Australian Forestry
Australian Forestry is published by the Institute of Foresters of Australia (IFA) for technical, scientific and profes sional communication relating to forestry in Australia and adjacent geographic regions. The views expressed in this journal are not necessarily those of the IFA. The journal is included in the Register of Refereed Journals maintained by the Australian Government Department of Education, Science and Training, and as from the first issue in 2008, in Science Citation Index Expanded (SCIE). Managing Editor: Dr Colin Matheson Production Editor: Mr Alan Brown Editorial Panel: Dr Stuart Davey Dr Brian Turner Mr Jack Bradshaw Mr Neil (Curly) Humphreys Dr Ian Bevege Dr John Herbohn Dr Grant Wardell-Johnson Dr Humphrey Elliott Dr Russell Washusen Contributions Contributions to this journal are sought covering any aspect of forest ecology, forest management, forest policy and land use related to Australia and the South Pacific region. Contributions related to the performance of Australian tree genera elsewhere in the world are also welcome. Instructions to authors are given inside the back cover of each issue. Contributions should be sent to the Executive Director at the address below. Past issues Full-text copies of the journal for the years 2001–2006 are available to the public on our website, http://www.forestry.org.au/ifa/f/f15-ifa.asp Journal subscriptions 2010 A$268 including GST within Australia A$298 per year in all other countries The above prices are for hardcopy. For information about options for electronic access and pricing, please check the website or contact the Executive Director at the address below. All correspondence relating to subscriptions should be addressed to: Executive Director, The Institute of Foresters of Australia, PO Box 7002, Yarralumla, ACT 2600, Australia. Phone: 61 2 6281 3992. Fax: 61 2 6281 4693. Email: [email protected] Web: http://www.forestry.org.au Cover The front cover features Corymbia hybrids and their parents Corymbia torelliana (cadaga) and C. citriodora subsp. variegata (spotted gum). The hybrids are promising plantation varieties for subtropical and tropical environments in the 700 to 1300 mm MAR zone. The development of these hybrids was described by David Lee in Australian Forestry Vol. 70, pages 11–16. David also selected the photos on the cover. Main photo (No. 6): In the foreground is one replicate (a five-tree row plot) of a six-year-old Corymbia hybrid family, with a mean height of 14.8 m and a diameter at breast height over bark (dbhob) of 13.8 cm, on a black earth soil near Gympie, Queensland. Behind David (standing) is a C. torelliana family with a mean height of 7.7 m and a dbhob 8.5 cm. The C. citriodora subsp. variegata controls, which were badly damaged by frost, had a mean height of 7.2 m and a dbhob of 5.8 cm at the same age. (Photo courtesy of Elizabeth Lee.) Photo 1: Corymbia citriodora subsp. variegata buds developing on a grafted tree in a clone bank at Gympie in January 2008. (Photo courtesy of David Lee.) Photo 2: Grafted, potted clone bank established to aid production of Corymbia hybrids. These C. citriodora subsp. variegata clones are 18 months old and already yielding good quantities of pollen. (Photo courtesy of David Lee.) Photo 3: Healthy three-year-old Corymbia hybrids (on left of photo) with frost-killed C. citriodora subsp. variegata on the right. (Photo courtesy of David Lee.) Photo 4: One of many selected Corymbia hybrid trees harvested at age 3.6 years for evaluation of a range of wood properties. Coppice that developed on the stumps has been propagated and clonal testing of the hybrids is underway, with two-year-old clonal trials continuing to demonstrate the early good potential of the hybrids. (Photo courtesy of David Lee.) Photo 5: Ends of butt logs from selected 3.6-year-old Corymbia hybrids (diameter range 12–23 cm). These logs were used to study the wood properties of early-age Corymbia hybrid selections. Properties evaluated included sawing, splitting and pulping. (Photo courtesy of David Lee.)
ISSN 0004-9158 ACN 083 197 586
Australian Forestry
The Journal of the Institute of Foresters of Australia
Volume 72 Number 4, 2009
ISSN 0004-9158 blank page
blank Ron Adams 147
Guest editorial
Forestry and wood products research and promotion moves with the times
Australia’s dynamic forestry and wood products industry is roles of the national industry representative bodies of NAFI, A3P, embracing change as it responds to emerging opportunities and Australian Forest Growers (AFG), Engineered Wood Products economic uncertainties. Research and promotion for our industry Association of Australasia (EWPAA) and the Timber Importers are rising to meet these challenges. As the world grapples with Federation. It became clear that evolution into a privately owned greenhouse issues, we are taking a more muscular approach to research and services company was an essential step. With the publicising wood’s many advantages. exception of pulp and paper, FWPA spans all forest industry sectors, including imports. As an engineer with over 40 years experience in this industry in Western Australia, and a family involvement spanning three The transition reflected the industry’s desire to undertake generations, embracing the hardwood and softwood sectors, coordinated promotion of forest and wood products in an I have had the privilege of serving on the boards of Bunnings increasingly competitive marketplace, and to more effectively Forest Products and Wespine (with its links to Wesfarmers and promote its environmental credentials. Developing an industry WESFI (now Laminex)), National Association of Forest Industries promotion campaign for timber and wood products has therefore (NAFI), and Australian Plantation Products & Paper (A3P). In been high on our agenda. that time I have seen enormous changes: We aim to improve the competitiveness and sustainability of the • Forty years ago green hardwood was the norm for timber- Australian forest and wood products industry by investing in framed houses in many Australian states. Framing is now innovative and relevant R&D, and promotion of the industry’s dominated by softwood. products, services and values. We undertake industry-level • Not so long ago, plantations were mostly established on cleared activities where a collective approach delivers more effective and natural forest. They are now typically on former agricultural valuable commercial outcomes than individual action can. And land. we deliver services that support the industry’s capacity as a major • The popularity of managed investment schemes over the past exporter and its vital role as a mainstay of regional communities ten years has seen a quadrupling of the area under hardwood in many parts of Australia. plantations to 900 000 ha. As a statutory authority, FWPRDC was controlled by the • Forestry has now intensified to a small number of regions across Commonwealth Public Service Act. It maintained industry contact the country. It is significant that Albany (in WA) has become a through NAFI and A3P. In contrast, FWPA has direct interaction major forest industry centre; not many years back the forestry with its member companies, and a contractual relationship with presence there was minimal. the Commonwealth Government. We are not alone — about half • Larger softwood mills are now processing around 400 000 to of Australia’s ‘rural’ R&D corporations have become private 800 000 cubic metres annually, three to four times what an companies. average mill was handling when I first entered the industry. All industry members who pay domestic processor levies, • And we are rapidly adopting forest certification — most large grower levies or wood import charges are eligible to apply for industrial forest growers are already certified or are well on free membership of FWPA. I think it’s significant that the state the way to certification. This is possible and desirable because government forest services chose to be levy payers. We have our industry needs to continue to address many issues around revised the structure of the industry advisory committees that sustainability. provide feedback to the board and management on the research programs to better reflect the wood value chain. Our focus on Over the past few years I’ve been able to assess industry progress the value chain also reflects our stronger commercial focus and from a unique vantage point: I’ve had the immensely satisfying emphasis on services and outcomes for the industry. responsibility of chairing the Forest and Wood Products Research and Development Corporation (FWPRDC) and also more recently its New research directions successor, Forest and Wood Products Australia Limited (FWPA). The aggregate annual R&D expenditure across the whole industry I believe the formation of FWPA has opened a new chapter for is about $100 million, of which our share is an investment of our industry. As an industry-owned services company established $7–8 million. On the face of it we are not a big player, but we in September 2007, it is an unlisted public company. In contrast, invest strategically in projects with the potential to influence FWPRDC was a Commonwealth statutory authority that had been research, development and adoption by other organisations. a key investor in forest and wood products industry research and We seek to expand our influence in such a way that the benefits development since 1994. will be multiplied by partnering organisations, yielding greatly In the two-year lead-up to the change we consulted widely with expanded gains for modest outlays. Wherever possible we choose stakeholders to identify expectations and requirements of an ‘cornerstone’ activities with a multiplier effect. Partners could industry services body with an expanded role beyond research and include CSIRO or an international consortium. development. This role was supportive of the critically important
Australian Forestry 2009 Vol. 72 No. 4 pp. 147–148 148 Forestry and wood products research and promotion moves with the times
With our industry-wide perspective we can contribute solutions The message that wood is better for the environment than other to critical issues using the expertise of our board and the advisory materials will be especially important for people who influence the committees. Examples of the lead role FWPA and its predecessor choice of building materials. The CRC for Greenhouse Accounting have played include the following: developing a comprehensive calculated that by choosing wood products wherever possible in life-cycle inventory for the industry for use in life-cycle assess house construction, emissions equivalent to up to 25 tonnes of ments; a range of tree breeding and improvement programs (such carbon dioxide per house could be saved in Australia. as the Juvenile Wood Initiative, Treeplan and Seedplan); and the scientific protocols and principles underlying the Australian Wood. Naturally Better.™ is much more than an advertising Forestry Standard. campaign. By the time you read this close to 400 organisations will have signed up to be licensed to use the Wood. Naturally Over the past 12 months we have been reshaping and consolidating Better.™ logo. This acts as a multiplier — driving awareness of our research investment priorities to avoid the problem of the brand and our message to benefit both individual licensees spreading our funds too thinly. Thirteen research priorities have and our industry as a whole. been agreed: five deal with forest issues, three with processing and five are concerned with market access and consumption. Another area in which Wood. Naturally Better.™ is active is in identifying and developing strategic partnerships. These may involve sponsorships, joint activities or other actions which benefit Advisory groups restructured both parties. Examples of this are relationships with Archicentre We have restructured our four industry advisory groups to (the Australian Institute of Architects) and Planet Ark. encompass the forest and wood products industry value chain We launched the program to address the many misconceptions — and to better reflect the broader group of levy payers represented confirmed by a Newspoll survey of more than 1200 Australians in our new structure. Members bring extensive industry expertise — about Australia’s timber industry. The Newspoll survey to the advisory groups. We are most grateful to their businesses commissioned ahead of the launch showed that more than half and employers who enable them to make this vital contribution. of Australians don’t know that using wood (as opposed to other The four advisory groups cover sustainability and resources, materials) is good for the environment. An astounding one-third solid wood processing, engineered wood products, and market of respondents thought that the concrete, steel or plastic industries knowledge and development. produced fewer carbon emissions. These groups ensure that the industry can provide advice about Four in ten Australians did not know that wood stores carbon research needs and directions and knowledge gaps. Feedback on and 56% thought more trees were being cut down than planted technology transfer is also an important part of their role. in an average year in Australia. The strength of these fallacies amazes me. Wood. Naturally Better.™ campaign Throughout this campaign the opinions and knowledge of professionals The Wood. Naturally Better.™ program is a bold industry and consumers is being tracked by regular, independently conducted initiative designed to increase the use of wood and wood products research. We hope to see these many misconceptions corrected over by creating a recognisable brand and logo that will become time as the campaign gains momentum. associated with the advantages of wood at both a trade and a As part of improving our understanding of consumer attitudes consumer level. The ability to create and deliver this initiative and behaviour, we are using a proprietary research method was a significant driver in the establishment of FWPA. and associated surveys. Known as Lifestyles Of Health And This multi-million dollar program is expected to run over at least Sustainability (LOHAS), this is an internationally recognised five years. It is a first for the Australian market and positions wood system that gives us the ability to insert our own questions while as a preferred material in the eyes of consumers and the building accessing a wider range of results. industry — based on its environmental, aesthetic, economic, renewable and ‘architecturally smart’ characteristics. Timber standards The first key message of the Wood. Naturally Better.™ campaign In September this year, following extensive industry consultation, in magazines and newspapers is highlighting the environmental FWPA assumed the role of coordinator of standards for timber and benefits of wood and the important role wood products play in timber design. This role is critically important to the successful tackling climate change because of wood’s carbon storage, and marketing of timber products predominately into the building its renewable and sustainable qualities. industry. Development of a strategy for standards review, develop ment and renewal will be the first task within this new role. As is well known to readers of Australian Forestry, trees absorb and store carbon dioxide from the atmosphere. This carbon Finally, I would like to close with an invitation. I encourage readers remains locked away even after the wood has been turned into to become actively involved and to have a say in the development manufactured products. Moreover, production of wood requires of research priorities, promotion and the new industry services. substantially less energy, and hence produces fewer carbon FWPA is committed to organisational transparency and to working dioxide emissions, than other building materials generate. with our members, other levy payers and stakeholders to develop effective industry support services. The Wood. Naturally Better.™ campaign plans to drive home these positive industry messages. The campaign’s first target were consumers. Ron Adams Later stages will address professionals, including architects, specifiers, Chairman policy makers and others in the construction industries. Forest and Wood Products Australia Limited
Australian Forestry 2009 Vol. 72 No. 4 pp. 147–148 Jonathan E. Brand 149
Effect of different Acacia acuminata variants as hosts on performance of sandalwood (Santalum spicatum) in the northern and eastern Wheatbelt, Western Australia
Jonathan E. Brand
Forest Products Commission, Locked Bag 888, Perth BC, WA 6849, Australia Email: [email protected]
Revised manuscript received 6 August 2009
Summary Introduction The effect of seven ‘jam variants’ (Acacia acuminata typical Western Australian sandalwood (Santalum spicatum (R.Br.) variant, A. acuminata narrow-phyllode variant, A. acuminata A.DC.) is a root hemi-parasite (Hewson and George 1984) that small-seed variant, A. acuminata/burkittii variant 1, A. acuminata/ occurs naturally with a range of vegetation types, including many burkittii variant 2, A. burkittii and A. oldfieldii) on the survival Acacia species (Loneragan 1990). Sandalwood produces valuable and growth of sandalwood (Santalum spicatum) was examined aromatic oils within its heartwood and, due to its commercial in two relatively low-rainfall locations, Dowerin and Morawa, in value, silvicultural trials have been established over the past 30 y to the northern and eastern Wheatbelt of Western Australia. During determine suitable methods to grow plantations (Loneragan 1990; the course of the trial (2000–2008) the mean annual rainfall Brand et al. 2000; Fox 2002; Woodall and Robinson 2002). One of was only 326 mm at Dowerin and 259 mm at Morawa. In July the preferred host species identified from these trials is called ‘jam’ 2000, seedlings from 84 families from 18 separate populations (Acacia acuminata Benth.), which has consistently supported and representing each of the seven different jam variants were relatively fast sandalwood growth on a range of soil types (Brand planted on cleared farmland at both locations. A total of 4032 2002; Brand et al. 2004). Acacia acuminata is a suitable host host seedlings were planted at each site. At host age 2 y (April because it is able to supply nutrients to sandalwood, including N, 2002), S. spicatum seeds were sown near 1807 host plants at the K, Ca and Cu (Struthers et al. 1986). Acacia acuminata is also first site and near 1397 hosts at the second. relatively long-lived and should support the sandalwood through to harvesting at age about 20–25 y. The different jam variants were associated with clear differences in sandalwood performance at both sites. At age 1 y, mean survival Although A. acuminata has been extensively used as a host, this of sandalwood was significantly greater nearA. acuminata small- species is variable in at least form and habit, and this variation seed variant (64%) and A. acuminata narrow-phyllode variant may have the potential to affect sandalwood performance in (50–54%) than near A. acuminata typical variant (21–34%). This plantations. To aid sandalwood research, Maslin et al. (1999) pattern of sandalwood survival between the different jam variants during 1998–1999 reviewed the taxonomy of the ‘jam group’, was similar at age 6 y. consisting of A. acuminata and two of its close relatives in Western Australia (WA), A. burkittii Benth. and A. oldfieldii F.Muell. At age 6 y, the mean stem diameter (at 150 mm above ground) of They divided the jam group into seven taxa: A. acuminata typical sandalwood was the greatest near A. acuminata small-seed variant variant, A. acuminata narrow-phyllode variant, A. acuminata within each site: 62 mm at Dowerin and 47 mm at Morawa. The small-seed variant, A. acuminata/burkittii variant 1, A. acuminata/ mean stem diameter of sandalwood near A. acuminata typical burkittii variant 2, A. burkittii and A. oldfieldii. The different taxa variant was 51 mm at Dowerin and only 25 mm at Morawa. or ‘jam variants’ were separated on differences in morphological Stem borers were more common in A. acuminata typical variant characteristics including phyllode width, spike length, and pod than in the other jam variants at Dowerin. No borers were observed and seed morphology, as well as patterns of genetic variation. The in the jam variants at Morawa. At both sites, gall rust was observed identification of the different jam variants and their distribution in each of the jam variants except A. oldfieldii. within WA (Fig. 1) provided the opportunity to collect seed from the seven different variants and grow seedlings that could be used Keywords: plantations; parasites; host plants; host parasite relationships; to examine their effect on sandalwood performance. arid climate; families; provenance; jam variant; sandalwood; Santalum spicatum; Acacia acuminata; Acacia burkittii; Acacia oldfieldii
Australian Forestry 2009 Vol. 72 No. 4 pp. 149–156 150 Effect of Acacia acuminata variants on performance of Santalum spicatum in WA
114°00' 120°00' 126°00'
Warburton community Meekatharra Wiluna
Kalbarri Yalgoo 28°00' Geraldton Mullewa Morawa Paynes Find Mingenew Kalgoorlie Range extended to South Australia Range extended to South and New South Wales Moora Wongan Hills Forrest 0 200 Dowerin Kellerberrin km Norseman 32°00' Corrigin Williams Ravensthorpe Acacia acuminata (typical variant) Acacia acuminata (narrow-phyllode variant) Borden Acacia acuminata (small-seed variant) MAP AREA Acacia acuminata/burkittii (variants 1 and 2) Acacia burkittii Acacia oldfieldii
Figure 1. Distribution of the seven different jam variants in Western Australia (From Maslin et. al. 1999)
Of the seven defined jam variants (Maslin et al. 1999), Methods A. acuminata typical variant, which occurs naturally mainly in the region of medium mean annual rainfall (400–700 mm) of Jam variant description and distribution the western ‘Wheatbelt’ (Fig. 1)1, has been extensively planted throughout the Wheatbelt as a sandalwood host in plantations over Maslin et al. (1999) provided a detailed description and distribu the past 10–15 y (pers. obs.). However, other jam variants may tion of each of the seven different jam variants; for the purposes be better suited as a sandalwood host, especially in regions of of this study a summary of their review is given below: relatively low mean annual rainfall (300–400 mm) of the northern 1. A. acuminata (typical variant): occurs in the western part of and eastern Wheatbelt. Observations of sandalwood plantations the Wheatbelt from near Three Springs to Borden and Ravens throughout the Wheatbelt indicate that some jam variants appear thorpe (Fig. 1). Tall tree or shrub 3–10 m tall; phyllodes flat to be more stressed than others and are more prone to insect attack and 3–10 mm wide; seeds 3–4.5 mm long. and diseases including stem borers and gall rust (Uromycladium 2. A. acuminata (narrow-phyllode variant): common in the genus). Stem borers include cossid moth larvae that feed on northern, central and eastern Wheatbelt and into the arid zone the cambium of A. acuminata stems and can ringbark the trees from near Morawa to Kalgoorlie and Norseman. Rounded (The Avon Sandalwooder 2005). Gall rust is present in many shrubs or small trees 2–7 m tall; phyllodes flat and 2–4 mm forms of A. acuminata (Maslin et al. 1999), and Uromycladium wide; seeds 3–4 mm long. tepperianum has been shown to reduce population densities of 3. A. acuminata (small-seed variant): occurs near Yalgoo and A. saligna in South Africa (Morris 1997). It may be possible to Mullewa and also south of Borden. Rounded shrubs or small select jam variants that are more resistant to pests and diseases, and trees 2–5 m tall; phyllodes flat and 3–6 mm wide; seeds to use these jam variants as hosts to grow sandalwood plantations 2–3 mm long. in the Wheatbelt and other areas of WA. 4. A. acuminata/burkittii (variant 1): small distribution near This study aimed to plant representatives of the seven different Mullewa. Tree form and height are variable; phyllodes flat jam variants together on two separate sites as sandalwood hosts and 2–3 mm wide; seeds 2.5–3.5 mm long. in the northern and eastern Wheatbelt of WA. The specific aims 5. A. acuminata/burkittii (variant 2): small distributions near of this study were to compare: Eradu and Northampton. Tree form and height are variable; • the effect of the different jam variants on S. spicatum survival phyllodes flat and 3–8 mm wide; seeds 4–5 mm long. and growth 6. A. burkittii: widespread in the arid zone from Yalgoo to South • the level of disease (stem borers and gall rust) within the Australia. Multi-stemmed shrubs or small trees, 1.5–8 m different jam variants. tall; phyllodes terete and 0.7–2 mm wide; seeds 3.5–6 mm long. 7. A. oldfieldii: small distribution along the Murchison River near Ajana. Rounded spreading shrubs or small trees 2–4 m 1 See also http://www.Wheatbelt.wa.gov.au/files/Wheatbelt_maps/map_ tall; phyllodes flat and 3–5 mm wide; seeds 1.5–2 mm Wheatbelt.pdf long.
Australian Forestry 2009 Vol. 72 No. 4 pp. 149–156 Jonathan E. Brand 151
Jam variant seed collection Dowerin During 1998–1999, Maslin et al. (1999) collected seeds from 400 Morawa Dowerin mean individual parent trees (families) growing in various populations Morawa mean to represent each of the seven jam variants. From this collection 300 seeds were obtained from 10–15 families and 2–4 provenances of each variant for use in this study (Table 1). In total 84 families 200 growing in 18 separate populations or provenances were used.
Annual rainfall (mm) 100 Site description 0 To examine the effects of jam variants on sandalwood perform 2000 2001 2002 2003 2004 2005 2006 2007 ance, sites were selected near Dowerin (31°06′S, 117°02′E) and Morawa (29°05′S, 115°50′E). The Dowerin site is about Year 10 km north of Dowerin and 40 km south-east of Wongan Hills Figure 2. Annual rainfall during 2000–2007 near Dowerin (townsite) (Fig. 1); the Morawa site is about 12 km west of Morawa and and Morawa (2000–2004 townsite; 2005–2007 airport). Long-term 45 km east of Mingenew. Both sites are near the border where mean annual rainfalls for Dowerin (364 mm, 1904–2007) and Morawa the geographic ranges of A. acuminata (typical variant) and (334 mm, 1911–2007) are also shown. A. acuminata (narrow-phyllode variant) meet, in the northern and eastern Wheatbelt. Each site was about 5 ha in area and was farmland previously used for cropping and grazing. The soil at Host establishment Dowerin consists mainly of grey sand over clay loam at 0.1–0.5 m, while at Morawa it is mainly a red sandy loam over red clay at Before planting, both sites were ripped to a depth of 0.5 m in lines 0.1–0.3 m. spaced 4 m apart, and sprayed with knock-down and residual herbicides to control weeds for the first 2–4 months after estab During the trial (2000–2008) the annual rainfalls were generally lishment. In July 2000, jam seedlings (age six months) were low, averaging 326 mm at Dowerin (range 277–391 mm) and only planted along the rip-lines at 3-m intervals (833 stems ha–1). The 259 mm (range 199–308 mm) at Morawa (Fig. 2). These falls were seedlings from each of the 84 families were grown separately well below the long-term annual means of 364 mm at Dowerin in ‘Colmax’ trays at the Department of Conservation and Land (1904–2007) and 334 mm at Morawa (1911–2007). Besides Management nursery in Narrogin, WA. having a relatively low mean annual rainfall, these locations also have relatively high mean annual evaporation rates of about 2300 mm at Dowerin and 2500 mm at Morawa.
Table 1. Provenance, latitude, longitude and number of families from which seeds were collected from the seven jam variants
Number of Jam variant Provenance Latitude Longitude families 1. A. acuminata (typical) Bolgart 31°13′S 116°29′E 4 York 32°00′S 116°48′E 2 Borden 34°05′S 118°13′E 4 Brookton 32°24′S 116°53′E 5 2. A. acuminata (narrow) Mukinbudin 30°44′S 118°21′E 5 Latham 29°43′S 116°25′E 5 Kalannie 30°26′S 117°23′E 5 3. A. acuminata (small seed) Binnu 28°02′S 114°59′E 5 Yalgoo 28°23′S 116o20′E 5 4. A. acuminata/burkittii (var. 1) Murchison 27°49′S 114°42′E 5 Mullewa 28°35′S 115°35′E 5 Nerren Nerren 27°12′S 114°37′E 4 5. A. acuminata/burkittii (var. 2) Northampton 28°18′S 114o37′E 5 Greenough 28°43′S 115°02′E 5 6. A. burkittii Wiluna 26°40′S 120°02′E 5 Ninghan 29°24′S 117°14′E 5 7. A. oldfieldii Murchison River 27°43′S 114°40′E 5 Hawkes Head 27°47′S 114°28′E 5
Australian Forestry 2009 Vol. 72 No. 4 pp. 149–156 152 Effect of Acacia acuminata variants on performance of Santalum spicatum in WA
At both sites, each jam family was planted separately in single- Results line plots of eight seedlings over a length of 24 m. Each of the 84 family plots was randomised within a block, and each site Hosts contained six blocks or replications. Although the jam seedlings were planted in separate family plots, they were also grouped Survival into their seven respective jam variants (Table 1) for analysis. At At age 8 y, mean host survival differed significantly (P < 0.001) each site, 480–720 seedlings from each of the seven different jam between different jam variants at both Dowerin and Morawa. variants were planted, with a total of 4032 seedlings planted. At Dowerin, the mean survival of A. oldfieldii (51.1 ± 4.0%, Fig. 3a) was significantly lower than that of the other jam Sandalwood establishment variants (80.2–85.3%). At Morawa, the mean survival of A. burkittii (70.0 ± 4.5%, Fig. 3b), A. acuminata/burkittii At host age 2 y (April 2002), four S. spicatum seeds (obtained variant 2 (68.1 ± 3.1%) and A. acuminata narrow-phyllode variant from a sandalwood plantation near Narrogin, WA) were sown (67.5 ± 4.8%) were significantly greater than that ofA. oldfieldii per ‘spot’ near every second surviving host plant, to obtain an (29.8 ± 3.6%) and A. acuminata typical variant (48.2 ± 4.5%). approximate sandalwood-to-host ratio of 1 : 2. At each spot, the sandalwood seeds were buried 2–3 cm below the surface and Overall, the mean host survival at Dowerin (77.8 ± 2.0%) was 0.5 m from the base of each host stem. The seeds had no pre- significantly (P < 0.001) greater than at Morawa (57.8 ± 2.5%) treatment. Sowing spots were sprayed with glyphosate in early at age 8 y. July 2002 to control weeds before the S. spicatum seedlings emerged in August–October 2002. A total of 1807 sowing spots Height were established at Dowerin and 1397 at Morawa. Mean host height at age 8 y also differed significantly (P < 0.001) Germination, survival and growth between jam variants at both Dowerin and Morawa. At Dowerin, the mean height of A. acuminata typical variant (3.6 ± 0.1 m, Sandalwood germination was assessed at age three months Fig. 3c) was significantly greater than that of the other jam variants (November 2002), and recorded as germinated if at least one (1.9–2.5 m). At Morawa, the mean height of A. acuminata typical seed germinated per sowing spot. At sowing spots where variant (2.7 ± 0.1 m, Fig. 3d) was again significantly greater than sandalwood seeds did not germinate or the seedlings died within that of all of the other jam variants except for the narrow-phyllode a year, they were replaced with seeds in April 2003 to maintain form (2.5 ± 0.1 m). the approximate sandalwood-to-host ratio of 1 : 2, but these new plantings were not included in the analysis. At sandalwood age Overall, the mean host height at Dowerin (2.4 ± 0.05 m) was 1 y, the sowing spots were thinned to leave only one sandalwood significantly (P < 0.001) greater than at Morawa (2.2 ± 0.05 m). seedling at each spot. Diseases Of the sandalwood seeds that germinated during August–October 2002, survival was recorded at age 1 y (November 2003) and 6 y Stem borers (August–September 2008) at both sites. At age 6 y, sandalwood stem diameter at 150 mm above the ground and height from the At Dowerin, the mean incidence of A. acuminata typical variant ground to the tallest green leaf was measured for 768 trees at trees with stem borers (8.1 ± 1.9%) was significantly (P < 0.001) Dowerin and 385 trees at Morawa. greater than in the other jam variants (0 – 0.4%) at age 5 y.
Host survival and height were assessed at age 8 y (August– The mean fraction of host trees with stem borers at Dowerin was September 2008). Height was measured for 388–664 trees at only 1.2 ± 0.5%, but was significantly greater than at Morawa each site. (P < 0.001) where no borers were observed.
Diseases Gall rust
Stem borer holes and gall rust (Uromycladium genus) were Gall rust was present at a low incidence (1.1–3.0%) in all of the recorded as either present or absent from each jam tree at age jam variants growing at Dowerin except for A. oldfieldii, which 5 y (November 2005). had no visible signs. At Morawa, the mean incidence of gall rust differed between jam variants (P < 0.001), with significantly more present in A. burkittii (25.1 ± 4.7%) than in A. acuminata Statistical analysis small-seed variant (9.7 ± 2.0%). The other variants had means of Plot means for germination, survival, growth and the presence of 11.2–23.0%, except for A. oldfieldii which had no visible galls. disease, from six replications, were compared between treatments The mean fraction of host trees with gall rust at Morawa (seven jam variants and two sites) using two-way analysis of (14.3 ± 1.7%) was significantly (P < 0.001) greater than at variance (ANOVA). Proportions were angular transformed before Dowerin (1.6 ± 0.3%). analysis and Tukey’s test was used to compare means. SYSTAT version 10.2 was used for all statistical analysis.
Australian Forestry 2009 Vol. 72 No. 4 pp. 149–156 Jonathan E. Brand 153
(a) Survival: Dowerin (b) Survival: Morawa
100 100 90 a a a a a a 90 80 80 a ) a 70 ) a
% 70 ab
%
l ( ab 60 b l ( 60 50 b
viva
viva 50
ur
40 ur
S 40 30 S c 30 20 20 10 10 0 Typ. Narr. Small V ar. Var. Ac. Ac. 0 seed 1 2 burk. old. Typ. Narr. Small V ar. Var. Ac. Ac. seed 1 2 burk. old. Variant Variant
(c) Height: Dowerin (d) Height: Morawa a 3.5 3.5 3 3 a ab b b b 2.5 bc b 2.5 cd
m) c c m)
( c 2 d d
t 2
ht (
igh 1.5 ig 1.5
He He 1 1 0.5 0.5 0 0 Typ. Narr. Small V ar. Var. Ac. Ac. Typ. Narr. Small V ar. Var. Ac. Ac. seed 1 2 burk. old. seed 1 2 burk. old. Variant Variant
Figure 3a–d. Mean survival (% ± standard error) and tree height (± standard error) of the seven jam variants (age 8 y) planted at Dowerin and Morawa. Values with the same letter are not significantly different, using Tukey’s test (P > 0.05).
Sandalwood Morawa
Germination At Morawa, the mean survival of 1-y-old sandalwood seedlings differed between jam variants (P < 0.001), with survival Mean sandalwood germination per spot near the different jam near A. acuminata small-seed variant (64.0 ± 6.7%), narrow- variants was 92–96% at Dowerin and 94–99% at Morawa. phyllode variant (50.5 ± 4.4%) and A. oldfieldii (50.2 ± 5.5%) There was no significant (P = 0.150) difference in sandalwood being significantly greater than near A. acuminata typical germination between sites. variant (20.6 ± 2.2%) and A. acuminata/burkittii variant 2 (29.7 ± 2.4%). Survival The 6-y-old sandalwood displayed a similar pattern of survival Dowerin (Fig. 4b), except that sandalwood survival next to each of the variants dropped by 6.9–12.0%. The mean survival of At age 1 y, the mean survival of sandalwood seedlings at Dowerin sandalwood trees near A. acuminata small-seed variant differed between jam variants (P < 0.001), with survival near (55.7 ± 7.3%), A. oldfieldii (40.9 ± 6.3%) and narrow-phyllode A. acuminata small-seed variant (64.4 ± 2.9%) and narrow- variant (34.5 ± 4.4%) was significantly greater than near phyllode variant (53.9 ± 2.1%) significantly greater than near A. acuminata typical variant (10.6 ± 1.9%). A. acuminata typical variant (34.5 ± 2.4%) and A. oldfieldii (32.2 ± 6.4%). Mean survival near the other A. acuminata variants Between sites was 35.7–47.4%. Sandalwood mortality was highest in the first year after germina At age 6 y, the pattern of sandalwood survival near the different tion, with a mean overall survival of 43.6 ± 2.3% at Dowerin A. acuminata variants (Fig. 4a) was the same as at age 1 y. The and 40.6 ± 2.6% at Morawa, with no difference between sites mean survival of sandalwood trees was significantly greater (P = 0.38). By age 6 y, however, mean sandalwood survival at near A. acuminata small-seed variant (63.1 ± 3.2%) and narrow- Dowerin (42.5 ± 2.2%) was significantly (P < 0.001) greater than phyllode variant (53.3 ± 2.3%) than near A. acuminata typical at Morawa (29.9 ± 2.7%). variant (33.7 ± 2.4%) and A. oldfieldii (30.9 ± 6.4%).
Australian Forestry 2009 Vol. 72 No. 4 pp. 149–156 154 Effect of Acacia acuminata variants on performance of Santalum spicatum in WA
(a) Survival: Dowerin (b) Survival: Morawa
70 a 70 a 60 ab 60 abc 50 50 ab bc bc abc 40 c c 40 30 30 bcd bcd cd Survival (%) 20 Survival (%) 20 d 10 10 0 0 Typ. Narr. Small Var. Var. Ac. Ac. Typ. Narr. Small Var. Var. Ac. Ac. seed 1 2 burk. old. seed 1 2 burk. old. Variant Variant
(c) Diameter: Dowerin (d) Diameter: Morawa
a 60 60 b ab b b b b 50 a 50 ab abc 40 40 bc bc c 30 30 c
Diameter (mm) 20 Diameter (mm) 20
10 10
0 0 Typ. Narr. Small Var. Var. Ac. Ac. Typ. Narr. Small Var. Var. Ac. Ac. seed 1 2 burk. old. seed 1 2 burk. old. Variant Variant
(e) Height: Dowerin (f) Height: Morawa
2.5 2.5 a b b b 2 b b 2 c a ab abc 1.5 1.5 bc bc c c 1 1
Height (m) Height (m)
0.5 0.5
0 0 Typ. Narr. Small Var. Var. Ac. Ac. Typ. Narr. Small Var. Var. Ac. Ac. seed 1 2 burk. old. seed 1 2 burk. old. Variant Variant
Figure 4a–f. Mean survival (± standard error), stem diameter at 150 mm (± standard error), and tree height (± standard error) of Santalum spicatum (age 6 y) growing near the seven different jam variants, at Dowerin and Morawa. Values with the same letter are not significantly different, using Tukey’s test (P > 0.05).
