreferredManagement Priorities (PMPs) in 1981.

The assessment of the MAF currently used for theArea was carried out in 1987. At this time, theEMA included three Districts. This division ofresponsibility resulted in differences in approachand contributed to a lack of adequatedocumentation.

The merchantable volumes of both sawlog andpulpwood available on all uncut coupes wereestimated in the 1987 assessment by comparisonwith actual yields from the neighboring utilisedcoupe. A cruise of the uncut coupe was carriedout by an experienced supervisor to confirm ormodify the estimate that had been based onutilisation yield. Where yield information from autilised coupe was not available to assistestimation, an estimate was made from cruisingalone. In work since, the estimated yields forunpaired coupes have been amended asutilisation data became available.

In operations up to about 1985, selective loggingwas carried out in some (a total of 198 out of the3076 coupes cut at that time) adjacent uncutcoupes to supplement yields from integratedlogging. Adjustment for this reduction in yieldwas made during the cruise of the uncut coupe.

The information obtained was stored in a DbaseIII+ coded database, since updated to the currentAccess coded ’Cricket’ database.

Regrowth forest

The limited time available for the completion ofthe Eden CRA meant that existing data was used.Inventory of the regrowth by SFNSW was alreadyunder way and this was used as the basis for thisproject. This data was supplemented bymeasurement of additional plots in strata withlimited information.


8

INVENTORY OF THE REGROWTH FOREST

A major problem with evaluating the regrowthinventory was the lack of documentation onmethodology for both the API stratification andthe methods used in plot measurement. An outlineof the proposed work dated August 1993 wasexamined. This provided general instructionsonly; it did not, for example, describe proceduresfor selecting which stands of regrowth should besampled and how plots should be located in thestands. Guidance as to methods was alsoobtained from the State Forests Field MethodsManual. While this document is excellent forbasic methodology, it does not provide sufficientdetail for specific tasks such as this inventory.The following description of methods wasobtained by discussion with staff at Eden.

Objectives

As given in the August 1993 outline:

n To measure the volume of regrowth availablefor thinning, and the volumes of sawlog andpotential sawlog available in each stratum, inorder to determine the priority of regrowthstands for thinning.

n To determine whether enough thinningresource exists to sustain another treeharvester in Southern region.

An unstated objective that has become moreimportant since 1993 and is implied by projectswithin the current RFA process is the need forinformation to calculate a reliable sustainableyield for the resource as a whole.

Aerial photo interpretation (API) was used tolocate and delineate boundaries of regrowth.Judging from the strata chosen (categories ofnumber of stems/ha, stand mean diameter, maturetree density and shrub density), it was expectedthat some stand data could also be obtained fromAPI. Plot measurements were carried out bothbefore and after the strata were delineated.

Selecting stands for sampling

Colour aerial photographs (1:15,000 scale) takenin 1993, were available for the southern part ofthe EMA. Two different techniques were used toselect stands for plot sampling; one techniqueused before the API stratification of the resourcewas complete and a different techniquesubsequent to that.

The first technique concentrated sampling effortinto regrowth known to be the oldest available inthe eastern part of the EMA, because, as specifiedin the 1994 EIS, these would be the first stands tobe thinned. Maps showing the regrowth in thenominated compartments were prepared from the1993 airphotos and the sampling crew wasdirected to sample these stands.

When the API stratification became available, theplots measured to date were allocated to the APIstrata and sampling effort was then directed tostrata that had not been sampled or where thecoefficient of variation of mean total standvolume was greater than 30% of the mean.

Refer to full report in Appendix 2 for discussionon plot measurement procedure and analysis.

Over 1,000 field plots in fire and harvestingregrowth were measured for basal area, stockingrate and height, and tree diameter for a proportionof plots. For those plots, where bole height (notdominant height) for typical codominant treesnear the sample point was estimated, meandominant heights were estimated using arelationship between tree diameter and tree height(Bi unpublished).

The extent of fire regrowth was derived fromaerial photograph interpretation (API) ofcompartments and from compartment historyrecords elsewhere. The area of harvestingregrowth was derived from compartment historyrecords.

3.1.3 Results

Multi-aged forest

The method has the advantages that it is cheap,requiring minimal field work compared with atraditional plot-based inventory, and is directlyassociated with the real world, throughcomparison with actual utilisation of theneighboring coupe. The disadvantages are thatthe method assumes the same products will beremoved, using the same technology and in thesame conditions of market demand, as the firstcoupe, all conditions that may have considerableinfluence on the yields obtained. The considerabletime separation between the two operations (up tofourteen years) accentuates these disadvantages.

The method also depends heavily on the presenceof an adjoining utilised coupe for reference


9

purposes. When a nearby coupe is not available.the method becomes just another cruise estimate.

A further disadvantage is the requirement that thecruise be carried out by a person, usually alogging supervisor, who must have sufficientexperience to make a reliable estimate ofmerchantable volumes in the highly variableforests concerned.

The reliability of the data has been assessed usinga comparison of the predicted and actual yieldsper hectare obtained using the method over along time period, shown in Figure 1 [referAppendix 2] for sawlog and Figure 2 [referAppendix 2] for pulpwood. For the periodbetween about 1977 and 1994, there is aconsistent relationship, particularly for sawlogvolumes, of underestimation. For that period,sawlog volume was underestimated by 8% andpulpwood volumes by 29%. However, from 1994,the relationship appears to break down, withsawlog yields dropping to about the same level aspredictions and pulpwood production alsodropping substantially. A range of theories havebeen proposed for these changes includingincreases in stream buffer width, the beginning ofutilisation in the uncut coupes of alternate coupes(which are the very small coupes of the late1970s), changes in market conditions etc.However, there is no way to determine which ifany of these alternatives is the real reason for thechange.

While the graphs of mean yield per hectare forthe many coupes (about 50) utilised per yearshown in Figures 1 and 2 are relatively smoothand level, they represent relatively consistentmeans that have large variations. Confidencelimits (95%) have been calculated for the actualand estimated yields and they are shown asvertical error bars on Figures 1 and 2. Theyindicate that the sawlog estimates weresignificantly different in only one year (1982) andthat pulpwood yields were significantly greaterthan estimated for five years (1982-93). Theconfidence range in most years was around 30%of the mean, although individual years variedfrom 20% to 100%. These results are good forforests of this type.

Refer to the full report (Appendix 2) fordiscussion on the alternatives to the abovemethod. The report indicates that alternativeprocesses could have been used but withsignificantly greater costs and more time required.

Conclusion

Given that the alternative methods available mayproduce some extra benefits, but at significantextra cost, without appreciably improving theprecision of estimates of merchantable yields, thecurrent method is considered to be anappropriate way to estimate wood volumes in themulti-aged forest concerned.

However, attention needs to be directed towardsfinding the reasons why the yields obtained since1994 have been less than those in previous years.If the yields continue to be low, there may be aneed to refine the existing predictions. A reliableestimate of the remaining utilisable volume in theMAF is a crucial requirement in estimating thesustainable yield for the Regional ForestAgreement.

Cruising estimates made in future could beimproved by the application of simple basal areacount methods that allocate count trees tocategories such as regrowth, advance regrowth,overwood trees and merchantability classes. Themethodology for these surveys needs to bedeveloped, tested and documented beforewidespread application.

Regrowth forest

API stratification

The inclusion of stand variables into the APIstratification provides the opportunity to deriveestimates of area for the different strata directly.However, if in fact the strata cannot be effectivelyseparated using API, then a significant amount ofwork will have been done with no useful result.The separation of strata was examined (duringthis review), by comparing mean stand variablesobtained from plot measurements within thedesignated strata. Table 1 gives the results of thiscomparison[refer to Appendix 2 for Table 1].

The table shows that the stem stocking results inall strata except one (4a), were outside thenominated target ranges and that the mean valuesare mostly not significantly different (theconfidence ranges overlap). For mean diameter,six out of eight sample mean diameters felloutside the target ranges and only one mean, inthe 1a stratum, was different from the otherranges in the same stocking class. While retrievalof useful data may be possible through pooling ofsome categories, the API stratification can onlybe regarded as unsatisfactory in providing useful


10

stand data. It does however effectively delineatethe boundaries of regrowth stands.

