The development of an effective methodology to capture forest type, condition and volume data for privately managed forests

Cris Brack, Ian Ferguson, Claire Howell, Jerry Leech, Phil Tickle and Ian Wild

Report of a Scoping Study for the Forest and Wood Products Research and Development Corporation on behalf of the School of Forestry, University of Melbourne.

Published May 1998

Executive summary

This report was commissioned by the FWPRDC to develop effective methodologies to capture forest type, condition and volume data for privately managed forest throughout Australia. The objectives of the study were to identify current techniques, scope the development of a cost-effective, standard method suitable for use throughout Australia and, where new methods cannot be developed, identify research priorities.

Australia's forests cover a wide range of categories. Privately managed forests, as indicated by the brief, can be defined as privately owned and publicly owned but privately leased. These forests total approximately 69 per cent of Australia's total forest estate. Adequate information is available for the private plantation resource but little is known about the privately managed native forests. Appropriate and consistent data at the regional level would greatly assist planning activities at both the regional and national level. For the relatively few individual growers collecting inventory data, use of a standard protocol would enable integration of the local data into the regional and national level frameworks. A more consistent approach to inventory across the three levels would therefore be advantageous even if varied in intensity. However it is recognised that a practical application of the proposed protocol may be difficult and a number of issues need to be worked out to ensure that it can be applied.

In developing a methodology, the measurement data and/or determination of protocols should be consistent with the Montreal Criteria and Indicators. Some of the indicators are imprecise in their definition and need refinement at less than regional level. Furthermore, this scoping study does not address the issues required for monitoring the Montreal Criteria, growth being an important omission, which is vital in the determination of sustainability.

Productivity, an indicator of growth potential, is usually calculated from a number of measurable parameters and its definition, determination and subsequent consistent mapping depend on a whole range of site and stand characteristics such as soils, species, past treatments. Similarly, biometric techniques, especially those for estimating assortment volumes by product and quality, need considerable research. All of these issues are extremely important both for owners and for all levels of management but are beyond the bounds of this study, which is involved with inventory methodology. Therefore, it is strongly recommended that a separate study be initiated to investigate these areas of research and development.

Management intent is also an important issue for any inventory. Management intent of privately managed forests varies considerably with size of the resource, the owners' objectives, and prevailing government regulations. The inventory methodology finally adopted will therefore have to include a survey of management intent in order to translate the estimates of standing volumes into volumes likely to be made available for sale at a particular point of time.

In general, inventory methods applied to privately managed forests have been adapted from those used by Government agencies.

Those used for privately managed plantations appear adequate for the larger organisations and/or cooperatives but are proving expensive to maintain. Furthermore, they are not readily applicable for small privately managed estates.

Very little inventory is carried out relative to the total area of privately managed native forests. However, broad scale (1:100,000) mapping of forest areas has been carried out and does address some of the regional and national needs.

Area and volume estimates are available for some privately managed forests and plantations, but are mostly restricted to the larger plantation or native forest estates or co-operatives.

Operational inventories of privately managed native forests at the more intensive level are generally opportunistic, fragmented varied in design and protocols, and of varying levels of precision. They are therefore limited in their use at the regional and national levels.

Existing methodologies relevant to this study are those developed for the CRA/RFA process in South East Queensland and the private forest inventory and management intent survey used in Tasmania.

One common inventory procedure cannot be implemented for all privately managed forests throughout Australia due to organisational differences, historical reasons, the intensity and related cost of inventory work for different objectives, and the differences in the owners intents.

Thus a general protocol is suggested to provide relevant information and address both the operational and regional needs for sustainable forest management. A systematic and consistent storage of information, at regional and national levels is required, with appropriate indices of precision. Given such a system and protocol, the design of inventories can be tailored to meet the particular objectives of the owner as well as providing appropriate data at regional and national levels.

The protocol deals with three distinct stages

• mapping of forest areas 
• attribute data collection using both remote sensing and field-based sampling, and 
• surveying of management intent.

To gain a more quantitative understanding of Australia's forest resource, spatial and non-spatial data needs to be accumulated for the entire forest estate for a number of attributes including; forest cover, land tenure, forest type, age (age class), stocking, height, standing volume, and condition.

A multi-level approach needs to be adopted based on the use of both remote sensing and field-based sampling, subject to a minimum sampling level considered acceptable for regional and national purposes.

For the entire forest estate, spatial and non spatial data needs to be accumulated and then further subdivided according to owner's long term objectives and short term intent. The following subdivision is suggested.

• Non-forest or forest land, with forest land further categorised by forest type and condition. 
• Apparent objectives. For example: government including reserves, privately managed commercial, privately managed non-commercial. 
• Management intent. For example: commercial harvesting intended in next five years, commercial but withheld from harvesting for next five years.

This subdivision will need refinement in the light of a separate survey to ascertain apparent objectives and management intent, but the principles will be apparent from the examples above.

Area data can be collected through a variety of tools ranging from satellite imagery to ground based survey. Research is required to address their integration in forest planning, how the measurements should best be conducted, and to define the goals for spatial precision and freedom from bias.

GIS is recommended for the management of large spatial data sets, especially for growers or cooperatives with estates greater than 1,000 hectares, and for regional and national level data. However, standards need to be specified concerning scales for data entry to ensure spatial compatibility when aggregating data to the regional and national levels.