Stem diameter Mean stem diameter of 6-y-old sandalwood at Morawa (Fig. 4d) was significantly greater near A. acuminata small-seed variant Mean stem diameter of 6-y-old sandalwood differed significantly (46.6 ± 2.3 mm) and A. oldfieldii (42.4 ± 3.3 mm) than near (P < 0.001) between jam variants at both Dowerin and Morawa. At A. acuminata typical variant (24.8 ± 4.2 mm) and A. acuminata/ Dowerin, the diameter of sandalwood (Fig. 4c) was significantly burkittii variant 1 (27.2 ± 3.3 mm). greater near A. acuminata small-seed variant (61.6 ± 1.4 mm) than near the other variants (50.6–54.3 mm), except for A. acuminata/ burkittii variant 2 (55.5 ± 1.6 mm).
Australian Forestry 2009 Vol. 72 No. 4 pp. 149–156 Jonathan E. Brand 155
Overall, the mean sandalwood stem diameter was significantly were more pronounced at Morawa. At Dowerin the mean stem (P < 0.001) greater at Dowerin (54.0 ± 0.7 mm) than at Morawa diameter of sandalwood near the small-seed variant (62 mm) (34.3 ± 1.0 mm) at age 6 y. was only 22% greater than near the typical variant, whereas at Morawa the mean stem diameter of sandalwood near A. acuminata Height small-seed variant (47 mm) was 88% greater than near the typical variant. The observed differences in sandalwood growth rate Mean height of 6-y-old sandalwood differed significantly between sites is not likely to be due to soil type because both (P < 0.001) between jam variants at both Dowerin and Morawa. sites contained sandy-loam over clay soils which are considered At Dowerin, the mean height of sandalwood (Fig. 4e) was to be suitable for growing both jam hosts and sandalwood (Anon. significantly greater near A. acuminata small-seed variant 2007). The difference in sandalwood growth between sites is more (2.23 ± 0.04 m) than near the other variants. The mean height of likely to be related to differences in rainfall and evaporation. sandalwood near A. acuminata typical variant (1.73 ± 0.04 m) During the years 2000–2007, the mean annual rainfall at Dowerin was significantly less than near the other variants. was 326 mm while at Morawa it was only 259 mm. The annual evaporation is also about 200 mm greater at Morawa than at At Morawa, the mean height of 6-y-old sandalwood (Fig. 4f) was Dowerin. The lower rainfall and greater evaporation at Morawa significantly greater near A. acuminata small-seed variant most likely explains why the sandalwood trees were smaller and (1.70 ± 0.07 m) and A. oldfieldii (1.64 ± 0.10 m) than near why there was also a greater difference in sandalwood growth A. acuminata typical variant (1.09 ± 0.12 m) and near between host treatments. Acacia acuminata typical variant A. acuminata/burkittii variant 1 (1.24 ± 0.10 m). appeared not to be a suitable host for sandalwood at Morawa, Mean sandalwood height was significantly (P < 0.001) greater but was a relatively better host at Dowerin after the sandalwood at Dowerin (1.97 ± 0.02 m) than at Morawa (1.41 ± 0.03 m) at were a year old. This again appears to be related to A. acuminata age 6 y. typical variant being less suited to the more arid conditions at Morawa than some of the other jam variants.
Discussion Besides survival and growth, there were also some differences in the incidence of diseases within the different jam variants. At Sandalwood survival was significantly greater nearA. acuminata host age 5 y, there were significantly more borers inA. acuminata small-seed and narrow-phyllode variants than near the typical typical variant (8%) than in the other jam variants at Dowerin. variant at both Dowerin and Morawa. The marked difference in Although not present in large numbers, the borers did appear to sandalwood survival between the different host variants occurred target the A. acuminata typical variant more than the other jam early, at age 1 y, and this pattern was maintained at age 6 y. The variants, with some trees completely ringbarked. At a similar age, initial difference in sandalwood survival between variants may borers were not observed in any of the jam variants at Morawa. have been related to the below-average rainfall at both sites Gall rust was also present in each jam variant except A. oldfieldii, during the establishment year, 2002. The total annual rainfall in but did not appear to be greatly reducing phyllode production or this year was only 260 mm at Dowerin and 202 mm at Morawa, health of the trees at age 5 y. well below their respective long-term averages of 364 mm and 334 mm. More importantly, the rainfall was low at both sites in the To more effectively compare sandalwood performance near first winter, a critical stage of sandalwood germination and initial A. acuminata typical variant with that near the other variants, growth. The total rainfall in the 2002 winter was only 129 mm at a trial should also be established in the medium mean annual Dowerin and 90 mm at Morawa, again well below their respective rainfall region (400–700 mm) of the western Wheatbelt, where long-term winter averages of 172 mm and 153 mm. During this A. acuminata typical variant occurs naturally (Maslin et al. 1999, dry year the A. acuminata typical variant, which is derived from Fig. 1). Unfortunately this current study did not include a site a region with greater mean annual rainfall than the A. acuminata in the medium annual rainfall region, but in previous trials in small-seed and narrow-phyllode variants (Maslin et al. 1999), the western Wheatbelt A. acuminata typical variant has proven may have been under more stress and less capable of supporting to be a suitable sandalwood host. In sandalwood trials planted a parasite during this critical establishment phase. with A. acuminata typical variant near Dandaragan, Narrogin and Katanning, sandalwood survival was 77–86% and mean At age 6 y, there were also significant differences in sandalwood stem diameter (at 150 mm) was relatively high (26–37 mm) at growth near the different jam variants. At each site, sandalwood age 3 y (Brand et al. 2000, 2003). The combined results of this stem diameter and height were significantly greater near current study and previous trials indicate that A. acuminata typical A. acuminata small-seed variant than near A. acuminata typical variant is more suited for sandalwood plantations in the medium variant. Interestingly, the differences in mean sandalwood growth mean annual rainfall region of the western Wheatbelt than in near these two different jam variants did not appear to be related the low mean annual rainfall regions of the northern and eastern to host growth during the trial. At host age 8 y, the mean height Wheatbelt. Acacia acuminata small-seed variant appears to be of A. acuminata typical variant was significantly greater than that the best suited of the jam variants to support good sandalwood of the small-seed variant. survival and growth in relatively low-rainfall regions of WA, Although sandalwood performed best near A. acuminata small- and should be further investigated for its use in the sandalwood seed variant at both sites, the sandalwood trees were significantly plantation industry. larger at Dowerin and the relative differences between jam variants
Australian Forestry 2009 Vol. 72 No. 4 pp. 149–156 156 Effect of Acacia acuminata variants on performance of Santalum spicatum in WA
Acknowledgements Fox, J.E.D. (2002) Silviculture of sandalwood: a semi-parasitic species for family farm enterprises in Western Australia. In: International I thank Bruce Maslin for identifying the different Acacia Symposium on Contributions of Family-Farm Enterprises to acuminata variants and their close relatives, and for coordinating Sustainable Rural Development. Proceedings. IUFRO 3.08.00, 28 the seed collection for this project. Thanks also go to Warren and July – 1 August, Gengenbach, Germany. Forstwissenschaftliche Fakultat der Universitat, Freiburg, pp. 304–316. Joy Carslake for providing a field site near Morawa; to Barry, Hewson, H.J. and George, A.S. (1984) Santalaceae. In: Flora of Australia, Betty and John Morrell for providing a field site near Dowerin; and Volume 22, pp. 29–67. Bureau of Flora and Fauna, Canberra. to David Mickle, Robert Archibald, Edward Lim, Ivan Zirngast, Australian Government Publishing Service, Canberra. Brian Haines and John Enklemann who helped to establish and Loneragan, O.W. (1990) Historical Review of Sandalwood (Santalum measure the trials. The WA Bureau of Meteorology supplied spicatum) Research in Western Australia. Research Bulletin No. 4. rainfall records for Dowerin and Morawa. This study was funded Department of Conservation and Land Management, Western by the Forest Products Commission, WA. Australia. Maslin, B., Byrne, M., Coates, D., Broadhurst, L., Coleman, H. and Macdonald, B. (1999) The Acacia acuminata (jam) Group: An References Analysis of Variation to Aid Sandalwood (Santalum spicatum) Plantation Research. Department of Conservation and Land Anon. (2007) Sandalwood (Santalum spicatum) Guide for Farmers. Tree Management, Western Australia, 60 pp. Facts No. 1, Forest Products Commission, WA. Morris, M.J. (1997) Impact of the gall-forming rust fungus Uromycladium Brand, J.E. (2002) Review of the influence of Acacia species on tepperianum on the invasive tree Acacia saligna in South Africa. establishment of sandalwood (Santalum spicatum) in Western Biological Control 10, 75–82. Australia. Conservation Science Western Australia 4, 125–129. Struthers, R., Lamont, B.B., Fox, J.E.D., Wijesuriya, S. and Crossland, T. Brand, J.E., Crombie, D.S. and Mitchell, M.D. (2000) Establishment (1986) Mineral nutrition of sandalwood (Santalum spicatum). and growth of sandalwood (Santalum spicatum) in south-western Journal of Experimental Botany 37, 1274–1284. Australia: the influence of host species. Australian Forestry 63, 60–65. The Avon Sandalwooder (2005) The Avon Sandalwood Network [Perth, WA]. Issue 4, Spring 2005. Brand, J.E., Robinson N. and Archibald R.D. (2003) Establishment and growth of sandalwood (Santalum spicatum) in south-western Woodall, G.S. and Robinson, C. J. (2002) Same day plantation Australia: Acacia host trials. Australian Forestry 66, 294–299. establishment of the root hemiparasite sandalwood (Santalum spicatum (R.Br.) A. DC) and hosts. Journal of the Royal Society Brand, J.E., Jones, P.J. and Donovan, O.H. (2004) Current growth of Western Australia 85, 37–42. rates and predicted yields of sandalwood (Santalum spicatum) plantations in the Wheatbelt, Western Australia. Sandalwood Research Newsletter 19, 4–7.
Australian Forestry 2009 Vol. 72 No. 4 pp. 149–156 Rebecca M. Ford, Kathryn J. H. Williams, Ian D. Bishop and John E. Hickey 157
Public judgements of the social acceptability of silvicultural alternatives in Tasmanian wet eucalypt forests
Rebecca M. Ford1,4, Kathryn J. H. Williams1, Ian D. Bishop2 and John E. Hickey3
1Department of Resource Management and Geography, University of Melbourne, 221 Bouverie St, Parkville 3010, Australia 2Department of Geomatics, University of Melbourne, Parkville, Victoria 3010, Australia 3Forestry Tasmania, 79 Melville St, Hobart, Tasmania 7000, Australia 4Email: [email protected]
Revised manuscript received 12 August 2009
Summary community. With the aim of identifying more ecologically and socially acceptable harvest and regeneration methods, Forestry This research compared individual’s social acceptability ratings Tasmania is investigating silvicultural systems at the Warra of six harvest and regeneration systems that could be applied in Long Term Ecological Research site (Warra LTER) to compare wet eucalypt forests: the clearfell, burn and sow system and five the existing clearfell, burn and sow system with five alternative alternatives. A smaller calibration study also tested the effect of treatments (Hickey et al. 2001). Although there have been several providing participants with a wider range of alternative systems silvicultural systems trials in Australia, there is as yet no published including a no-harvest option and conversion to blue gum information about general community responses to alternative plantation. A survey method was complemented by qualitative systems here. follow-up interviews. About 300 Tasmanians with different affiliations in relation to forest harvesting were asked to judge Most previous studies into public perceptions of forestry have the acceptability of the harvest systems. These were presented measured the scenic beauty of forested scenes based on the to them in two ways: as still images showing the forest in the assumption that negative visual perceptions are a primary cause first year after harvest and as animated sequences showing the of public reactions against forest harvesting (Gobster 2001). forest regenerating over time. On average, non-affiliated and The research described here has a broader approach, taking the conservation-affiliated people rated the clearfell system least view that focusing on visual perceptions alone ignores a more acceptable and a selective logging system most acceptable. For cognitive (thoughtful) basis for people’s concerns in the things these two groups, as a general trend, acceptability increased with that they value and their beliefs about the effects of logging. In the proportion of forest retained at the first harvest. Aggregated the USA, researchers have argued for the investigation of the harvest patterns were more acceptable than dispersed. Follow-up broader concept, ‘social acceptability’ (Ribe 2002; Shindler et al. interviews provided examples of aesthetic and more thought- 2002). This research addresses a lack of previous Australian social based responses leading to these trends. The trend for large-scale research into alternative silvicultural systems by measuring the timber-industry-affiliated participants was opposite: the more social acceptability of harvest systems based on those investigated forest retained on the site, the less acceptable the result. In the at the Warra LTER. calibration study, which included a wider range of systems and mostly non-affiliated participants, a no-harvest option was rated Landscape assessment research is not regularly applied to most acceptable and the selective-harvest system was rated Australian forestry, but can provide benefits to managers and significantly less acceptable than when the no-harvest option was policy-makers. Decisions in forest management are typically not available. Viewing animated sequences led to higher ratings informed by biophysical and economic research, and social issues of the more intensive harvest systems and lower ratings of the are often left to be dealt with at the political level, in an ad hoc selective harvest system than those based on still images. While way, after economic and biophysical data have been integrated and significant, these differences were small. considered by policy-makers (Endter-Wada et al. 1998). There are advantages in taking a more systematic approach. In the same way Keywords: silvicultural systems; landscape; aesthetic value; acceptability; that economic and biophysical research can identify unintended social values; beliefs; Tasmania consequences that might not otherwise be obvious, social research can help to identify unanticipated social responses before they Introduction occur. Silvicultural trials, such as those being conducted at Warra and in other parts of the world, provide an excellent opportunity Clearfelling followed by burning and sowing is considered by to investigate the social implications of alternative silvicultural forest managers and the large-scale timber industry in Tasmania systems, for example Ribe (1999, 2005). In the Warra silvicultural to be an effective method of regenerating wet eucalypt forests, but trials, one of the drivers for seeking alternatives to clearfelling the impacts of the practice concern many people in the general is general community concern with that practice. It is important
Australian Forestry 2009 Vol. 72 No. 4 pp. 157–171 158 Social acceptability of silvicultural alternatives in Tasmanian eucalypt forest to find out whether proposed alternatives will actually be more to an open-ended question about issues of concern, ‘logging’ socially acceptable than clearfelling and to identify whether some was the most commonly reported issue at 46%. The ‘logging’ alternatives are more acceptable than others. category included clearfelling, cable logging and woodchipping. Responses to a further question ‘how do you think these changes The research reported here is part of a larger project, Social might be better managed’ included ‘better’ or ‘selective logging’ Acceptability of Forest Management Systems (see Ford et al. (including no clearfelling) at 22% and ‘prohibit logging’ at 14%. 2005 for an overview of the entire project). The three elements Participant observation of stakeholder groups identified concerns of the larger project are: with clearfelling including visual impacts on the landscape, habitat • participant’s social acceptability judgements of forest manage reduction and impacts on waterways (Tasmanian Public Land Use ment systems based on still, simulated images of coupes in the Commission 1996). first year after harvest (reported in this paper) In 1999 the Tasmanian government appointed a Community • a comparison of participant’s social acceptability judgements Leaders Group to run a wide-ranging community consultation with and without two types of additional information: animated process aimed at developing a vision for the state. One of the sequences showing forests regrowing and information about visions was related to native forest management, to ‘sustainably consequences of harvesting for fauna habitat, forest products, manage old-growth forests and to phase out clearfelling in those soil and water, forest products and safety (see Ford et al. in forests’. Two targets were set: ‘(a) to end clearfelling in areas of press, animation results repeated in this paper) high conservation value old-growth forest by January 1, 2003’; • an exploration of the basis for social acceptability judgements and ‘(b) complete phase out of clearfelling in old-growth forests in participants own values and beliefs (reported in Ford et al. by 2010’. Consensus was not achieved on the second target, with 2009). five members of the twenty-two member Community Leaders Group dissenting (Tasmania Together Community Leaders Group Background 2001).
Changing silviculture and social values in Australia In March 2004, the Examiner newspaper reported the results of a poll that found 57% of participants thought Tasmania’s forests The Tasmanian public’s views about clearfelling can be considered were ‘very well’ or ‘quite well’ managed and 38% thought they in the context of changing silviculture of wet eucalypt forests were ‘quite poorly’ or ‘very poorly’ managed (Hocking 2004). and changing social values. Silvicultural research in the 1950s The main reasons given for the belief that they were poorly found that clearfelling followed by high-intensity burning and managed were ‘I don’t agree with clearfelling’ and ‘it is destroying sowing was the most effective method of regenerating this forest old-growth forests’. Fifty-six per cent of the people interviewed type (Hickey et al. 2001). However, this understanding of how wanted some policy changes. The main reason given for wanting to achieve good regeneration, and the emergence of new export policy change was ‘End the clearfelling of old-growth forests’ woodchip markets for residues in the 1970s, coincided with (33% of participants). The main reason given for continuing changing social values (Dargavel 1995). The book The Fight for current policies was ‘They support employment and local the Forests (Routley and Routley 1974) articulated a difference in communities’ (14% of participants) (Hocking 2004). ethical approach between an emerging concern to retain forests for their own sake and the wood production emphasis then underlying These studies did not aim to achieve academic standards of the practices of management agencies. The aesthetic impact of reliability, but a theme emerges that a significant proportion of large clearfalls was associated with destruction of valued aspects Tasmanians do not support current forest policy and that the of the forest. From the late 1970s forestry agencies responded main concerns relate to the practices of clearfelling and logging to public pressure by emphasising multiple use and changing in old-growth forests. On the other hand, a significant proportion practices to reduce environmental and visual impacts, for example of the population appears to support current forest policy and by implementing Codes of Practice (Forest Practices Board practices because of the employment and other benefits to local 2000). In Tasmania, partial harvesting silviculture was applied communities. From these results it is clear that Tasmanians are in many other eucalypt forest types. The clearfell, burn and sow divided about current practices such as clearfelling, but it is not silvicultural system has however continued to be applied in wet possible to draw any conclusions about the kinds of practices that eucalypt forests. might be more broadly acceptable. For this it is insightful to look to the international literature. The few studies, polls and community consultation processes that have been carried out in Tasmania suggest that the concerns of Visual aesthetic quality of managed forests those who would like to see changes in forest management focus on the current practice of clearfelling, particularly when carried Most of the knowledge internationally about public perceptions out in old forests ( Tasmanian Public Land Use Commission of forestry is in the form of research that sought to measure the 1996; Tasmania Together Community Leaders Group 2001; ‘scenic beauty’ or ‘visual landscape aesthetic quality’ of forest Hocking 2004). conditions and relate it to physical characteristics of the forest ( Daniel and Boster 1976; Ribe 1989; Daniel 2001a,b). General A telephone survey of the wider community undertaken for the findings from this body of research are useful in considering Regional Forest Agreement process found that 30% of participants how the public might respond to silvicultural alternatives in were concerned with changes to native forests due to human use Tasmania, but there are limitations in trying to make predictions (Tasmanian Public Land Use Commission 1996). In response
Australian Forestry 2009 Vol. 72 No. 4 pp. 157–171 Rebecca M. Ford, Kathryn J. H. Williams, Ian D. Bishop and John E. Hickey 159 for Tasmania as many of the more detailed findings are specific for open stands with preferences for larger trees, thinning from to particular forest types in the USA or Scandinavia. Only the below may produce the most scenic result (Ribe 1989). most general findings are summarised here. Only one major study has been undertaken in Australian forests (Williamson and A few studies have shown that scenic beauty changes over the Chalmers 1982). harvest and regeneration cycle. Hull and Buhyoff (1986) modelled scenic beauty for each year over a period of 40 years. Their At the most general level, humans prefer unmanaged or more results showed strong increases in scenic beauty in about the first natural-looking forest scenery when compared to intensively ten years following harvest. Scenic beauty then increases more managed, recently harvested or intensively thinned areas (Daniel slowly as the forest ages until the next harvest event, when it and Boster 1976; Williamson and Chalmers 1982). In comparing drops again. Brunson and Reiter (1996) found that scenic quality the scenic effects of silvicultural techniques, the most essential increased slightly two years after harvesting. Shelby et al. (2003) finding is that the greater the proportion of forest cover removed found similar increases in scenic beauty over a ten-year period in from a logging area or coupe, the lower the scenic beauty of the a range of silvicultural systems. They also observed that the rate result in the first years after the operation (Vodak et al. 1985, at which scenic beauty improved varied with harvest system and cited in Ribe 1989). Consistent with this, large clear-cuts are that the ratings for more and less intensive systems converged least preferred. Smaller clear-cuts are preferred to larger ones, but over time. After about ten years, a clearfell system was given are still less attractive than shelterwood cuts (Daniel and Boster almost the same rating as less-intensive systems that had initially 1976; Schweitzer et al. 1976 cited in Ribe 1989). In the Pacific been judged more scenic. In a study of radiata pine plantations Northwest, Brunson and Shelby (1992) found that ‘new forestry’ in New Zealand, however, Thorn et al. (1997) found very little treatments that retained some trees had acceptance in between difference in visual quality at different points in time in a 20-year that of clearcuts and unharvested old-growth. In a recent study in clearfell cycle. Pacific Northwest forests, Ribe (2005) used computer simulations to control for differences other than the harvest systems and found Ribe (1989) identified some gaps in this body of knowledge. One that vista views of dispersed-retention patterns were considered is the absence of research comparing the scenic quality of different more beautiful than similar views of aggregated patterns of the silvicultural treatments applied to similar forests at the same age. same intensity. Tonnes et al. (2004) arrived at a similar conclusion Such comparisons would be particularly useful for choosing for Finnish forests, finding that the scenic quality of clear-cutting between silvicultural systems (Ribe 1989). In his more recent can be substantially improved by retaining at least 3 m3 ha–1 of research in the Pacific Northwest, Ribe (2005) began to address trees, and that scenic quality was most enhanced when the retained this gap, as did Tonnes et al. (2004) in Finland. Ribe (2005), trees were larger, in good condition and retained as individuals however, noted that further research is needed with different rather than in groups. participant cultures and forest types. The Australian research described here contributes to filling this knowledge gap. Research into regeneration treatments suggests slash and ground treatments are important to aesthetic judgements in the first few The social acceptability of managed forests years after harvest. Slash is universally disliked (Ribe 1989). Arthur (1977 cited in Ribe 1989) found slash volume to be the There is evidence that when people are asked to rate forest principal determinant of scenic beauty estimates. Burning slash management, their responses include both aesthetic reactions piles detracts from scenic beauty as long as the charred piles and more thought-based considerations such as values, beliefs remain evident (Schweitzer et al. 1976; Yeiser and Shilling and attitudes (Ribe 1999, 2002, 2006; Ford et al. 2009). The 1978 cited in Ribe 1989). However, after a year, clear-cuts with ‘social acceptability’ concept was chosen to be measured in this burned slash had higher scenic beauty than unburned sites due to research as it encompasses both aesthetic and more cognitive enhanced ground vegetation growth. After about three years the (thoughtful) responses based in people’s values and beliefs. The effects of different slash treatments converge as ground vegetation term ‘social acceptability’ can have a range of meanings (Brunson regrows (Anderson et al. 1982 cited in Ribe 1989). 1996; Shindler et al. 2002). For this research, it is defined as an attitude held by individual people that includes aesthetic and In contrast to the findings reported above that removal of forest cognitive components: cover decreases scenic beauty, thinning treatments that make the forest more open, while maintaining cover, can increase scenic The social acceptability of a forest practice, or condition, is beauty. Important visual elements of forests are the presence of an attitude (a bipolar evaluation) held by individuals. Social large trees and forest density. Scenic beauty research confirms acceptability has an aesthetic component, which is primarily intuitive preferences for mature and old-growth trees of all kinds affective, and a more cognitive component, which consists (Brush 1979). Ribe (1989) suggests one explanation for a general primarily of people’s underlying values and beliefs. As with other landscape evaluations, social acceptability is shaped by the preference for mature forests is that the high density of young physical characteristics of the forest, the characteristics of the forests detracts from scenic beauty (Daniel and Boster 1976; Brush people assessing it and both physical and social context. 1979; Hull and Buhyoff 1986). Measures of visual penetration have been found generally to be good predictors of scenic beauty An important aspect of this definition is that social acceptability in forests (Ruddell et al. 1989). Thinning treatments can increase of harvest systems measured for this research is the aggregate scenic beauty by altering forest density and the proportions of of many individual ratings, rather than for example the more small and large trees. For example Brush (1979) found that complex socially constructed group judgements of interest groups, Massachusetts woodland forest owners preferred the spaces professional bodies or policy-making organisations. We do not created by thinning well-stocked stands. Combining preferences
Australian Forestry 2009 Vol. 72 No. 4 pp. 157–171 160 Social acceptability of silvicultural alternatives in Tasmanian eucalypt forest seek to make claims about the acceptability of the systems to Research approach, design and method society overall, but present the mean judgements of different groups of individuals. In investigating these judgements, we are A comparative survey design was used to compare social interested in the role of both aesthetic responses and people’s acceptability ratings of eight silvicultural systems made by people existing values and beliefs that they bring to their judgements in three affiliation groups: large-scale timber-industry-affiliated; about forestry issues. conservation-affiliated and non-affiliated. Data were collected in two separate studies, a Main Study which presented six native Ribe (2002) compared scenic beauty and acceptability responses to forest silvicultural systems and a Calibration Study which the same forest scenes and found that for non-aligned participants, presented three of the native forest systems and two more extreme scenic beauty was a reasonable proxy for acceptability. However, systems. Follow-up qualitative interviews were conducted with for people with strong existing environmental attitudes to six people to help interpret quantitative trends. either resource protection or production, there were significant differences between acceptability and scenic beauty judgements. The research began with selection of silvicultural systems and People with stronger attitudes tend to make judgements that are development of simulated images of these systems. The simu more cognitive and those without strong attitudes tend to make lated images were then used as visual stimuli in the collection more aesthetic responses (Ribe 2002). These observations can be of quantitative social acceptability ratings from about 300 explained by Abelson’s (1988) argument that people’s attitudes participants. and beliefs are strengthened by repeated exposure. When an attitude is important to a person, it is associated with more Choice of management systems knowledge and stronger beliefs that are more quickly accessed and easier to apply than if the attitude is not important. Native forest harvest system selection for the Main Study drew on the range of feasible silvicultural systems chosen for the trial The small amount of research comparing the ‘social acceptability’ at the Warra LTER (Hickey et al. 2001). Various studies point to of alternative harvest systems shows similar trends to general harvest intensity and pattern as important characteristics that affect trends in scenic beauty research. Acceptability appears to increase public responses (Ribe 1989, 2005; Burchfield et al. 2003) and with the proportion of forest retained (Ribe 1999; Burchfield et so harvest systems were selected to provide a range of intensities al. 2003). and patterns. The systems (Fig. 1) were: The aim of the research described in this paper was to compare • a clearfell, burn and sow system that is routinely applied in the social acceptability of alternative silvicultural systems wet eucalypt forests investigated at the Warra LTER. The research draws on well- • aggregated-retention systems at two intensities:15% and 30% established approaches in landscape assessment research. Such retention research generally aims to understand people’s judgements as • dispersed-retention systems at two intensities: 15% and 30% these occur in everyday life, for example while driving through a retention forest or looking at pictures of forests. Some forestry professionals • a small-group or single-tree selection system with 80% retention have questioned this focus on ‘naïve’ judgements, made without at each harvest. the benefit of specific information about silvicultural practices. In contrast, we believe it is vital to understand uninformed public It was recognised that a wider range of management alternatives judgements, as these are types of responses that people make to would be of interest to many participants. The aim of the forestry in their everyday life. Studying these responses is more Calibration Study was to set the Main Study results in the context useful for predicting public responses to management actions of a wider range of options. The Calibration Study included two than studying responses based on the kinds of information that additional management systems, un-harvested forest and conver trained foresters use to make decisions. Another part of the overall sion of native forest to blue gum plantation. Conversion of native research project focused on a comparison of ‘naïve’ and ‘informed’ forest to blue gum plantation involves clearfelling and burning responses (Ford et al. in press).
Clearfell, burn and 15% aggregated 15% dispersed 30% aggregated 30% dispersed Selective (80% sow retention retention retention retention retention)
Figure 1. Main Study harvest systems — light-grey shading indicates the area where trees have been harvested and dark-grey shading indicates retained trees
Australian Forestry 2009 Vol. 72 No. 4 pp. 157–171 Rebecca M. Ford, Kathryn J. H. Williams, Ian D. Bishop and John E. Hickey 161 followed by windrowing and planting. The Calibration Study and regeneration systems, but representation of 90 or 100 years also included the clearfell system, the 30% aggregated-retention in a few minutes is very different from the way people normally system and the selection system to facilitate comparison of the experience forests. results with those of the Main Study (Fig. 2). Development of the images was driven by an aim to achieve realism, at least in those elements of the forest harvesting systems Development of simulated images that were important to acceptability judgements. Preliminary As the aggregated, dispersed and selection harvest systems studies were used to identify salient elements of the forest for described above were unfamiliar to many Tasmanians, visual simulation and to test the representational validity of the images. images were used to represent all of the systems. There were The processes used to develop these images, including technical only one or two field examples of the aggregated, dispersed and psychological considerations, have been published elsewhere and selective systems, and these had been harvested at different (Bishop et al. 2003, 2005; Williams et al. 2007). times and harvest intensities, so it was not possible to take comparable photographs. Simulated images were used instead, Data collection which meant the harvest systems could all be illustrated with the same topography, lighting conditions and time since harvest. This A mix of quantitative (questionnaire) and qualitative (interview) approach also meant that animated sequences could be prepared methods was used in the main data collection. While quantitative showing the forest regenerating over time. Much landscape research can distil people’s complex and individual mental states assessment research undertaken overseas has been conducted with into a few measured concepts of interest (Mertens 1998), quali close-up or vista-views of logging coupes shortly after harvest tative research can provide examples of some of the complexity (Ribe 1989, 1999, 2005; Brunson and Shelby 1992; Tonnes et underlying the quantitative trends. al. 2004). Such images are considered to have some validity A sample was selected from a population of Tasmanians over as research stimuli, as foreground views of recent harvests are the age of 18 years. Four main sampling locations were chosen: highly noticeable in the landscape and have been found to be an Greater Hobart, the Huon Valley, Launceston and Burnie with important factor in people’s judgements of forestry within the the aim of sampling people from urban and rural localities in wider landscape (Ribe et al. 2002). These are the types of views roughly the proportions as they occur in Tasmania. For each of that people are likely to notice if driving through a forest or these locations, publicly available lists of community and industry reading reports in the media or other sources. A simulated image organisations were located and combined to form a sampling was prepared for each harvest system representing the coupe frame (Bryman 2001). Community organisations such as sporting in the first year after harvest (see Fig. 3). Still images do not clubs, hobby clubs and service organisations were offered a small provide the complete spatial or temporal context in which forest donation of $5–10 per participant to the organisation’s funds. management takes place. To address the temporal aspect, each of Organisations were selected systematically from the sampling the harvest and regeneration systems was also represented as an frame to reduce researcher bias. animated sequence showing changes through time (see Fig. 4). These sequences began with an image of mature forest followed The Main Study sample was selected with the aim of including by the image of the first year after harvesting described above. about 150 non-affiliated participants representative of the Tas Subsequent images showed the forest in each of the following manian population and about 60 participants each for the less years to 90 years for the clearfell and aggregated-retention diverse large-scale timber-industry- and conservation-affiliated systems and 100 years for the selective systems. To make up the groups. The large-scale timber-industry-affiliates includes people animation, each still image was shown for one second, with longer whose work is in industrial-scale forest harvesting, and members pauses at key times since harvest (years 0, 3, 10, 25 and 89) to of organisations such as Timber Communities Australia. It does enable participants to compare systems at similar ages. These not include members of organisations such as Timber Workers for sequences provide more complete information about the harvest
Conversion to blue Clearfell, burn and 30% aggregated Selective (80% Unharvested forest gum plantation sow retention retention)
Figure 2. Calibration Study harvest systems — light-grey shading indicates the area where trees have been harvested and dark-grey shading indicates retained trees. The grey lines on the blue gum plantation diagram indicate windrows of piled logging residue.
Australian Forestry 2009 Vol. 72 No. 4 pp. 157–171 162 Social acceptability of silvicultural alternatives in Tasmanian eucalypt forest
Clearfell, burn and sow system
Small-group or single-tree selection (SGS) system
30% aggregated-retention system
30% dispersed-retention system
Figure 3. Examples of still, simulated images of forest in the first year after harvesting
Forests which support smaller-scale harvesting including selective For the Main Study four separate presentations were developed logging for specialty species. An additional 50 non-affiliated with different systems and presentation orders. Participants participants were sought for the Calibration Study. Large-scale viewed the clearfell and selective systems and either the 15% timber-industry-affiliated participants proved difficult to recruit, or 30% variants of the aggregated and dispersed systems. Some and despite additional efforts the number of participants in groups viewed the selective system first, and others saw the this group was smaller than intended. The reasons for this are clearfell, burn and sow system first. Two Calibration Study unclear, but could include the smaller numbers in this group and presentations were developed which showed the five systems in their geographic dispersal. Final numbers of participants were different orders. Groups of participants were randomly allocated large-scale timber-industry-affiliated — 30 people, conservation- to a presentation by a coin toss. For non-affiliated groups, another affiliated — 66 people and non-affiliated — 156 people for the coin toss decided whether they participated in the Main Study or Main Study and an additional 48 people for the Calibration the Calibration Study. Study.