An alternative approach would be to selectsimpler API categories that were demonstrablyable to be delineated successfully. This willachieve the main task of API - deriving a reliableestimate of location and area with a simplesubdivision into density categories. Standvariables can then be estimated using plotmeasurements. Using simpler API strata may alsomake possible the use of smaller scale airphotos,that may provide complete coverage of the area ofinterest without the delay involved in newphotography.

Refer to full report in Appendix 2 for discussionon plot measurement techniques.

Effectiveness of the plot sampling scheme inrepresenting the resource

The resource includes a range of different foresttypes, age classes, productivity classes, types ofregrowth origin and geographic locations. Anappropriate sampling scheme must efficientlysample the variation in each of these categories.There has been no formal consideration of theserequirements, sampling has been directedtowards the older stands because it is recognisedthat information is most urgently required forthese stands, since these will be available forutilisation first.

An indicative examination of the coverage of theresource by the current set of regrowth plots wascarried out using maps of forest type, date ofcoupe utilisation and plot location (Figures 3 to 5respectively)[Refer Appendix 2 for figures]. Thecompartments sampled to date are indicated onFigure 3, shaded red. The samples cover thelimits of geographic spread of the resource.However, there are gaps in the coverage withinthese limits, for example, in Yambulla, NullicaState Forests. Forest types in the resource includeDry Shrubby, Dry Grass, Moist and IntermediateShrubby (Figure 3). Dry Shrubby, dominated byE. sieberi, covers the largest area, occurring inthe south east of the EMA and including themajority of plots established so far. Moist forestoccurs in the northwest of the EMA and includesthe next largest number of plots. Dry Grass forestoccurs in the south west of the EMA and containsonly a few plots.

Figure 4, indicating regrowth age, shows whencompared with Figure 5, that stands logged

between 1980 and 1990 are not well representedby sample plots. This, and other gaps inrepresentation, should be corrected by thecontinuing measurement of sample plots.

The site productivity layer is not yet available anddistribution of plots within it cannot bedetermined. This layer was available subsequentto this analysis; see Section 3.4.

Table 2 [refer to Appendix 2 for Table 2 -Number of plots in different types and ages ofregrowth] indicates a majority of plots in the fireregrowth, in accord with the expressed priority togather data in those stands first. Future samplingshould redress this imbalance, since all standswill eventually take part in production and thismust be considered in the calculation ofsustainable yield. The distribution of regrowthplots to stands needs to be done in proportion toresource area, with some modification ifnecessary to achieve similar precision formeasurements in different strata. This allocationof plots to strata cannot be done until a detailedarea statement for the regrowth resource hasbeen prepared. It is currently assumed that,because regeneration is usually successful,logged area is equal to new regrowth area. Thismay or may not be an accurate assumption.

Quality of baseline inventory informationprovided by the regrowth inventory

In this context, 'baseline inventory information’ isinterpreted to mean data that is necessary forpredicting the development of regrowth andestimating merchantable outturn from it for aminimum of one rotation. This will provide datato determine how long the MAF resource needs tolast before production from the resource can bereliably provided from regrowth stands.

To meet this criteria, inventory plots need to beestablished in appropriate strata, as describedabove, and the plot measurement techniques needto provide appropriate and reliable data.

It is difficult to quantify the effect on datareliability that the shortcomings outlined abovewill cause. Because the methodology has not beenclearly specified, operators have the latitude todevelop their own techniques that may or may notbe appropriate (it is not suggested that they have,merely that it is unknown). An assessment carriedout with trialed and documented methods willhave greater precision, but it is impossible to sayhow much greater. With the qualifications


11

expressed above, subsets of the existing data setare suitable for use, provided they aresupplemented with new data that is representativeof currently unsampled strata. The existing datacan be reallocated to any new API stratificationthat is developed for the resource as a whole.

3.2 RETAINED TREE INVENTORY

Retained tree inventory



Timber harvesting during some periods hasresulted in significant numbers of trees leftstanding in harvested coupes to provide seedtrees, fauna habitat and some future sawlogs.Retained trees can affect future yields in that:

n some future sawlog volumes may be availablefrom previously retained trees; and

n there may be an impact on regrowth growthrates.

3.2.2 Methods

This project involved capturing the harvestingprescriptions which related to each harvestingevent.

As the records in compartment histories oftengive only a general description of what wasretained and not an actual figure of how manytrees or what volume was retained, the followingdata was collated and reviewed:

n actual measurements from compartmenthistories and Permanent Growth Plots (PGP)in harvested areas was used to determine thestocking rate and basal area of trees retainedafter harvesting;

n Post logging surveys carried out in the past.These have been:

1. Regeneration surveys. These surveysfocussed on the stocking levels ofregeneration. Some of these surveysincluded a basal area sweep for a measureof retained trees. There were about 560coupes that fell into this category, theycovered the logging years from 1972 to1996.

2. Specific retained tree surveys. Due to thevarious objectives of the surveys at thetime, different parameters were measured.About 100 coupes were measured, 25 withstocking of retained trees only and 75 withbasal area measured. The logging yearcovered ranged from 1977-1996.

3. As part of the IFA, field validation of asmall number of coupes for retained treeswas carried out. All retained trees weremeasured with stocking, basal area andvolume recorded.

3.2.3 Results

After the data had been collated, there was not ameasure of statistical significance through time tobe able to provide a statistically valid measure ofretained trees for each logging year. However, thedescriptions of harvesting prescriptions combinedwith actual measures of tree retention provided atrend in harvesting events. These trends show anincrease in tree retention rates over time.

Retained tree rates were derived from differentsources: actual basal area measurement frominventory, actual number per hectare frominventory, or retained tree prescriptions applied tothe stand and assumed to be implemented. Theretained tree rates were entered into the EdenFRAMES database (refer Section 5.1).

In the conclusions of the IFA retained sawlogreport (refer Appendix 3) it states that futuresawlog trees were only retained in any large wayfor areas logged during the period from the mid1980s to 1990. The analysis of the retained treedata showed that retained basal area for thisperiod varied from 3 to 8m2/ha. For the period1990 to 1996 the range was also 4 to 8m2/ha.These figures add weight to the conclusion thattrees retained at the time of logging would berequired to meet current and future prescriptions.

Therefore, previously retained trees are likely tobe required for ongoing habitat, thus are unable tosignificantly contribute to sawlog volume duringfuture harvesting operations.

The impact of retained trees on regrowth growthrates is dealt with in the discussion ofSTANDSIM (Section 4.1).


12

3.3 DEVELOPMENT OF NETHARVESTABLE AREA

Future Net Harvestable AreaProject No. NE/12 FRA


The objective of this project was to develop aquantitative and repeatable assessment of the netarea available for harvesting under currentprescriptions at the coupe level and across theEden CRA Region.

This assessment was required to convert grossharvestable area into net harvestable area so thatfuture yields from thinning and harvestingoperations could be predicted. The project alsoallows for an assessment of the variation thatfuture management prescriptions might cause tothe net harvestable area.

3.3.2 Methods

The net harvestable area project generated adigital layer of areas available/unavailable forharvesting within State forests under currentmanagement prescriptions. The following sectionsdescribe the method used.

3.3.3 Identification of exclusions fromharvesting

In order to determine the net harvestable area, thearea of forest in which harvesting is excluded wasspecified and mapped. The following data layersand attributes were used to generate the net areaavailable for harvesting under currentprescriptions.

These exclusions constitute ‘hard exclusions’ toharvesting (ie those exclusion areas that can bereadily mapped); ‘soft exclusions’ (exclusionareas that cannot be easily mapped) includeadditional areas for soil and water protection thatdo not show on maps of the area.

PMP

Excluded - SFNSW Preferred ManagementPriority (PMP) categories:

n 1.1.2 - Special Emphasis (Recreation);

n 1.1.7 - Flora and Fauna Protection;

n 1.1.9 - Aboriginal Sites;

n 1.2 - Undeveloped Natural Forest;

n 1.3 - Preserved Natural Forest;

n 3.1 - Cleared; and

n 3.2 - Special Development.