A protocol that facilitates the collection of standard resources data for different management planning levels needs to be developed which would encourage owner input to the scheme even where the owners' management intent is to 'do nothing'.

A number of statistical techniques are available for the efficient collection of attribute data. A mutli-stage approach using remote sensing is likely to be useful for stratification. Different but appropriate technologies for sampling could be employed at each stage, for example satellite imagery for the first stage and aerial photographic interpretation for the second stage.

Multi-stage sampling has the advantage of concentrating work on specific samples of interest. The technique is particularly useful where primary sampling units vary in importance (for the parameter/s being estimated) and allows further sampling to be concentrated in areas of greater significance with respect to the parameters being estimated. The choice of technology and sampling intensity at each stage is related to available funds, scale and management intent.

Education and promotion of the overall scheme is of paramount importance to encourage local inventory programs and will require considerable funding and time. Costs can be reduced by targeting existing grower cooperatives (for example the Regional Plantation Committees) and encouraging the establishment of new ones.

Some economies of scale can be effected through the NFI, for example the purchase of regional or multi regional remote sensing data or the provision of suitable processing systems (spatial and attribute) where individual small owners are unable to provide their own systems.

Because Australia has a large area of forest it is not considered practical to implement separate inventories at all planning levels. An approach is required that integrates national, regional and operational forest management planning. A multi level approach is recommended that has the objectives of providing

• At the operational level, sufficient detail for an owner to meet the basic requirements for regional sustainable forest management, and, 
• At the regional or at a broader scale, consistent and suitably precise information, relatively free of bias, for regional sustainable forest management.

The protocol should also apply to government managed forest, including reserves and parks, which have sometimes been neglected in the past.

It is recommended that a pilot study be set up, to build on the findings of this study and more thoroughly design a mapping and sampling protocol for operational, regional and national inventory.

Research needs

The following research needs have been identified but are not in any order of priority because different aspects are likely to be funded and carried out by different agencies.

More efficient and cost effective methods for stand stratification are needed. Remote sensing technologies offer considerable scope but the following issues needs to be addressed

1. More efficient and cost effective methods for stand stratification are needed. Remote sensing technologies offer considerable scope but the following issues needs to be addressed
   • Remote sensing and GIS technologies need to be integrated fully into planning
   • No single remote sensing platform is capable of resolving inventory requirements (FWPRDC Application of Remote Sensing to Forestry study), and thus greater attention needs to be given to combining these technologies with other methods of stand stratification, including the possible integration with environmental data such as terrain, climate and soils, and
   • The use of radar technology to gather cloud-free imagery for forest mapping. 
2. Given the fragmented nature of the private resource, research is needed to investigate the use of new forms of remote sensing to supplement and/or replace data that has traditionally been collected by field sampling. Technologies such as the use of digital frame cameras, digital video, laser profiling and differential global positioning systems, all offer considerable scope. 
3. Reporting of the errors of estimates and the level of accuracy (or precision) needs to be standardised, both for spatial and attribute information. 
4. Research is needed into the mapping of forest communities, especially for biodiversity management. 
5. At the operational level, there is a need for product quality information. Research is needed to improve the efficiency of attaining product and product quality estimates. 
6. Field sampling techniques should continue to be developed and tested, including continuation of the development of p-3P and centroid sampling. 
7. The design of surveys to ascertain owners' objectives and management intents needs research. These are social surveys and need sociological input. 
8. Research is needed to develop generic growth models. The availability of data and the balance between different modelling techniques in order to meet completely different objectives needs to be considered in selecting appropriate models for different regions. 
9. A consistent approach to mapping plantation productivity is highly desirable at least on a region by region basis. 
10. Local cooperatives need to urge owners (in particular owners of smaller forest estates) to undertake more intensive inventory and management of their forest.

In addition, there are serious concerns about computing technology.

The development of computer systems is expensive. Are there better ways of minimising this expense through collaborative development? Suitable systems are generally not available for use by the small private owner. Many larger private owners are using 'out-of-date' systems because of the cost of updating and a hiatus in the availability of 'off-the-shelf' systems.

Thus, there is a need for

• a generic base level system that can provide all forest owners and managers with a method of calculating inventory and providing the generic processed information that is required at regional/national level, and 
• a system to manage the processed information across all forested land areas.

This will require substantial funding but would provide information currently not available for sustainable forest management and regional planning.

Table of contents

Executive Summary

Research Needs

1. Introduction

1.1 Preamble
1.2 Montreal Criteria and Indicators
1.3 Terminology and Definitions

2. Current situation

2.1 Current methodologies and technologies

2.1.1 Tasmania
2.1.2 Victoria
2.1.3 South Australia
2.1.4 Queensland
2.1.5 Western Australia
2.1.6 New South Wales
2.1.7 Summary Comments

2.2 Data processing for private forests

3. Evaluation and review of current practices

4. Management Objectives and Intentions

5. Common Elements to Future Inventory Design

5.1 The changing context of inventory information
5.2 Maintaining regional and national data
5.3 Inventory design

5.3.1 Area classification
5.3.2 Spatial data
5.3.3 Attribute data collection
5.3.4 Multi-stage inventories

6. Data Processing and Implementation

6.1 Data processing
6.2 Implementation

7. Differentiating between Native Forests and Plantations

8. Recommendations and Research Needs

8.1 Research Needs

9. Literature Cited

Appendix I : Montreal Criteria and Indicators considered relevant to this study
Appendix II : Consultation