Australian Forestry 2009 Vol. 72 No. 4 pp. 157–171 Rebecca M. Ford, Kathryn J. H. Williams, Ian D. Bishop and John E. Hickey 163
Small-group or single-tree selection system at year 2
Small-group or single-tree selection system at year 44
Small-group or single-tree selection system at year 86
Figure 4. Examples of single frames that were part of the animated sequence for the selective-logging system (republished from Ford et al. in press)
In late 2003 and early 2004, participants came in groups to a Before making acceptability judgements, participants were venue where the simulated images and instructions for filling in encouraged to make deliberate comparisons of the harvest the questionnaire were presented on a large screen. The screen systems. Near the start of the session, the still images were shown assembly was about 6 m wide × 1.5 m high and made of three in a block. One minute was allowed for initial viewing of each panels, with the end panels angled in 30° to provide a sense of of the systems in turn. The set of systems was then shown again immersion in a forest. Images were rear-projected onto the screen for half a minute each during which participants were asked to with three data-projectors located about 5 m behind each screen. make their rating. The well-known clearfell, burn and sow system The presentation had automated timing of the simulations and a was labeled, ‘Clearfell system’. The less-well-known alternatives voiceover. The voiceover set the context for the research as, ‘wet were not labeled, but designated, for example, ‘System B’ (Fig. 3). eucalypt forest available for harvesting’, and provided instructions Near the end of the session, the animated sequences were shown for making ratings, ‘When filling in the questionnaire I ask that in a block, first for three minutes each (pausing for 15 seconds at you answer the questions from your personal viewpoint. Please years 0, 3, 10 , 25 and 89 years), then again for two minutes at draw on your own values, knowledge, feelings and experience, each, pausing for 5 seconds at years 0, 3, 10, 25 and 89. in addition to the pictures that I show you’. Acceptability ratings were made on a seven-point scale in the The questionnaire asked participants to rate conviction and social questionnaire with points numbered from one to seven. The end acceptability among some other questions that are not reported and mid points of the scale were anchored with labels: ‘very here. Conviction ratings were made on a five-point scale using unacceptable’ one, ‘neutral’ four, and ‘very acceptable’ seven. Abelson’s (1988) six-item scale. These were: Six participants took part as individual interviewees. They were 1. How strongly do you hold your views about native forest recruited in a similar way to group participants, by contacting harvesting? community organisations. Interviewee recruitment targeted 2. How important are your views on native forest harvesting non-affiliated organisations, but of the six interviewees recruited, to you? three were actually conservation-affiliated. Interviewees were 3. How concerned are you about the issue of native forest asked to comment verbally on their questionnaire responses harvesting? immediately after they answered each question. Interviews were 4. How often do you think about native forest harvesting? tape recorded and later transcribed and loaded into the NVIVO software (QSR International Pty Ltd 2002). NVIVO facilitates 5. How often have you discussed your views on native forest the organisation of text in qualitative data analysis. The number harvesting with friends and family members? of interviewees is small and cannot be considered representative 6. Are you willing to work a day a month for a group supporting of Tasmanians, but was adequate to provide some examples of your views on native forest harvesting?
Australian Forestry 2009 Vol. 72 No. 4 pp. 157–171 164 Social acceptability of silvicultural alternatives in Tasmanian eucalypt forest the complexity underlying the general trends obtained from the been connected by lines to make the graphs easier to interpret. It much larger number of quantitative questionnaires. may not be meaningful to interpolate responses between points. The longer line for each affiliation connects the clearfell, 15% The non-affiliated Main Study sample demographics were aggregated, 30% aggregated and small-group selection (SGS) compared with those of the Tasmanian population to identify systems. The second shorter line for each affiliation connects the any sampling bias. The sample was found to be unbiased with 15% and 30% dispersed-retention systems. regard to gender and labour force participation, but had some bias towards older, tertiary-educated people and those working in the services sector. The most significant bias was in the number of people who had tertiary education — that is a diploma, bachelor degree or postgraduate degree. Just over half of the 7 sample was tertiary-educated, compared with about 15% of the Tasmanian population. To test how it might affect acceptability 6 results, weighted mean acceptability ratings were calculated. The 5 weightings were calculated by taking the Tasmanian proportion in each category and dividing it by the sample proportion. These 4 weightings are included in Table 2. Mean acceptability ratings 3 were then calculated for each education category. These were 2 multiplied by the weightings and added to provide an overall weighted average acceptability for each harvest system. These Acceptability (4 is neutral) 1 are presented in the next section to facilitate comparison with the 0 other acceptability results. 0 10 20 30 40 50 60 70 80 90 Forest retention (% at first harvest) Results non-affiliated (aggregated pattern) non-affiliated (dispersed pattern) Main Study acceptability judgements conservation (aggregated pattern) conservation (dispersed pattern) industry (aggregated pattern) industry (dispersed pattern) In Figure 5 social-acceptability ratings based on still images of the harvest systems in the first year after harvest have been Figure 5. Mean harvest system acceptability by affiliation of participants graphed on a scale that represents harvest intensity as per cent (still representation in the first year after harvest). Dispersed retention is of forest retained. The graphed points for each affiliation have shown as a separate line.
Table 1. Affiliation differences in Main Study harvest system acceptability (maximum acceptability rating 7) and Kruskal–Wallace test statistics (harvest systems viewed as still images in the first year after harvesting)
Affiliation Statistics Industry Nonaffiliated Conservation Harvest system Chisquare Mean (N) Mean (N) Mean (N) Pvalue (df = 2) Clearfell 6.3 (30) 2.0 (156) 1.1 (66) 115.8 <0.01 15% dispersed 5.8 0(8) 2.8 0(80) 1.5 (38) 035.1 <0.01 30% dispersed 3.6 (22) 2.3 0(76) 1.4 (28) 009.8 <0.01 15% aggregated 5.5 0(8) 3.8 0(80) 2.2 (38) 030.5 <0.01 30% aggregated 4.5 (22) 3.7 0(76) 2.6 (28) 008.8 <0.05 Smallgroup selection 2.4 (30) 5.3 (156) 5.1 (66) 037.9 <0.01
Table 2. Nonaffiliated Main Study harvest system acceptability weighted by education (maximum acceptability rating 7) (harvest systems viewed as still images in the first year after harvesting)
Unweighted Weighted Harvest system Mean (N) Mean (N) Clearfell 2.0 (156) 2.3 (156) 15% dispersed 2.8 0(80) 3.1 0(80) 30% dispersed 2.3 0(76) 2.3 0(76) 15% aggregated 3.8 0(80) 3.8 0(80) 30% aggregated 3.7 0(76) 3.4 0(76) Smallgroup selection 5.3 (156) 5.1 (156)
Australian Forestry 2009 Vol. 72 No. 4 pp. 157–171 Rebecca M. Ford, Kathryn J. H. Williams, Ian D. Bishop and John E. Hickey 165
Clearfelling is unacceptable to the conservation and non-affiliated Are differences between affiliations significant? groups (Table 1 and Fig. 5). The aggregated-retention system is more acceptable than the clearfell, and the SGS system is most The Kruskal–Wallis test was used to identify whether there were acceptable to these two groups. There appears to be an inverse significant differences between the judgements by the three relationship between acceptability and harvest intensity. This groups of affiliates of the harvest systems (see Table 1). Affiliation relationship should be viewed with caution as the pattern or type differences were significant for all of the harvest systems. They of harvesting is also important. The aggregated-retention pattern were significant to the 0.01 level, except for the 30% aggregated is more acceptable than dispersed retention at the same harvest system, for which affiliation differences were significant at the intensity. The large-scale timber-industry-affiliation showed an 0.05 level. opposite trend, judging clearfelling acceptable, and the SGS system unacceptable. There appears to be a linear relationship Are differences between harvest systems significant? between acceptability and harvest intensity. We do not know, Harvest intensity however, whether this relationship would hold between 30% and 80% forest retention. Each Main Study participant viewed only four of the six harvest systems. For comparisons in which the same participant viewed Non-affiliated acceptability weighted by participant education both systems (clearfell and 15% retention, 30% retention levels and SGS), the Wilcoxon test was used to test for significance (Table 3). The 15% and 30% variants of the aggregated and As noted previously, the main discernable difference between the dispersed systems were viewed by different participants. The non-affiliated part of the sample and the Tasmanian population was Mann–Whitney test was used to test for significance differences that the sample had higher education levels. To assess the effect between them (Table 4). that education might have on mean acceptability judgements, weighted mean acceptability of the still images was calculated These tests show that for the conservation and non-affiliated for each of the harvest systems (Table 2). participants, differences in acceptability between the clearfell and 15% retention systems were significant. Similarly, differences Weighted ratings are slightly higher for the clearfell system between the 30% retention systems and the SGS were significant. and slightly lower for the SGS system than unweighted ratings. The differences were not significant for the large-scale timber- However, the effect of the weighting is quite small and the industry affiliation. The small size of the industry sample may weighted ratings follow the same general trend as the unweighted have meant that potentially significant differences were not data; harvest systems which retain more forest are judged more detected. acceptable. Throughout the rest of the analysis the results are used in unweighted form. Some care is still needed in generalising Differences between the 15% and 30% retention variants of the from these results. aggregated and dispersed systems were not significant for any of
Table 3. Wilcoxon test statistics for differences in the intensity of harvest system pairs viewed by the same participant (harvest systems viewed as still images in the first year after harvesting)
Affiliation Harvest system pair Industry Nonaffiliated Conservation Zvalue Pvalue Zvalue Pvalue Zvalue Pvalue Clearfell and 15% dispersed –0.6 <0.52 –3.9 <0.01 –2.9 <0.01 Clearfell and 15% aggregated –1.1 <0.29 –6.6 <0.01 –4.1 <0.01 30% dispersed and SGS –1.7 <0.09 –6.9 <0.01 –4.5 <0.01 30% aggregated and SGS –2.4 <0.05 –4.7 <0.01 –4.2 <0.01
Table 4. Mann–Whitney test statistics for differences in the intensity of harvest system pairs viewed by different participants (harvest systems viewed as still images in the first year after harvesting)
Affiliation Harvest system pair Industry Nonaffiliated Conservation Zvalue Pvalue Zvalue Pvalue Zvalue Pvalue 15% dispersed and 30% dispersed –1.9 0.06 –1.8 0.07 –0.3 0.73 15% aggregated and 30% aggregated –0.9 0.38 –0.5 0.60 –0.9 0.36
Australian Forestry 2009 Vol. 72 No. 4 pp. 157–171 166 Social acceptability of silvicultural alternatives in Tasmanian eucalypt forest the affiliations, though for the dispersed system they are close to Calibration Study acceptability judgements significant for the industry- and non-affiliated groups. The fact that different participants viewed the 15% and 30% variants made The Calibration Study was designed to check whether the harvest it more difficult to detect any real differences, because statistical systems would be judged differently when presented in a different tests for independent samples are less sensitive. It is also possible context, that is, within a broader range of systems. It is striking that that, if the same participants had viewed both the 15% and 30% Main and Calibration Study acceptability ratings for the clearfell variant and been able to compare them directly, they may have and aggregated-retention systems are similar (Table 6). The only noticed the small difference in harvest intensity and reflected it significant difference in viewing a wider range of systems is that in their ratings. the SGS system was judged less acceptable by Calibration Study participants. Harvest pattern Conviction Each Main Study participant viewed both a dispersed and an aggregated variant at either the 15% or 30% harvest intensity. The Mean conviction ratings varied by affiliation, as might be expected Wilcoxon test was used to test for significant differences between (Table 7). Participants with conservation affiliation had the responses to these two harvest patterns (Table 5). strongest conviction, followed by those with industry affiliation. Non-affiliated participants had a lower mean rating again, with For the conservation and non-affiliated groups, the aggregated- a greater standard deviation. Differences in conviction between retention system was judged significantly more acceptable than affiliations were significant. the dispersed-retention system of the same harvest intensity. For participants with large-scale timber-industry affiliation there was no significant difference between the two harvest patterns.
Table 5. Wilcoxon test statistics for differences in harvest pattern (harvest systems viewed as still images in the first year after harvesting)
Affiliation Harvest system pair Industry Nonaffiliated Conservation Zvalue Pvalue Zvalue Pvalue Zvalue Pvalue 15% dispersed and 15% aggregated –0.4 0.68 –5.1 <0.01 –3.6 <0.01 30% dispersed and 30% aggregated –1.8 0.07 –5.7 <0.01 –3.7 <0.01
Table 6. Comparison of Main Study and Calibration Study harvest system acceptability (nonaffiliated group only) and Mann–Whitney test P values for differences in acceptability (harvest systems viewed as still images in the first year after harvesting)
Presentation Statistics
Harvest system Calibration study Main study Nonaffiliated Nonaffiliated Zvalue Pvalue Mean (N = 48) Mean (N) Conversion to blue gum plantation 2.6 – Clearfell 2.1 2.0 (156) –0.9 <0.38 15% dispersed – 2.8 0(80) 30% dispersed – 2.3 0(76) 15% aggregated – 3.8 0(80) 30% aggregated 3.5 3.7 0(76) –0.4 <0.66 Smallgroup selection 4.5 5.3 (156) –2.9 <0.01 Unharvested forest 5.9 –
Table 7. Mean conviction ratings by affiliation (maximum rating 5), including Kruskal–Wallis test statistics
Affiliation Statistics Industry Nonaffiliated Conservation Attribute Chisquare N = 30 N = 156 N = 66 PValue (df = 2) Mean SD Mean SD Mean SD Conviction 3.9 0.75 3.2 0.88 4.1 0.70 55.6 <0.01
Australian Forestry 2009 Vol. 72 No. 4 pp. 157–171 Rebecca M. Ford, Kathryn J. H. Williams, Ian D. Bishop and John E. Hickey 167
Does viewing animated sequences make a difference to 7 social acceptability ratings? 6 Responses to animated sequences are shown in Figure 6. Accept ability ratings of animated sequences and the corresponding still 5 images in the first year after harvest are compared in Table 8. There 4 is a general trend among non-affiliated and conservation-affiliated participants to increased acceptance of the more intensive 3 systems when viewed as animations, and decreased acceptance 2
of the selective system. These effects are quite small, with most Acceptability (4 is neutral) participants giving still images and animated sequences the same or similar ratings. Despite being small, the animation effects are 1 significant or near significant for most or all harvest systems for 0 the non-affiliated and conservation-affiliated groups. They were 0 10 20 30 40 50 60 70 80 90 most significant for the clearfell and selective systems. Forest retention (% at first harvest)
non-affiliated animated non-affiliated animated Qualitative responses to harvest systems (aggregated pattern) (dispersed pattern) conservation animated conservation animated Interviewees’ self reports of their responses to the still images (aggregated pattern) (dispersed pattern) industry animated industry animated illustrate some of the complexities underlying the quantitative (aggregated pattern) (dispersed pattern) trends. These qualitative data are not representative of the quantitative sample, nor the general population. Nonetheless they Figure 6. Mean harvest system acceptability by affiliation of participants provide some insights into the minds of some members of the (animated sequences showing change over a period of 90 years for the public and the reader is asked to take a leap of the imagination clearfell, dispersed and aggregated systems, and 100 years for the small- and try to see the world from these lay perspectives. group selection system). Five of the six interviewees rated the silvicultural systems consistent with the general non-affiliated trend, that is systems that retained more forest were rated more acceptable. Interviews Oh I think it was just … the look, you know, the visual aspect I showed different combinations of aesthetic and thought-based think, that I was looking at. I thought the first one, the clearfell, responses to the harvest systems. Two interviewees reported their I thought that was dreadful … that was worse than the others responses in terms of the aesthetics or ‘the look of’ the forest, I think. for example:
Table 8. Animation effects on harvest system acceptability by affiliation and Wilcoxon test statistics. ‘Still’ images are of the first year after harvesting while ‘animated’ sequences cover a period of 90 years for the clearfell, dispersed and aggregated systems, and 100 years for small group selection system.
Affiliation Harvest system Industry Nonaffiliated Conservation Mean Z P Mean Z P Mean Z P Clearfell (N = 30) (N = 156) (N = 66) Still 6.3 –1.5 0.25 2.0 –5.9 <0.01 1.1 –3.5 <0.01 Animated 6.4 2.6 1.5 15% dispersed (N = 8) (N = 80) (N = 38) Still 5.8 –1.5 0.13 2.8 –3.8 <0.01 1.5 –1.5 <0.14 Animated 5.1 3.3 1.6 30% dispersed (N = 22) (N = 76) (N = 28) Still 3.6 –0.6 0.52 2.3 –3.1 <0.01 1.4 –2.9 <0.01 Animated 4.0 2.8 2.1 15% aggregated (N = 8) (N = 80) (N = 38) Still 5.5 –1.6 0.10 3.8 –2.8 <0.01 2.2 –0.4 <0.71 Animated 5.0 4.3 2.3 30% aggregated (N = 22) (N = 76) (N = 28) Still 4.5 –1.3 0.18 3.7 –2.2 <0.05 2.6 –2.7 <0.01 Animated 5.1 4.0 3.5 Small-group selection (N = 30) (N = 156) (N = 66) Still 2.4 –0.4 0.72 5.3 –4.2 <0.01 5.1 –3.6 <0.01 Animated 2.5 4.8 4.4
Australian Forestry 2009 Vol. 72 No. 4 pp. 157–171 168 Social acceptability of silvicultural alternatives in Tasmanian eucalypt forest
These two interviewees were non-affiliated and had lower Only four interviewees viewed both the dispersed and aggregated conviction scores than most of the other interviewees (2.3 and 3.5 systems as two interviewees viewed the Calibration Study respectively out of a maximum conviction score of 5). presentation, which did not include the dispersed retention. Of the four who viewed both, two gave the same ratings to the Three interviewees reported a mix of aesthetic and more cognitive aggregated and dispersed systems and two rated the aggregated- responses, for example: retention higher. One of these interviewees stated explicitly that … As an overview I think I was looking at, of course the aesthetic his judgements of these systems were based on aesthetics, ‘I’ve appeal, that’s the first thing, that’s open to your mind. But more rated according to like the aesthetic appeal’; the other used a mix important was: were there corridors of wildlife preservation able of aesthetic and cognitive considerations to judge the systems: to be represented in those particular images? So that pushed me only to one acceptable mechanism, which was [small-group The [dispersed-retention] I ticked very unacceptable, but I started selection]. to think well, those trees that are still standing, are they even going to stay up there? … because, you’re going to get wind and The conviction scores of these three interviewees were towards … its going to affect what’s already there so I wondered about … the higher end of the scale (all over 4). In the example above, the wind and those sorts of things. The [aggregated-retention], well I interviewee articulated an initial aesthetic reaction followed by a ticked sort of two for, not as unacceptable as the others, but still unacceptable because there were just large patches and … I’ve more cognitive assessment of the presence of wildlife corridors. noticed nearby forests, there’s been a lot of wind damage in those This illustrates how visual images provide both stimulus for areas. So I was concerned I guess about wind damage, and also scenic-beauty responses and additional information that people just how much of the remaining ecosystems would still be there can assess for consequences of concern to them, such as presence in the remaining parts because of those areas being logged. or absence of wildlife habitat. This interviewee’s final criterion for judging, the amount of This more complex way of responding to the simulations again remaining ecosystem, indicates one possible reason why the led to the general trend of judging more intensive systems less aggregated pattern was generally preferred over the dispersed. acceptable. It is not possible to tell whether the aesthetic or the Participants viewed both systems at nominally the same harvest cognitive response is more powerful, as there is little conflict intensity, either both at 15% or both at 30% retention depending between them. Visual appeal increases with the proportion of on which presentation they saw. However, the labelling of harvest intact vegetation (Ribe 1989), and assessments of remaining systems as 15% or 30% retention is based on eucalypt tree wildlife habitat are also likely to be more favourable where more retention, not retention of the whole forest ecosystem, which in forest is retained. this forest type contains a tall dense understorey. If understorey is considered, then the aggregated-retention system, which retains One interviewee reported only cognitive concerns. He was 15% or 30% of the understorey as well as the eucalypts, actually the only interviewee whose ratings were not consistent with has a higher proportion of retained forest as a whole. Preference the general trend for non-affiliated and conservation-affiliated for this system is therefore consistent with the general trend. responses. Finally the interviews provide some clues as to the basis for It’s the biodiversity, plus also the complete and utter waste of a resource … with clearfelling, the sawlog component is about animation effects. Qualitative data relating to animation effects 10%, and the chiplog is about 90%. have been presented elsewhere (see Ford et al. in press). The general trend in which animated sequences representing the more This interviewee had a reasonably high conviction score (4.17) intensive systems are judged slightly more acceptable than still and was conservation-affiliated. His cognitive criteria for judging images is consistent with the older forest having greater scenic the systems were initially stated quite simply, biodiversity and beauty. However, the size of the effect is small for a range of waste. He later expanded by drawing on pre-existing beliefs reasons. Some interviewees based their judgements on their about the consequences of clearfelling. Beliefs about sawlog to own beliefs about the way forests grow back after harvesting pulpwood ratios were clearly prior knowledge, and not gleaned rather than the regrowth shown in the simulated images. Other from viewing the images. Aesthetics or visual appeal were not participants noted that the forest grew back but considered the mentioned in this interview. timescale too long. The most significant finding, that the selective system was rated less acceptable based on the animated sequences, Considering these six interviews together, there appears to be is due to the animations containing new information, that there a pattern of interviewees with lower conviction scores making are multiple harvest events in one rotation of this system (Ford more aesthetic judgements, and those with higher conviction et al. in press). scores including a cognitive component in their acceptability judgements. This pattern is to be expected, as some of the indicators of conviction used in the questionnaire, such as Discussion and conclusions ‘thought’, are cognitive. This research has estimated social acceptability ratings of a range One quantitative result appears at first glance to be inconsistent of silvicultural alternatives to the clearfell, burn and sow system with the general trend of increased forest retention leading to for wet eucalypt forests presented as visual images. The general greater acceptability. This is that the aggregated-retention pattern finding that, for conservation-affiliated people and the general was rated significantly higher than the dispersed-retention pattern public, acceptability increases with the proportion of forest for the same harvest intensity. Qualitative examples assisted in retained is consistent with previous studies that have found that looking at this result more closely. both scenic beauty and acceptability ratings increase with the
Australian Forestry 2009 Vol. 72 No. 4 pp. 157–171 Rebecca M. Ford, Kathryn J. H. Williams, Ian D. Bishop and John E. Hickey 169 proportion of forest retained ( McCool et al. 1986; Ribe 1989, systems are named according to the percentage of eucalypt trees 2005; Burchfield et al. 2003; Tonnes et al. 2004). Qualitative retained, and their names do not reflect the percentage of whole observations that these acceptability judgements were based on forest retained, which is greater for the aggregated system since a aesthetic evaluations by some people and either partially or wholly percentage of tall dense understory vegetation is retained similar cognitive considerations by others illustrate the aesthetic and to that of eucalypts. The greater acceptance of aggregated systems cognitive components of social acceptability (Ribe 1999, 2002; is therefore consistent with the general trend of preferring harvest Shindler et al. 2002). The observation that interviewees with systems that retain a greater proportion of the forest at first harvest. lower levels of conviction were more inclined to base their ratings Because understorey and regeneration methods such as burning on aesthetics and that people with higher levels of conviction vary so much between forest types, there may be no broad general made more belief-based cognitive judgements is consistent rule to be discovered concerning the relative acceptability of with Ribe’s (2002) finding that scenic beauty and acceptability harvest patterns. judgements are similar only for people without strong attitudes to the environment. The finding that more intensive systems were more acceptable when the full cycle of forest harvest and regenerating was viewed For those with large-scale timber-industry affiliation, the is consistent with other studies in which scenic beauty was trend runs in the opposite direction, with clearfelling judged found to increase over time after harvesting (Hull and Buhyoff acceptable. It is little wonder that there is conflict in Tasmania 1986; Brunson and Reiter 1996). However, the small size of the over forestry. This finding differs from much previous research difference was surprising. It is possible that the strong visual in that it shows a fundamental difference between the responses impact of the initial year after harvest is weighted more strongly of industry-affiliated and other participants, not just a matter by participants in making their judgements and colours their of degree. Scenic-beauty studies have found consistently that reactions to the remainder of the animated sequences. It is also industry-affiliated participants rate recently harvested sites as possible that participants have less feeling for, or confidence in, more beautiful than the general public rates them, but the trend these predictions about how the forest will look a long way into line is nonetheless in the same direction as for the general public, the future. While a still image of a forest soon after harvest can with systems that retain more forest rated more beautiful. Ribe be easily compared with harvested coupes seen in the field, it (2002) found that acceptability judgements of forest scenes made is more difficult to compare an animation with reality when it by people with attitudes in favour of production were higher represents a 90–100-year time frame in a few short minutes. The than scenic beauty judgements, but nonetheless followed this finding that selective logging is considerably less acceptable when same general trend. Ribe (2002) ascribed this result to scenic the full 100-year cycle is viewed suggests that most people are beauty being a contributor to ‘productionists’ acceptability generally not aware that selective logging involves several harvest judgements, but also being something that could be discounted events over the rotation period, and that the animated sequences if what they considered to be good management produced ugly provided this important new information to them. results. It is possible that the Tasmanian industry-affiliated group was more experienced and knowledgeable than Ribe’s (2002) What do these findings mean for forest management and policy? ‘productionists’. In this research it appears that aesthetics has They confirm that clearfelling is disliked by many members been more strongly discounted to the point where it does not of the public, and show that retaining even a small proportion seem to have played a part in the industry affiliate’s acceptability of trees on a coupe, as low as 15%, can lead to a significant judgements at all. Ford et al. (2009) show that beliefs that increase in social acceptability. This is consistent with findings clearfelling is safer, more economically efficient and will in the in the Pacific Northwest of the USA (Ribe 2005) and Finland longer term lead to a good quality regrowth forest for timber (Tonnes et al. 2004) and should provide managers considering production are important underlying factors in social acceptability a shift away from clearfelling with some cause for optimism. ratings made by this group. The research also provides clear direction to managers that aggregated-retention patterns are more acceptable than dispersed This research found that aggregated-retention systems were retention in wet eucalypt forests, though this may not hold for more acceptable to all participants than dispersed-retention other forest types. systems of similar harvest intensity. This finding differs from the latest findings of Tonnes et al. (2004) and Ribe (2005) in The significant differences found between the three affiliation which dispersed patterns were preferred over aggregated. But groups’ acceptability ratings confirm that there is no easy policy our finding is consistent with Ribe’s (1999) preliminary research. solution, as there is no harvest system that is acceptable to One explanation for these differing results is that factors such as everybody. Mean ratings made by the three groups were closest for viewpoint, the condition of retained trees and understorey can the 30% aggregated-retention system. It is interesting to consider affect scenic beauty (Tonnes et al. 2004). Viewpoint is important how these mean acceptability ratings might be interpreted. as understorey or retained trees can hide some of the evidence Brunson and Kruger (1996) ask whether social acceptability is of harvesting (Tonnes et al. 2004). The oblique aerial view a preference or a threshold. Which systems do members of the used in the Ribe studies may have hidden much of the negative general public prefer, and which will they just be willing to put visual impact below the dispersed tree crowns. With the in-stand up with? We do not know exactly where the threshold for this view used in this research, there was nothing to hide the burnt kind of acceptance lies on the scale and it is difficult to know seedbed, understorey vegetation and tree stems resulting from the whether the 30% aggregated-retention system, if it replaced dispersed-retention treatment. The aggregated-retention pattern, clearfelling, would be accepted by any of the affiliation groups by contrast, contained areas of unburnt forest with understorey without protest. intact that served to hide some of the harvested coupe. These
Australian Forestry 2009 Vol. 72 No. 4 pp. 157–171 170 Social acceptability of silvicultural alternatives in Tasmanian eucalypt forest
Adding to the difficulty of making such predictions is the fact that to the many people who contributed to this project, particularly participants tended to use the acceptability scale at least partly as Daniel Loiterton and Trevor Webb. The authors acknowledge two a relative, rather than an absolute, measure of acceptability. The anonymous examiners for their helpful comments on the PhD comparison of Calibration Study and Main Study results showed thesis of Ford (2006) that formed part of this research. that people tend to use all of the scale in making their judgements, placing their least- and most-preferred systems at the ends, and References systems that they viewed as being intermediate somewhere in the middle of the scale. The range of available options presented, Abelson, R.P. (1988) Conviction. American Psychologist 43, 267–275. particularly inclusion or exclusion of the more extreme options, Anderson, L.M., Levi, D.J., Daniel, T.C. and Dieterich, J.H. (1982) can affect ratings, as was demonstrated for the SGS system, which The Esthetic Effects of Prescribed Burning: A Case Study. USDA was rated significantly higher when the ‘unharvested forest’ option Forest Research Note RM-413 Rocky Mountain Forest and Range Experiment Station, Fort Collins, Colorado. was not available. One can speculate that if, at the other end of the Arthur, L.M. (1977) Predicting scenic beauty of forest environments: scale, the clearfell system were not included as an option in the some empirical tests. Forest Science 23, 151–159. research, the aggregated- and dispersed-retention systems might Bishop, I.D., Fasken, G., Ford, R.M., Hickey, J.E., Loiterton, D. and appear more extreme and perhaps be rated lower. This could also Williams, K.J.H. (2003) Visual simulation of forest regrowth under happen in reality if clearfelling were to be phased out in some or all different harvest options. In: Trends in Landscape Modelling. of Tasmania’s forests. An alternative such as aggregated-retention Proceedings at Anhalt University of Applied Sciences, Dessau, might then be judged less acceptable because there was nothing Germany, pp. 46–55. worse for comparison. Bishop, I.D., Ford, R.M., Loiterton, D. and Williams, K.J.H. (2005) Studying the acceptability of forest management practices using The application of these social findings to forest management visual simulation of forest regrowth. In: Bishop, I.D. and Lange, is not straightforward. For example, the results indicate that E. (eds) Visualization in Landscape and Environmental Planning. a selective harvesting system would be most acceptable to Taylor and Francis, London, pp. 112–119. non-affiliated people. However, the general public may not be Brunson, M. (1996) A definition of ‘Social Acceptability’ in ecosystem management. In: Brunson, M., Kruger, L., Tyler, C. and Schroeder, aware of some of the consequences of this system (Ford et al. S. (eds) Defining Social Acceptability in Ecosystem Management: in press). Clearly the full range of characteristics of the harvest A Workshop Proceedings. US Department of Agriculture Forest systems (social, ecological and economic) should be considered Service, Portland, Oregon, pp. 7–16. in decision-making. It is important for managers to communicate Brunson, M. and Kruger, L. (1996) Introduction: social acceptability to the public the reasons for silvicultural choices that may not in ecosystem management. In: Brunson, M., Kruger, L., Tyler, C. otherwise be obvious to them. Ford et al. (in press) found that, and Schroeder, S. (eds) Defining Social Acceptability in Ecosystem at least in a research setting, information about the consequences Management: A Workshop Proceedings. US Department of for fauna species habitat, soil and water, quantity of forest Agriculture Forest Service, Portland, Oregon, pp. 1–5. products, economics, forest worker safety and fire safety led to the Brunson, M., Reiter, D.K. (1996) Effects of ecological information on judgements about scenic impacts of timber harvest. Journal of selective harvest system being rated less acceptable and the 30% Environmental Management 46, 31–42. aggregated-retention system being judged more acceptable than Brunson, M. and Shelby, B. (1992) Assessing recreational and scenic they were judged without the information, as reported here. For quality. Journal of Forestry 90(7), 37–41. the selective system the information about consequences and the Brush, R.O. (1979) The attractiveness of woodlands: perceptions of forest animated sequences reported here had a similar effect; both types landowners in Massechusetts. Forest Science 25, 495–506. of information led to lower acceptability ratings. There is clearly Bryman, A. (2001) Social Research Methods. Oxford University Press, room to provide the public with more information about this Oxford. system in comparison to the more intensive systems. In general it Burchfield, J.A., Miller, J.M., Allen, S., Schroeder, R.F. and Miller, T. is important to consider the role of communication and education (2003) Social Implications of Alternatives to Clearcutting on the in formulating policy regarding silvicultural practices. Tongass National Forest — An Exploratory Study of Residents’ Responses to Alternative Silvicultural Treatments at Hanus Bay, Additional research would improve the match between the Alaska. USDA Forest Service, Pacific Northwest Research Station, artificial research context and the Tasmanian landscape and policy PNW-GTR-575, Missoula, Montana. environment. For example, measuring the social acceptability of Daniel, T.C. (2001a) Aesthetic preference and ecological sustainability. In: Sheppard, S.R.J. and Harshaw, H.W. (eds) Forests and alternative harvest systems applied in different types of forests, Landscapes: Linking Ecology, Sustainability and Aesthetics. such as in old-growth and regrowth native forest within a IUFRO Research Series No. 6. CABI Publishing in association landscape context, could be of direct relevance to forest managers. with the International Union of Forestry Research Organisations: Additional qualitative research may be helpful to explore in more Wallingford, Oxon, pp. 15–29. detail the ways in which people judge social acceptability. Daniel, T.C. (2001b) Whither scenic beauty? Visual landscape quality assessment in the 21st century. Landscape and Urban Planning Acknowledgements 54, 267–281. Daniel, T.C. and Boster, R.S. (1976) Measuring Landscape Esthetics: The The research described in this paper was part of a larger project, Scenic Beauty Estimation Method. USDA Forest Service, Rocky Mountain Forest and Range Experiment Station, Research Paper Social Acceptability of Forest Management Systems, funded by RM-167, Fort Collins, Colorado. the Australian Research Council with industry contributions Dargavel, J. (1995) Fashioning Australia’s Forests. Oxford University from Forestry Tasmania and the Bureau of Rural Sciences. It Press, Melbourne. was approved by the University of Melbourne Human Research Ethics Committee, Arts and Humanities sub-committee. Thanks
Australian Forestry 2009 Vol. 72 No. 4 pp. 157–171 Rebecca M. Ford, Kathryn J. H. Williams, Ian D. Bishop and John E. Hickey 171
Endter-Wada, J., Blahna, D., Krannich, R. and Brunson, M. (1998) Ribe, R. (2006) Perceptions of forestry alternatives in the US Pacific A framework for understanding social science contributions to Northwest: information effects and acceptability distribution ecosystem management. Ecological Applications 8, 891–904. analysis. Journal of Environmental Psychology 26, 100–115. Ford, R.M. (2006) Social acceptability of forest management systems. Ribe, R.G., Armstrong, E.T. and Gobster, P.H. (2002) Scenic vistas and PhD thesis, The University of Melbourne. the changing policy landscape: visualising and testing the role of Ford, R.M., Williams, K.J., Bishop, I.D. and Webb, T.J. (2005) Social visual resources in ecosystem management. Landscape Journal acceptability of forest management systems: Project overview. The 21, 42–68. University of Melbourne, Parkville, http://www.landfood.unimelb. Routley, R. and Routley, V. (1974) The Fight for the Forests: The edu.au/research/social/psychology/projects.html. Takeover of Australian Forests for Pines, Woodchips and Intensive Ford, R.M., Williams, K.J., Bishop, I.D. and Webb, T.J. (2009) A value Forestry. Research School of Social Sciences, Australian National basis for the social acceptability of clearfelling in Tasmania, University: Canberra. Australia. Landscape and Urban Planning 90, 196–206. Ruddell, E.J., Gramann, J.H., Rudis, V.A. and Westphal, J.M. (1989) Ford, R.M., Williams, K.J., Bishop, I.D. and Hickey, J.E. (in press) The psychological utility of visual penetration in near-view forest Effects of information on the social acceptability of alternatives scenic-beauty models. Environment and Behaviour 21, 393–412. to clearfelling in Australian wet eucalypt forests. Environmental Schweitzer, H.W., Ullrich, J.R. and Benson, R.E. (1976) Esthetic Management, DOI 10.1007/s00267-009-9392-7. Evaluation of Timber Harvesting in the Northern Rockies: A Progress Forest Practices Board (2000) Forest Practices Code. Forest Practices Report. USDA Forest Research Note INT-203, Intermountain Forest Board, Hobart, Tasmania. and Range Experiment Station, Ogden, Utah. Gobster, P.H. (2001) Foreword. In: Sheppard, S.R.J. and Harshaw, H.W. Shelby, B., Thompson, J., Brunson, M. and Johnson, R. (2003) Changes in (eds) Forests and Landscapes: Linking Ecology, Sustainability scenic quality after harvest: a decade of ratings for six silvicultural and Aesthetics. IUFRO Research Series No. 6. CABI Publishing treatments. Journal of Forestry 101, 30–35. in association with the International Union of Forestry Research Shindler, B.A., Brunson, M.W. and Stankey, G.H. (2002) Social Organisations, Wallingford, Oxon, pp. xxi–xxviii. Acceptability of Forest Conditions and Management Practices: A Hickey, J.E., Neyland, M.G. and Bassett, O.D. (2001) Rationale and Problem Analysis. US Department of Agriculture, Forest Service, design for the Warra Silvicultural Systems Trial in wet Eucalyptus PNW-GTR-537, Pacific Northwest Research Station. obliqua forests in Tasmania. Tasforests 13, 155–182. Tasmania Together Community Leaders Group (2001) Tasmania Hocking, T. (2004) Forestry management — Enterprise Marketing and Together. The Group, Hobart. Research Services poll report card: room for improvement. The Tasmanian Public Land Use Commission (1996) Tasmanian– Sunday Examiner, Launceston, Australia, 21 March 2004, pp. 1, Commonwealth Regional Forest Agreement, Social and Economic 12–15. Report. Vol. 1. Background Report, The Commission, Hobart, Hull, R.B. and Buhyoff, G.J. (1986) The Scenic Beauty Temporal Tasmania. Distribution Method: an attempt to make scenic beauty assessments Thorn, A.J., Daniel, T.C., Orland, B. and Brabyn, N. (1997) Managing compatible with forest planning efforts. Forest Science 32, forest aesthetics in production forests. New Zealand Journal of 271–286. Forestry 42(2), 21–29. McCool, S.F., Benson, R.E. and Ashor, J.L. (1986) How the public Tonnes, S., Karjalainen, E., Lofstrom, I. and Neuvonen, M. (2004) Scenic perceives the visual effects of timber harvesting: an evaluation impacts of retention trees in clear-cutting areas. Scandinavian of interest group preferences. Environmental Management 10, Journal of Forest Research 19, 348–357. 385–391. Vodak, M.C., Roberts, P.L., Wellman, J.D. and Buhyoff, G.J. (1985) Mertens, D.M. (1998) Research Methods in Education and Psychology: Scenic impacts of eastern hardwood management. Forest Science Integrating Diversity with Quantitative and Qualitative Approaches. 31, 289–302. Sage Publications, Thousand Oaks, California. Williams, K.J., Ford, R.M., Bishop, I.D., Loiterton, D. and Hickey, QSR International Pty Ltd (2002) QSR NVIVO Version 2.0.161. In: QSR, J.E. (2007) Realism and selectivity in data-driven visualisations: http://www.qsrinternational.com/ (accessed 4 Nov 2009). a process for developing viewer-oriented landscape surrogates. Ribe, R. (1989) The aesthetics of forestry: what has empirical preference Landscape and Urban Planning 81, 213–224. research taught us? Environmental Management 13, 55–74. Williamson, D.N. and Chalmers, J.A. (1982) Perceptions of Forest Scenic Ribe, R. (1999) Regeneration harvests versus clearcuts: public views of Quality in Northeast Victoria: A Technical Report of Research the acceptability and aesthetics of Northwest forest plan harvests. Phases I and II. Forests Commission Victoria, Melbourne. Northwest Science 73, (Special Issue), 102–117. Yeiser, J.L. and Shilling, C.L. (1978) Student responses to selected terms Ribe, R. (2002) Is scenic beauty a proxy for acceptable management? and scenes in natural resource management. Journal of Forestry The influence of environmental attitudes on landscape perceptions. 76, 497–498. Environment and Behaviour 34, 757–780. Ribe, R. (2005) Aesthetic perceptions of green-tree retention harvests in vista views: the interaction of cut level, retention pattern and harvest shape. Landscape and Urban Planning 73, 277–293.