Rainforest and a protective buffer

Excluded - All areas within the NSW NationalParks and Wildlife Service (NPWS)‘FLORISTICS’ coverage where the itemSYMBOL contained R or RE; plus a 20 metrebuffer.

Rocky area & buffers

Excluded - All areas within the NSW NPWS‘FLORISTICS’ coverage where the itemSYMBOL contained Qb01, WI02, Wr, Ws,Sw01, or Td01; plus a 20 metre buffer where areawas between 0.1-0.5 hectares or 40 metre bufferwhere area was greater than 0.5 hectares.

Swamp (wetlands) & buffers

Excluded - All areas within the NSW NPWS‘FLORISTICS’ coverage where the itemSYMBOL contained Qf01, or Wh01; plus a 10metre buffer where area was between 0.1-0.5hectares or 40 metre buffer where area was greaterthan 0.5 hectares.

Heath & buffers

Excluded - All areas within the NSW NPWS‘FLORISTICS’ coverage where the itemSYMBOL contained Sr01, Sk01, Td02, Td03,Tt01, Tw01, or Tw02; plus a 20 metre bufferwhere area was between 0.2-0.5 hectares or 40metre buffer where area was greater than 0.5hectares.

Slope

Excluded - All areas greater than 30 degrees.

Fauna stream buffers

Excluded - 1:25 000 scale streams were bufferedaccording to Conservation Protocols (29November, 1996). Buffers on streams dependedupon their stream order : first order stream buffer,10 metres; second order stream buffer, 20 metres;third and greater order stream buffer, 40 metres.Note, only 80% of third and greater order streamsare required to have a 40 metre buffer, the


13

remainder may have a 20 metres wide buffer. Inthis instance, a 40 metre buffer was applied to allthird and greater order streams.

PCL stream buffers

Excluded - As per the Pollution Control Act1970. See Section 3.3.4 for a description of themethod.

Erosion hazard category 4

Excluded - All area of hazard category 4 wereexcluded from harvesting.

3.3.4 Derivation of erosion hazardcategories and stream buffers

Soil regolith stability

1:100 000 mapscale soil landscapes units wereavailable for the Craigie and Bega mapsheets. Forareas outside these two mapsheets, dominantlithology was used to define soil landscape unitsconsistent with the Bega and Craigie mapsheets.Each soil landscape was reclassified by theDepartment of Land and Water Conservation(DLWC) into soil regolith cohesion and soilregolith sediment delivery potential according to“Soil erosion and water pollution hazardassessment for logging operations” (EPA / SFNSW / DLWC, August, 1997).

Slope

Slope was calculated from the 1:25 000 DigitalElevation Model (DEM).

Rainfall erosivity

A rainfall erosivity surface was provided byDLWC according to Rosewell and Turner (1992)

Calculation of erosion hazard for eachcompartment

n Slope, rainfall erosivity and soil regolithstability grids were overlayed and classifiedaccording to “Soil erosion and water pollutionhazard assessment for logging operations”(EPA/SFNSW/DLWC, August, 1997) in orderto obtain a hazard class on a cell by cell basis.A hazard class was then calculated for eachcompartment based on the proportions of eachhazard category.

n Compartment level hazard classes were thenoverlayed with streams attributed with theirstream order and specific catchment area inorder to reclassify the hazard class accordingto a threshold of 100 hectares.

n Using rules defined by the Pollution ControlAct (1970) streams were differentially bufferedaccording to the compartments hazard class,stream order and specific catchment area.

The above data allows determination of theimpact of changes in prescriptions or moratoriumareas on available areas for harvesting.

3.3.5 Results

An accredited fully documented geographicinformation system (GIS) database was generatedat 1:25 000 scale. The above exclusions werecombined within the GIS to produce thefollowing outputs:

n a combined coverage including all harvestingexclusions;

n a binary coverage of Net Harvestable Area;

n database tables summarising the area of eachexclusion on a coupe by coupe basis accordingto current prescriptions; tabular summarieswere also generated for the whole CRARegion (Table 3a); and

n database tables summarising the cumulativeincrease in excluded area per prescriptiveexclusion on a coupe by coupe basis.

It should be noted that Fauna Moratorium Areaswere not included in calculations.

Current net harvestable area without excludingfauna moratorium areas was calculated as around75% of the total State forest area. The areas ofexclusion are as follows in Table 3a.

TABLE 3A: EXCLUSIONS IN THE EDEN CRAREGION

Exclusions Area Cumulative Area

(hectares) (hectares)* (%) PreferredManagementPriority

22 914 22 914 11.56

Rainforest andbuffers

5 061 26 731 13.49

Rocky areasand buffers

2 122 27 711 13.99


14

Exclusions Area Cumulative Area

(hectares) (hectares)* (%) Swamp(wetlands) andbuffers

754 28 320 14.29

Heath andbuffers

2 166 29 590 14.93

Slope>30o 3 456 31 746 16.02

Fauna streambuffers

14 350 40 987 20.69

Pollutioncontrol streambuffers

16 804 44 416 22.42

High erosionrisk

5 236 45 493 22.96

* Due to overlap of areas these figures are not simplya sum of exclusion areas.

A table of net harvestable area was generatedfrom the GIS layers. The net harvestable areainformation was entered into the Eden FRAMESdatabase (see Section 5.1).

The net harvestable area model used to derive thenet harvestable area layer can be used torecalculate net harvestable area under alternativearea based prescriptive exclusions.

3.4 SITE PRODUCTIVITY INDEX

Development of a Site Productivity Index forEden



The site productivity index for Eden aimed todevelop a quantitative and repeatable siteproductivity index that would identify thepotential productivity of the regrowth forest forgrowth modelling.

It sought to provide a mechanism to quantifyvariability in relation to productivity withincompartments in order to more accurately assessthe impacts current or future prescriptions mayhave on potential yields.

3.4.2 Methods

The project was based around the mapping ofenvironmental factors that contribute to forestgrowth. These factors were broadly: nutrients,moisture, temperature and light.

The project was split into three components:

n The development of the generalised siteproductivity index (GSPI) which incorporatedmonthly radiation, rainfall, evaporation,minimum temperature, and topographicposition.

n The re-classification of forest types into yieldassociations.

n The development of a low, average and highmean annual increment for each yieldassociation. These were qualitative estimatesbased on limited plot data.

Within each yield association, the GSPI was thennormalised to assign a low, average or high meanannual increment (MAI) by reclassifying thefrequency distribution into three equal intervalclasses. This generally resulted in exposednortherly aspects and ridges being assigned alower MAI and sheltered southern aspects andgullies being assigned a higher MAI. A weightedaverage MAI was calculated for each coupe,based on the area of each MAI class within thecoupe.

MAIs ranged from 2m3/ha/yr to 12m3/ha/yr. TheseMAIs were converted to an index between 0 and1.2, with an index of 1 equivalent to 12m3/ha/yr.Based on a combination of inventory and researchdata providing MAI and site productivityestimates at known locations, the index was splitinto classes as follows:

n for fire regrowth, two site productivity stratawere identified, comprising ≤0.6 and >0.6index values;

n for logging regrowth, three site productivitystrata were identified, comprising <0.5, 0.5-0.7and >0.7 index values.

These site productivity indices (SPI) are used inthe stratification set out in Table 4a.

3.4.3 Results

The project produced a SPI for each coupe andregrowth forests were stratified using this SPIvalue, whether the regrowth was derived fromharvesting or fire and whether the coupe waslocated within the coastal or tablelands part of themanagement area..

Different STANDSIM yield tables applied to eachstratum. The stratum for each coupe was includedin the Eden FRAMES database (Section 5.1).Yield information was passed to the yield


15

estimating and scheduling toolbox (see Section5.2)

This dataset has not been validated through fieldchecking and is only an interim dataset.

3.5 SAWLOG SPECIES MIX

Sawlog Species Mix

Project No: NE 20/FRA


The objective of the project was to determine, forthe multi-aged forest, proportions of species andsize class within each compartment and coupe.