Australian Forestry 2009 Vol. 72 No. 4 pp. 157–171 172 Low-rainfall species trials in New South Wales Part 1
Low-rainfall species trials in New South Wales Part 1: Survival, growth and form over the first 57 months
I.G. Johnson1,2, A. Kathuria1, J. Parekh1 and C. Barton1
1Forest Science Centre, NSW Department of Industry and Investment, PO Box 100, Beecroft, NSW 2119, Australia 2 Email: [email protected]
Revised manuscript received 14 August 2009
Summary Introduction A trial of potentially commercial tree species, planted over 45 The extensive clearing of natural woodlands and forest on the sites across the 500–800 mm annual rainfall zone of New South western slopes and the eastern section of the western plains of Wales, was assessed for survival and height at ages 10–57 months, New South Wales (NSW) has led to changes in the water balance, and for stem form at 57 months. The sites were located from the accompanied by a greatly increased risk of dryland salinity. North-west Slopes and Plains to the South-west Slopes, covering Dryland salinity or shallow water-tables already affect large areas about six degrees of latitude, and were designated as occurring (about 152 000 ha) within the Murray–Darling Basin in NSW; in North, Central or South region. The plantings overall included 93% of this is agricultural land. The area at risk is predicted to nine species, with six of these planted at each individual site. increase to 1.3 million ha by 2050. Salt loads in streams in some A mixed-model analysis of variance was applied to height and catchments may exceed international drinking water guidelines survival data for all ages. Species, age and species × region by this time (ANRA 2008). interaction effects were highly significant for both traits, while region was also significant in the case of survival. High survival One method to reduce the spread of salinity is to re-establish trees rates of over 80% were maintained for 57 months across sites by in strategic locations in the landscape. In the main agricultural Eucalyptus camaldulensis, E. sideroxylon and Saltgrow™ hybrid belt in south-eastern Australia, an area that has a medium to low –1 eucalypt clones. Survival of Pinus pinaster and Acacia mearnsii rainfall of about 500–800 mm y , tree plantations of suitable declined rapidly within 2–3 y to 17–34% overall. Pinus pinaster species may be established successfully without irrigation. survived especially poorly in the North region, possibly due to Trees planted in recharge and break-of-slope zones, alone or in clay soils at several sites. Frost was considered to have reduced conjunction with shrubs or perennial pastures, can be useful for survival of two spotted gum species (Corymbia maculata and controlling the expression of salt-laden groundwater to the surface. C. citriodora subsp. variegata) in the North and Central regions The effectiveness of such plant-based solutions is governed by relative to the South. Overall superior species for height included many factors including hydrogeology, location of plantings, E. camaldulensis, C. maculata, SaltgrowTM clones and A. mearnsii species planted, climate, watertable salinity and soil type (Kuginis (4.8–5.5 m at 57 months). Pinus pinaster was much the poorest and Daly 2001). There are several documented cases in Australia (overall mean 1.6 m), growing best in the South region. Eucalyptus of trees lowering water tables under a wide range of landscape cladocalyx of southern origin and E. argophloia of northern origin conditions, although the magnitude of response has varied (e.g. appeared, from limited data, to be equally adapted to the North and Reid 1995; Stolte et al. 1997; Raper 1998). South regions. From analyses of the proportions of trees with no Tree plantings in medium- to low-rainfall areas of south-eastern forking and with superior stem straightness, species, region and Australia may have a range of additional environmental and species × region effects were significant.Pinus pinaster tended to commercial benefits apart from, or in conjunction with, reducing have the straightest and least-forked stems, while E. cladocalyx, the threat of salinity. Notably there is potential for financial returns represented by a known poor provenance, was poorest for these to landowners from wood production. Small-scale, private farm traits. The spotted gums also had high proportions of straight trees forestry aimed at producing sawn wood, poles, firewood and (75–77%) but had somewhat lower numbers of unforked trees. charcoal is presently in its infancy in the 500–800 mm rainfall Proportions of unforked trees were significantly greater in Central zone. The current area of plantations classified as ‘farm forestry’ compared with other regions; the reason for this is not known. in NSW is about 28 000 ha, but only a minor proportion is Keywords: plantations; arid zones; species trials; growth; survival; stem located in the low-rainfall zone of interest here (URS Forestry form; spotted gums; Eucalyptus species; Acacia mearnsii; Pinus pinaster; 2008). ABARE and BRS (2001) estimate that there are several New South Wales million hectares of land in western NSW capable of growing
Australian Forestry 2009 Vol. 72 No. 4 pp. 172–183 I.G. Johnson, A. Kathuria, J. Parekh and C. Barton 173 timber plantations, but most is classed as 4 or 5 for productivity accuracy of predictions. More detailed information can be found (‘conventional plantation development may be possible with at http://www.greenhousegas.nsw.gov.au/. improved establishment techniques and management practices’; and ‘may be capable of tree planting to provide environmental The Australian Federal Government plans to establish a national benefits’). Indicative mean annual increments (MAI) for these Carbon Pollution Reduction Scheme. The details of this scheme classes are 8–12 and 3–8 m3 ha–1 y–1 respectively, at age 20 y are yet to be finalised but the ‘White Paper’ has identified (ABARE and BRS 2001). reforestation as an eligible offset.
Market opportunities for more traditional wood products such Other markets considered to have longer time scales and medium as sawn timber, pulp, plywood and LVL (laminated veneer market opportunities are for bioenergy and biofuels (URS lumber), posts and poles in lower-rainfall areas are considered Forestry 2008). These would probably depend on large areas of to be limited at present, due to the small volumes available, dedicated biomass plantation, or thinnings from large areas of inconsistency and uncertainty in the wood resource and distance timber-producing plantations, and so may be of limited relevance from existing production facilities (Blackwell and Stewart in lower-rainfall zones. 2003). The SMARTimbers Cooperative in south-western and Rain-fed tree growth in the 500–800 mm y–1 rainfall zone may central Victoria has successfully developed the milling and be slow and highly variable from year to year. Hence careful marketing of E. cladocalyx (sugar gum) timber from an existing species selection is important, particularly if commercial wood plantation resource of 3000 ha, much of it in scattered broad production is an aim of planting. There is limited quantitative windbreaks. Important factors in this success were establishing knowledge of which tree species might be successfully grown market acceptance for a hitherto little-known timber, including for wood production and or environmental benefits on rain-fed the identification of ‘niche’ markets, and developing sawing and sites in this rainfall zone in NSW, especially on the North-West drying methods to maximise value. A net return of up to $400 m–3 Slopes and Plains. The Australian Low-rainfall Tree Improvement on the stump has been realised for larger sawlogs, depending Group (ALRTIG) considered several species planted fairly on product (Lang et al. 2007). This example demonstrates that widely in low-rainfall areas of the southern Australian mainland smaller-scale timber production in low-rainfall areas can be before 2000 as candidates for more extensive rain-fed plantings profitable, particularly if growers can cooperate to establish and genetic improvement programs aimed at commercial wood markets and promote value-adding. If portable sawmills can be production. These species had been planted mainly by private used, distance to markets is not a significant factor affecting the farmers in shelterbelts or small woodlots and some had a history cost structure of producing and supplying timber (Blackwell and of plantation use in drier Mediterranean countries. Species Stewart 2003). proposed for improvement programs by ALRTIG (Harwood Some ‘emerging’ markets have potential to make larger-scale et al. 2001) included E. cladocalyx (planted widely in western plantations in low-rainfall areas financially viable. Foremost Victoria and south-eastern South Australia, including some quite among these is carbon sequestration, in which the opportunity large older plantations, and having potentially valuable timber); for farm forestry to participate is rated highly (URS Forestry E. sideroxylon and E. tricarpa (used already in small-scale farm 2008). Plantations could be developed for sequestration alone forestry and considered candidates for plantations overseas); or as long-rotation plantations for timber. It seems likely that E. camaldulensis (a long history of planting in Australia and sequestration rates in drier-area plantations would be in the low overseas, often for ‘environmental’ purposes; also useful as a or medium ranges as defined by the Australian Greenhouse Office parent in commercial hybrid crossing); spotted gums, especially (2001), with 16–30 t of carbon (equating to about 60–110 t of C. maculata (widely used in farm forestry plantings and having good wood quality); P. pinaster (already widely planted in the CO2-equivalent) stored per hectare in trees in mixed plantings at age 20 y. NSW Department of Primary Industries (DPI) has 400–600 mm rainfall zone in Western Australia, with an ongoing produced a model, the Carbon Sequestration Predictor (CSP), to breeding program and having good sawn wood quality). assist landholders to determine the amount of carbon sequestered The inland Queensland species E. argophloia, a newcomer to under different land-use change scenarios. For example, the model plantation forestry in southern Australia in 2000, was not included estimates that a slow-growing 20–25-y-old commercial plantation, in the ALRTIG program but it was showing promise by then in established on annual pasture on Chromosol soil and with a mean Queensland trials. It was described favourably as a potential annual rainfall of 500 mm, will sequester about 53 t of carbon low-rainfall plantation species with promising timber properties (including soil carbon) per hectare (Fortunaso et al. 2008). by Allworth (2000). New South Wales has a state-based Greenhouse Gas Reduction In order to investigate the performance of some potentially (Abatement) Scheme (GGAS) whereby electricity providers commercial tree species in the 500–800 mm rainfall zone, State can offset their emissions through Abatement Certificates Forests of NSW (SFNSW) and CSIRO established a series of trial (NGACs) created from reforestation. These certificates may plantings in 2000 across 51 sites spanning about 750 km in latitude be created by plantation owners or managers, with a maximum and ranging from 515 to 780 mm mean annual rainfall (Barton number from a formula based on carbon stock change. The and Parekh 2004). (CSIRO established two of these sites, which number of certificates created is calculated from the estimate from late 2000 onwards were managed by SFNSW and then by of sequestration, but this estimate is discounted according to the NSW Department of Primary Industries.) By age 57 months its uncertainty: the number created is limited to that number for (2005), 45 of the original 51 planted sites were considered to be which there is 70% probability that actual sequestration exceeds this value. Thus there is a strong incentive to improve the
Australian Forestry 2009 Vol. 72 No. 4 pp. 172–183 174 Low-rainfall species trials in New South Wales Part 1 still viable and worth assessing at that age. Six sites have failed sections of land — rocky, less fertile or steep. The area of each for various reasons and have been abandoned. site was limited to less than 2 ha.
This paper (Part 1 of a series of two) includes analyses of growth Nine sites, three in each of the North, Central and South regions, and form for all species across the 45 sites viable in 2005, using were identified for additional measurements and known as data collected between ages 10 months and 57 months after ‘research’ sites, as opposed to ‘standard’ sites (the remaining 36 planting. Part 2 will present analyses of effects of environmental sites). A criterion for selection of the ‘research’ sites was that they factors (climatic and edaphic or geological) on survival, volume were ‘well prepared’ and it is possible that this factor gave some and tree biomass of species at 57 months, concentrating on species extra growth advantage to trees planted there. that were represented at most sites. Selection of tree species and trial design Methods The relatively small area of each trial site (< 2 ha) and a desire Selection of sites to apply later silvicultural treatments to fairly large plots with adequate replication limited the number of species that could Trial sites were chosen to represent three broad climatic regions be tested. in NSW within the 500–800 mm annual mean rainfall zone — North (tendency to higher average summer than winter rainfall), Eight tree species were used in the trial (Table 1). These were Central (about equal average summer and winter rainfall), and considered to have potential for good survival, growth and solid South (tendency to higher average winter than summer rainfall). wood or pulp production when grown in low-rainfall environments Of the 45 sites viable in 2005, 19 were in North region, 17 were without irrigation. The choice of species was based heavily on those in Central region and nine were in South region, in latitudes selected by ALRTIG for tree improvement work — Eucalyptus 29°36′S–35°39′S (Fig. 1). camaldulensis, E. cladocalyx, E. sideroxylon, P. pinaster and spotted gums (Corymbia maculata and C. citriodora subsp. The general intent was to plant the trials on slopes or ridges where variegata) (Harwood et al. 2001) — and some species such as the trees were likely to intercept aquifer re-charge, sometimes spotted gums and E. cladocalyx with encouraging results from above salt-affected areas. Local experts and organisations wood quality testing of low-rainfall plantation-grown material interested in natural vegetation re-establishment recommended (Washusen et al. 1998). E. argophloia had been fairly intensively suitable locations on private properties with cooperative owners. studied as a plantation tree in Queensland before 2000 and is Six sites in Central and North regions were on government-owned considered to be one of the species best suited to low or irregular land, for example in dam catchments. Many sites used on private rainfall, frosty conditions and moderately saline, alkaline or property were on poorer quality (from an agricultural viewpoint) medium to heavy clay soils. It has a high-quality, durable timber and is expected to give acceptable recovery of high-grade products if grown under optimum management (Allworth 2000; Lee et al. 2006). Eucalyptus argophloia was therefore another species suitable for testing. Acacia mearnsii was included because of its QLD fast growth and suitability for pulp and charcoal production.
Seedlings of the species in the trial were grown from wild
NORTH single-provenance seed in most cases (Table 1). Seedlots were obtained from the CSIRO Australian Tree Seed Centre (four NSW lots) and Forests NSW (two lots), and E. argophloia seed was 30°0'00''S Tamworth provided by the Queensland Department of Primary Industries. Two seedlots of E. camaldulensis (Lake Albacutya, Victoria
Dubbo CENTRAL and Cowra, NSW) were used. Seedlot details for the Cowra provenance of E. camaldulensis are not known. Genetically improved seedlings of P. pinaster, with CGF (growth and form) ratings of 22 and 32, were supplied by the Western Australian Sydney Department of Conservation and Land Management (CALM). Wagga SOUTH 34°0'00''S Some seed used for later refilling was also obtained from a New Wagga Zealand P. pinaster improvement program.
Hybrid clones, produced from controlled crosses among select E. camaldulensis and E. globulus subsp. globulus or E. grandis VIC TM N parent trees and marketed by Saltgrow , were used as a ninth ‘species’ (Table 1). The clones were developed for adaptability
0 50 100 200 km 38°0'00''S and fast growth on saline, heavy clay and low-rainfall sites (Dale 2002). 148°0'00''E 152°0'00''E At each site, six species out of the pool of nine were planted. Figure 1. Locations of clusters of species trial sites in North, Central The species used at each trial site are shown in Table 2. The and South regions of New South Wales Lake Albacutya provenance of E. camaldulensis was planted at
Australian Forestry 2009 Vol. 72 No. 4 pp. 172–183 I.G. Johnson, A. Kathuria, J. Parekh and C. Barton 175 a total 31 sites (including all North region and some Central and sites. Glyphosate, as ‘Roundup’ at rates of 1.2–4.0 L ha–1, was South region sites) while the Cowra provenance was planted at 14 used at nearly all sites, usually in combination with pre-emergent Central and South region sites, but no site had the two provenances herbicides such as Simazine, ‘Surpass’ (active ingredient 2,4-D) planted together. Reasons for these different allocations to sites and ‘Visor’ (active ingredient thiazopyr). A follow-up spray was are not known. applied at most Central region sites and at M01 and M02 (see Table 2 for the location of these and other particular sites). Five randomised complete-block designs were used containing different species mixes perceived as suitable for the different Initial planting occurred at most sites in August–September 2000, regions (characterised by predominantly summer, winter and with a few exceptions: MA03, in December 2000 (except for evenly spread rainfall) and to accommodate the clones. This E. camaldulensis, planted in May 2001); MA13, in June 2001. Site resulted in species being planted at differing numbers of sites, G03 was completely replanted in July 2001 after almost complete ranging from 45 sites for E. camaldulensis to 23 sites for both failure of the original September 2000 plantings. Seedlings at all E. argophloia and E. cladocalyx and only six for the clones sites were hand-planted, generally with spades. Soil moisture at (Table 2). planting varied over sites. In North region about half the sites had dry soil and at six sites the seedlings were watered in soon after The species plots at all sites normally contained 72 trees (six rows planting. In the other regions soil moisture was adequate to high of 12 trees) planted at 4.0 m spacing between rows and 2.5 m at planting. Fifty grams of DAP fertiliser (N: 18.0% P: 20.0% within rows. Plots of non-clone species included an inner 40-tree S: 1.6%) was buried near each seedling soon after planting at all measurement plot (four rows of 10 trees). In clone plots, all sites except M01 and M02 (South region, established by CSIRO), 72 trees were measured, including the outer rows (see below). A where no fertiliser was applied. fairly large plot size was employed since one aim of the plantings was to later apply silvicultural treatments such as thinning and Herbicides were applied within a few months after planting at pruning in suitable plots. A small number of measurement plots most sites, but were considered unnecessary at three South region contained up to 44 trees due to ‘extras’ being planted in one or sites. ‘Roundup’ alone or ‘Roundup’ and Simazine were used at more rows, but maintained the correct plot area. In a few plots most sites, supplemented by hand slashing at a few sites. there were some missing or extra buffer trees. Some replanting of failed seedlings with new stock (refilling) The plots of the six species were replicated four times at each was carried out at most sites following the initial planting. Early site, except at MA02. The trial there was split over two sub-sites failure was caused by a variety of agents, both climatic and biotic (A and B) about 4 km apart, of different geology, with only two (see below for more detail). Refilling as reported in establishment replicates at each. These are treated as two separate sites in this records, not including details of positions replanted in plots, was paper. carried out at about 75% of sites, mostly in the Northern and Central regions. In most cases the refilling was done within one Establishment to three months of the original planting across several species and to differing extents among sites, varying from a few dozen Most sites were deep-ripped along contours to depths of at least to several hundred seedlings. The most consistently affected 60 cm using bulldozers of Caterpillar D6–D7 size. Mounding (to species was P. pinaster, whose plots were refilled, once or twice, about 30 cm high) was mostly carried out using two- or four-disc on at least 12 sites. mound-ploughs and a profiler to give a shallow depression along the mid-line to help retain water. Except at some very rocky sites Assessment methods and some Central region sites where the soil was very wet at the time of site preparation, the mounds were re-cultivated to reduce Seven rounds of assessment were carried out on the trial within its cloddiness of the soil. first 5 y, each on differing numbers of sites. The most complete of these, in terms of site coverage, were at 10 months (45 then-viable All sites received one or two herbicide applications about one to sites were assessed, including 43 of the 45 discussed in this paper), three months following soil preparation. These were sprayed over and at about 57 months (all 45 then-viable sites). Intermediate the mounds in most cases but broadcast over whole plots at five
Table 1. Species and provenances included in the species demonstration trial. Seedlot details are not available for all species.
Species Provenance Seedlot Acacia mearnsii North of Bungendore, NSW 18975 (CSIRO) Corymbia maculata Curryall State Forest, NSW 19663 (CSIRO) Corymbia citriodora subsp. variegata Richmond Range, NSW 323 (Forests NSW) Eucalyptus argophloia Queensland Unknown (Qld DPI) Eucalyptus camaldulensis Lake Albacutya, Victoria and Cowra, NSW 424 (Forests NSW) and unknown Eucalyptus cladocalyx Marble Range, SA 19349 (CSIRO) Eucalyptus sideroxylon West Wyalong, NSW 19677 (CSIRO) Pinus pinaster Improved stock ex WA breeding program Clones of Eucalyptus camaldulensis × Mixed clones, supplied by Saltgrow™ E. globulus subsp. globulus, and × E. grandis
Australian Forestry 2009 Vol. 72 No. 4 pp. 172–183 176 Low-rainfall species trials in New South Wales Part 1
Table 2. Species planted at each site in each of the five trial designs (present = ×)
Species planted
Design Site (regiona) Acacia mearnsii Corymbia maculata variegata C. citriodora Eucalyptus argophloia camaldulensis E. E.cladocalyx E. sideroxylon Pinus pinaster Clones
‘Standard winter’ B01 (C) × × × × × × B03 (C) × × × × × × G03 (N) × × × × × × M01 (S) × × × × × × M02 (S) × × × × × × MA01 (C) × × × × × × MA03 (C) × × × × × × MA05 (C) × × × × × × MA09 (C) × × × × × × MA10 (C) × × × × × × MA13 (C) × × × × × × MU03b (S) × × × × × × MU04 (S) × × × × × × MU07 (S) × × × × × × N05 (N) × × × × × × N15 (N) × × × × × × N19 (N) × × × × × × N21 (N) × × × × × × ‘Standard summer’ B02 (C) × × × × × × G02b (N) × × × × × × G04b (N) × × × × × × MA02c (C) × × × × × × MA06 (C) × × × × × × MA14 (C) × × × × × × MA16d (C) × × × × × MU01 (S) × × × × × × N06b (N) × × × × × × N09 (N) × × × × × × ‘Standard summer’ N10 (N) × × × × × × N12 (N) × × × × × × N13 (N) × × × × × × N14 (N) × × × × × × N17b (N) × × × × × × N18 (N) × × × × × × N20 (N) × × × × × × ‘Winter clonal’ MU02 (S) × × × × × × MU05 (S) × × × × × × ‘Summer clonal’ MA07 (C) × × × × × × MA11 (C) × × × × × × N04 (N) × × × × × × N16 (N) × × × × × × ‘Rainfall distribution’ MA08 (C) × × × × × × MU06 (S) × × × × × × N11 (N) × × × × × × a C = Central; N = North; S = South b Shortrow layout sites c MA02 was planted over two subsites A and B (two replicates, with the same species, at each subsite) d MA16 — E. argophloia replaced in field with E. camaldulensis
Australian Forestry 2009 Vol. 72 No. 4 pp. 172–183 I.G. Johnson, A. Kathuria, J. Parekh and C. Barton 177 measurements covered subsets ranging from the nine ‘research’ to be very suppressed or grossly deformed (e.g. very bushy trees sites only (measured on all occasions including at 4 y) to about with no dominant stems). 60% of ‘standard’ sites (measured at each of 22 and 34 months, a different group on each occasion). Up to age 57 months each Statistical analyses surviving ‘standard’ site was assessed two or three times. The data from this experiment are multilevel; sites were selected Height was measured at each assessment up to 57 months. at random, and each site had a block design with the six ‘species’ All living trees were measured in the earlier assessments. The treatments replicated four times. Measurements were taken over protocols for the 57-month height assessment were complex: a period of 5 y. The data therefore have a hierarchical structure • If trees were small (nominally < 3 m average height, but broadly with different levels of variation. Analysis of variance — except interpreted) or if there were 10 or fewer in a plot, all trees were in balanced or nested designs — is difficult to apply to data with assessed. a multilevel structure. Mixed modelling is becoming a standard • In all clone plots all trees were assessed. approach for analysing such data, particularly because it can deal • In other plots not covered above, only tree numbers 1, 5, 9, with complicated unbalanced structures. 13, 17, 21, 25, 29, 33 and 37 were to be assessed (ten trees), Exploratory analysis of data for the two E. camaldulensis prove with the next tree number in the row substituting if one of the nances at different ages, using scatter plots for each region and specified trees was dead or missing. In plots with relatively tall provenance within region, did not show any obvious differences trees, some (up to four or five) larger-diameter trees were often in height or survival between the two provenances. As there were measured for height in addition to the prescribed 10-tree sample, no sites where the two provenances were planted together the to allow calculation of another plot statistic (top height — not provenance effect was confounded with the site effect. As data reported in this paper). The data from these ‘extra’ trees were from all sites were collected at age 57 months, analysis of variance included in the present analyses. was carried out to test for the provenance difference within region • The tallest stem on each assessed tree was measured from at this age. No significant difference was found between the two ground to tip. provenances for survival or height.
At the 57-month assessment, two measurements of stem form were All tree data were averaged at the plot level; plot height was the recorded as a broad guide to the timber production capabilities of average height of live trees and survival was the proportion of the different species. These are listed below: live trees in each plot. Preliminary analysis of the relationship of 1. A forking score (FORK), on an objective scale of 1 to 4, best height and survival with age showed that the relationship was non- to worst, designed to be applied across species and sites: linear. The linear mixed model with smoothing splines method of Score 1: tree without forks Verbyla et al. (1999) was used with age included as a covariate. REML (residual maximum likelihood) was used for estimation, Score 2: forking in the top half of the stem and analysis was conducted using ASREML (Gilmour et al. Score 3: forking in the second quarter of stem above 1999, 2002). Wald statistics were used for significance testing of ground the fixed effects. The Wald test, in a mixed-model framework, Score 4: forking in the first quarter of stem above investigates the same hypothesis as the F-test in the analysis of ground. variance. The fixed effects in the model were Species, Region, Scores excluded fresh, twiggy growth at the top of the tree. Type (i.e. ‘standard’ or ‘research’ — see earlier) and lin(Age). In trees with two or more definite stems (usually of similar Here lin(Age) is the linear effect of the factor age, taken to be a diameter) originating from bifurcation within 20 cm of continuous variable; this contributes a single regression parameter the ground, the two stems of largest diameter were scored (slope). All two-way and three-way interactions were included separately. At two Central region sites this type of forking in the initial model. All the interactions that were not significant habit was abundant due to regrowth of stems after a wildfire (P > 0.05) were removed from the final model. The random terms in December 2001. for which the variance components were close to zero were also removed from the final model. The final model, where the fixed 2. A straightness score (STR), on a semi-objective scale, 1 to and the random components of the model are written using the 4, best to worst, designed to be applied across species and symbolic form proposed by Wilkinson and Rogers (1973), was: sites: Score 1: stem straight or with very minor deviations Response variable: Average height or survival (at plot level). Score 2: stem retaining a generally vertical axis, but with fixed ~ constant + Species + Region + Type + lin(Age) + Species. minor to moderate shifts or kinks or a minor overall bow Region + Species.Type + Region.Type + Species.Region.Type. Score 3: moderate overall bow and or repeated moderate kinks or wobbles; no definite vertical axis random ~ Site + spl(Age) + Species.Site + Species.spl(Age) + Score 4: severe bowing, kinking or twisting. site.rep.plot. As for the FORK score, the largest two stems of trees with a bifur The term spl(Age) denotes the corresponding random spline cation within 20 cm of the ground were scored independently. component. These scores were assigned to all trees 1.3 m or more in height, except for a very small number of such trees that were considered
Australian Forestry 2009 Vol. 72 No. 4 pp. 172–183 178 Low-rainfall species trials in New South Wales Part 1
The percentage of scored trees in each plot with 10 20 30 40 50 60 no forks (score 1) at 57 months was extracted. Eucalyptus argophloia E. camaldulensis E. cladocalyx The percentage of trees per plot with ‘superior’ 6 stem straightness (defined as having STR score 5 1 or 2) was also extracted. In both traits, data for 4 the stems of greater diameter were used in the 3 cases of trees for which two stems originating 2 close to ground level had been scored. Exploratory 1 analyses carried out for the two E. camaldulensis 0 provenances showed no significant differences E. sideroxylon Corymbia maculata C. citriodora variegata between them in the forking and straightness 6 traits, as for survival and height (see above). 5 4 The model used for analysing the forking and 3 straightness data was a mixed-effects model with 2 species, region and species × region interactions 1
as fixed effects, and site as a random effect. The Average height (m) 0 numbers of trees differed in different plots, so Clones Acacia mearnsi Pinus pinaster this variable was included as a covariate in the 6 analysis. 5 4 3 Results 2 The mean values of height and survival over 1 0 the first 57 months are presented in Figures 2 and 3. Different groups of sites were assessed 10 20 30 40 50 60 10 20 30 40 50 60 on different occasions (e.g. all sites in 2001 Age (months) and 2005, and subsets in other years). The loess Figure 2. Loess plot of average height vs. age in months for all species. Points on graphs curves suggest a non-linearity in the relationship represent mean values for sites assessed between 2000 and 2005, with slight differences of height and survival with age for some of the in time of measurement between different groups of sites within some years. nine species, although this is confounded with the sites measured. Eucalyptus camaldulensis, 10 20 30 40 50 60 C. maculata and the clones showed rapid growth Eucalyptus argophloia E. camaldulensis E. cladocalyx to 57 months, while that of A. mearnsii appeared to be declining with age. Pinus pinaster grew 0.8 particularly slowly, at about only half of the rate 0.6 of other species. 0.4 Most species declined only slowly in overall 0.2 survival within the first 57 months (Fig. 3). Very high survival rates were maintained for E. sideroxylon Corymbia maculata C. citriodora variegata E. camaldulensis, the clones and E. sideroxylon, which remained above 80%. The survival rates 0.8 of P. pinaster and A. mearnsii were very poor. 0.6 The highly variable survival of A. mearnsii and 0.4 P. pinaster in 2001 (at about age 9 months) was 0.2 due in each case to a small group of sites with high survival of 83–90%, assessed in April 2001, and a Clones Acacia mearnsi Pinus pinaster group with relatively poor survival (57 and 18%) Average survival (proportion) assessed in June 2001. Most sites were assessed 0.8 in May 2001 and had moderate (P. pinaster) or 0.6 high (A. mearnsii) survival. Mean values for these different months were plotted as separate 0.4 points (as was the case for all the graphs). Note 0.2 that a lift in survival by P. pinaster, particularly, and some other species in 2004 (48 months) was 10 20 30 40 50 60 10 20 30 40 50 60 due to only ‘research’ sites being assessed in that Age (months) year, some of which had much higher survival rates than most sites. Figure 3. Loess plot of average survival vs. age in months for all species. Points on graphs represent mean values for sites assessed between 2000 and 2005, with slight The results of the mixed-model analysis are differences in time of measurement between different groups of sites within some presented in Table 3. For height, the species years.
Australian Forestry 2009 Vol. 72 No. 4 pp. 172–183 I.G. Johnson, A. Kathuria, J. Parekh and C. Barton 179
Table 3. Analysis of variance (height and survival) for the mixed significant (P = 0.003). This was a result of contrasting mean model height values for some species in different regions, for example considerably greater growth of E. camaldulensis and the clones Variable and source of variation F value P value in North region compared with the others, and superior growth of Height C. c. variegata in South compared with Central region (Table 5). 00Species 163.04 <0.001 The southern-origin species E. cladocalyx (from south-east South 00Region 1.70 0.20 Australia) and also the northern species E. argophloia (from inland 00Type 0.23 0.63 south-east Queensland) showed similar mean height growth in 00lin(age) 16016 <0.001 all regions; 3.0–3.5 m and 3.4–3.7 m respectively. Eucalyptus 00Species.Region 2.45 0.003 sideroxylon, C. maculata, clones and A. mearnsii were also stable in height across regions. Pinus pinaster, a species originating in 00Species.Type 0.51 0.83 a Mediterranean climate, performed best in the South region and 00Region.Type 1.36 0.27 its mean height there was significantly superior to that in North 00Species.Region.Type 1.49 0.13 region (Table 5). Survival 00Species 32.32 <0.001 Mean proportions of surviving trees varied from 0.86 ± 0.05 for E. camaldulensis to 0.17 ± 0.03 for A. mearnsii, a range of 00Region 11.96 <0.001 0.69 (Table 4). Four other species had mean values which were 00Type 8.95 0.005 not significantly different from E. camaldulensis, ranging from 00lin(age) 991.36 <0.001 0.72 to 0.84 (E. argophloia, E. sideroxylon, C. maculata and 00Species.Region 2.44 0.003 clones). Pinus pinaster had very poor survival (0.34 ± 0.05). The 00Species.Type 1.71 0.112 region effect for survival was highly significant, indicating that 00Region.Type 2.38 0.107 different regions had different overall survival rates. The survival 00Species.Region.Type 0.25 0.998 proportions for the regions varied from South (0.70 ± 0.05) to North (0.54 ± 0.04) with Central region having an average survival of 0.64 ± 0.04. The significant species× region effect (P = 0.003) effects (P < 0.001), age effect (P < 0.001) and species × region reflected considerably higher mean survival byC. c. variegata and interaction (P = 0.003) were highly significant. For the survival E. argophloia in South region compared with the other regions, data, species (P < 0.001), region (P < 0.001), trial type (P = 0.005), and poor survival of P. pinaster and the clones in the North age (P < 0.001), and species × region interaction (P = 0.003) were compared with other regions (Table 5). Survival of E. cladocalyx all highly significant. was essentially similar across the regions, though slightly higher in the South. E. camaldulensis was very stable in survival across Species mean height across sites, after 57 months of growth, regions, with high mean survival of > 83% in all. ranged widely, from 5.45 ± 0.29 m for E. camaldulensis and 5.42 ± 0.33 m for A. mearnsii, down to 1.58 ± 0.12 m for P. pinaster The results of the forking and straightness analyses are presented (Table 4). The first two species, together withC. maculata and the in Table 4. For forking, the species, region and species × region clones, formed a superior subset, and P. pinaster was significantly interaction effects were highly significant (P < 0.001 for all). For poorer than all other species. The region effect for height was straightness, species and region effects were highly significant not significant (P = 0.20), but the species × region effect was
Table 4. Predicted means and standard errors (SE) for height, survival and form traits (incidence of unforked trees and trees of superior straightness (STR)) of species and regions overall, at age 57 months. Means followed by the same letter are not significantly different at P = 0.05; paired comparisons were made using a ttest.
Height Survival Unforked Superior Factor SE SE SE SE (m) (proportion) trees (%) STR (%) Species Eucalyptus argophloia 3.55 bc 0.26 0.80 ab 0.08 74.3 b 2.7 45.4 d 3.9 E. camaldulensis 5.45 a 0.29 0.86 a 0.05 63.8 d 1.9 35.4 e 3.2 E. cladocalyx 3.28 c 0.23 0.66 b 0.07 52.7 e 2.2 18.0 f 3.4 E. sideroxylon 4.37 b 0.26 0.84 ab 0.06 70.0 bc 1.9 45.6 d 3.2 Corymbia maculata 4.83 ab 0.34 0.72 ab 0.07 66.1 cd 2.3 74.7 b 3.5 C. citriodora variegata 4.57 b 0.33 0.63 b 0.07 68.3 cd 2.7 77.2 b 3.8 Clones 4.82 ab 0.46 0.83 ab 0.11 66.7 cd 3.6 46.7 d 5.2 Acacia mearnsii 5.42 a 0.33 0.17 d 0.03 67.9 cd 2.1 59.4 c 3.5 Pinus pinaster 1.58 d 0.12 0.34 c 0.05 93.8 a 2.4 100.0 a 3.6 Region North 4.01 a 0.27 0.54 a 0.04 58.4 b 2.4 47.9 b 4.3 Central 3.70 a 0.25 0.64 ab 0.04 88.1 a 2.5 57.6 ab 4.5 South 4.26 a 0.32 0.70 b 0.05 61.3 b 3.5 62.3 a 6.2
Australian Forestry 2009 Vol. 72 No. 4 pp. 172–183 180 Low-rainfall species trials in New South Wales Part 1
Table 5. Predicted means and standard errors (SE) for height, survival and form traits (incidence of unforked trees and trees of superior straightness (STR)) of species within regions, at age 57 months. This table shows the regional differences within each species. Within a species, region means followed by the same letter are not significantly different at P = 0.05; paired comparisons were made using a ttest.
Region Height Survival Unforked Superior Species SE SE SE SE (m) (proportion) trees (%) STR (%) Eucalyptus argophloia North 3.50 a 0.31 0.69 b 0.08 064.99 b 3.00 043.87 a 4.79 Central 3.41 a 0.33 0.78 ab 0.10 091.38 a 3.70 041.06 a 5.53 South 3.74 a 0.61 0.96 a 0.20 066.62 b 6.47 051.36 a 8.98 E. camaldulensis North 6.31 a 0.55 0.88 a 0.09 051.52 b 2.78 028.85 b 4.61 Central 4.93 b 0.43 0.83 a 0.09 087.19 a 2.95 043.79 a 4.87 South 5.21 ab 0.51 0.87 a 0.10 052.56 b 4.04 033.66 b 6.69 E. cladocalyx North 3.04 a 0.40 0.58 a 0.12 037.48 b 3.95 018.92 a 5.71 Central 3.35 a 0.42 0.68 a 0.13 074.49 a 3.49 014.04 a 5.34 South 3.46 a 0.35 0.72 a 0.09 046.03 b 4.15 021.15 a 6.76 E. sideroxylon North 4.45 a 0.43 0.76 a 0.10 061.12 b 2.81 039.57 a 4.62 Central 4.40 a 0.43 0.87 a 0.11 091.62 a 3.02 048.94 a 4.93 South 4.26 a 0.45 0.89 a 0.11 057.26 b 4.14 048.21 a 6.78 Corymbia maculata North 4.91 a 0.64 0.69 a 0.13 058.73 b 4.06 052.49 b 5.79 Central 4.17 a 0.53 0.64 a 0.13 087.70 a 3.61 081.42 a 5.46 South 5.51 a 0.55 0.83 a 0.10 051.85 b 4.15 090.23 a 6.78 C. citriodora variegata North 4.71 a 0.42 0.58 b 0.07 063.19 b 3.09 063.54 b 4.88 Central 3.57 b 0.33 0.56 b 0.08 085.99 a 3.59 074.64 b 5.43 South 5.68 a 0.93 0.78 a 0.18 055.82 b 6.47 093.49 a 8.98 Clones North 5.63 a 0.88 0.69 b 0.17 053.16 b 6.04 038.55 b 7.91 Central 4.54 a 0.71 0.91 a 0.20 088.28 a 6.09 045.29 ab 8.25 South 4.39 a 0.70 0.91 a 0.20 058.52 b 6.43 056.33 a 9.29 Acacia mearnsii North 5.05 a 0.49 0.12 a 0.04 043.46 c 3.35 034.37 b 5.18 Central 5.30 a 0.52 0.23 a 0.05 086.05 a 3.25 070.09 a 5.17 South 5.96 a 0.64 0.16 a 0.05 074.23 b 4.44 073.83 a 7.07 Pinus pinaster North 1.36 b 0.17 0.19 b 0.07 092.15 a 4.20 100.00 a 5.99 Central 1.48 ab 0.18 0.41 a 0.10 100.00 a 3.45 098.73 a 5.34 South 1.96 a 0.26 0.46 a 0.11 089.07 a 4.67 092.84 a 7.22
(P < 0.001), while the species × region interaction was significant species tested, P. pinaster thus tended to have the straightest and (P = 0.021). Overall, P. pinaster had by far the highest percentage least-forked stems. Eucalyptus cladocalyx was poorest for both of unforked trees (93.8 ± 2.4), followed by E. argophloia and of these traits. Percentages of trees with superior straightness E. sideroxylon (74.3% and 70%). E. cladocalyx was significantly were similar between regions in E. argophloia, E. cladocalyx, the poorest, at 52.7 ± 2.2. Corymbia c. variegata, C. maculata, E. sideroxylon and P. pinaster. Corymbia maculata and the clones and A. mearnsii were not significantly different from A. mearnsii were much poorer in North region compared with one another. Forking across species was most prevalent in the other regions, E. camaldulensis was best in Central region North region, with only 58.4 ± 2.4% of trees without forks, and and C. c. variegata and the clones were significantly straighter lowest in Central region with 88.1 ± 2.5% of trees without forks. in South compared with North region (Table 5). Additionally, all individual species apart from P. pinaster had significantly more unforked trees in Central compared with the Discussion other regions. Acacia mearnsii had significantly fewer unforked trees (43.5%) in North region compared with both Central and Survival over all species was poorer in North region compared South (Table 5). with the South. This was also the case for individual species except for E. camaldulensis. Possible causes for this are discussed In the case of straightness, the Region effect was also significant, below. though less so than for forking, with South region having the highest percentage of straight trees (62.3 ± 6.2), followed by Eucalyptus camaldulensis showed generally good survival at Central with an average of 57.6 ± 4.5, and then North (47.9 ± 4.3). 57 months (mean = 86.6%). At only a few sites was the survival Percentages of trees with superior stem straightness varied of this species below 80%. It also had the greatest mean height greatly among species, from 100% for P. pinaster to 18% for overall, with the best growth in North region. However, it showed E. cladocalyx. The next-best species after P. pinaster were the comparatively poor stem straightness. Eucalyptus camaldulensis spotted gums (77.2% and 74.7%), which were significantly is regarded by Eldridge et al. (1993) as being especially tolerant superior to the next best species, A. mearnsii (Table 4). Of the of drought and high temperatures, water-logging and soil salinity,
Australian Forestry 2009 Vol. 72 No. 4 pp. 172–183 I.G. Johnson, A. Kathuria, J. Parekh and C. Barton 181 which give it superiority over many other species as a plantation Relatively poor survival of spotted gum (particularly by species for hot, dry sites. The Lake Albacutya provenance used C. c. variegata) in North and Central regions compared with on most sites in the present trial has been shown, in many trials South region was probably due to frost damage recorded at many both in Australia and overseas, to have a superior growth rate sites, particularly during a severe drought in 2001–2003. South (e.g. Eldridge et al. 1993). This provenance ranked top for region sites were little frost-affected, since they were mostly sited height, diameter, and volume mean annual increment but only in higher parts of the topography. There was little difference in 11th for stem straightness out of 20 provenances in a trial in the survival between the species C. maculata and C. c. variegata. The Wellington catchment (Western Australia; MAR 700 mm) at Curryall State Forest provenance of C. maculata used in this trial age 9 y (Mazanec 1999a). There are no known prior data for the was the most frost-tolerant within that species in laboratory tests at performance of the Cowra provenance, which in the present trial –6.1°C conducted by Larmour et al. (2000), ranking third overall generally performed similarly to Albacutya in the regions where out of 16 provenances of spotted gum. The Richmond Range both provenances were planted. provenance of C. c. variegata used was also relatively tolerant in these tests, ranking just below Curryall. Under the severe frost Eucalyptus sideroxylon and E. argophloia also had high mean conditions experienced in the field trial, both species had many survival rates of over 80%. Eucalyptus sideroxylon has been noted plots where tree crowns were killed back once or twice, despite for its tolerance of, and acceptable growth in, dry conditions in their frost tolerance in the laboratory. other parts of the world (Jovanovic and Booth 2002). Survival of E. argophloia was very high in South region and significantly The spotted gums showed good height growth overall among the poorer in North region, although it was planted at only two sites tested species, with C. maculata slightly (though not significantly) in South region and so was not comprehensively tested there. superior to C. c. variegata. Early frost damage to spotted gums In North region E. argophloia had mediocre to poor survival was noted at 14 sites, of which 10 were in North region and four in (under 60%) at seven out of 14 sites, including basalt and Central region. In the north, more sites with C. c. variegata (eight sedimentary geology. Overall, basalt sites had considerably lower out of 13) were frost-affected than were sites with C. maculata mean survival than sedimentary sites in this region. Eucalyptus (two out of six), and this helps to explain the overall slightly argophloia is reported as suited to soils with a medium to high poorer height growth of C. c. variegata. The Richmond Range clay content (Lee et al. 2006), so the predominant clay texture provenance of C. c. variegata has shown relatively good growth in of the basalt soils per se was not likely to be a factor limiting a provenance trial in the Wellington catchment (Western Australia; survival. The mean height growth of E. argophloia was similar 680 mm MAR) (Mazanec 1999b). It is interesting that Jovanovic in both North and South regions. From very limited evidence, and Booth (2002) did not consider the area climatically suitable this northern-origin species appeared adaptable to southern areas. for C. c. variegata to include the southern inland areas, nor the Jovanovic and Booth (2002) considered the climatically suitable Central-west, of NSW. Our trial provides some evidence of at zone for planting E. argophloia to include the eastern Riverina least satisfactory survival and growth at several sites in these in NSW (in South region). areas, and performance comparable with that of C. maculata, particularly in the south. Eucalyptus cladocalyx was poor in height growth overall (second-poorest after P. pinaster). It also displayed poor stem Many spotted gum plots in the South region had > 40% forked form, often with a low, bushy habit. The provenance used in our trees, proportionately more than in other regions. The reason for trial, Marble Range, was noted (along with another provenance this is not clear but was not considered to be due to frost damage at from Eyre Peninsula, SA) as having relatively poor growth and these sites. Spotted gums are characteristically straight-stemmed poor form in trials in south-western Victoria at ages 3 and 6.5 y even in the wild state, especially compared with species such as (Harris 2000, 2005). McDonald et al. (2003) also advised caution E. camaldulensis and E. sideroxylon, which are often crooked in using populations from the Eyre Peninsula as sources of (Boland et al. 1984; Marcar and Crawford 2004). The relatively material for timber plantations, and considered the possibility that poor straightness of these latter species was evident in this trial the dwarf sugar gum variant in cultivation (var. nana) originated and not surprising in trees from wild, unimproved seedlots. from there. Our trial was thus not a reliable test of the potential South region plots of both C. maculata and C. c. variegata had of E. cladocalyx in plantations. Provenances of E. cladocalyx consistently high proportions of superior straightness trees, with superior growth and form have been identified from South contrasting with plots of these species in the other regions, Australian trials (Bush 1999). There was limited evidence from the particularly North. Again, this was probably due to the lack of present trial that this southern-origin species was better adapted to frost damage to South region plots, which were mostly in higher South region sites than to North region — its mean survival rate parts of the topography. Young re-growing stems in frost-affected was about 14% greater and its mean height about 0.4 m greater plots (most numerous in North region) tended to be fairly crooked, in the south, though these figures were not significantly different as were such stems at two Central region sites previously burnt from those in the north. It had poor survival (under 60%) at three in late 2001. out of five North region sites, but was poor (54%) at only one out of eight South region sites. ‘Frost damage’ was noted on this Pinus pinaster had poor survival overall. It suffered high mortality species at two sites (one in North and one in Central region) in soon after planting at several sites (mostly in North and Central 2002, and may have contributed to poor survival at 57 months regions) due to a range of causes, including drought, weed (41%) at the former site. The western slopes area of NSW south competition and animal browsing. Browsing of this species by of about 31° latitude is considered by Jovanovic and Booth kangaroos, sheep and hares was noted for a few sites, but there were (2002) to be within the zone climatically suitable for planting insufficient data to indicate whether it was preferentially browsed E. cladocalyx. more than other species in the trial. Refilling ofP. pinaster plots
Australian Forestry 2009 Vol. 72 No. 4 pp. 172–183 182 Low-rainfall species trials in New South Wales Part 1 was recorded as being carried out, twice in some cases and up to 12 trial) was superior for volume at one site and for stem form at months following original planting, on at least 12 sites across these both sites. The data of Searle et al. (1998) indicate the potential regions, and was probably necessary in several more. In contrast, of A. mearnsii for short-rotation wood production in low- to seven of the nine South region sites were established without medium-rainfall areas better than do the present trial results and the need to refill P. pinaster or other species. Despite refilling identify N Bungendore as a superior provenance. efforts, by 2005 P. pinaster was entirely dead at 11 sites out of 39 where planted, most (nine) of these being in North region. Its The reason for the significantly higher average proportions of survival in this region was correspondingly significantly poorer unforked trees (overall and in most species) in the Central region than that for the other regions. This may have been due partly to is unclear. Trees in each region were assessed by different teams, the presence of clay soil texture in the top- and or subsoil at 10 and operator bias, while possible, is considered unlikely since of the 13 North region sites that displayed <10% survival. Five it implies an inability to see forking in trees of several species. out of six sites in Central region with similar clay-textured soils Later observation of large numbers of Central region plots during also had poor P. pinaster survival of 0–33%, while survival rates marking for thinning confirmed that many of them did in fact on lighter-texture soils were mostly over 65%. Pinus pinaster is contain high proportions of unforked trees. On two sites (eight regarded as relatively drought- and frost-tolerant, but is not suited plots per species) severely damaged by wildfire at about age to clay soils (Marcar and Crawford 2004; Harper et al. 2008). It 16 months, new stems had grown from near ground level. Under was also very slow in growth in our trial, with the poorest mean the assessment protocols, two dominant stems per tree were often height of all the species. However, it was a very straight-stemmed scored on these trees; these stems were mostly unforked. species and had by far the lowest occurrence of forking, which reflects the genetic improvement for form applied to P. pinaster Conclusions and embodied in the select seed used. 1. There were large variations in survival and mean height of The relatively much better survival and growth of P. pinaster at the species tested, with E. camaldulensis generally perform South region sites compared with those in the north may indicate ing well and P. pinaster poorly, to 57 months of age. Most poorer adaptation of this Mediterranean-origin species to the more other species had moderate to high survival overall, except sub-tropical climate of the north. Jovanovic and Booth (2002) for A. mearnsii, which had mostly died by 57 months. considered that preferred low-rainfall areas for the species may 2. The southern origin species E. cladocalyx showed better be in the southern part of NSW with > 500 mm annual rainfall. growth and survival at southern sites compared with northern However, their climatically suitable zone included some western sites, while the northern origin E. argophloia grew equally slopes areas north to about 30° latitude in NSW. well at both northern and southern sites. 3. Differences between regions were significant for survival, Acacia mearnsii had the overall lowest survival rate. Marcar with South region superior to the North and Central. Higher and Crawford (2004) state that this species is sensitive to severe survival and or growth by several species in the south was drought and frosts of 4° or lower. Ongoing drought conditions in associated with a lack of severe frosts there, in contrast to the first few years of the trial led to a sharp decline inA. mearnsii records of repeated frosting at some North and Central region survival by May 2003. The year 2002 was especially dry in all sites in the early years of the trial. regions, with autumn, winter and spring rainfalls well below median levels at most sites. Frost probably also played a role in 4. Tree form was highly variable among species. P. pinaster mortality on some sites. The species is also prone to borer and tended to be straight and unforked and the spotted gums were mealybug attack, especially on sites where it is stressed (Marcar also notably straight-stemmed. E. cladocalyx was most af and Crawford 2004). Insect damage was not assessed in detail fected by forking and had the poorest stem straightness (which in the present trial, but the presence of suspected insect damage may have been largely due to the Marble Range provenance was noted. In 2001, only about 9% of assessed A. mearnsii trees used, now known to generally have poor form). showed signs of insects (26% had fungi), and in 2002–2004 poor 5. At the relatively young age of 5 y, the spotted gums appeared health of the trees was ascribed overwhelmingly to water stress; promising as potential log-producing species in medium– over half the trees were dead in 2003. By 2005, A. mearnsii was low-rainfall areas, but would require reliably frost-free sites. completely dead at five sites out of 40 where it was planted; Eucalyptus sideroxylon also showed promise for commercial three of these were in North region, which had the lowest mean use, with a high survival rate and fairly good height growth value for survival in this species. Acacia mearnsii grew rapidly and freedom from forking; the crookedness of many trees was in the early years, hence its height mean was relatively large (not its main drawback. Considering growth and survival alone, significantly below that ofE. camaldulensis) at 57 months, even E. camaldulensis was superior. when its survival was poor. A few large trees alone remained in some plots. Acknowledgements Results of a species–provenance trial of wattles at two sites in A large number of people did much good work in establishing the Australian Capital Territory (ACT) showed that A. mearnsii and measuring the trials reported here. In the establishment phase, had relatively high mean volume and relatively few stems per we particularly acknowledge the efforts of Brian Royal, Steve tree at 32 months of age, compared with most of the other 24 Dobson and Steve Lacey of SFNSW. Two of the southern sites Acacia species tested (Searle et al. 1998). Survival of A. mearnsii were established by Murray Brown from Murray–Riverina Farm and most other species was high, since they apparently did not Forestry. Trial assessments were ably carried out by David Bell, suffer prolonged severe drought. Among seven provenances of Cath Carney, Jason Crooks, Gabby Cusack, Glenn Dale, Helen A. mearnsii, the N Bungendore provenance (as used in the present
Australian Forestry 2009 Vol. 72 No. 4 pp. 172–183 I.G. Johnson, A. Kathuria, J. Parekh and C. Barton 183
Engel, Brian Fisher, David Giles, Piers Harper, Ian Hides, Brad Jovanovic, T. and Booth, T.H. (2002) Improved Species Climatic Profiles. Jarrett, Darrel Johnstone, Tracey Kerin, Warrick Moore, Jim RIRDC project No. CSF-56A. Publication No. 02/095, RIRDC, O’Hara, Col Wilkinson and Sam Wood. We also greatly appreciate Canberra, 68 pp. the cooperation of the many owners and managers of the trial Kuginis, L. and Daley, J. (2001) Plant based solutions for dryland sites at all phases of establishment and measurement. Thanks salinity management. In: Bathgate, A. and Malden, J. (eds) Salinity Economics: A National Workshop. NSW Agriculture, Orange, NSW, are also due to Drs Annette Cowie and Rod Kavanagh for useful Australia, 22–23 August 2001, pp. 24–33. comments during the drafting of this paper. Lang, A., Fisken, D. and Wettenhall, G. (2007) Realising the Potential of Farm Forestry. Making Money out of Farm Trees. SMARTimbers References Cooperative, Ballarat, Victoria, 37 pp. Larmour, J., Whitfeld, S.J., Harwood, C.E. and Owen, J.V. (2000) ABARE and BRS (2001) An Assessment of the Potential for Plantation Variation in frost tolerance and seedling morphology of the Development in New South Wales. ABARE and BRS report to the spotted gums Corymbia maculata, C. variegata, C. henryi and C. New South Wales Plantations Taskforce Steering Committee on citriodora. Australian Journal of Botany 48, 445–453. Plantation Capability and Suitability, 109 pp. Lee, D., Lawson, S., Dickinson, G. and House, S. (2006) Species Allworth, D. (2000) Western Queensland White Gum. Greening Australia Profile — Western white gum (Eucalyptus argophloia). Queensland Fact Sheet. 4 pp. Department of Primary Industries and Fisheries Hardwoods Advice. ANRA (2008) Australian Natural Resources Atlas web-site, at: http:// Website www2.dpi.qld.gov.au/hardwoodsqld/11172.html. www.anra.gov.au/topics/salinity/overview/nsw.html. Marcar, N.E. and Crawford, D.F. (2004) Trees for Saline Landscapes. Australian Greenhouse Office (2001)Growing Trees as Greenhouse Sinks Publication No. 03/108, RIRDC, CSIRO, Canberra, 235 pp. – an Overview for Landholders. The Office, Canberra, 13 pp. Mazanec, R. (1999a) Nine year results from a Eucalyptus camaldulensis Barton, C. and Parekh, J. (2004) Establishment and Early Growth Data Denh. provenance trial in the Wellington catchment of Western of Species Demonstration Trials in the 500–700 mm Rainfall Zone Australia. Australian Forestry 62, 166–172. of NSW. Report to the Australian Greenhouse Office, Canberra, Mazanec, R.A. (1999b) Thirteen year results from a spotted gum 28 pp. plus appendices. provenance trial in the Wellington catchment of Western Australia. Blackwell, P. and Stewart, M. (2003) Using Portable Sawmills to Australian Forestry 62, 315–319. Produce High Value Timber from Farm Trees in the Semi-arid Zone. McDonald, M.W., Rawlings, M., Butcher, P.A. and Bell, J.C. (2003) Publication No. 03/046, RIRDC, Canberra, 128 pp. Regional divergence and inbreeding in Eucalyptus cladocalyx Boland, D.J., Brooker, M.I.H., Chippendale, G.M., Hall, N., Hyland, (Myrtaceae). Australian Journal of Botany 51, 393–403. B.P.M., Johnston, R.D., Kleinig, D.A. and Turner, J.D. (1984) Raper, G.P. (1998) Agroforestry Water Use in Mediterranean Regions Forest Trees of Australia. Nelson/CSIRO, 687 pp. of Australia. Water and Salinity Issues in Agroforestry No. 2. Bush, D. (ed.) (1999) National Low-rainfall Tree Improvement Workshop. Publication No. 98/62, RIRDC, Canberra, 54 pp. plus appendix. Proceedings, Adelaide, November 1998. RIRDC Project No. SAR- Reid, M. (1995) Burke’s Flat — a salinity treatment success story. 17A. Publication No. 99/66, RIRDC, Canberra, 89 pp. In: Murray–Darling 1995 Workshop Extended Abstracts, 11–13 Dale, G. (2002) Salt tolerant eucalypts for commercial forestry: progress September,Wagga Wagga. Record 1995/61, Australian Geological and promise. In: Proceedings of Australian Forest Growers 2002 Survey Organisation, Canberra, pp. 205–210. National Conference, Albany, Western Australia, October 2002: Searle, S.D., Jamieson, D.T. and Cooper, N.K. (1998) Growth and Session Paper 351. form of 25 temperate Acacia species in two trials near Canberra, Eldridge, K., Davidson, J., Harwood, C. and van Wyk, G. (1993) Eucalypt Australia. In: Turnbull, J.W., Cropmton, H.R. and Pinyopusarerk, K. Domestication and Breeding. Clarendon Press, Oxford, 288 pp. (eds) Recent Developments in Acacia Planting. Proceedings Fortunaso, K.L., Barton, C.V.M., Cowie, A.L., George, B.H., Montagu, of an international workshop held in Hanoi, Vietnam, 27–30 K.D., Rawson, A. and Wilson, B.R. (2008) Carbon Sequestration October 1997. ACIAR Proceedings No. 82. Australian Centre for Predictor for Land Use Change in Inland Areas of New South International Agricultural Research, Canberra, pp. 66–79. Wales. Background, user notes, assumptions and preliminary model Stolte, W.J., McFarlane, D.J. and George, R.J. (1997) Flow systems, tree testing. Version 3.1. NSW Department of Primary Industries, 24 pp. plantations, and salinisation in a Western Australian catchment. plus appendices. Australian Journal of Soil Research 35, 1213–1229. Gilmour, A.R., Cullis, B.R., Welham, S.J. and Thompson, R. (1999) URS Forestry (2008) Australia’s Farm Forestry Industry — Evaluation ASREML Reference Manual. Biometric Bulletin No. 3. NSW of Resource, Market and Development Prospects. RIRDC Project Agriculture, Orange Agricultural Institute, Forest Road, Orange, URS-2A. Publication No. 08/106, RIRDC, Canberra, 32 pp. NSW, Australia. Verbyla, A.P., Cullis, B.R., Kenward, M.G. and Welham, S.J. (1999) Gilmour, A.R., Gogel, B.J., Cullis, B.R., Welham, S.J. and Thompson, The analysis of designed experiments and longitudinal data by R. (2002) ASReml User Guide Release 1.0. VSN International Ltd, using smoothing splines (with discussion). Applied Statistics 48, Hemel Hempstead, HP1 1ES, UK. 269–311. Harper, R.J., Booth, T.H., Ryan, P.J., Gikles, R.J., M’Kenzie, N.J. and Washusen, R., Waugh, G. and Hudson, I. (1998) Wood Products from Lewis, M.F. (2008) Site Selection for Farm Forestry in Australia. Low-rainfall Farm Forestry. Final Report, FWPRDC Project Publication No. 08/152, RIRDC, Canberra, 101 pp. PN007.96. FWPRDC, Melbourne, 78 pp. Harris, S. (2000) Sugar gum provenance trial results. Agroforestry News Wilkinson, G.N. and Rogers, C.E. (1973) Symbolic descriptions of 9(2), 18–19. factorial models for analysis of variance. Applied Statistics 22, Harris, S. (2005) Two decades of trials trying to determine which species 392–399. will grow best. Australian Forest Grower 28(2), 25–26. Harwood, C., Bulman, P., Bush, D., Mazanec, R. and Stackpole, D. (2001) Australian Low-rainfall Tree Improvement Group: Compendium of Hardwood Breeding Strategies. RIRDC Project CSF-58A. Publication No. 01/100, RIRDC, Canberra. 138 pp. plus appendix.