3.5.2 Methods

The method used in this project was as follows:

n Sawlog volumes since 1979 were extractedfrom State Forests’ sales systems (FORSALEand FORPRAC) by compartment, species anddiameter classes. Some data prior to this datewas available from paper records.

n For compartments with the alternate coupeharvested, the percentage of species of the totalcoupe volume was calculated. Thesepercentages were included in the EdenFRAMES database (refer Section 5.1) andapplied to the predicted total volumes (referSection 3.1).

n For compartments with the alternate coupesnot harvested (approximately 15% ofcompartments), the percentage of species ofthe management section’s total volume wascalculated. These percentages were included inthe Eden FRAMES database (refer Section5.1) and applied to the predicted total volumes(refer Section 3.1).

n For diameter classes, the percentage diameterclasses were calculated for each managementsection. These percentages were included inthe Eden FRAMES database (refer Section5.1) and applied to the predicted total volumes(refer Section 3.1).

3.5.3 Results

The expected species and size class sawlogpercentages for each coupe were included in theEden FRAMES database (refer Section 5.1). This

data enables volume information to be subdividedinto species and indicates the volume of thespecies and size class of sawlogs that have value-adding potential (refer to Table 6b).

17

4. MODEL DEVELOPMENTAND APPLICATION

4.1 GROWTH AND YIELD MODEL

Growth modelling



The objective of this project was to developgrowth and yield models for regrowth standsregenerated from harvesting and/or fire in theEden CRA Region. The project also aimed to testthe model using growth data from permanent ortemporary growth plots measured in the EdenCRA Region and modify the model further ifnecessary.

4.1.2 Methods and Results

It was assumed that the multi-aged forest resourcehas a zero net growth in quota sawlogs, as anygrowth is offset by decline (this has been verifiedby a small scale study undertaken by SFNSW(Jurskis and Hudson, 1994; refer Appendix 4).

For the regrowth resource, estimation of growthrates (by developing growth models based onactual forest growth) was required to estimatefuture yields of wood products from the regrowthcomponent of the forest. STANDSIM, a standsimulator developed for Victorian stands ofEucalyptus sieberi (Silvertop Ash), and otherspecies, was modified and adopted for use in theEden CRA Region.

Data for growth models are drawn from ResearchGrowth Plots (RGPs). RGPs are plots in whichall trees are permanently identified and are subjectto remeasurement, usually every five years. RGPs(and Permanent Growth Plots elsewhere in thestate) are designed to enable prediction of growthrates under normal forest conditions (refer to Bi,Chikumbo and Jurskis, unpublished (Appendix 6)for details of how this data was used).

This other plot data used in Eden was inventoryplot data for regrowth stands. This inventory isdetailed in Section 3.1.2. The inventory data wasclassified into strata based on:

n regrowth type (ie fire or logging regrowth);

n site productivity (refer Section 3.4);

n location (ie cost or tablelands); and

n thinning history.

Within each of these strata, inventory data wasanalysed to derive input parameters forSTANDSIM. These parameters were stockinginformation, basal area, site productivity indexand age.

The uses of this RGP and inventory data in modeldevelopment are noted in the following text initalics which has been extracted from the report“Modification of the STANDSIM model for theEden Management Area” Forest Essentials PtyLtd, 1997b report (refer Appendix 5)

4.1.3 Review of the STANDSIM model

STANDSIM is a deterministic FORTRAN codedgrowth simulation model originally designed(Opie 1972) for evenaged, single species stands(it has since been modified to accommodate twoaged stands). It is unnecessary to specify thearrangement of trees, making it possible to usestandard inventory information as inputs to themodel, although to make this possible, it isassumed that tree spacing is reasonably even(with the exception of response to strip thinning).The FORTRAN program includes a set of fifteengrowth and mensurational functions that apply toone species. The program structure is not speciesspecific, thus to modify the model for a newspecies, it is only necessary to insert the functionsnecessary for that species. Functions currently inthe model apply to Victorian stands of E.regnans, E. delegatensis, E. nitens and E. sieberi.

The model can simulate growth from age zero,requiring specification of stem stocking and site


18

index (stand height at age 20). To simulategrowth from ages greater than 15, it is necessaryto specify stem stocking, site index and standingbasal area, or instead of basal area, adistribution of stem diameters. The estimation ofgrowth is most accurate if diameter distributionsare specified.

The model can estimate the consequences ofthinning from below or thinning in strips, fromage fifteen onwards. For E. sieberi, the effects offour intensities of burning on mortality and standgrowth can be estimated.

The conversion of gross bole volume tomerchantable product volumes can be simulatedby relationships between product cutoff pointsand tree DBHOB, tree height or product smallend diameter and the specification of defectpercentage. Maximum and minimum dimensionscan be specified for product pieces.

A more detailed description of the model and itsoperation is given by Incoll (1983).

Refer Appendix 5 for Figure 1 - Maximum heightvs age curves for E. sieberi, Eden.

Functions used in the SFNSW model

The functions available for use came from workcarried out by the South East Regrowth ForestGrowth and Yield Modelling Project (Bi 1994)since 1991, listed as follows:

1. Single tree volume functions for a range oftablelands and coastal species (Bi undated,probably early 1997). This function, withcoefficients for E. sieberi, was usedunaltered.

2. Taper functions predicting stem diameterat any height on the tree stem. The taperfunctions described (Bi 1997a) areformulated to estimate diameter at anypoint on the tree stem as a function ofrelative height at the desired point andtotal height. For use in STANDSIM, it isalso necessary to estimate height for anominated diameter on the tree stem. Bi’sfunction could be used iteratively toachieve this, but there was not enough timeto code this capability in the current work.

3. Tree height as a function of tree diameterfor a range of species in tablelands andcoastal forests. This function is describedin Bi 1997b, but because the functions do

not include stand variables, it is unclearwhether they are suitable for inclusion inSTANDSIM. This will need to be clarifiedwith Dr Bi.

4. Maximum tree height/age functions fortablelands forests. This function was notincluded during the current work becauseE. sieberi forests were of higher priority.

5. Maximum tree height/age functions forcoastal forests (Bi et al 1997c) [referAppendix 6 of this report]. This functionwas used in a modified form. When plottedon the same axes as function 4, themaximum height of E. sieberi between ageten and age 33 was expected to be greaterthan for tablelands forests, thereafterbecoming less again. Additionally, themaximum height at age 20 was expected tobe about 33m, considerably in excess ofthis author’s experience in high site qualityOrbost stands. The function was thereforemodified to provide a maximum height atage 20 of 25m and this function was usedin the model. A further reason formodifying this function is that the variablerequired by STANDSIM is mean dominantheight, not the maximum tree heightprovided by Bi’s function. The relationshipbetween maximum and mean dominantheight is unclear, but mean dominantheight would be expected to be less thanmaximum height, as in the modifiedfunction. The functions concerned areshown on Figure 1, including the Victorianfunction (for a site index of 25m).

6. Total stocking as a function of age andoverwood stocking for tablelands forests.This function was not included during thecurrent work because E. sieberi forestswere of higher priority.

7. Stocking of upper canopy trees as afunction of age, overwood stocking and siteproductivity index for coastal forests (Bi etal 1997c). This function was insertedunchanged in the model, but wassubsequently modified as a result of testruns (see later).

8. Stand volume of age, site productivityindex, overwood stocking, and regrowth


19

stocking for tablelands for tablelandsforests. This function was not includedduring the current work because E. sieberiforests were of higher priority.

9. Stand volume as a function of age, siteproductivity index, overwood stocking, andregrowth stocking for coastal forests. Thisfunction was not necessary sinceSTANDSIM requires a single tree volumetable, provided as function 1 above.

10. Stand basal area as a function of age, siteproductivity index and overwood stockingand mean annual increment of basal areaas a function of the same variables.STANDSIM requires a function estimatingcurrent annual increment of gross basalarea, thus this function was notappropriate.

11. Current annual increment of stand volumeas a function of age, site productivity indexand overwood stocking for tablelandsforests. This function was not includedduring the current work because E .sieberiforests were of higher priority.