Australian Forestry 2009 Vol. 72 No. 4 pp. 172–183 184 The Australian Farm Forestry Financial Model
The Australian Farm Forestry Financial Model
John Herbohn1,2,4, Nick Emtage1,2, Steve Harrison1,2 and David Thompson3
1School of Integrative Systems, The University of Queensland, St Lucia, Queensland 4072, Australia 2Marine and Tropical Science Research Facility, Cairns, Queensland 4870, Australia 3Care Pty Ltd, Armidale, NSW 2350, Australia 4Email: [email protected]
Revised manuscript received 24 August 2009
Summary time and effort to compile. A further motivation for developing the AFFFM was to provide users with ready access to the available, The Australian Farm Forestry Financial Model (AFFFM) has been albeit incomplete, data in a central repository (i.e. through the developed as a tool for the financial appraisal of farm forestry incorporation in a financial model of default data sets). investments. It is a whole-farm financial model developed in modules using Visual Basic as the programming language, and The design and structure of the AFFFM was based on the provides estimates of key financial criteria including net present Australian Cabinet Timbers Financial Model (ACTFM) (Herbohn value, land expectation value and internal rate of return. It also et al. 1999) developed through the Rainforest Cooperative provides details of the effects of forestry options on cash flows and Research Centre. The ACTFM is a spreadsheet-based financial the business cash position. Included in the AFFFM are data sets model designed to predict financial returns from plantations of for plantation establishment and maintenance costs, growth rates mixed rainforest species in north Queensland. The AFFFM also and potential timber products of a number of tree species, for three incorporates the simplified form of the whole-farm financial regions. This paper outlines the scope of the AFFFM including its model used in the AGROFARM model developed by the Centre structure, assumptions, input parameters and key variables. The for Agricultural and Regional Economics (Care) Pty Ltd. paper also reports the process by which the model was validated through replication of previous studies, the development of case A number of choices were faced when developing the AFFFM. studies and testing by undergraduate students. The first decision to be made was the platform or programming language to be used in the construction of the model. The Keywords: models; farm forestry; financial planning; investment; predecessors to the AFFFM (the ACTFM and AGROFARM) profitability; growth were both developed in Excel, and people wishing to use these models needed access to this spreadsheet package. The ACTFM Background to the development of the AFFFM made extensive use of Visual Basic macros, button bars and forms to aid users in moving around within the spreadsheet and Small-scale forestry or farm forestry offers many benefits to to automate a number of functions. The ACTFM was originally farmers, including revenue generation, income diversification, developed with Excel for Windows 3.11. Major problems were collateral for borrowing, a superannuation facility and a legacy experienced when the ACTFM was subsequently converted into for children. An assessment of the potential financial performance newer versions of Excel. Some functionality was lost and the of farm forestry is important for decision-support (Herbohn and package became unstable when it was transferred to computers Harrison 2000; Harrison et al. 2001). The mechanics of calculating using newer operating systems. As increased functionality was financial performance are relatively straightforward and the task added to the ACTFM, there was a marked increase in the time is easy using spreadsheet packages such as Excel (Herbohn and required by the model to recalculate financial variables. In effect, Harrison 2002; Dayanandra et al. 2002). Unless the underlying the ACTFM exceeded the capacity of Excel to undertake the financial analysis theory is understood and applied correctly, required calculations. The decision was made to program the new however, the analysis will provide misleading results. The model, the AFFFM, in Visual Basic. This overcame the problems principles of investment analysis are often incorrectly applied to of portability between versions of Excel and of inadequate farm forestry. Common mistakes include the inappropriate use capacity of Excel. of real or nominal discount rates and the inappropriate inclusion of sunk costs and interest. A prime motivation for developing the The following sections of this paper outline the scope of the AFFFM was to provide a sound framework within which farm AFFFM including its structure, assumptions, input parameters forestry investments could be analysed. and key variables, along with the process used for testing the model. Any financial analysis of farm forestry requires data on a range of financial and biophysical parameters. These data may have to be collected from many disparate sources and take considerable
Australian Forestry 2009 Vol. 72 No. 4 pp. 184–194 John Herbohn, Nick Emtage, Steve Harrison and David Thompson 185
Scope of the AFFFM been added in the sixth release using financial support from the AgForest program1. The AFFFM is a whole-farm model that captures the essential physical and financial elements of the whole-farm system. It has been constructed using a modular approach, with separate Input parameters, assumptions and key variables modules being developed for each of the main components of a Because the AFFFM was designed to assess the private benefits farm forestry operation (i.e. plantations, native forests, agriculture of farm forestry to landholders, a financial (as distinct from and farm finances). Modules have also been developed to present economic) model was developed. Key outputs include net present the outputs of the financial analysis. value (NPV), internal rate of return (IRR), business cash position Being a whole-farm model, the AFFFM can be used to investigate (equals bank balance), annual cash flows, an equivalent annual the financial effects of adding a forestry enterprise to an return value and land expectation value (LEV). The model will existing farm business. This is regarded as being a considerable display indicators of financial performance for the ‘with’ and improvement on models that consider only the forestry investment ‘without’ forestry situations so that users can ascertain if adding in isolation. The AFFFM can also be used to predict financial a forestry enterprise improves financial performance over that of performance of plantations, native forests and agricultural the current farm business structure. activities in isolation. The AFFFM was initially developed with a Discussions with potential users and previous experience with focus on three geographic regions, namely far north Queensland, farmers interested in forestry has revealed that many farmers the Darling Downs and the New England Tablelands. are more comfortable with information showing simple costs Early prototypes of the model included considerably more detail and returns and the timeframe in which they occur (as opposed on the overall farm structure than the final version. Potential users to the standard forestry financial measures NPV, IRR, LEV and who examined early versions commented that there was too much annuity equivalents). For this reason, the model generates a series detail for their purposes and too much data entry was required, of cashflow estimates in a text file which can also be displayed so the data required to operate the grazing enterprise module was graphically. The figures in the text file include an overall business simplified. In particular, an ‘Advanced Livestock’ data screen cash position (essentially, this tracks all costs and revenues and was removed and the livestock herd and flock dynamics were therefore the business’s bank balance), as well as cashflows simplified. Similarly, the model output options and sensitivity (equivalent to enterprise gross margins) for the various enterprises analysis screens were simplified. (agriculture, plantations and native forestry). Providing these data in a text file allows users to export the data into a spreadsheet. The The AFFFM includes plantation forestry and native forestry provision of some simple graphical functions allows the user to options as well as the ability to model planting of trees in either visually investigate the effect of the ‘with’ and ‘without’ forestry a woodlot or a shelter-belt configuration. The inclusion of native scenarios on the farm business cash position (i.e. a cumulative forestry is important because native forestry is a key land-use measure of performance) and the annual cashflows (i.e. a non- option for two of the three geographic regions for which the model cumulative measure of performance in each year of the forestry was initially developed (i.e. the New England Tablelands and the planning horizon). Darling Downs). Native forest management for conservation alone Provision is made for the inclusion of other cost and revenue can be simulated within the current structure simply by computing categories including finance costs, capital expenditure, overhead the cost implications for the farm business. expenses and other farm and off-farm income. The capability The inclusion of estimates of timber yield for a broad range of also exists to include an allowance for ‘living expenses’, which native tree species planted in various soil and climate conditions essentially represents the landholder’s withdrawal of funds in alternative locations was viewed as an important step forward from the farm business for family and personal reasons. A basic in farm forestry financial analysis because most models developed assumption of the AFFFM is that all prices (e.g. establishment and to date — except the Agroforestry Estate Model (AEM) of continuing maintenance costs, timber and other revenues) change Middlemiss and Knowles (1996) — rely on users having mean over time due to inflation. This means that a ‘real’ discount rate annual increment (MAI) estimates and yield tables at their is used when discounting cash flows; use of a real discount rate disposal. The AEM incorporates yield table generation software means that present-day costs and returns do not need to be adjusted called STANDPAK. for the effects of inflation if they are projected to future dates.
The possibility of adding tree growth estimation code to The effects of taxation have been included in the model. Due to AFFFM was examined, but this step was not feasible due to a the complexity and changeable nature of the taxation environment lack of appropriate growth equations. Alternative approaches this module is relatively simplistic. to estimating growth were therefore used, including the use of The input parameters required for each of the farm forestry estimates generated from the Plantgro software package (Hackett activity modules in the AFFFM are listed in Table 1. Considerable 1991) and growth rates published in the literature. flexibility has been built into the forestry modules which allow Simple MAI estimates are used in AFFFM to generate native users to enter data in a number of alternative forms. forest yields. If a landholder has sections of native forest with differing growth potentials, this is catered for in AFFFM through 1 The AgForest program is a joint project between Timber Queensland and the capacity to specify up to five separate native forest areas for AgForce which is designed to assist landholders to improve the management the farm business. The grazing of cattle in native forest areas of native forest areas so these can provide sustained timber harvests as well as was not included in the first five releases of the model, but has sustained cattle grazing.
Australian Forestry 2009 Vol. 72 No. 4 pp. 184–194 186 The Australian Farm Forestry Financial Model
An underlying assumption of the AFFFM is that the costs and are based on information for the New England region and the returns to the various agricultural enterprises remain static. Community Rainforest Reafforestation Program (CRRP) in north Exceptions to this assumption can occur in the case of the Queensland, as set out in Tables 2 and 3. The New England data interactions between grazing and plantation establishment. The are based on information supplied by a local contractor (Farm model assumes by default that where both agricultural enterprises Forestry Plantations), while the CRRP data are an average of the and plantations are assessed in a scenario, the plantations are costs per hectare of setting up 2000 ha of plantations of tropical established on grazing land (with the user also able to choose cabinet timbers. Stumpage prices, where provided, are based on the option to establish plantations on otherwise ‘unused’ land). Russell et al. (1993). Where establishment of plantations on grazing land reduces the farm’s total stock-carrying capacity below the level required to Default growth data included in the AFFFM maintain the stock levels specified by the user at the start of a scenario, the model calculates the yearly stock-carrying capacity When developing the AFFFM, to make the model useful to a wide and automatically adjusts stock numbers (and thus the aggregate range of users a decision was made at an early stage to include livestock gross margins) to levels within the farm’s carrying in the model as much growth data as possible. A module was capacity. Stock levels may also be increased automatically if the developed that contains information about potential growth rates user specifies that shelter from plantations will lead to a greater and harvest ages for 46 timber species for the Atherton Tableland, carrying capacity for the farm despite loss of grazing land to the Darling Downs and the New England Tableland. Data on the plantations. Calculation of the interactions between forest growth potential growth rates, harvest ages and product mixtures for tree and returns to grazing under native forests and returns to grazing species suitable for plantations on the New England Tablelands under plantations are formulated to account for variation in pasture and the Darling Downs were generated using the Plantgro program productivity during the growing cycle of the plantation estate. (Hackett 1991). Users of the model in these areas can choose to Users can specify a ‘normal’ carrying capacity for the forest areas, load the appropriate values for the various species by specifying then specify the proportion of this normal capacity that is available the location and soil type of the planned plantation. Data on the for cattle grazing for the first 30 y of a scenario. potential growth rates, harvest ages and timber prices for 31 species of cabinet timbers in far north Queensland were also Default financial data used in the AFFFM included. These data were obtained from a survey of forestry experts undertaken by Herbohn et al. (1999). The process through The AFFFM contains default data for costs of plantation which growth data used in the AFFFM were collected is outlined establishment that can be loaded by users. These default data in detail in Davidson and Herbohn (2008).
Table 1. Parameters of the Australian farm forestry financial model
Farm activity Basic parameters Advanced parameters Plantations Species used (names) Harvest cycles (three harvest ages per species) Plantation size(s) (ha) Plantation product mix per harvest cycle (as % of standing volume) Plantation density (stems ha–1) and product prices Percent of hectare per species Series of plantations Mean annual increment(s) (m3 ha–1 y–1) Time between plantation establishments Harvest age(s) (y) Number of plantations to be established Stumpage price(s) ($) Plantation design (woodlots or shelterbelts) Establishment costs ($) Maintenance costs ($) Harvest costs ($) Annual costs ($) Native forestry Forest area(s) (ha) Noncommercial thinning costs and timing Mean annual increment (m3 ha–1 y–1) Timber products and recovery rates Maintenance costs ($) Stumpage price(s) ($) Agriculture (livestock) Carrying capacity (DSE ha–1) Improvement in DSE with shelter (%) Stock numbers (No. of head) DSE ratings (DSE head–1) Gross margins ($ DSE–1) Agriculture (crops) Crop type(s) (names) Crop area(s) (ha) Gross margin(s) ($ ha–1) Farm finances Overhead costs ($) Loan(s) ($) Capital costs ($) Interest rates — loans, savings, overdrafts Other farm income ($) Offfarm income ($)
Australian Forestry 2009 Vol. 72 No. 4 pp. 184–194 John Herbohn, Nick Emtage, Steve Harrison and David Thompson 187
Table 2. Default data for New England hardwood plantations Table 3. Default data for north Queensland hardwood plantations
Year of Amount Year of Amount Type of cost Type of cost operation ($ ha–1) cost ($ ha–1) Non-recurrent costs Non-recurrent costs Road and firebreaks 00 0100 Planning and design 00 0075 Clearing residual vegetation 00 0105 Incidental clearing 00 0160 Burning 00 0010 Site preparation and cultivation 00 0265 Fencing 00 0020 Cover crop establishment 00 0090 Precultivation weed control 00 0070 Preplanting weed control 00 0090 Soil preparation 00 0195 Cost of plants 00 0450 Weed control preplant 00 0070 Planting and refilling 00 0645 Planting stock 00 0350 Postplanting weed control 00 0540 Planting 00 0160 Fertiliser 00 0085 Initial fertiliser 00 0110 Fencing 00 0560 Refilling 00 0150 Postplanting weed control 01 1310 Weed control postplant 01 0070 Postplanting weed control 02 0810 Youngage fertiliser application 03 0110 Postplanting weed control 03 0215 Overspray 00 0190 First pruning (plus certification) 04 0880 Slashing 00 0020 Second pruning (plus certification) 08 0650 Prethinning 1 inventory 10 0020 Third pruning (plus certification) 12 0865 Prethinning 1 marking 10 0020 Thinning 08 0500 Prethinning 2 inventory 20 0020 Total nonrecurrent costs 8245 Prethinning 2 marking 20 0020 Recurrent costs Preclearfall inventory 30 0020 Harvest marking and inventory Varies with species 0055 Preclearfall marking 30 0020 Protection and management 0–30 0040 Total nonrecurrent costs (y 0–30) 1850 Recurrent costs Protection (fire, pests etc.) 0–30 0015 Administration 0–30 0005 Insurance 0–30 0010
Structure of the AFFFM Initial testing of AFFFM The AFFFM consists of modules for each of the main activities, Members of the model development team undertook initial testing with each of these modules having forms (screens) linked by of the AFFFM, as reported in Emtage (2008). The tests involved the use of button bars. Screenshots of the main screens within assessment of the calculations made by the model through the AFFFM are provided in Figure 1 and are labelled with an comparison of the results with those of other models where uppercase letter. After the AFFFM program is executed, users possible, evaluation of simple investment scenarios to verify data are required to click on the ‘Start’ button on the ‘Australian manipulation within the model, and tests of the effects of various Farm Forestry Financial Model’ in order to access the functions functions and combinations of functions on the outputs of the of the model (see button marked A in Fig. 1). Clicking on the model. The model was further tested by the developers replicating ‘Start’ button brings up the ‘Farm Structure’ screen (B). From the results of published studies of small-scale forestry enterprises this screen, the main modules of the program are accessed. These (e.g. Ward 1995) and by evaluating case studies of real-life farms modules are ‘Agriculture’ (C), ‘Native forests’ (D), ‘Plantations’ using the model (e.g. Maczkowiack et al. 2008). (E), ‘Farm finances’ (F) and ‘Activity options’ (G). From these screens, the main parameters of the model are set, by entering Various people not connected with the project also tested the the figures directly, accessing other screens through button bars AFFFM user interface. The model was presented at a series or loading default data or saved scenarios from drop-down of seminars and workshops held at the Gatton campus of The menus. The additional screens that may be accessed from the University of Queensland and the Cairns campus of James Cook ‘Agriculture’, ‘Native Forests’, ‘Plantations’ and ‘Activity University in 2004 for the purposes of publicising the AFFFM Options’ are summarised in Table 4. The ‘Farm structure’ screen software package and demonstrating the types of information that also has a button bar that allows users to graph business cash can be generated using this package. position and cashflows (H). Basic information is also provided Researchers who requested and received early versions of the on this screen on the results of the financial analysis ‘with’ and model include: ‘without’ forestry, together with a summary of areas of the farm used for various activities. Further details are available in the • Queensland Forestry Research Institute and Department of AFFFM Users Manual (Emtage et al. 2002). Primary Industry personnel in far north Queensland
Australian Forestry 2009 Vol. 72 No. 4 pp. 184–194 188 The Australian Farm Forestry Financial Model F C E B Modules of the AFFFM accessible from the farm structure screen 1. Modules of the Figure G D A
Australian Forestry 2009 Vol. 72 No. 4 pp. 184–194 John Herbohn, Nick Emtage, Steve Harrison and David Thompson 189
Table 4. Additional features of the AFFFM that are accessible via button bars in the ‘Native forests’, ‘Plantations’, ‘Agriculture’ and ‘Farm structure’ screens
Module screen Additional screens accessible Description of function from module screen Native forests Native costs by area Screen to allow users to specify the costs associated with native forestry enterprises Native forest grazing Screen to allow users to specify the carrying capacity for cattle grazing in native forest areas Native forest grazing and tree Allows users to specify changes in the carrying capacity of native forest areas growth interactions over time in response to management activities Native forest products Screen to allow users to specify the timber product mixtures, their prices and the ‘recovery rate’ associated with native forestry enterprises Plantations Suitability ratings Screen allows users to specify the relevant climate and soil for the scenario being considered then run through the species list for the area to assess the suitability rating for each combination. Users can then choose to load the relevant data into the plantation module. Plantation costs Screen to allow users to specify the costs associated with plantation enterprises Plantation products Screen to allow users to specify the timber product mixtures and their prices for up to three harvests for plantation enterprises Sensitivity of NPV to timber growth Allows users to set a percentage rate to increase and decrease timber growth and price rates and prices to assess the effects on the NPV of a scenario Agriculture Understorey grazing Allows users to specify the percent of normal grazing carrying capacity available from grazing under plantations for the first 29 y of a plantation Farm structure Options Allows users to alter the various options for the different activities included in a scenario
• CSIRO personnel in Brisbane involved in plant physiology • landholders and timber industry workers who visited a stall at studies the agricultural exhibition (FarmFest) held in Toowoomba in • personnel from NSW Agriculture involved in development of 1999, which was staffed by members of the project research private forestry team. • officers of Private Forests North Queensland (PFNQ) and In general, the extent of feedback from researchers was Private Forests South Queensland (PFSQ), these being the disappointing. Despite requests for feedback about the AFFFM, regional plantation committees for far north Queensland and those who assessed early versions of the model provided little, southern Queensland respectively if any, useful feedback. In part this may have been due to the • researchers based in Western Australia investigating crop informal nature of the requests for feedback, the instability of diversification strategies. early versions of the model, and the lack of on-line help. A further explanation is that many of the people requested to test the model The model has been demonstrated at a number of conferences, did not have a reason to use the model as part of their work. seminars and meetings, including those involving: • researchers from the Rainforest CRC at the CRC’s annual The most critical assessment of the AFFFM resulted from it being conference in Cairns in August 2000 exposed to two student cohorts in the Agribusiness Program at The University of Queensland, Gatton Campus, in 2001 and • RIRDC staff and researchers interested in farm forestry financial 2002. Students enrolled in ABUS 3002 (Project Appraisal II) modelling, Canberra, July 2000 were required to complete an assessable project in which they • researchers in the Department of Forestry at the Visayas State had to develop a scenario using the model, involving agriculture, Agricultural College (now College of Forestry and Natural some forestry and farm financing details. The students spent Resources at Visayas State University), Leyte, Philippines, considerable time collecting and entering data to run the model, February 2001 something that most others who tested the model appear not to • landholders, researchers, politicians and timber industry have tried. The students then analysed the results of the scenarios personnel at a workshop Farm Forestry: What’s in it for me in and assessed the functionality of the model. They were asked to Mareeba in far north Queensland, December 2001 list problems with the model and to recommend ways to overcome • members of the RIRDC-sponsored Wholefarm and Regional the problems. Agroforestry Decision Making project research group and The students identified a number of problems with the AFFFM that invited guests at regular project meetings over the course of critically limited functionality in the version of model as presented the research project (between 1998 and 2002)
Australian Forestry 2009 Vol. 72 No. 4 pp. 184–194 190 The Australian Farm Forestry Financial Model basic o o being o, risk version o o o south orthern NSW, Researchers, farm advisers advisers farm Researchers, and some landholders Assess investments in farm forestry Farm forestry assessment tool AFFFM Yes, but simple treatment treatment simple but Yes, N N N N N N N developed developed northern far and eastern Qld Yes Yes Yes, Yes manual o o o o o ational Yes, forest managers and forest managers researchers Map, measure, manage Map, measure, and grow forest farm General purpose expert system for plantation forestry farm Farm Forestry v5 Toolbox Partially Partially Partially Partially Advisers, landholders, Advisers, landholders, Yes, but simple treatment N N N N N N Yes ? Yes o o o o o Researchers Researchers Assess bestguess species best Identify guess tree options for further RandD ? Western Australia N N N N N Yes IMAGINE o o o o o o o o orth Landholders, Landholders, farm advisors and landholders Assess in investments forestry farm Farm forestry assessment tool ACTFM Yes, in versions Yes, in modified for modified for research vsns Yes, mod in for research ified N N N N N N N N N Queensland Yes Model basic o o o o o o o o Western Western Researchers, Researchers, farm advisers Assess in investments forestry farm Farm forestry assessment tool Agroforestry calculator Australia Australia N N N N N N N N Yes Yes, Yes variables variables o o o o ew South Wales Researchers, Researchers, farm advisers Assess in investments forestry farm Farm forestry assessment tool AGROFARM Risky added can be Yes, if risky added variables added variables N N N N N Yes Yes Yes Yes (some vsns) Yes, if risky d detailed detailed o o o o o o o ew Zealan Farmers, farm advisers Yes, Assess in investments forestry farm Farm forestry assessment tool Agroforestry Estate Model (AEM) N N N N N N N N Yes Yes Yes o o o o o o o o o Farmers, farm advisers, researchers Victoria Assess in investments forestry farm Farm forestry assessment tool N N N N N N N N N Farmtree Yes Yes o o o o o o o Compare alterna Yes advisers, researchers Yes Australia Australia tive farm plans plans farm tive (might include forestry) farm ment tool Farmula N N N N N N N Yes Yes Farm manage Farm manage Western Western Farmers, farm for: Comparison of financial models available in Australia and New Zealand New Zealand and in Australia available models of financial Comparison production production degradation degradation Wholefarm Likely users Likelyusers Risk Risk Accounts typeSeason Taxation understorey Tree simula Iterative capacity tion Geographic focus Table 5. Attribute Key purpose use Intended shelter Stock invest Alternative ment options Menudriven structure Erosion/land Stochastic
Australian Forestry 2009 Vol. 72 No. 4 pp. 184–194 John Herbohn, Nick Emtage, Steve Harrison and David Thompson 191 o o values present et Default physical and for financial data forestry and agricultural (including enterprises native not but plantations forestry) can supplied but be modified estimate to capability Has rating suitability species and soil location based on yield type to and import with associated tables Plantgro the ratings using package software 300 y 300 N N N Can graph Can numerous graph key model variables return of rates Internal values expectation Land Business cash position model key output Can for file text to variables Excel in manipulation analysis Sensitivity o User generated but but generated User be used can defaults climatic, for set and and edaphic, financial management data. Userdefined up to 100 up Userdefined y N Each of the 15 of sub Each programs has output. contextual Manager Stand The outputs NPV, IRR, cash and product flows. can be data All tabular Excel to exported sensitivity Simple analysis Maps Yes o et present ? All data provided by users Basic — blocks blocks Basic — N N values values o et present ? Yes, in versions Yes, in Default physical physical Default and financial data for agricultural and forestry enterprises supplied but modified can be N research for Cashflow N values of Internal rate return expectation Land value Sensitivity analysis o o et present 50 y y 50 Default physical physical Default and financial data for agricultural and forestry enterprises supplied but modified can be Cashflow N N N values Annuity values ratio Benefitcost Sensitivity analysis o et present 60 y y 60 Can be modified be modified Can Cashflow Default physical physical Default and financial agricul data for forestry tural and enterprises sup be can but plied modified taxation Farm arrangements Probability if distributions using in stochas mode tic timber Existing mgt. and area options N N values values of Internal rate return Annuity values income Taxable Farm equity loss Profit and statement to optimise to optimise o o et present Unlimited Unlimited N N N Cashflow values values Area planted Enterprises displaced bytrees Woodflows Labour requirements Sensitivity analysis Physicaland for financial data and agricultural entered forestry |byAllows user. crop detailed understorey choices. option Labour requirements Cutting right options Joint venture options ls available in Australia and New Zealand Zealand New and Australia in available ls o o et present Unlimited Unlimited Cashflow values values of Internal rate return Sensitivity analysis Annuity values N N N Default data for Default data for and agricultural forestry enterprises can but supplied be modified o et present 20 y y 20 Electronic map Yes N N values values Equivalent annuityvalues Breakeven analysis of property Defined mgt units (e.g. based type) on soil Crop plans Stock plans Financial plans Cashflow (continued) Comparison of financial mode of financial Comparison (continued) Inputs Timeframe Farm plan plan Farm graphics Table 5. Optimising capacity Outputs
Australian Forestry 2009 Vol. 72 No. 4 pp. 184–194 192 The Australian Farm Forestry Financial Model to them. As a result of this rigorous testing, a number of changes and costs involved in establishing, maintaining and harvesting have been made that have substantially improved the model. plantations. The lack of data is perhaps most acute for native forestry operations (Taylor 1997; Taylor and Nester 1998; Ryan Target users of the AFFFM and Taylor 2006) and worse still for the interactions between these operations and grazing (Schulke undated; Stunzner and The target users for the model identified by the project team Stephens undated). Most available forestry data are fragmented, changed over the course of model development. While the project cover limited ecological and biophysical ranges, lie in often team initially thought the model would be used primarily by hard-to-access (at least for many landholders) publications such landholders, the complexity of the model and users’ reactions led as scholarly journals or departmental records and are in a format to a decision that the model should be designed for and targeted not suitable for use in the model. The compilation of suitability at farm financial advisors and those involved in extension in ratings for species for plantation development in the New England forestry and agriculture. It was reasoned that the model was best Tablelands and Darling Downs has created a valuable source of suited to those who could potentially use it repeatedly rather than data for people in these areas. for one-off assessment of a proposed forestry enterprise. This conclusion was subsequently supported by findings of research Comparison of the AFFFM with other Australian farm into the low adoption and use of software models to assist forestry models landholders’ decision-making by McCown (2002). McCown (2002: 180) concluded that there are four promising directions A range of farm forestry models are currently in use in Australia for the development of decision support systems in agriculture: that aim to provide estimates of the economic impact of trees on a ‘small’ tool for aiding farmers’ tactical decisions; a versatile farms (e.g. see Warner 2007). Several of the more widely-used simulator as a consultant’s tool; a versatile simulator as the core farm forestry models are compared in Table 5. These models differ of a facilitated ‘learning laboratory,’ (and) a formal framework that considerably in complexity and scope. For instance FARMTREE supports regulatory objectives in constraining and documenting (Loane 1994) contains considerable detail on tree growth rates, farming practice. The developers of the AFFFM aimed to create product recovery and stock shelter effects. The outputs from ‘a versatile simulator for consultants to use’ (McCown 2002). FARMTREE are calculated on a per-hectare basis. Others models including the Agroforestry Estate Model (Middlemiss and The information requirements for the AFFFM can be demanding, Knowles 1996) operate at the whole-farm level, and estimates and it is expected that users will require at least a few hours to of timber yields are provided by the user (or calculated via other become proficient in operating the model. Once the data for software) rather than being calculated directly by the model. typical agricultural and forestry enterprises in a region have been The AEM, which was developed originally for use in New collated and entered to develop ‘scenarios’ (i.e. sets of preloaded Zealand, has the capacity to model various combinations of tree data in the model), the effort required to modify parameter values understorey enterprises and to track farm labour requirements. to reflect the circumstances of landholdings in a given region will AEM is commercial software for use under licence and is regularly be far less than that needed if all the data for a scenario must be maintained. Farmula (Kubicki et al. 1991) provides a graphical collected. interface for farm planning purposes and was originally designed The student assessments identified a number of factors in relation as a tool for examining land management practices including to the data requirements of the model that have implications for erosion-amelioration effects of tree planting. The AGROFARM potential adoption. One is that many users are unlikely to be models are a suite of spreadsheet models developed for research expert in or even familiar with all the enterprises represented in and extension purposes by Care Pty Ltd and have a focus on the model. Users from different backgrounds with different skills northern NSW. These whole-farm models take account of taxation, (e.g. foresters, agronomists and landholders) are likely to be able and several versions have been developed to cater for plantations, to reliably estimate certain parameter values but be less able to native forestry and joint venture arrangements. Risk assessment reliably estimate others. The agribusiness students who assessed functions can be added to the model structure using the @RISK™ the AFFFM were well informed about the concepts involved in add-in for Excel. modelling agricultural investments. In terms of the aspects of the The Australian Cabinet Timbers Financial Model or ACTFM model relating to forestry investments the students were almost (Herbohn et al. 1998) is a spreadsheet model developed with totally reliant on the data embedded in or linked to the model. This funding from the Rainforest CRC. It estimates NPV and LEV was partly due to their lack of investigation of forestry growth, for forestry investments in northern Australia involving non- yield and price parameters, but also indicated that users with a traditional eucalypt and rainforest timber species. The model predominantly agricultural background will be less able to use the contains a large amount of default data including harvest age model without the inclusion of case-specific forestry data. and growth estimates for 31 species obtained from a survey of Resources on the web sites of state agricultural agencies experts (Herbohn et al. 1999), and estimates of costs and revenues provide sufficient information about returns to agricultural obtained from sources including the CRRP planting scheme that enterprises for use in the agricultural module of the model, and operated in north Queensland, and published cost data. The model many landholders and their financial agents possess detailed includes a series of button bars, pull-down menus and Visual Basic information about the costs and returns for their agricultural macros. Users have the option of selecting default data contained enterprises. The same cannot be said for forestry data. There is in the model or inputting their own data. a paucity of published data about many species of trees suitable for plantation development, in terms of growth rates, harvest ages
Australian Forestry 2009 Vol. 72 No. 4 pp. 184–194 John Herbohn, Nick Emtage, Steve Harrison and David Thompson 193
The AFFFM involves a combination of the features of the CARE functionality and ease of use. Throughout the development, input AgroFarm models and the ACTFM developed by Herbohn et al. has been sought from potential users to guide model development. (1998). The developers seek additional feedback and would welcome comments from users of the model (contact nick.emtage@ Discussion uq.edu.au). It was crucial to identify the target audience when developing Acknowledgments the AFFFM. The AFFFM is pitched primarily at farm forestry advisers, researchers and computer-literate farmers, with the The authors wish to thank the Rainforest CRC, the Rural Industries explicit purpose of improving their ability to estimate the returns Research and Development Corporation and AgForests for from farm-forestry investments. The AFFFM is differentiated from funding support for the development of the AFFFM. The authors other similar tools through the inclusion of growth projections would also like to thank the many people who have provided for a number of species in the New England Tablelands region, feedback on the AFFFM during its development. south-east Queensland and north Queensland. The inclusion of these data was considered critical to making the model useful to References the target audience, many of whom do not have ready access to reliable tree growth data. Davidson, J. and Herbohn, J.L. ( 2008) Growth components of the AFFFM. In: Harrison, S.R. and Herbohn, J.L. (eds) Agroforestry Other models have recognised the link between physical and Farm Forestry: Support Systems to Assess the Viability of variables in a farming system and forestry (e.g. the effect of Whole-Farm and Regional Agroforestry Enterprises. RIRDC trees on land degradation or stock shelter), but have ignored Publication No. 08/097, Canberra, pp. 137–147. important financial links such as the impacts on cashflow, debt Dayanandra, D., Irons, R., Harrison, S., Herbohn, J. and Rowland, P. and equity which result from forestry investment. Discussions (2002) Capital Budgeting: Financial Appraisal of Investment Projects. Cambridge University Press, UK. with farmers in the New England Tableland have revealed that Emtage, N.F. (2008). Development, testing and presentation of the they are comfortable with the physical effects of trees on their AFFFM. In: Harrison, S.R. and Herbohn, J.L. (eds) Agroforestry farm business and have embraced environmental plantings, but and Farm Forestry: Support Systems to Assess the Viability of uncertainty about financial effects is a disincentive to commercial Whole-farm and Regional Agroforestry Enterprises, 1st edn. tree investment. In particular, the large up-front establishment Publication No. 08/097, RIRDC, Canberra, xviii + 285 pp. costs associated with do-it-yourself forestry can have important Emtage, N., Harrison, S., Herbohn, J. and Thompson, D. (2002) The debt and cashflow implications. No financial model, including the Australian Farm Forestry Financial Model — Users Guide. The AFFFM, can eliminate uncertainty about financial impacts. What University of Queensland, Brisbane, Australia. financial models can do, however, is to allow the users (potential Hackett, C. (1991) Plantgro: A Software Package for Coarse Prediction investors) to better understand the potential quantum of cashflows of Plant Growth. CSIRO, Melbourne, Australia. Harrison, S.R., Herbohn, J.L. and Emtage, N.F. (2001) Estimating and variability of cashflows (risk) associated with investment in investment risk in small-scale plantations of rainforest cabinet forestry. As such the AFFFM can help farmers and their advisers timbers and eucalypts. In: Harrison, S.R. and Herbohn, J.L. to identify better and understand the key risk factors and their (eds) Sustainable Farm Forestry in the Tropics. Edward Elgar, likely effects in forestry investment. Cheltenham, pp. 47–60. Herbohn, J.L. and Harrison, S.R. (2000). Assessing the financial per If landholders are to claim deductions for farm forestry formance of small-scale forestry. In: Harrison, S.R., Herbohn, J.L. under Australian taxation legislation it is critical that they and Herbohn, K.F. (eds) Sustainable Small-scale Forestry. Edward can demonstrate that they are carrying on a business with the Elgar, Cheltenham, pp. 39–49. reasonable expectation of profit. As such, one further application Herbohn, J.L. and Harrison, S.R. (2002). Introduction to discounted cash of the AFFFM is as a tool to demonstrate the likely profitability flow analysis and financial functions in EXCEL. In: Harrison, S.R., of farm forestry for taxation purposes in order to demonstrate that Herbohn, J.L., Mangaoang, E. and Vanclay, J. Socio-Economic the forestry activity is carried on with ‘a view to a profit’. Research Methods in Forestry: A Training Manual. Rainforest CRC, Cairns, pp. 109–118. The AFFFM also has considerable scope as an educational and Herbohn, J.L., Harrison, S.R. and Emtage, N.F. (1998) Australian Cabinet extension tool in workshops held to encourage farm forestry. In Timber Financial Model. Computer software based on Microsoft Excel, including manual. Rainforest CRC, Cairns, 28 pp. particular, it can be used to illustrate the financial performance and Herbohn, J.L., Harrison, S.R. and Emtage, N.F. (1999) Potential cashflow patterns of alternative forestry options for landholders performance of rainforest and eucalypt cabinet timber species in with an interest in farm forestry. It could be used as a useful plantations in North Queensland. Australian Forestry 62, 79–87. adjunct to the technical information usually delivered in farm Kubicki, A., Denby, C., Haagensen, A. and Stevens, M. (1991) Farmula forestry courses. The developers’ experience has been that such User Manual. Western Australia Department of Agriculture, South courses generally have limited economic content and an ability to Perth, Australia. experiment with the financial outcomes of alternative plantation Loane, B. (1994) The FARMTREE model: computing financial returns and native forestry scenarios would add value to these courses. from agroforestry. Faces of Farm Forestry. Proceedings of Australian Forest Growers Conference, Launceston, Australia, 2–4 The development of the AFFFM has been a challenging task and May 1994, pp. 275–285. the product that has resulted represents considerable effort by McCown, R.L. (2002) Changing systems for supporting farmers’ many people. It has been a complex task to develop the conceptual decisions: problems, paradigms, and prospects. Agricultural framework for the model and then translate this into a useable Systems 74, 179–220. product. Much thought has gone into the tradeoffs between
Australian Forestry 2009 Vol. 72 No. 4 pp. 184–194 194 The Australian Farm Forestry Financial Model
Maczkowiack, R., Herbohn, J.L., Emtage, N.R., Slaughter, G. and Taylor, D.W. (1997) Enhancement of timber production through Harrison, S.R. (2008). Farm Forestry Experiences: Development silvicultural activities in native forest in south-east Queensland: of the Australian Farm Forestry Financial Model in South-East a preliminary analysis of two experiments. Internal report to the and North Queensland. Publication No. 08/096, RIRDC, Canberra, Queensland Forestry Research Institute, Gympie, 10 pp. ix + 160 pp. Taylor, D.W. and Nester, M. (1998) An overview of two forest harvesting Middlemiss, P. and Knowles, L. (1996) AEM Agroforestry Estate Model, experiments in spotted gum – ironbark forest in south east User Guide for v. 4.0. New Zealand Forest Research Institute, Queensland. Internal report to the Queensland Forestry Research Rotorua, New Zealand. Institute, Gympie, 16 pp. Russell, J.S., Cameron, D.M., Whan, I.F., Beech, D.F., Prestwidge, D.B. Ward, J. (1995) An economic analysis of small-scale hardwood and Rance, S.J. (1993) Rainforest trees as a new crop for Australia. plantations. Honours thesis, Faculty of Environmental Sciences, Forest Ecology and Management 60, 41–58. Griffith University, Gold Coast, Australia. Ryan, S. and Taylor, D. (2006) Sustainable Native Forest Management: Warner, A. (2007) Farm Forestry Toolbox Version 5.0: Helping Case Studies in Managing Private Native Forest in South- Australian Growers to Manage Their Trees. Publication No. 07/135, East Queensland. Department of Primary Industries, Brisbane, RIRDC, Canberra. 117 pp. Schulke, B. (undated) Native forest management: implications for grazing. Accessed from Queensland Forestry Research Institute, Gympie. 13 pp. Accessed on 26 February 2007 from http://www.privateforestrysthnqld.com.au/downloads/ GrazingAndForestManagement.pdf. Stunzner, A.E. and Stephens, M.L. (undated) Silvopastoral potential in central Queensland: a case of looking for the wood in the trees. 4 pp. Accessed on 13 March 2009 from http://www.southwestnrm. org.au/information/downloads/Stunzner_paper.pdf.
Australian Forestry 2009 Vol. 72 No. 4 pp. 184–194 Ian Ferguson 195
Fires, Forests and Futures: The ANU Westoby Lecture
Ian Ferguson
Forest and Ecosystem Science, University of Melbourne, Parkville, Victoria 3010, Australia Email: [email protected]
Revised manuscript received 26 August 2009
Summary Mother Nature, especially relating to fire, and this means that a joint approach to addressing the risks is desirable. This paper addresses management and policy issues arising from the 2009 and other recent fires that burnt the iconic ash-type The ash-type conservation reserves are as much about multiple forests of eastern Victoria. After sketching a general framework uses as are state forests. Some uses are common to both, some for an analysis of sustainability, the paper elaborates the results excluded from one and not the other, and some that vary by of a simulation study on seed collection and storage (to be degree rather than absolute exclusion in one or the other. We face published separately), especially in relation to the risks posed to a future in which there is a material risk that we will lose most or the ash-type state forests by major fires. These forest types are all of these ash-type forests, not over millennia, but in the next notably fire sensitive and easily killed by fires of moderate or 60 y. We can decide to do nothing and accept that loss, or we can even mild intensity. They are also cyclical in the extent of seed engage in active management to conserve their values and uses production and do not set seed in commercial quantities for the by reducing the risks across both tenures. first 20 years of life. If seed collection and storage is restricted Keywords: fire; risk; seed production; seed collection; forest management; to ‘low’ levels until 2050, the risks are substantial, involving planning; Monte Carlo method; Eucalyptus regnans; Eucalyptus probabilities over 90% for alpine ash, implying insufficient seed delegatensis to permit artificial regeneration of those burnt areas that have not regenerated naturally, whether climate change occurs or not. Those risks can be reduced substantially by higher levels of seed Introduction collection, but even at ‘median’ levels the probabilities may This lecture was dedicated to Jack Westoby, the former head of approach 40% under high rates of climate change. The simulated the FAO Forestry Division. As is well known, Jack underwent outcomes for mountain ash state forests are much less severe but an epiphany late in his career, leading him to question the the impacts of high rates of climate change are still great enough fundamental creed that he and his organisation had pursued in to cause discomfort. the years following World War II. He argued that forests were Data for the ash-type conservation reserves were not available for people, not for their intrinsic worth, and captured the nub but descriptive reports indicate that the risks for both forest of environmental and sustainability concerns in his argument. types in the conservation reserves may be very high, as many of Few who heard his ‘Kangaroo’ speech that used a quote from the old-growth stands in the national parks (often closed water the D.H. Lawrence novel as an opener in the 1974 FORWOOD catchments) were severely burnt in the 2003, 2006–2007 or 2009 conference will forget the power of his argument. fires. The risks posed by fire in the next 20 y to water quality and Alf Leslie was a close colleague of Jack Westoby. In my view, biodiversity are very likely to be material. he was as much responsible for this conversion as was Jack. I This leads to advocacy of active management measures that can recollect the discussions with Alf in Rome in which he was include: (1) a well-equipped workforce, (2) removal and salvage advancing similar views prior to Jack’s epiphany. So this lecture is of fire-killed trees along access roads, (3) seed collection and also dedicated to Alf, as he was a much-respected teacher, mentor, artificial regeneration in conservation reserves, (4) strategic swaps colleague, friend and sceptic over many years. Alf loved a contrary of age classes to the advantage of both tenures, (5) regrowth argument. While I could not emulate his approach or style, I tried thinning for habitat management and water production, (6) better to illuminate some current issues in dealing with ‘fires, forests access for fire management and more fuel reduction burning and and futures’ in the light of recent history in Victoria. (7) strategic orientation of harvest areas. Sustainability In the past, the conservation versus development debate about forests has been a ‘zero-sum game’. Whatever one side gained Much attention has rightly been focused on sustainability in recent the other lost, as areas were transferred from one designation years and on sustainable forest management in particular. The most to the other. That approach neglects the role of a third player,
Australian Forestry 2009 Vol. 72 No. 4 pp. 195–205 196 Fires, forests and futures widely cited definition of sustainability rests on the definition of erratic, and the long-distant future is unpredictable. Central to sustainable development by the Bruntland Commission (1997): this dilemma is determining, as a society, what values we are seeking to maintain in different forest ecosystems in both the Sustainable development is development that meets the needs immediate future and the longer term. Anticipating the needs of the present without compromising the ability of future of future generations, or assuming that they will value the same generations to meet their own needs. environmental attributes as does our generation, is a continuing challenge, not only for economics but for the political process Sustainable forests and forestry as well.
To an economist, the central theme of this definition amounts to To address this dilemma, three features were introduced in a concern about intergenerational equity — how we balance the constructing a simulation model for analysis of futures in the preference for present consumption against our concern for that ash-type forests of Victoria. of future generations. To some economists, but not all, the equity tag places it outside the direct province of economic theory. The first is that forest planning over time is best restricted to a planning horizon of 40–50 y, because attempting to forecast Some time ago (Ferguson 1996), I tried to rationalise that desire beyond that time is a futile exercise. That is not to say the long- for intergenerational equity in valuing utility over time through distant future should be neglected. Rather, it forms the linchpin consideration of the social rate of time preference, the discount of the second feature. rate that measures the relative preference for present over future consumption. I posited a discount rate that followed commercial The second feature is that we should focus on the state of the forest values over the investment time horizons that are common (i.e. (i.e. the structural and spatial properties) at the end of planning generally less than 50 y), but then declined progressively thereafter horizon to ensure that we have left it in a better condition for those until it reached a steady low state founded on our unwillingness future generations, that is, a system that is capable of providing to discriminate between the consumption of successive future the things that we currently value, or that future generations generations in the long-distant future. might value. In other words, the terminal state at the end of the planning horizon becomes another constraint in the model. This Argument over the social discount rate has a long history and sidesteps the issue of the longer-term path of the discount rate is continuing. Koopmans (1960) showed that under certain but also provides a sharper focus on the goals we are setting for assumptions characterising utility functions that the discount rate the longer-distant future. Ideally, of course, the model should should be constant. Beltratti et al. (1994) introduced an approach also be spatial in content, which this model is not. Nevertheless, that was based on achieving the highest indefinitely sustainable we are slowly moving towards that goal as computing power and welfare — the so-called ‘Green Golden Rule’, in which the long- speed increases. This will enable the issues of variability and term future is the sole dictator of the social discount rate, which erratic changes due to natural disasters such as fire to be better therefore resides at a constant rate of zero. Chichilnisky (1996) addressed. showed that with the addition of further rigorous assumptions about the long-term future, a time path for the social discount The third feature concerns recognition of the impact of natural rate similar to the one I have described could be justified. Heal disturbances, and fires in particular, on forest planning. These (1998) describes this latter approach as the balance between a impacts are much broader than those of harvest scheduling or ‘dictatorship of the present’ and a ‘dictatorship of the future’. planning horizons, but the limit on the planning horizons and consideration of the condition of the forest at the end of the horizon But if any one of these three starkly differing approaches is an are still appropriate. Indeed, such an approach can be applied to appropriate device for valuing resource consumption issues over advantage to the broader forest estate that includes conservation time, how come it is so little adopted? Economic studies that reserves as well as the smaller areas of state forests now available choose one or the other in analysing major issues such as climate for timber harvesting. change have not been widely accepted, and choice is left to the political process. Ash-type forests Part of the reason is that other factors mask whatever path the The prompt for the title of this lecture stemmed from a study social discount rate takes. Skidelsky (2008), the biographer of of seed-storage requirements following the 2009 fires for the Keynes, quotes him as describing economics as ‘a polite technique ash-type forests in Victoria, commissioned by the Department which tries to shield us from the fact that we know very little of Sustainability and Environment. Technical details of the study about the future’. Keynes (1936) was forthright in arguing that are reported in Ferguson (2010), so this lecture described and the long-term future is unpredictable and went on to underscore illustrated the relevant material in general terms and avoided that point many times, for example, ‘The outstanding fact is the detailed repetition of source material, methods, equations and extreme precariousness of the basis of knowledge on which our results, except for those germane to the thesis of the lecture. estimates of prospective yield have to be made.’ Of course, a sceptic might extend his barb directed at economics to many other Much has been written about the ash-type forests of Victoria. disciplines, including ecology, forestry and climate sciences. Two major and distinct forest types are involved; the mountain ash (Eucalyptus regnans) forests and alpine ash (E. delegatensis) We are stuck with the dilemma that we should look to a compromise forests. There are smaller areas of related forest types such as between the dictatorship of the present and a dictatorship of the shining gum (E. nitens), but to keep the presentation simple I distant future, knowing that the immediate future is variable and
Australian Forestry 2009 Vol. 72 No. 4 pp. 195–205 Ian Ferguson 197 refered only to these two forest types and the dominant species national parks) in which no directly exploitive uses are permitted, that characterise them. The differentiation in altitude that the and state forests in which multiple uses including wood production respective names imply results in a predominantly separate spatial and certain other exploitive uses are permitted. geography, as shown in the Figure 1.
Mountain ash forests are the archetypal tall forests of Australia Fires and are mainly concentrated in one large patch in the Central Let us restrict our attention initially to state forests. Most of the Highlands of Victoria. The patch is interspersed by higher- mountain ash in state forests originated from a very extensive elevation ridges carrying alpine ash, snow gums and alpine fire in 1939, but some smaller stands of regrowth from 1919 and meadows and by foothill ridges carrying mixed species such 1926 fires were either not burnt or not much damaged by the 1939 as messmate stringybark and peppermint. Alpine ash forests fire. The 1939 Black Friday fire was a fire that by any measure generally have a less complex and diverse understorey than was intense — possibly more intense in places than the recent mountain ash and are scattered more widely in patches across 2009 Black Saturday fire. In 1939, about 85% of mountain ash the Eastern Victoria highlands from the eastern end of the Central forests now in state forests was burnt at such high intensities that Highlands, through the Great Divide, and into the far eastern the overwhelming majority of trees were killed. Fortunately that Alpine area — to use some names that recur in formal reports was a good seed year and in most places prolific regeneration (Wareing and Flinn 2003; Flinn et al. 2009). followed on a seedbed in which competition from other plants Ash-type forests are an iconic part of our natural heritage. They had been much reduced by the intensity of the fire. There were are aesthetically attractive, highly productive of market values exceptions, such as patches where earlier fires had produced such as wood and biomass, and provide important non-market regrowth that in 1939 was too young to bear seed, or where the catchment and biodiversity values and wildlife habitat. They seed crop was locally insufficient, resulting in dense regeneration are divided by formal tenure into conservation reserves (mainly of silver wattle or other species.
20 10 0 20 40 60 80 Kilometres
Figure 1. Location of ash-type forests in eastern Victoria
Australian Forestry 2009 Vol. 72 No. 4 pp. 195–205 198 Fires, forests and futures
Many of the alpine ash state forests were also burnt and 1939 fires killed extensive areas of trees, the resultant even-aged regenerated in 1939, but they have a wider geographic dispersal regeneration provided useful information that has been used in the and more varied fire history, not least through successive fires in estimation of the areas of age classes (Ferguson 2010, Table 1). 2003 and 2006–2007 as well as 2009. Figure 2 is a composite picture of fire boundaries for the Central Highlands (2009), Great The issue of fire frequency, both in history and in the future, is Divide (2006–2007) and Alpine (2003) fires. important given the projections of future climate change. Past research is scanty and has highlighted the uncertain state of our Figure 2 does not distinguish between forest types, but the areas knowledge of past fire frequency, let alone the issue of climate are so widespread and overlapping of those in Figure 1 that the change. magnitude of the impact on both the forest types under exam ination will be obvious. The mean intervals between fires were used to characterise the probability distributions in Monte Carlo processes used to model the occurrence of fires in state forests. Because the estimates of Fire occurrence and extent the mean intervals between fires for mountain ash spanned such Ash-type forests are moist forest types, generally found at medium a wide range, a range of values was used in this study based on to high elevations, with relatively high rainfall, deep soil and, discussion with field observers. For mountain ash, values of 100, droughts permitting, abundant soil water reserves. Under most 80 and 60 y were used. For alpine ash, the corresponding estimates conditions they are seldom dry enough to support an intense were 60, 50 and 40 y. fire. This makes make fuel reduction burning impractical in all A trinomial probability distribution was used to differentiate but the mixed forest margins or ecotone sites. Burning of the ash between the years in which no fires occurred, those in which forests is in any event undesirable because of their sensitivity to minor fires less than 2000 ha occurred, and those in which fire and widespread mortality following fire. Under conditions of major fires larger than 2000 ha occurred. As with all probability prolonged drought, however, they carry heavy loads of ground- distributions discussed in this study, the trinomial refers to the based and elevated fuels, capable of supporting fires of high parent distribution. It is applied in an inverse cumulative form in intensity. As Tolhurst and McCarthy (2004) have pointed out, fire Monte Carlo processes to enable a random value of the variable intensities and rates of spread are more weather-dependent than in question to be drawn. That value becomes a sample observation fuel-dependent under extreme conditions, such as those above a that is combined with those for other variables drawn similarly, Forest Fire Danger Index of 50. enabling a simulation of successive years up to the planning Before the 1980s, our knowledge of the frequency and extent horizon in a single run. The process is repeated many times, in of past fires in these forest types was poor. Most of the prior this case for 2000 runs, for each set of decision or key variables, descriptions were unable to differentiate forest types and estimates enabling probability distributions of the final economic and risk of areas burnt were imprecise. However, where the 1919, 1926 or criteria of interest to be estimated and assessed.
Figure 2. Areas burnt in 2003, 2006–2007 and 2009 fires in Eastern Victoria (Source A. Haywood, DSE)
Australian Forestry 2009 Vol. 72 No. 4 pp. 195–205 Ian Ferguson 199
The areal extent of minor fires was estimated using a truncated intervals in 2050, they show a progression that looks sensible in (at about 2000 ha) negative exponential probability distribution. terms of the likely impact of climate change on fire occurrence. That for major fires was estimated from a uniform probability distribution ranging from 2000 ha to the total area carrying The mean intervals between fires were adjusted on a linear basis to sufficient fuel. account for climate change, such that by 2050 they were reduced by dividing the mean interval in 2009 by 1.2 for the lowest projected rate of climate change and 2.32 for the highest. No Climate change climate change can be represented by a ratio of 1.0 throughout. But what of climate change? Nitschke and Hickey (2007) initiated a study of the implications of climate change for the Seed production future distribution of forest types. However, so far this research has not included a detailed analysis of the impact of disturbances Both mountain and alpine ash are cyclical in seed production with such as fire. marked peaks — on average every 4 y in mountain ash and 5 y in alpine ash. Commercial seed collection is feasible only when Fortunately, Lucas et al. (2007) provided a historical analysis there is a sufficiently heavy crop of capsules, because trees have of past trends in the Forest Fire Danger Index (among others) to be climbed and the branches cut, involving both skill and safety over the period 1973–2007, together with projections to 2050. issues, as Figure 3 shows. This index is a composite of rainfall and the time since rain, air temperature, relative humidity, wind speed and drought index. It Using the data for mountain ash as a basis, and recognising that was originally calibrated to a maximum of 100, corresponding regional estimates across the entire forest type would tend to notionally to general conditions of the Black Friday fire in 1939. moderate the fluctuations apparent in that data, a stochastic model Any value over 50 is said to constitute extreme forest fire danger of the occurrence of good seed production years was formulated, and over 75, catastrophic danger. involving an immediate decline after the peak and progressive increase to the next peak year. Finally, a stochastic overlay was Notwithstanding the obvious limitations of any such index, the imposed on those trends. Samples of the simulated time patterns results of Lucas et al. (2007) provided a means of taking climate of the proportions by area of natural regeneration are shown in change into account using the change in the number of days of Figure 4. extreme and catastrophic levels to 2050. The ratio of the values at the beginning and end of this period was 1.2 for their lowest Simulation of seed collection was related to the proportion of projected rate of climate change (corresponding to an expected natural regeneration, assuming that this proportion was also a increase in mean temperature of around 1°C), while that for reflection of the amount of seed held, on average, in capsules in the their highest rate of climate change was 2.32 (corresponding to remaining areas of living trees over 29 y of age. The relationship an increase of about 2.5°C). If the mean intervals chosen earlier was determined using a non-linear function that took the target are divided by these values to obtain new estimates of the mean level of seed collection for a ‘good’ seed production year as an upper limit. A stochastic component was also added to moderate the exact relationship. Figure 5 shows the simulated time patterns of seed production that were derived by applying this process to the data in Figure 4.
Economic criterion and risks The discounted value of net costs was the economic criterion used to examine alternative seed supply storage and collection
Mountain ash 1
0.5
0 5 10 15 20 25 30 35 40 Proportion by area Year
Alpine ash 1 0.5 0 5 10 15 20 25 30 35 40
Proportion by area Year
Figure 4. Sample of simulated proportions of area naturally regenerating after major fires Figure 3. Seed collection (Source: Justin McManus, FairfaxPhotos)
Australian Forestry 2009 Vol. 72 No. 4 pp. 195–205 200 Fires, forests and futures facilities, but this criterion and decision is not germane to this across all runs of the particular simulation to provide a probability discussion. However, possible trade-offs between the costs and criterion for assessing unacceptable risks. For example, if, taken risks involved also had to be considered and the risks are very over 2000 runs, 60% of the outcomes for a particular criterion were much germane to this study. undesirable, the overall outcome might be deemed to constitute an unacceptable risk. One undesirable risk was the biological criterion where, put simply, there was not enough forest of seed-bearing age plus seed in storage in 2050 to supply the average future demand for Key variables artificial regeneration over the ensuing 20 years. The simulation study also had to account for several other Another risk was the timber production criterion, where the area variables besides mean intervals, climate change, seed-production that had reached rotation age or greater in 2050 was insufficient years and extent of natural regeneration after major fires for each to maintain the harvest level over the ensuing 20 years. forest type, and some deterministic constraints. Other stochastic variables included the years of occurrence of minor and major fires Figure 6 is based on the results of two simulation runs deliberately respectively and the areal extent of those fires. The deterministic selected to illustrate the risks visually. The upper panes of Figure 6 constraints included the allocation of burnt areas to age classes, show a low-risk situation for seed production in 2050, because the confining seed collection to areas over 20 y of age, and restricting age-class distributions for the two forest types are well spread and successive fires to areas of forest of at least 4 y of age that have relatively even. The lower panes of Figure 6 illustrate a patently sufficient fuel to burn. high risk, because the overwhelming majority of the area of the forest type is under 40 y of age and most of it is under 20 y in Three key variables seemed likely to affect the outcomes the case of alpine ash. materially, so a range of values was tested to explore the sensitivity of the outcomes. These were: These two criteria can be defined quantitatively in terms of the • seed collection and storage targets variables available in the simulations and thus it is possible to • mean interval between fires summarise the outcomes in terms of probabilities over each set of 2000 simulation runs. Undesirable risks were then summarised • climate change. The major decision variable was obviously the targets set for seed collection in good seed production years, the other two being outside the control of DSE. Mountain ash 3000 Results
State forests 0
Seed collected (kg) 5 10 15 20 25 30 35 40 Year Because the results from varying mean interval parallel those for varying climate change, attention can be restricted to the latter Alpine ash in this account1. 4000
2000
0 1 Seed collected (kg) 5 10 15 20 25 30 35 40 Attendees at the ANU Westoby Lecture in Canberra on 26 August 2009 may notice Year differences between Figures 7 and 8 and their analogues in that presentation. This reflects a subsequent inspiration to model the proportion of natural regeneration Figure 5. Simulation of seed collection based on data in Figure 4 and seed collection as stochastic variables, rather than deterministic. )
–1 Mountain ash: low seed-production risk Alpine ash: low seed-production risk 100 150
0 Seed yield (kg ha 1 2 3 4 5 6 0 1 2 3 4 5 6 )
–1 Mountain ash: high seed-production risk Alpine ash: high seed-production risk 100 150
0 0
Seed yield (kg ha 1 2 3 4 5 6 1 2 3 4 5 6 Age class: 1 = 0–19 y, … Age class: 1 = 0–19 y, …
Figure 6. Examples of simulations showing low and high seed-production risks, based on age class distributions in 2050
Australian Forestry 2009 Vol. 72 No. 4 pp. 195–205 Ian Ferguson 201
A probability of 0.1 represents a risk of violating the criterion Alpine ash, on the other hand, is at material risk (i.e. above a of one in ten, which is probably tolerable. A probability of 0.6 probability of 0.2) by 2050 if seed collection is low (2000 kg in represents a risk of six in ten, which is not. More importantly a good seed production year), and at perilous risk (probability for this analysis, an incremental increase in probability of 0.2 or above 0.9) if climate change is low or greater. Increasing seed greater constitutes a material unacceptable change. collection to the median 4000-kg level reduces the risk for all rates of climate change but there is nonetheless a material risk Figures 7 and 8 summarise the results for the probabilities of for a high rate of climate change. unacceptable risks for two scenarios of Low and Median targets for seed collection in good seed-production years. The results Figure 8 shows that the viability of timber production in 2050 for the High scenario target are not shown because they are only both forest types is not greatly affected by the target level of seed slightly less than those for the Median target and a High target is production. Both target levels show material and increasing risk probably unattainable or inordinately costly in practical terms. as the rate of climate change increases, the risk levels for the two forest types under high rates of climate change and median seed Figure 7 shows that, for mountain ash, increasing seed collection collection levels being uncomfortably high (above 0.4). targets from 1000 to 2000 kg in good seed production years reduced the 2050 risk of having insufficient area of seed-producing Like all simulations, the results are open to interpretation based age plus seed in storage to meet average artificial regeneration on views about various assumptions, probabilities and models needs in the ensuing 20 y. However, this level of seed collection involved. The actual outcomes could be worse if good seed did not eliminate the risk, which is at the borderline of materiality production years are less frequent and fires more frequent and or if the rate of climate change is high.
Mountain ash Alpine ash
1 Median seed collection 1 Median seed collection Low seed collection Low seed collection 0.8 0.8
0.6 0.6
0.4 0.4 Probability Probability 0.2 0.2
0 0
No climate change No climate change
Low climate change Low climate change
High climate change High climate change
Figure 7. Seed production risks
Mountain ash Alpine ash
1 Median seed collection 1 Median seed collection Low seed collection Low seed collection 0.8 0.8
0.6 0.6
0.4 0.4 Probability Probability 0.2 0.2
0 0
No climate change No climate change
Low climate change Low climate change
High climate change High climate change
Figure 8. Timber production risks
Australian Forestry 2009 Vol. 72 No. 4 pp. 195–205 202 Fires, forests and futures extensive. Likewise they could be better if the converse holds. We current policy. Protecting that regeneration for the next 20 y, have no way of knowing how the cards will fall. Quantitatively however, is a matter that merits urgent consideration. expressed risks provide the most appropriate guidance. The linkage of mean interval with climate change enabled a relatively Active management simple stochastic analysis and a simple exposition of results based on a range of values for each, rather than the so-called ‘info-gap’ Franklin (2002) and others have stressed the need for active approach (see McCarthy and Lindenmayer 2007). management of temperate natural forests, both conservation reserves and state forests. They are not referring to an open slather Conservation reserves of exploitation but to a limited and conservative application of a range of activities that includes some salvage logging Some will ask, ‘What about the conservation reserves?’ It would (Lindenmayer et al. 2008), some fuel reduction measures, be interesting to run a similar exercise but the data were not some thinning of young stands to accelerate the development available at the time of undertaking this study. Present Parks of structures characteristic of late succession, and measures to Victoria policy is to not allow intervention through artificial reduce the risks posed by exotic pests and diseases under climate natural regeneration and this greatly increases the risk to the change. survival of these forest types. The National Conservation Reserve system is a very important The mountain ash conservation reserves are mainly associated with achievement of the Regional Forest Agreement process and a Melbourne’s closed water catchments and many carried iconic much-needed component of management of Australian native old-growth forests prior to 2009. In addition to their biodiversity forests. But the ash-type forests in conservation reserves (mainly values, these old-growth forests played an important role in national parks) are at least at much at risk as those in state forests, increasing domestic water yields and decreasing turbidity. The as the 2009 fires have shown. Hence the need to consider active O’Shannassy, Wallaby Creek2, Maroondah, Bunyip and Tarago management measures that cross over the public perception of catchments carried old-growth mountain ash not previously burnt the tenure boundaries. Ash-type forests, whether state forest or severely for over 150–250 y. Significant areas were burnt severely conservation reserve, have especially important roles for a wide in 2009 (Melbourne Water 2009), early reports indicating that 93, range of uses. If we continue to neglect the risks fire poses, we 100, 76, 100 and 100% respectively of those catchments were may not be able to sustain the uses and values we currently use burnt and much of the mountain ash forest killed.. These forests or ascribe to them. will probably regenerate naturally given that 2009 was a good seed production year, but another fire in the next 20 y might be The question, then, is how to effectively manage such risks. Seed less kind, so the future risk to these forests is material, especially collection for artificial regeneration of areas burnt but that are until the regenerating stands set seed. unable to regenerate naturally are only one part of the solution. While the past reluctance (if not legal inability) of the managers The succession of fires in 2003 and 2006–2007 has also created of conservation reserves to countenance some of these measures an age-class distribution in the alpine ash conservation reserves in national parks and related reserves is entirely understandable, a that puts them at risk, probably much more so than the results major review of management and policy is needed. Like Franklin, reported for state forests, because artificial regeneration is not Lindenmayer and many other ecologists and foresters, I believe currently allowed and large areas are less than 20 y old. active management of ash-type forests is essential and needs to include: What are the implications if regenerating stands of mountain ash and alpine ash are burnt prior to reaching 20 y of age and those species do not subsequently regenerate in the conservation 1. A well-equipped workforce reserves? There is, of course, great variation in the plant species While helicopter and fixed-wing water tankers and helicopter composition in the ash-type forests, making detailed generali crews are extremely useful for first attack on fires, they generally sations about future composition difficult. Much also depends on have to be backed by the construction or renovation of tracks the subsequent fire frequency (Florence 1996). But the general and associated measures to contain the fire on the ground, using tendency would be to promote more open, less rich and more vehicular water-tankers and trained personnel. DSE mostly fire-prone forest types with different animal and plant biodiversity contracts in bulldozers to assist and likewise Parks Victoria and less attractive properties for domestic water supply, especially and Melbourne Water, who contribute to fire suppression on in relation to turbidity. Such changes are not an argument for Melbourne’s water supply catchments in national parks. But all artificial regeneration of ash-type forests across all conservation are limited in the trained staff available and in the bulldozers and reserves, but there may be a cogent argument, especially relating other resources that can be brought in quickly for first attack. to domestic water supply, for artificial regeneration on specific areas. Provided regeneration from the 2009 and earlier fires is Timber harvesting contractors form a potentially important and adequate, as seems likely, this is an issue for future rather than somewhat under-utilised component in this workforce in an otherwise very sparsely populated environment, in which local knowledge and experience in tall forest is critical. Swing blades and water tanks can be fitted to wheeled skidders and forwarders, possibly on a removable basis, as Jaffe and O’Brien (2009) have 2 Ashton (2000), a much respected ecologist, provided a detailed account of ‘The Big Ash’ in Wallaby Creek in 1949 compared with 1997. Sadly, his closing words shown. The formation of more mobile harvesting crews, both in the abstract — ‘Whether the Big Ash will be spared from fire in future centuries forwarders and tree-fallers, could assist timely first attack, as is very doubtful’ — proved all too accurate.
Australian Forestry 2009 Vol. 72 No. 4 pp. 195–205 Ian Ferguson 203 well as later containment of major fires in moderate topography, 4. Strategic swaps of age classes to the advantage of both following the cessation of extreme weather conditions. Attention tenures and possibly subsidies need to be given to the design and use of harvesters that are better-suited to dual usage for harvesting Marginal swaps of regrowth areas between the conservation and fire fighting, with as small a soil disturbance footprint as reserves and state forests can potentially reduce the risks to both. possible. Recruiting some older regrowth into the conservation reserve in exchange for salvage of fire-killed regrowth can reduce the Few harvesting contractors have large bulldozers, so fire- totality of harvest removals plus mortality to the advantage of management agencies will continue to draw mainly on the future timber production, carbon footprint, biodiversity and earth-moving contractors for these. In common with all vehicles water production. Adding some older regrowth to conservation used in fire suppression, these need to be fitted with automated reserves can provide a more robust distribution of age-classes, vehicle location systems for personnel safety and identification successional stages and biodiversity. This is part of broader of containment boundaries. philosophy of looking at vegetation types across the landscape in a holistic way and creatively considering the best way to 2. Removal and salvage of fire-killed trees along access maintain patterns of age-distributions and ecosystem resilience, roads irrespective of tenure.
Fire-killed trees along roads and access tracks later become This would require legislative changes to permit timely resolution hazards in terms of torching and spotting in later fires, or impeding with respect to salvage logging and associated protocols to protect access when they fall over roads or access tracks. They should be both forms of tenure from exploitation. felled and where possible sold for timber, substituting for harvest volume from living forest. Agencies presently fell trees that are 5. Regrowth thinning for habitat management and water dangerous along tourist roads, move the trunks away from the production road where they impede access and constitute additional fuel that is difficult to extinguish, and even mechanically slash under Lindenmayer (2002) has made the case that some thinning of the residual trees in or adjacent to some conservation reserves. regrowth stands may be desirable in order to maintain populations Utilisation of sound timber hardly has a devastating impact under of Leadbeater’s possum. Given the extent of the recent fires in these circumstances. state forests and conservation reserves, the situation warrants review in order to achieve the best outcomes. Similar legislative Similarly, felling of dead trees on ridges may be necessary to and related changes to those outlined in the previous point may reduce potential spotting if the subsequent regrowth is ignited. be needed. Likewise felling of fire-killed trees on aesthetically important landscape units may be required. Thinning of mountain ash regrowth stands to increase water production was the subject of earlier research (O’Shaughnessy et Salvage logging3 needs to be subject to transparent and timely due al. 1993). The experiments are being remeasured and assessed but process, based on some well-established protocols (Lindenmayer only preliminary results are available to date (Hawthorne 2008). et al. 2008). For example, Lindenmayer and Ough (2006) have Because of the preponderance of mountain ash in Melbourne’s mounted a persuasive argument that salvage logging of fire- water catchments, thinning to increase water production in killed old growth ash-type forest should not be permitted, and conservation reserves warrants consideration in any new the protocols should reflect this argument, except in prescribed protocols. exceptional circumstances. 6. Better access for fire management and more fuel 3. Seed collection and artificial regeneration in conservation reduction burning reserves Further investment in the provision of access tracks and the Seed collection is a strategically vital matter if the risks posed reduction of hazards along them by mechanical slashing is by fire are to be offset. Access to conservation reserves for seed needed. DSE had commenced such a program prior to the 2009 collection is currently restricted except for purely botanical fires — a move I viewed with some scepticism prior to the 2009 studies. Greater access needs to be accorded to seed collection, fire, but no longer. If the risks to the ash-type forests are to be given the magnitude of the task. reduced, better access in state forests and conservation reserves is essential. Maintenance of all roads and tracks in or serving the Biodiversity values are as much at risk as others, so the general ash-type forests is equally important: it is a practice that often stricture against the use of artificial regeneration also needs to be tends to slip when budgets are tight. These measures will not reconsidered, especially on water catchments where that use has at stop a head fire burning under extreme conditions but they may least as great a priority as maintaining biodiversity through natural assist containment of flank fires and assist in attacking head fires processes. This is not an argument for universal use of artificial when conditions moderate. They also may reduce the likelihood regeneration in the ash-type forests but rather a limited and of linking of nearby lighting strike fires and hence reduce the strategic program using sensible protocols and mechanisms. overall potential for a major fire front to develop.
More strategic fuel-reduction burning of peripheral mixed 3 DSE has guidelines for salvage logging but these do not extend to many of the forest types is needed. DSE has already started to implement an measures advocated here.