The functions described above were inserted inthe STANDSIM source program, SSIMNU.FOR.Comments have been inserted in the source codewhere alterations have been made and these canbe located by searching on part or whole of theline:

"C E. sieberi SFNSW aug97"

Testing the Modified program

Inserting functions

Testing of the single tree volume function wascarried out externally to the STANDSIM model. Aseries of DBHOB classes and their estimated treeheights were obtained from a typical fireregrowth STANDSIM run. Volumes werecalculated for the sets of diameters and heightsusing the Victorian and Eden functions. Thedifference between the estimates was expressed asa percentage of the Eden estimate. Figure 2 [referAppendix 5 for Table 2 – Comparison of Victoriaand Eden volume tables] shows the results of thiscomparison, with the percentage differenceplotted against DBHOB2*Height. The figureshows that the differences are appreciable forsmall, short trees, but fall below 5% when D2H isgreater than 10,000 (ie DBHOB about 14 cm)and less than 2% when D2H is greater than

22,000 (DBHOB about 22 cm). The differencesshould not be regarded as an error in eitherfunction, since the trees may well be differentshapes. Whatever the source of the differences,they are regarded as unimportant for thepurposes of estimating yields of merchantablematerial.

Further testing was carried out following theinsertion of each function into the STANDSIMmodel and the effects evaluated by comparisonwith runs of an unaltered version of STANDSIM.The model was run using an initial stocking ofl0,000 trees/hectare at age 5, with a site index of20m, equivalent to a stand productivity index of0.6. Functions were inserted as follows.

1. The height age function. Figure 3 [referAppendix 5 for Figure 3 – Effect of replacingheight age and volume function on gross bolevolume prediction] indicates the gross bolevolume (gbv) estimated by the two differentSTANDSIM models. The figure shows thatdifferences are minor until about age 50, whengbv for the Eden model has fallen about 10%behind. By age 80, the difference is about 20%.These differences approximate the differences instand height, see Figure l.

2. The maximum stocking function. Figure 4[refer Appendix 5 for Figure 4 – Effect ofreplacing height age, volume and maximumstocking functions on simulation of gross bolevolume prediction] indicates the gross bolevolume (gbv) estimated by the two differentSTANDSIM models. The figure shows thatdifferences now remain minor throughout the agerange simulated. However, when the stemstocking predictions are examined (Figure 5)[refer Appendix 5 for Figure 5 – Effect ofreplacing maximum stocking functions on stemstocking predictions], it is apparent that the Edenfunction results in very heavy mortality up to age20, followed by very little subsequent mortality.This seems an anomalous situation, sinceregrowth stands in the real world can beobserved to have continuing mortality well pastage 20. Various modifications to the Eden modelwere tried without changing the situation muchand eventually a different model form was used,resulting in a function intermediate in shapebetween the Eden and Victorian models. Themodel used was:

N = 25000 (1-e (-4.7*spi/age))


20

where

N = maximum stocking, stems per hectare frominventory data as specified above

spi = site productivity index from Section 3.4

age = stand age in years recorded as starting agefor regrowth type

This model resulted in prediction of stemstockings intermediate between the two curvesshown in Figure 5.

Testing with temporary plot data

While data from permanent plots were beingprepared) the model was tested against theavailable temporary plot data by running themodel from typical initial stem stockings throughto ages similar to the temporary plots Thesimulated volumes were then compared with thevolumes on the temporary plots. When this wasdone simulation of typical fire regrowth stockings(5,000 stems/ha) produced standing gross bolevolumes within the top half of the spread oftemporary plot standing volumes at age 45. Thisis considered a sensible result, considering thatmost real world stands will have been burntseveral times severely enough to have retardedtheir growth rates. Time was not available toinvestigate the effects of burning in the simulationruns, although the model could haveaccommodated this. Simulation of the growth oflogging regrowth (initial stocking 1,000 stems/ha)resulted in higher standing volumes at age 45,and although there were no temporary plot dataof this age resulting from logging, it seemssensible that the lower density (but still fullystocked) logging regrowth stands would producehigher volumes than fire regrowth.

Testing with permanent plot data - standvariables only

The STANDSIM model was initially tested usingdata from a report describing the response to arange of early thinning for a series of plots in theMumbulla State Forest. Input data (stem stockingand standing basal area) were obtained from thereport for stem stocking at age four. Thesimulation was then run from age four to age 28,the last age for which data were given in thereport. Diameter distributions were not given inthe report, making the simulation task less likely

to be accurate, and the unthinned plots were ofrestricted area.

The estimates made by the STANDSIM modelconsiderably underestimated growth in the realworld, probably because the functions estimatingbasal area at age 15 and the correspondingdiameter distributions are inappropriate for thetest material. The (Victorian) function estimatingbasal area at age fifteen was modified in anattempt to correct this problem, but there wasinsufficient time available in the present study tocomplete the modifications.

The model must therefore be regarded ascurrently unreliable in estimating growth forunderstocked stands from initial ages beforefifteen years.

Testing with permanent plot data - diameterdistributions

The model was used to estimate growth betweenthe initial and final measurements of thecommercial thinning trial in East Boyd StateForest (trial L374). The trial included sixunthinned plots and six plots thinned from below.These two sets of plots were pooled to providecomposite thinned and unthinned sample plots,each 0.36 ha in area. The growth of the unthinnedplots was simulated from age 25 to 36 and thethinned plots from age 27 to 36. The factor ofmost interest was how accurately the modelwould simulate the final diameter distributionsand this is illustrated in Figures 6 and 7 [referAppendix 5 for Figure 6 – Simulation of growthon unthinned plots trial L374 and Figure 7 –Simulation of growth on thinned plots, trial L374]for the unthinned and thinned plots respectively.

Figure 6 indicates that the simulation of growthon unthinned plots has been reasonablysuccessful. However, for thinned plots, the growthof small trees appears to have beenunderestimated in that the simulated distributionhas more small trees than the actual stand. Thesimulated stand also has more live trees than thereal stand, so it may also be that mortality hasbeen underestimated. While the accuracy of themodel in this area can be improved by furtherwork, the conservative nature of the inaccuracymeans that the model can still be effectively usedin its current state for the current planningexercise.

Future modifications


21

Further work is necessary on the model to refinethe simulation of growth and mortality to betterreflect real world observations. This work wasnot possible in the current study because of timerestrictions. Testing also needs to be extended toinclude the other permanent plot data available.

Further modifications, when carried out, need tobe inserted into the source program, compiledand the object code renamed SSIMCORE.EXE.This file should be used to replace the current fileof the same name in the STANDSIM directory.The program STANDSIM.EXE callsSSIMCORE.EXE as necessary to carry outsimulations designed by the user.

The database (refer Section 5.1) classifies allregrowth into strata by:

n regrowth type (ie fire or harvesting regrowth);

n site productivity (refer Section 3.4); and

n location (ie coast or tablelands).

The STANDSIM growth model uses site index asan input parameter. This site index relates to thestand height at age 20. Analysis of research plotsin coastal regrowth had shown that a medium tohigh site productivity index had a STANDSIMsite index of 17. Thus, the site productivity indexvalues were converted to STANDSIM site indicesranging from 13 to 22. The upper index of 22 wasapplied to high site productivity index tablelandsareas following discussions with Bill Incoll(Forest Essentials Pty, Ltd).

The results of the stratification of regrowth areasare shown in Table 4a.

TABLE 4A: STRATIFICATION OF REGROWTHAREAS

Strata RegrowthType

SiteProduct-ivity Index

STANDSIMSite Index**

Location

1 Fire ≤0.6 13 Coast 2* Fire >0.6 17 Coast 3* Fire >0.6 17 Coast 4 Logging 0.5-0.7 16 Coast 5 Logging >0.7 18 Coast 6 Fire Thinned 17 Coast 7 Fire ≤0.6 13 Coast 8 Fire >0.6 17 Coast 9 Logging <0.5 13 Coast 10 Logging <0.5 15 Tablelands 11 Logging 0.5-0.7 18 Tablelands 12 Logging >0.7 22 Tablelands 13 Fire Thinned 18 Tablelands

* These strata are separated because inventoryinformation shows a significant difference between

these strata and they are related to two different fireevents.

For each stratum, a yield table is generated bySTANDSIM which predicts the volume byproduct by predicted year of harvest for eachregrowth coupe. The yield tables are provided inTable 4b.

TABLE 4B: STANDSIM PREDICTED YIELDTABLES

Stratum Operation Age Sawlogs(m3/ha)

Pulpwood(t/ha)

1 t1 35 0 100 1 t2 50 5 80 1 rh2 70 50 120 2 t1 30 0 100 2 t2 45 5 80 2 rh2 70 60 100 3 t1 32 0 120 3 t2 45 10 80 3 rh2 65 60 100 4 t1 35 0 100 4 t2 45 10 70 4 rh2 70 50 100 5 t1 30 0 100 5 t2 40 15 65 5 rh2 60 60 100 6 rh2 67 50 100 7 t1 47 2 90 7 rh2 68 50 100 8 t1 46 5 105 8 rh2 68 60 100 9 t1 40 0 100 9 rh2 70 50 100 10 t1 35 0 120 10 t2 50 10 70 10 rh2 70 60 130 11 t1 30 0 110 11 t2 45 12 70 11 rh2 70 70 160 12 t1 25 0 100 12 t2 40 15 80 12 rh2 60 70 170 13 rh2 66 75 150 Note: t1 means first thinning, t2 means secondthinning, rh2 means regeneration harvest.

4.1.4 Conclusion

The independent expert (Forest Essentials 1997b)recognised that STANDSIM, as modified, wasmost suitable for the E. sieberi dominated standsin the Eden CRA Region and would produce lessaccurate, but estimated as conservative outputs(particularly when yield tables were varied by siteproductivity class) for other forest types.


22

The available plot data gives a reasonablecoverage of age and thinning response in fireregrowth stands and of age in harvesting regrowthstands.

The limited time available for work meant that:

n functions for tablelands species have not beenincluded;

n taper functions have not been included;

n the model underestimates growth forunderstocked stands younger than fifteen yearsof age; and

n estimation of growth in stands thinned afterage fifteen is conservative.

Despite these shortcomings, the project hasproduced a growth model that can be used toprovide conservatively based yield tables for usein the yield scheduler (see Section 5.2) for theperiod up to around 2040. This will allow anestimation of sustained yield for the CRA/RFAprocess.

Further work is recommended to improve theaccuracy of the model and to fully utilise theSFNSW research programs in progress in theEden CRA Region.

4.2 MODIFIER MODELS

Development of Modifier Models



This project aimed to develop a suite of modifiermodels that would take into account the effects ofthinning of the regrowth forest, and other forestmanagement practices that could affect growthand future yields in all forest types.

4.2.2 Methods

The development of a range of modifier modelsas outlined in the project objectives is dependenton the availability of a range of data from trialscovering the proposed modifications. Forregrowth forest in Eden (and elsewhere) this datais limited. As a result a conservative approachwas used, which included testing against availabledata.

STANDSIM (which included growth models asadopted for Eden) has the ability to evaluate

yields from various thinning regimes for thestands. The input parameters for STANDSIM asdescribed in Section 4.1.2 were used for the stratadescribed in Table 4a. The input parameters werederived by analysis of inventory plot data andestimation based on experience and a stockingprediction function (Bi, 1994 and 1997).

STANDSIM was then run and yield tablesdetermined for commercial thinning(s) frombelow of each of the strata.

The data limitations outlined above precludedconsideration of management intensificationpractices such as fertilising and the impacts of fireregimes on sawlog yields.

4.2.3 Results

The input parameters to STANDSIM were variedto investigate the possible effects of differentinitial stockings and fire regimes. STANDSIMdoes not provide for this easily; the model doesestimate the effects of fire on growth, but firescannot be included in runs that also includeproduction thinnings.

STANDSIM simulated sawlog production in therange of 150 to 200m3/ha at final harvest at age70, based on the scheduled commercial thinningregimes for each stratum. Experience at Eden andelsewhere led to the conclusion that these resultswere unrealistic.

Further, increasing the initial stocking orsimulating a wildfire could also reduce sawlogprediction at final harvest at age 70. Given theabove, the output yields from STANDSIM weremodified to adopt sawlog volumes of 50 to75 m3/ha at final rotation that were conservative.It is felt that these estimates are acceptable for thecurrent process that requires estimation of sawlogproduction in a defensibly conservative manner(B. Incoll pers comm.).

23

5. DATA COMPILATIONAND ANALYSIS

5.1 COMPILE AND DOCUMENTDATABASE

Compilation and Documentation of theDatabase



The aim of this project was to collate the spatialand tabular data on the resource so that it could beutilised by growth and yield models andintegration projects.

5.1.2 Methods

This project collated the outputs from otherprojects; resource data (Section 3.1), retained treeinformation (Section 3.2), net harvestable area(Section 3.3), site productivity (Section 3.4) andavailable volumes by species (Section 3.5).

The collected data was compiled into a MicrosoftAccess © database.

5.1.3 Results

Multi-aged forest

The database stores:

n the sawlog and pulpwood volume per hectarefor each coupe (refer Section 3.1);

n the net harvestable area for each coupe (referSection 3.3).

This information is then multiplied together todetermine the total volume of sawlogs andpulpwood for each coupe. This occurs as part ofthe yield estimating procedure (refer Section 5.2).

A separate analysis further breaks down the totalvolume into diameter and species class (referSection 3.5). Coupes can be amalgamated into

management sections for modelling of supplyzones.

Regrowth forest

The database classifies all regrowth into strata asdescribed in Section 4.1.3).

Each regrowth coupe was assigned to a stratum.In instances where multi-aged forest coupes hadbeen partially affected by fire, these fire-affectedportions of the coupe were assigned to a regrowthstratum.

For each stratum a yield table was generated fromSTANDSIM (refer Section 4.1.3, Table 4b for theyield tables). The yield tables predict the volumeby product for predicted years of harvest.

For example, a coupe that falls within stratum 1 isa 1980 fire regrowth coupe that has lower siteproductivity and is located on the coast.According to the yield table, this coupe wouldreceive its first thinning at age 35 (ie in 2015)producing a predicted 100 t/ha of pulpwood. Thesecond thinning at age 50 (ie in 2030) is predictedto produce 5 m3/ha of sawlogs and 80 t/ha ofpulpwood. The regeneration harvest at age 70 (iein 2050) is predicted to produce 50m3/ha ofsawlogs and 120 t/ha of pulpwood.

Net harvestable area and net thinnable area werealso derived for each regrowth coupe (referSection 3.3).

The volume by product by predicted year ofharvest for each regrowth coupe was calculated bymultiplying the volume per hectare at thepredicted time of the operation from the yieldtable by the net thinnable area of regrowth. Thisoccurs in the yield estimating toolbox (seeSection 5.2).

Sampling of the regrowth resource is an ongoingprocess and plot measurement is continuing sothat more detailed information will be availablefor post RFA reviews.


24

In summary, sawlog and pulpwood volumes arecalculated for all multi-aged coupes and predictedfor all regrowth forest coupes. The yieldestimating and scheduling toolbox (refer Section5.2) analyses this data though a set of queries inorder to calculate a pattern of harvesting toproduce a non-declining annual volume ofsawlogs.

Metadata sheets have been prepared for all data.

5.2 YIELD ESTIMATOR ANDSCHEDULING TOOLBOX

Development of a yield estimating andscheduling toolbox


5.2.1 Objectives

The objective of the project was to integrate theforest resource data and growth models into apackage which estimated yields from each coupein the forest and enabled a program of futureharvesting to be simulated, with volumesscheduled during each year to ensure a predictableresource supply to wood-based industries duringthe RFA period.

5.2.2 Methods

Yield estimator

The yield estimation procedure is described inSection 5.1.

As the multi-aged forest is assumed to have zeronet sawlog growth, the estimated volumes onlyrequire scheduling (ie they are assumed toproduce the same volume if harvested now as ifharvested in the future).

The regrowth forest uses STANDSIM to estimateyields at various ages and types of operations.This is described in detail in Chapters 4 and 5.

Yield scheduler

The yield scheduler combines the wood flow fromthe multi-aged forest and regrowth forests(including thinnings) into a nominal annual woodflow.

Multi-aged forest

The multi-aged forest was scheduled to meetrequirements until sufficient sawlog volumesbecome available from the regrowth forests.Initially multi-aged forest coupes were scheduledup to 2019. Each of the uncut multi-aged forestcoupes were given a nominal year of harvestfollowing these principles.

1. Dispersion geographically. The number ofcoupes required for a particular year wasproportioned for each Management Section.

2. Dispersion through time. The uncut multi-aged forests coupe were scheduled so the timesince logging of the alternate coupes was keptas great as possible, generally greater than 15years.

Regrowth forest

The regrowth sawlog wood flow for the period2016 to 2040 is drawn primarily from harvestingof 1952 and 1960s fire regrowth and secondthinning of later fire or logging regrowth. Theseareas are scheduled according to the year woodbecomes available from thinning or regenerationharvest according to the yield tables derived fromSTANDSIM as described in Chapters 4 and 5.

5.2.3 Results

A series of tables were built into the EdenFRAMES Microsoft Access © database. Thiscomprises a set of base data tables and a range ofqueries to generate the various components ofwood flow and to join them together. Adescription of the sets of queries is providedbelow:

Multi-aged forest - This query predicts the totalsawlog volumes (m3) and total pulp volumes (t),disaggregated by species and size class, for eachunlogged coupe.

Fire regrowth - This query calculates thepredicted future sawlog and pulp volumes fromareas of regrowth resulting from past wildfireevents.

Harvesting regrowth - This query calculates thepredicted future sawlog and pulp volumes fromareas of regrowth resulting from harvestingevents.

Regrowth summary - This query compiles thefire and harvesting regrowth volumes into onetable.


25

Eden CRA Region summary - This querycompiles the regrowth and multi-aged forestvolumes into one table.

Scenarios - This query excludes coupes andcompartments for different scenarios/options andcalculates the remaining volumes for each yearscheduled by operation type.

A detailed outline of the Eden FRAMESMicrosoft Access © database and queries iscontained in Appendix 7.

The outputs of the yield scheduling for Edenshow:

n when coupes should be/are best harvested infuture years;

n expected yield by species, volume and sizeclass for the multi-aged forest;

n volume by products available from theregrowth forest; and

Analysis of the annual flows from the schedulerindicated that wood flows were not even. Thisuneveness reflects the wildfire and logginghistory, but particularly of large fires in the years1952, 1968, 1972 and 1980 that caused significantareas of forest to be converted to regrowth. Inparticular a large volume of 1952 fire regrowthfinal sawlog harvest was scheduled in 2020.Peaks of woodflow from both thinning andregeneration harvesting of fire and loggingregrowth are not operationally practical asphysically harvesting this much volume in anyone year is not possible.

An investigation was undertaken to determine thefeasibility of bringing forward harvest of the 1952regrowth. The STANDSIM outputs indicate thatat the nominal rotation age of 70, sawlog volumeper hectare is increasing rapidly and cutting thestands early will reduce long term sawlog yields.Reducing the rotation age from 70 years wouldreduce sawlog volume by 5% per year andharvesting earlier than 65 years greatly acceleratesthis rate of reduction (B.Incoll pers comm.).

For this reason, the minimum rotation age for1952 regrowth was set at 64 years for those standsthat have already received a first thinning and 65years for stands that will be thinned before theyear 2000. This assumption means that there issome guarantee the sawlogs demonstrably exist,rather than needing to develop from a thinningoperation has not yet been conducted. Theselimits mean that the earliest harvest years for

1952 fire regrowth is 2016 for thinned and 2017for stands currently unthinned.

A smoother has been developed for the purpose ofinforming FORUM (refer Section 6.1.2), toallocate harvesting years to individual multi-agedforest coupes to provide an indicative “order ofworking”. This produces a pattern of harvestingeach year, which facilitates assessment of howwell harvesting has been dispersed geographicallyand through time within the intent of the alternatecoupe harvesting system.

27

6. OUTPUTS

6.1 KEY OUTPUTS OF EDENFRAMES

Key outputs of Eden FRAMES are:

n resource data in terms of standing volume andvalue (the latter via current log price);

n prediction of future yields for multi-aged andregrowth forest; and

n gross and net harvestable areas of forest underspecified prescriptions.

6.1.1 Projected volumes

The timber currently available for harvesting isshown in Table 6a. The volumes have beencalculated using the current land tenure, andapplying current management prescriptions but donot take account of fauna moratorium areas whichwill be considered as part of the RFA.

There is potential impact from wildfire which wasnot taken into account in Tables 6a and 6b andapplying current management prescriptions. Table6b shows total harvestable volume for the multi-aged forest by species group and diameter class.This data has been extracted from the EdenFRAMES database.

The FRAMES Technical Committee recognisedthat the certainty of these estimates varies withtime period. The relevant time periods are:

n From 1997 to 2019 estimated yields are drawnprimarily from the multi-aged forest;

n From 2020 to 2040 estimated yields arelargely drawn from fire regrowth and thinningof logging regrowth.

There are many uncertainties in estimating timberyields from forests with high inherent variability,such as the Eden forests. In considering thescenarios, FRAMES estimates of mean sawlogyields per hectare from the multi-aged forest forthe period 1997-2020 have confidence limits of+/- 30%. Although some of this variabilitycancels out as compartments are aggregated toproduce an annual yield, that annual yield maystill have reasonably wide confidence limits.There is some uncertainty regarding areas and

growth rates for regrowth stands harvested up to2040 and no confidence limits can be assigned.

Beyond 2040 estimated yields are drawnprimarily from logging regrowth which has notbeen well sampled. Estimates are indicative only.

Based on this information allowable cutcalculations can be estimated for the currentplanning horizon (Forest Essentials Pty Ltd1997b).

6.1.2 Integration and development ofscenarios for Eden

As various scenarios were developed during theintegration phase of the CRA process, EdenFRAMES was used to predict the sawlog andpulpwood yields during the periods 1997 - 2019and 2020 - 2040, for the areas of forest availablefor harvesting and for the then currentprescriptions. These outcomes are provided inrelation to each of the scenarios.

Further analysis was applied to the FRAMES databy the ESFM Technical Committee to simulatechanges to forest management practices forparticular scenarios. These are reported for eachscenario in the reports “Towards an RFA forEden” and in ESFM project reports (Project Area5: Management Options and Scenarios toGenerate ESFM Targets for the Eden RFA).

For each scenario developed in the integrationphase of the CRA process, the breakdown ofvolume shown in Table 6b was generated, bylocation within the Eden CRA Region and foreach year of supply, and was supplemented byvolume of regrowth sawlogs and pulpwood. Thisdata was then provided to the Economic andSocial Technical Committee for use in the ForestResource Use Model (FORUM), to alloweconomic modelling of wood based production tobe carried out.

Alternative conservation protocols wereconsidered after integration for various scenarios.These will impact on available volumes and arediscussed in the report “Towards an RFA forEden”.


28

TABLE 6a ANNUAL WOOD RESOURCEVOLUMES FROM NATIVE STATE FORESTS IN

THE EDEN CRA REGION

Period Current land tenure includingany areas from which harvestinghas been deferred

Sawlog1 (m3)per annum

Pulp1 (t) perannum

1997-2019 28 300 403 000 2020-2040 28 300 444 000 2040+ >30 000 ~500 000 1 From the multi-aged forest and regrowth, includingthinning

TABLE 6b: TOTAL PREDICTEDHARVESTABLE VOLUME (m3 ) FROM THEMULTI-AGED FOREST BY SPECIES AND

DIAMETER CLASS

Sawlog diameter class (centrediameter under bark)

Species Group <40 cm (m3) ≥40 cm (m3) Silvertop Ash 10 000 134 000 Stringybarks 15 000 149 000 Messmate 3 000 56 000 Spotted Gum 3 000 12 000 Monkey Gum 5 000 68 000 Tablelands spp1 3 000 64 000 Specials2 1 000 3 000 Other 3 000 26 000 1 Mountain gum, brown barrel, shining gum, mannagum 2 Grey box, grey gum, ironbark, white box

6.2 AFTER THE REGIONALFOREST AGREEMENT

A number of issues are not likely to be addressedbefore the Eden RFA will be signed. The issuesmainly affect parts of the resource which wouldonly become available for sawlog harvestingbeyond around 2040 and include:

n large parts of the regrowth resource (mainlylogging regrowth) have not had adequate plotdata collected to enable estimation of currentstand variables;

n an objective statistical basis for growthpredictions for logging regrowth needs to bedeveloped;

n information on the growth of non-silvertop ashdominated State forests is very limited.

n the level of defect in regrowth sawlogs fromfire and logging regrowth needs to be betterunderstood;

n thinning and regrowth models need furtherrefinement.

The FRAMES Technical Committee thereforeconsiders that the data available beyond around2040 is not adequate to define the progressivelevels of sawlog yield that are available beyond2040.

Further, the Technical Committee highlights theneed for improved inventory and growthinformation on multi-aged and regrowth forestresources that become available before 2040.

In summary, Eden FRAMES identified a need formore inventory data for future evaluations of themulti-aged and regrowth resource. The FRAMESTechnical Committee considers an inventorysystem is necessary to more accurately predictwhen fire and logging regrowth will becomeavailable as sawlogs. An inventory of multi-agedand regrowth forest needs to be completed by2001. Initial results of the inventory should bereviewed two years after signing the RFA and thefive yearly RFA review should include a detailedreview of sustained yield calculations using thatinventory.

The FRAMES Technical Committee is of theview that considerable work on strategicinventory, development of growth models andyield scheduling should be continued by StateForests in the years following the signing of theEden RFA, and should be a requirement of theRFA. In particular, the over-estimation ofclearfall volume reported in Section 4.2.3 needsto be resolved in the first five years of the RFA.

Work should be carried out according toprinciples to be agreed by the TechnicalCommittee. This should be consistent with thegeneral FRAMES structure used elsewhere in theState which includes optimising timber flowswithin an ESFM framework. A commitmentshould be made by State Forests to fund thiswork.

These issues are addressed in more detail in theFRAMES Technical Committee submission tothe CRA/RFA Steering Committee immediatelyprior to negotiation. This submission is attachedas Appendix 8.

29

REFERENCES

Bi H. 1994 South-east Regrowth Forest Growth and Yield Modelling Design, Methods and Progress.Research Paper 24, State Forests of NSW.

Bi H., Chikumbo, O. and Jurskis V. (1997) Yield equations for regrowth forests regenerated from fire on thesoutheast coast of New South Wales. Draft report to the FRAMES Technical Committee.

Bi H. (unpublished) Total stem volume and bole volume equations for regrowth eucalypts of the SouthernRegion. Internal report to the Southern Region, State Forests of NSW.

EPA/SFNSW/DLWC, August, 1997. Soil erosion and water pollution hazard assessment for loggingoperations.

Forest Essentials 1997a Validation of Eden Wood Resources Data. A report to the Resource and AssessmentCouncil New South Wales and State Forests NSW.

Forest Essentials 1997b Modification of the STANDSIM model for the Eden Management Area. A report tothe Resource and Assessment Council New South Wales and State Forests NSW.

FRAMES Technical Committee 1997 Forest Resource and Management System Technical Framework

Opie, J.E.O. 1972 STANDSIM A general model for simulating the growth of even-aged stands. ThirdConference, Advisory Group of Forest Statisticians, IUFRO, Paris 1970 In Institute Nationale de laRecherche Agron. Publ. 72-3: 217-39.

Rosewell, C.J. & Turner. J.B, CALM, November 1992, Rainfall Erosivity in New South Wales.

RACAC 1996 Draft Interim Forest Assessment Report. New South Wales Government.

SFNSW / NPWS, November, 1996, Conservation Protocols for timber harvesting on State Forests for theduration of the IFA decision.

SFNSW, November 1994, Proposed Forestry Operations, Eden Management Area, Environmental ImpactStatement.

31

APPENDIX 1 – IFA WOODRESOURCE STUDYSUMMARY

The Interim Forest Assessment (IFA) was ascientific assessment of public forested areas inNew South Wales conducted by the Resource andConservation Assessment Council (RACAC) forthe New South Wales Government. The aim ofthe IFA was to identify, on a regional basis, thoseforested areas that may need to be deferred fromlogging because of their possible inclusion in aComprehensive Adequate and Representative(CAR) reserve system. It was a precautionarymeasure to avoid excluding conservation optionspending a more detailed assessment process (theCRA), while at the same time considering theneeds for wood production.

The IFA Wood Resources Study was used toprovide an assessment of the various outcomes fordeferred forest areas on the ongoing supply oftimber to industry.

The objectives of the Wood Resources Studywere to:

n define forest characteristics, particularly thoserelating to forest type and potential for growthand future yield, at the scale of sub-divisionsof forest compartments;

n use forest characteristics to define futuregrowth patterns and the options for futureyields;

n provide the capacity to attach measures offorestry importance to compartments and sub-compartments;

n develop the capacity to include (sub-)compartments in a deferred forest area andassess the implications of that inclusion onavailable yields over time;

n incorporate the mapping of disturbance history(originally a separate IFA project); and

n develop the capability to display forestcharacteristics (both originally measured and

subsequently assessed) on a map (either on acomputer or printed) (RACAC 1996).

Wood resources database

To enable the potential timber yields of a definedarea of forest to be predicted by a computermodel, forestry compartments were subdividedinto ‘resource units’. Each resource unit wasrelatively homogeneous in respect of tree speciesmix, standing harvestable volume and standstructure.

Growth and yield modelling

Growth and yield models were developed andapplied to resource units.

The rate of growth for given yield associationsbetween harvests was derived from analysis ofpermanent growth plot (PGP) monitoring carriedout by SFNSW over many decades.

The growth modelling process provided thevolume per hectare of high value, low value andpulpwood grade logs. It did not separately predictvolume of high value specialty products such aspoles, piles and girders. The volume estimationscould also be linked to forest types to estimatevolume by tree species.

Harvesting options

To enable an assessment to be made of the bestpattern and mix of resource units to harvest overtime, a range of harvesting options were generatedfor each resource unit.

This process produced a set of harvesting optionsfor each resource unit which essentially vary thefirst year in which harvesting could occur andassess the growth and yield from that pointonwards.


32

Value calculation

To calculate the value in commercial terms of aharvesting option, the net present value (NPV) ofthe set of yields nominated for a given harvestingoption was calculated using a seven percentdiscount rate of return.

Dollar values were attached to each yield overtime, with the resulting total returns fromharvesting discounted to current dollars to providea value for each harvest option.

Yield scheduling

A linear programming (LP) model was used to:

n maximise the value of the set of harvestingoptions selected; and

n ensure that the volume to be harvested in anygiven year is sustainable.

For each harvesting option a NPV was calculated.The model maximises the sum of NPVs for theharvesting options and resource units harvestedeach year.

The model was constrained so that the volume ofhigh value logs harvested in any year could not beless than the volume harvested in the previousyear. Hence, the volume of high value logsproduced over time cannot decline; in otherwords, the volume available for sale each year issustainable.

Because the IFA negotiations focused on quotasawlogs, the sustained yield of quota sawlogs wascalculated by filtering out quota sawlogs from themodel’s predicted sustainable yield of high valuelogs based on historical sales data.

The LP model was not intended to define an‘order of working’, even though harvesting yearsare assigned to each resource unit as an output ofthe model. The output identified a group ofresource units which contribute most to meetingthe sustained yield requirements over a specifiedperiod. Once a set of compartments or resourceunits is included in a deferred forest areacalculation, a re-run of the model identified newresource units possibly harvested in differentyears which contribute most to meeting thesustained yield requirements for the period.

33

APPENDIX 2 - VALIDATIONOF EDEN WOODRESOURCES DATA

35

APPENDIX 3 – IFARETAINED TREE REPORT

37

APPENDIX 4 – GROWTH OFSAWLOGS IN MULTI-AGEDFOREST AT EDEN

39

APPENDIX 5 -MODIFICATION OF THESTANDSIM MODEL FORTHE EDEN MANAGEMENTAREA

41

APPENDIX 6 – YIELDEQUATIONS FORREGROWTH FORESTSREGENERATED FROMFIRE ON THE SOUTHEASTCOAST OF NSW

43

APPENDIX 7 – EDENFRAMES USER GUIDE


44

APPENDIX 8 – FRAMESSUBMISSION TO CRA/RFASTEERING COMMITTEE

eden forest resource and management system report · eden forest resource and management system...

Documents