Australian Forestry 2009 Vol. 72 No. 4 pp. 195–205 204 Fires, forests and futures increased program of landscape-level fuel-reduction burning. populations and their habitat, notwithstanding the understandable However, the risks posed to the future of the ash-type forests distaste that some in the community have for the practice. We call for a strategic review of the program, specifically to protect sterilise koalas in conservation reserves and elsewhere for similar those forests. reasons and with similar concerns. And we remove dangerous limbs, fell trees in picnic and parking areas, fell dangerous trees A fruitless debate has raged over fuel reduction burning, especially along roadsides and even sometimes mechanically slash under in relation to the frequency, intensity and patchiness appropriate them in conservation reserves and related areas. The purist rhetoric to landscape-level burning. Gill’s (1999) dictum amounts to of one tenure – one use that is often featured in the media is a false varying the timing, frequency, intensity and patchiness as much as dichotomy that ignores the multiple uses and tradeoffs involved possible. But as Clarke (2008) has pointed out, pleas for burning and the stochastic variability of natural processes, and promotes to achieve a mosaic lack practical criteria for implementation. an unrealistic sylvan fundamentalism in the public domain. As with a number of the other active management measures, monitoring of and research on the biodiversity, timber and water Jack Westoby and Alf Leslie were very much of the view that production impacts will be needed to refine the protocols and managing forests was primarily about managing them for people, mechanisms — in some cases requiring major multi-disciplinary both now and in the future. If we are to maintain the values of research projects. these iconic forests for future generations, in a future climate that may be quite different from the current one, we will need to pursue 7. Strategic orientation of harvest areas active management, based on monitoring and research. Whatever ethical or philosophical position one might take on discount rates, Contrary to Gill’s (1999) view, there is ample anecdotal evidence this will involve costs and immediate and ongoing actions. that young (under 5 y) regrowth resulting from harvesting, together with the harvest road system, can assist suppression We face a future in which there is a material risk we could lose and containment, except under extreme conditions. While further most or all these ash-type forests, not over a period of several study is warranted, this opens up the issue of strategic location of millennia, such as that in which they evolved, but in the next 60 y. harvesting to protect key adjacent conservation assets. We can decide to do nothing and accept that loss, or we can engage in active management to conserve their values and uses.
Future policy In posing that choice, I hear Alf Leslie reminding me ‘that may be a cogent and compelling case you have argued, but don’t forget, Let me now summarise some of the implications for forest policy. when it comes to public policy, stupidity nearly always wins’ In the past, the conservation versus development debate about (Leslie 2000). That said, I know that Alf loved the ash forests and forests has been a ‘zero-sum game’. Whatever one side gained would be delighted to be proven wrong in this case. the other lost, as areas were transferred from one designation to the other. Acknowledgements That approach neglects the role of a third player, Mother Nature, especially relating to fire. We need to move away from the ‘one Thanks are due to the Department of Sustainability and tenure – one use’ thinking (Ferguson 2005) that underlies the Environment, Victoria, for permission to draw on a consultancy zero-sum game and recognise that there are multiple uses and report commissioned by them and to O. Bassett, L. Bren, L. Bull, values associated with the ash-type forests, many of which have A. Crisp, P. Fagg, F. Cumming, A. Haywood, R. Keenan, no special relationship with the nominal tenure. N. Ronan, M. Ryan, R. Sands, T. Sanders, P. West and A. Wilson for their assistance or comments, none of whom should be held The ash-type conservation reserves are as much about multiple responsible for the views expressed. uses as are state forests. Some uses are common to both, some are excluded from one and not the other, and some vary by degree References rather than absolute exclusion in one or the other. If we want to maintain the ash-type forests, we need active management Ashton, D.H. (2000) The Big Ash forest, Wallaby Creek, Victoria — that integrates across those tenure boundaries (Ferguson 2007) changes during one lifetime. Australian Journal of Botany 48, especially in relation to protecting against the risks posed by 1–26. fire, the biggest threat. Without radically altering present tenures Beltratti, A., Chichilnisky, G. and Heal, G. (1994) The green golden rule. Economic Letters 49, 175–179. or principles of management, co-operative arrangements and Bruntland Commission (1997) Our Common Future. World Commission trade-offs at the margin would enable both tenures to benefit, as on Environment and Development, New York. well as public forest ownership in general. Transparent and timely Chichilnisky, G. (1996) An axiomatic approach to sustainable processes and clear protocols are essential. Protocols will need to development. Social Choice and Welfare 13, 219–248. be reviewed in the light of further monitoring and research, which Clarke, M. (2008) Catering for the needs of fauna in fire management: need to be built into ongoing budget provisions. Without better science or just wishful thinking. Viewpoint article in Wildlife information, public interest and support for these measures will Research 35, 385–394. decline as memories of the recent fires fade. Ferguson, I. (1996) Sustainable Forest Management. Oxford University Press, Melbourne, 162 pp. There will be some who find some of the measures suggested Ferguson, I. (2005) Australian forestry: beyond one tenure, one use. In: distasteful, or worse. Yet we already countenance the culling Sayer, J. and Maginnis, S. (eds) Forests in Landscapes: Ecosystem of kangaroos in conservation reserves in order to protect the Approaches to Sustainability. Earthscan, London.
Australian Forestry 2009 Vol. 72 No. 4 pp. 195–205 Ian Ferguson 205
Ferguson, I. (2007) Integrating wood production with sustainable forest Lindenmayer, D.B., Burton, P.J. and Franklin, J.F. (2008) Salvage management: an Australian viewpoint. Journal of Sustainable Logging and its Ecological Consequences. Island Press, Washington Forestry 24, 19–40. D.C., 224 pp. Ferguson, I. (2010) Seed collection and storage facilities for Victorian Lucas, C., Hennessy, K., Mills, G. and Bathols, J. (2007) Bushfire Weather ash-type species. Australian Forestry 73, in press. in Southeast Australia: Recent Trends and Projected Climate Flinn, D., Wareing, K. and Wadsley, D. (2009) The Victorian Great Change Impacts. Consultancy report prepared for The Climate Divide Fires, December 2006 – February 2007. Fire and Emergency Institute of Australia. Bushfire CRC and CSIRO, Melbourne, Management, Department of Sustainability and Environment, 8 pp. Melbourne. 137 pp. Available on www.dse.vic.gov.au/fires. McCarthy, M. and Lindenmayer, D. (2007) Info-Gap Decision Theory for Florence, R.G. (1996) Ecology and Silviculture of Eucalypt Forests. assessing the management of catchments for timber production and CSIRO Publishing, Collingwood, 413 pp. urban water supply. Environmental Management 39, 553–562. Franklin, J.F. (2002) Challenges to temperate forest stewardship: focusing Melbourne Water (2009) Fire and Water: Challenges and Learning on the future. In: Lindenmayer, D.B. and Franklin, J.F. (eds) — Melbourne Water’s Experience, February 2009 Victorian Towards Forest Sustainability. CSIRO Publishing, Melbourne, Bushfires. Melbourne Water, Melbourne, 44 pp. Available on www. pp. 1–13. melbournewater.com.au. Gill, A.M. (1999) Biodiversity and bushfires: an Australia-wide Nitschke, C.R. and Hickey, G.M. (2007) Assessing the Vulnerability of perspective on plant-species changes after a fire event. In: Gill, Victoria’s Central Highlands Forests to Climate Change. Technical A.M., Woinarski, J. and York, A. (eds) Australia’s Biodiversity Report 1/2007, School of Forest and Ecosystem Science, University Responses to Fire. Biodiversity Technical Paper 1, Environment of Melbourne, Melbourne, 156 pp. Australia, pp. 5–53. O’Shaughnessy, P.J., Benyon, R.G. and Lucas, A. (1993) The Crotty Hawthorne, S.D. (2008) Predicting the long-term impact of forest Creek project: The effects of strip thinning Eucalyptus regnans on thinning on water yield. Proceedings U21 Postgraduate Research forest growth and water yield. In: Second Progress Report — Post- Conference, I: Water: How Need Drives Research and Research Treatment Phase. Melbourne Water Report No. MMBW-W-0020. Underpins Solutions to World-Wide Problems. 20–25 July 2008, Department of Conservation and Natural Resources and Melbourne University of Birmingham, Birmingham, UK, 6 pp. Water, Melbourne, Australia. Heal, G. (1998) Valuing the Future: Economic Theory and Sustainability. Skidelsky, R. (2008) What would Keynes have done? The Independent. Columbia University Press, New York, 226 pp. Saturday, 22 November 2008. Jaffe, V. and O’Brien, S. (2009) Mechanized Equipment for Fire and Fuels Tolhurst, K.G. and McCarthy, G. (2004) Effect of prescribed burning on Operations. The authors, 160 pp. Available on www.wildfirelessons. wildfire intensity — a case study from the 2003 fires in Victoria. net/documents/YellowBook_LowRes_Cover.pdf. School of Forest and Ecosystem Science, University of Melbourne, Keynes, J.M. (1936) The General Theory of Employment, Interest and 30 pp. Pers. comm. Money. Harcourt, Brace and World, Inc., New York, 403 pp. Wareing, K. and Flinn, D. (2003) The Victorian Alpine Fires, January – Koopmans, T. (1960) Stationary ordinal utility and impatience. March 2003. Fire Management, Department of Sustainability and Econometrica 28, 287–309. Environment, Melbourne. Six sections and appendices. Available Leslie, A.J. (2000) Stupidity nearly always wins. Wood Science Digest on www.dse.vic.gov.au/fires. No. 69, December 2000, 4–7. Lindenmayer, D.B. (2002) Integrating wildlife production and wood production. In: Lindenmayer, D.B. and Franklin, J.F. (eds) Towards Forest Sustainability. CSIRO Publishing, Melbourne, pp. 47–72. Lindenmayer, D.B. and Ough, K. (2006) Salvage logging in the montane ash forests of the Central Highlands of Victoria, south-eastern Australia. Conservation Biology 20, 1005–1015.
Australian Forestry 2009 Vol. 72 No. 4 pp. 195–205 206 Replacement of wooden shingle roofs on buildings at Norfolk Island
Reducing waste in timber procurement is critical to replacement of original wooden shingle roofs on heritage buildings at Norfolk Island
John Cokley
School of Journalism and Communication, Blair Drive, University of Queensland, Queensland 4072, Australia Email: [email protected]
Revised manuscript received 21 September 2009
Summary construct (data triangulation)’ that allows ‘mutual confirmation of measures and validation of findings’ and ‘a means of refining, Traditional timber cutting, recovery and installation processes broadening, and strengthening conceptual linkages’: used by loggers and builders on Norfolk Island are a major impedi ment to the continued and sustainable restoration and maintenance The important feature of triangulation is not the simple of historic buildings at the southern tip of the island. Sustainable combination of different kinds of data but the attempt to relate restoration and maintenance is likely to be a contributing factor them so as to counteract the threats to validity identified in each. in the expected decision by UNESCO about whether the island is included on the World Heritage List. In particular, use of case studies is supported by Berg (2001) who Keywords: shingles; roofs; durability; historic sites; heritage tourism; notes that ‘case study methods involve systematically gathering Araucaria heterophylla; Norfolk Island enough information about a particular person, social setting, event or group to permit the researcher to effectively understand how it operates or functions’ and that ‘the scientific benefit of the case Background study method lies in its ability to open the way for discoveries’. Norfolk Island, a self-governing Australian territory, is in the They may focus on ‘an individual, a group or an entire community Pacific Ocean east of the Australian mainland and has an area of and may utilise a number of data technologies such as life 34.6 km2. The Kingston and Arthurs Vale Historic Area (KAVHA) histories, documents, oral histories, in-depth interviews, and at the southern tip of the island is of exceptional historical and participant observation’ (Berg 2001). social value and is included on the Australian National Heritage List (Australian Department of Environment, Water, Heritage The site and the Arts 2008). It also formed part of the 2008 nomination by the Australian Government for World Heritage listing by the The vernacular buildings at KAVHA were built by hand using United Nations Educational, Scientific and Cultural Organisation materials collected and worked from the vicinity (KAVHA (UNESCO) for decision in 2009. The KAVHA site has four distinct Management Plan 2008). This document indicates that not all settlement periods: Pre-European, two European and a mixed these materials are available today. For example, endemic Norfolk British/European/Polynesian settlement. The precinct contains Island pine shingles were the original roofing material and hence a fine collection of colonial Georgian buildings that functioned the choice for authenticity. However, these shingles deteriorate as a convict settlement. The buildings have been stabilised by a quickly and require regular costly replacement that uses scarce program of restoration and conservation which is detailed in the resources. As a compromise Casuarina tortulosa shingles from KAVHA Management Plan (2008). This article addresses the most New South Wales on the Australian mainland have been used in significant maintenance program in the precinct to date, that of recent years to re-roof small prominent buildings. This shingle restoring and repairing the original Norfolk Island pine shingles was the original material only on the Surgeons Quarters building, a on the roofs of the historic buildings. This program commenced building constructed during the Second Penal Colony. The original in 1985 (P. Anderson, Norfolk Island, 2008, pers. comm.) and casuarina shingles were imported from Sydney, where casuarina effectively concluded 23 years later on 30 June 2008 (Vicky Jack, and ironbark shingles were commonly used. Presumably the Norfolk Island, 2008, pers. comm.). durability of these hardwood shingles had been well established for the Sydney region.
Method The KAVHA plan states that the casuarina shingles are available in limited supply and are expensive since they are imported. This study included a literature review, observation and case studies The expense increases because they are pre-cut in Australia (observation and interviews). The use of such a combination of (P. Anderson, ibid.) rather than being split on the island. The plan methods, identified by Berg (2001) as ‘triangulation’, is accepted, suggests that the shingles be monitored and alternative species indeed recommended, as a way to describe ‘multiple data- tested if the supply becomes inadequate. The plan also suggests collection technologies designed to measure a single concept or
Australian Forestry 2009 Vol. 72 No. 4 pp. 206–211 John Cokley 207 that present supplies of Norfolk Island pine are inadequate for As a roofing material, shingles are exposed to weathering from large-scale use for shingles. However, the management committee sun, wind, debris and precipitation, resulting in rough surfaces also concluded that should supplies of the pine become available and loss of dimensional stability (Salaita et al. 2008). Durable and or treatment methods be developed that extend the durability shingle roofs have been made by using durable timbers that are of the pine shingles then these should be used as the preferred able to repel water and resist decay. In the US, western red cedar roofing material. Given the high cultural heritage values of has provided such a timber as it is naturally water repellent, KAVHA it is essential that future conservation be carried out and gums in the heartwood inhibit the growth of lichens, mould in accordance with best conservation practice and methods. and fungi. The heartwood of old-growth western cedar is rated Generally, all authentic fabric must be conserved. As an aside, as extremely durable, although the generally small amount of shingle trimming and laying adds to the interpretation of the sapwood of this species is not. The heartwood and sapwood area and demonstrations are given at occasional events and for in a species can be very different (Taylor et al. 2002). There is tourists. Interpretation (writing and installation of signs) has general consensus (Cokley 1995; Nolan 2004; Meynink 2006) that largely replaced re-roofing in the KAVHA budget from July 2008 maturity, speed of growth (often rapid in regrowth and plantations) (Jack, ibid.). and source location (e.g. ridge-grown or valley-grown) are related to durability. Therefore the age and origin of Norfolk Island pine Norfolk Island pine wood trees considered for shingles may be an important factor in their selection. Norfolk Island pine (Araucaria heterophylla) is a conifer but not a true pine (Earl 1998). As its vernacular name suggests it is Western red cedar timber is a declining resource and wood from endemic to Norfolk Island and the name is mostly shortened by other species is being used as an alternative (De Groot 1994). The islanders to ‘Norfolk pine’. Norfolk Island pine trees grow to a wood may be durable heartwood and non-durable sapwood, or height of 50–65 m with straight vertical trunks and hard, extremely entirely non-durable wood. The durability of shingles and shakes durable limbs. The 18th century English explorer Captain James can be improved by wood preservation, and even durable woods Cook encountered these trees and he and others (Moorehead 1966; are treated to increase longevity (Feist and Hon 1984). A wide Lennon 2005) considered that they would provide ship masts, range of treatments are used to preserve timber generally and after which the trees were planted on many locations in the South are categorised into pressure treated and surface applied, either Pacific including Lord Howe Island, New Zealand and Hawaii oil- or water-borne (Dickey 2003). Wood from softwoods such as (Lennon 2005). Unfortunately the wood proved too weak for this western hemlock, pacific fir, grand fir, western white pine and red maritime purpose. In the late 1950s a trial shipment of Norfolk alder has been used successfully for shake production in Canada Island pine logs was sent to Sydney for plywood manufacturers and the US (De Groot 1994). These shakes were pressure-treated but the industry was deemed not sustainable by the Norfolk Island with chromated copper arsenate (CCA). Treated shakes from Advisory Council. The wood is light coloured and soft, with a southern yellow pine have a guaranteed lifespan of 50 years knotty pine character and strength characteristics similar to those (Horton et al. 1994). of Northern Hemisphere Douglas-fir (INTAD 2001). It is excellent In the past decade, a number of successful preservation treatments for turning and is used extensively by Hawaiian craftspeople have been banned for certain uses because of reported adverse (HFIA 2000). Another species of Araucaria, hoop pine, is grown effects on human health and environmental parameters (Dickey extensively in plantations in Australia, while parana pine once 2003; Evans 2003). The banned treatments include creosote, occurred in extensive natural forests in Brazil. These woods, of CCA and pentachlorophenol (PCP). As an example of the effects low natural durability, are used principally for timber framing, of these chemicals, Kalnis and Detroy (1984) showed that the interior woodwork, furniture and veneer (INTAD 2001) and not leaching of chemicals from treated softwood beehives had a for exterior applications. detrimental effect on bee colonies and honey production. The The Norfolk Island government passed the Trees Act 1997 to treatments used to treat shakes in the past decade can no longer preserve and manage protected trees such as Norfolk Island pine. be used. Current research on wood preservatives has focused There has been significant planting of pines and other native on the development of environmentally benign treatments trees on the island for the past ten years with the view of the (Termiz et al. 2008). Evans (2003) reviewed recent advances in island becoming self-sustaining for timber (Lennon 2005). Four wood protection. A large number of biocides originally used for hectares of pine were expected to be planted annually after 1985 agriculture are now being introduced as effective preservatives. (Benson 1985). Another area of advance is through the modification of wood by heat treatment which improves resistance to bio-corrosion and reduces hygroscopicity (Gosselink et al. 2004; Gunduz et al. Wooden shingle and shake roofs 2007). The resistance of carbonised wood to environmental decay Wood shakes and shingles are a traditional roof covering. The may be similar to or even better than that of creosote-impregnated difference between a shingle and shake is that a shake is made wood (Gunduz et al. 2007). The temperatures at which the wood by splitting a log longitudinally and it is normally thicker than is treated range from 180 to 200°C. This type of treatment could a shingle (Niemiec and Brown 1993). Today this type of roof be tested on Norfolk Island. is constructed frequently in the US and Canada and remains Wood finishes can provide protection against degradation from servicable for lengthy periods. Shingles (A. Evans, Norfolk ultraviolet light and absorption of water into the timber (Loferski Island, 2008, pers. comm.) are produced by cross-cutting logs, 1999). These materials usually contain a fungicide, a drying oil, shaping them into smaller sections and then splitting off the final a solvent and a small amount of water-repellent material such shingle.
Australian Forestry 2009 Vol. 72 No. 4 pp. 206–211 208 Replacement of wooden shingle roofs on buildings at Norfolk Island as wax or glycol. Crude tall oil, a by-product of the pulping of — Melbourne’s Royal Exhibition Building and Carlton Gardens resinous softwoods, has been used as an effective wood preserv (listed in 2004) and the Sydney Opera House (2007) — are listed ative (Termiz et al. 2008). Shingles traditionally have been as cultural properties, and a further four of which are listed as treated with copper or zinc naphthenate (Niemiec and Brown ‘mixed’. Australia has a current list of 11 more nominations, which 1993; MacLellan and Wolfson 2004). These are both low-toxicity it calls ‘The Australian Convict Sites’ and if these are included, compounds (Dickey 2003) and have been reported as useful for Australia will have 11 natural sites, 13 cultural sites and four bee colonies (M. Kennedy, Queensland Department of Primary mixed sites. At the time of writing the proposed ‘convict sites’ Industries and Fisheries, 2006, pers. comm.). The use of copper included four in or near Sydney, one (KAVHA) on Norfolk Island, compounds may not be acceptable for historic buildings as they five in Tasmania (including the Port Arthur historic site), and one may leave a green residue (Lokerski 1999). Zinc naphthenate in Western Australia. could prove an acceptable preservative for Norfolk Island as De Groot and Stroukoff (1988) reported that pine boxes in humid Observations and case studies jungle conditions resisted decay following zinc naphthenate treatment as well as or better than those treated with PCP. The author visited Norfolk Island in February 2006 and July 2008 to view the KAVHA restorations and to interview a range Norfolk Island pine wood is susceptible to borer and most pine of those involved for this research1. Interviewees included the used on the island is treated at the tanalith plant on the island Administration Minister responsible for the KAVHA precinct, the (Lennon 2005). Tanalith (Tanalised E) is a copper-based wood builders who had been conducting the shingles restoration, the preservative originally developed in Europe in the 1980s. The director of the KAVHA museum, tourism workers, a mill owner, preservative provides protection against most types of decay and a preservation plant manager and a local historian and shingles termites. A synthetic azole-type co-biocide provides protection expert who demonstrates Norfolk Island pine cutting, shingle against copper-tolerant organisms such as brown rot fungi. De manufacture and installation for tourists on the island. Groot and Nesenson (1995), in a review of preservatives for shakes, noted that copper azole was an effective treatment. Observations
Relevance to the World Heritage Convention In February 2006 and July 2008, the author inspected many of the buildings in the KAVHA precinct and photographed them. Discussion of the KAVHA restoration work in relation to the About 120 buildings have been re-roofed during the 23-year UNESCO World Heritage Convention (1972) is especially apt program and all were available for close external visual inspection since the first paragraph of the convention’s preamble notes that and photography. Among them were the Old Military Barracks, ‘the cultural heritage and the natural heritage are increasingly Munna’s, the engineers office, engineers stable, double boatshed, threatened with destruction not only by the traditional causes of single boatshed, pier store, surgeons quarters, surgeons kitchen, decay, but also by changing social and economic conditions which officers outbuilding, the hospital, kitchen, west officers quarters, aggravate the situation with even more formidable phenomena east officers quarters, the magazine, the privy, No. 10 Quality Row, of damage or destruction’. The subject of this article — shingle the Protestant chapel and the New Military Barrack magazine roofing of Norfolk Island pine at the KAVHA precinct — has (P. Anderson, ibid.). proved to be vulnerable to the causes of decay mentioned in the preamble: physical (rot and insect attack) and political (changing The figure of 120 is less than the total number of buildings in the social and economic conditions). precinct, since some now exist only in the form of ruins, without internal or roof fittings. It would be very difficult for an untrained Notwithstanding the decision of the World Heritage Committee observer to differentiate between shingles of Norfolk Island pine in 2009, the KAVHA precinct itself seems to be covered by the and casuarina, or even Hardiflex (P. Anderson, B. Hubber and list’s definition in Article 1 (UNESCO 1972) of cultural heritage: Vicky Jack, all Norfolk Island, 2008, pers. comm.). Indeed, this ‘groups of separate or connected buildings which, because of their proved to be the case during the author’s observations2. However, architecture, their homogeneity or their place in the landscape, are it was acknowledged (Hubber, ibid.) that a trained timber of outstanding universal value from the point of view of history, specialist or a heritage specialist might easily tell the difference art or science’. The aims of the Norfolk Island Administration and that this would effectively dilute the historic integrity and seem also to be well aligned with Article 5, paragraph 1 (KAVHA authenticity of the site. Another interviewee (K. Anderson, ibid.) 2008) that recommends the adoption of ‘a general policy which noted that the interpretative value of the site for tourists or school aims to give the cultural and natural heritage a function in the life children might be disrupted. of the community and to integrate the protection of that heritage into comprehensive planning programs’. This article is well As part of the observations process, the author also visited an aligned with UNESCO (1972) Article 5(c), which recommends island woodworker, the larger of the two island timber mills, the development of ‘scientific and technical studies and research several island homes as part of a regular island tourist excursion … to work out such operating methods as will make the state known as a ‘progressive dinner’, and the island Tanalith plant. capable of counteracting the dangers that threaten its cultural or natural heritage’. 1 This study was approved by a human ethics committee of the University of The World Heritage List includes 679 cultural sites, 174 natural Queensland in accordance with the Australian National Health and Medical sites and 25 mixed-site properties in 145 states. Cultural sites Research Council’s guidelines, Certificate 2008000814. outnumber combined natural and mixed sites by more than 2 Still and movie images were collected and are being assessed as part of ongoing three to one. In Australia there are 17 sites, only two of which research into this aspect of the restorations.
Australian Forestry 2009 Vol. 72 No. 4 pp. 206–211 John Cokley 209
The island woodworker, who agreed to be interviewed but did Assembly, its administrative offices, Government House and the not agree to be named in this article, demonstrated his methods group of intact buildings and ruins dating from Norfolk Island’s of turning Norfolk Island pine into tables, bowls, platters, inlaid 19th century convict days. clocks and thermometers, and other items for sale in island souvenir shops. He showed the author slabs of pine that had been The shingling program, initiated in 1985 by heritage architect 4 subject to insect attack and rot, and others which had been treated Robert Varman , later of the University of Sydney (P. Anderson, with various chemicals, such as the insecticides Metalex and Borid ibid.), initially specified the endemic Norfolk Island pine, but (boric acid powder). At the island mill, I viewed milling equipment the service life of untreated natural timber was only five years. that dated to World War II and I discussed the logging and milling (Records from the early settlements suggested those settlers of the island pine with the owner, Mr Howard Christian. Locals also both preferred Norfolk Island pine timber and experienced on the island referred the author to Mr Christian as the man who the resultant decay and replacement issues.) The builders tried knew most about logging and building with Norfolk Island pine, applying a Wattyl coating called ‘Combat’ but service life and indeed his mill compound — largely open-air with several extended only to 9–10 years (P. Anderson, ibid.). Treatment of portable and ‘forest mills’ and planers scattered about — showed pine shingles at the island’s Tanalith plant — commonly used to evidence of long-term use. During the progressive dinner, an add copper chromium arsenate (CCA) preservative to building evening event during which tourists are driven for meal courses in timbers cut on the island — was rejected for the KAVHA site a small coach between three houses owned by long-time islander because the chemicals turn the timber green and there was a families, the author viewed houses in which Norfolk Island pine concern that run-off from the roofs would contaminate the furniture and interior fittings were displayed prominently and ground around the site (see Cookson 2008). In 1990, installation discussed extensively. Mr Christian was mentioned often during began of casuarina shingles supplied by Shingles Australia of the night as the person responsible for felling, milling and carting Mullumbimby in northern New South Wales. These shingles had the timbers of which the islanders were evidently very proud. an expected service life of 80 years. However, the builders found His role in the community appeared to be one of overseer of the the casuarina very hard to work (P. Anderson, ibid.): endemic timber resource, and the person whom young families It’s a very hard timber … We had been nailing the Norfolk pine approached to have timber cut from their family blocks and milled on by hand but we needed nail guns to put up the casuarina. to order into timber for their homes. The pine, as an interior and There’s a natural grain through it, too, so it’s tough and doesn’t furniture timber, appeared to be a key tourism driver on the island, split easily. They have to use a hydraulic splitter. and an object of substantial pride among long-term residents. At the Tanalith plant, I met the manager, Mr Bevon Nicolai. He said The imported shingles cost about $AUD200 per square metre, that the plant was a second-hand cold vacuum pressure model machine cut and landed on Norfolk Island (32 shingles in a (1500 kPa) installed in 1948, using the CCA product Tanalith E square metre). Installation and equipment costs reached another manufactured by Koppers. This was sufficient to deter the endemic $AUD200 per square metre (P. Anderson, ibid.). pine beetle, the pine hole borer, the wood roach and various The island’s Tree Preservation Act includes strict rules for species of fungi. logging local Norfolk Island pine. Only dead or dying trees may be taken and the shingle team has very specific requirements: Case studies: interviews excellent quality, clean butt logs with at least 20–30 feet (6 m) from ground to the first limb. Nevertheless, ‘Norfolk pine was Subjects chosen for interviews were builders, politicians, a the only suitable timber to make roofing shingles in the early museum director and a history-tourism worker. days’ (Evans, ibid.) and so it was initially the preferred material. Restoration project manager Mr G.E. ‘Puss’3 Anderson, his After recovery (felling and dragging out), each butt log is cut into son Kane, and Mick Williams were the team of builders who 18-inch (about 450-mm) lengths called ‘juncts’ (Evans, ibid.). concluded the restoration project in June 2008. Others had worked These juncts are stood end-on and a pattern board, 4½ inches on the project from time to time since restoration commenced (about 110 mm) wide, is laid across the middle of the junct to in 1985. The principal task had been removal of asbestos tiles mark out what craftsmen call a ‘billet’. From the billet (using and decayed pine shingles and progressive replacement with a froe — an iron blade on a wooden handle, sharpened on one alternative materials. The Administration had made a long- side and blunt on the other — with a wooden mallet) are cut standing budget allocation of $50 000 a year towards the work shingles ½ inch (about 10 mm) thick and 18 inches long. Since but it was recognised that the actual costs of re-roofing would only a third of each shingle is exposed on the roof (6 inches by be more than that (Jack, ibid.). Removal of in-situ asbestos tiles 4½ inches wide) about 30 000 shingles are required for the roof of required shipping the hazardous material from the island to a an average three-bedroom house. The pitch of the roof should be suitable deep land-fill site in Australia. The cost of shingles alone more than 37° to reduce water damage and rot and thus maximise for an average-sized building in KAVHA was up to $AU50 000, service life (Evans, ibid.). The limbs of the pine are tough and compounded by the cost of other building materials and labour. durable but they make timber adjacent to their knots unsuitable The most recent estimate to completely re-shingle one building for shingling, causing substantial wastage of otherwise millable was $AU200 000 (Jack, ibid.). The allocated budget of $50 000 timber (P. Anderson, ibid.). will now be redirected to improving signs and other historical interpretation within the site, home to the island’s Legislative
3 Most men on Norfolk Island are more commonly known by nicknames 4 http://www.geocities.com/Paris/LeftBank/6559/scc3.html
Australian Forestry 2009 Vol. 72 No. 4 pp. 206–211 210 Replacement of wooden shingle roofs on buildings at Norfolk Island
Hardiflex, a cellulose fibre-cement product5, has also been used without the use of CCA-based treatments, which are rejected on on some buildings at the site but according to the builders this did environmental grounds. The second choice of roofing material, not supply an authentic timber shingle appearance. casuarina, is expensive, hard to work and reliant on suppliers in Australia (P. Anderson, ibid.) but its service life extends to ‘Timber is extensively used at the landing place,’ said Puss almost a century. Anderson. ‘We want to use traditional materials as far as possible.’ His fellow builder Kane Anderson said: ‘It’s part of our identity. The supply of island pine is not so tightly regulated as to prevent Tourists and residents always ask: “What are the roof shingles islanders from securing wood for wood turning for the tourist trade made of?” and comment on how good they look. These are the and for traditional house building. Nor is on-island regulation a most-asked questions.’ Locals use significant pines on hilltops and factor in KAVHA’s ability to secure supplies of Norfolk Island outcrops as navigation points for fishing, and as Puss Anderson pine for roofing from areas planted elsewhere, such as Lord said: ‘it’s the first thing you see when you approach the island’. Howe Island, Hawaii and New Zealand. The only foreseeable extra expenses are the costs of transport by sea and possibly The view that casuarina or Hardiflex disrupts the interpretive quarantine facilities, but these are mitigated by the identified value of the KAVHA site is challenged by the Administration’s significant present costs in milling, importing and installing the Environment Minister Mrs Vicky Jack: ‘I don’t think the majority very hard casuarina wood. of tourists who visit the site even notice or consider the particular aspect of the type of shingles used on the various roofs within the Convenient and inexpensive CCA treatment facilities already exist KAVHA site. Their consideration is given to the overall view, the on Norfolk Island at the Tanalith plant near the central town of panorama of the site. As to how the local community thinks about Burnt Pine. This is the standard treatment for the wood logged the type of shingle used … in my 17 years living on Norfolk, the and milled on the island for house-building by traditional island type of shingles on roofs within KAVHA has never made it into families. Other treatments are identified in this article, such as the any conversation outside my work environment.’ low-toxicity and environmentally benign compounds copper or zinc naphthenate, and carbonisation by heat treatment. Museum director Brian Hubber agreed with this view. ‘When visitors come to Norfolk Island they know that the island has a It is likely that Norfolk Island pine could provide durable shingles strong convict heritage and they have a general idea about the if the timber is treated appropriately. It would be best if this history of the place. But most people wouldn’t even know what treatment could be undertaken on site, thereby reducing costs. In kind of timber is used on the roofs.’ However, he acknowledged addition, other endemic species on Norfolk Island may provide a that ‘purists’ would argue that using anything other than authentic sustainable supply of wood for shingles. Alternatively, as Norfolk materials would be the thin end of the wedge, and that this ‘chips Island pine growing on Lord Howe Island is genetically the away at the authenticity of the buildings’. The best solution in same as that on Norfolk Island, timber from that island could be his opinion would be a roofing material that would last, would be considered as an additional authentic supply. financially sustainable and would use materials that are available on the island. Conclusions and recommendations Discussion The research conducted for this article strongly suggests that the issue of a sustainable resource to supply authentic timber shingles The literature suggests that one of the most significant aspects for the KAVHA historic precinct is one of social and political will of the convict site at KAVHA is that the buildings were made by rather than availability of materials or cost. Documents reviewed hand using materials collected and worked from the vicinity. The and interviews conducted suggest that the available resource of most suitable and readily available roofing material was shingles Norfolk Island pine on the island is sufficient to supply KAVHA of Norfolk Island pine (Evans, ibid.), a fact that suggests this needs into the future even with the expensive and apparently material should be chosen for authenticity during contemporary wasteful crafting methods of the traditional roofing process — and restoration (KAVHA 2008). Only one building, the surgeons even with the existing exploitation for tourist souvenirs and quarters (Second Penal Colony), was originally fitted with traditional islander housing and the legislated need to conserve imported casuarina shingles (P. Anderson, ibid.). The case for the existing forests. A less wasteful method of logging, milling and use of the pine is strengthened by the island economy’s reliance on crafting the pine into shingles would improve the effective size tourism, much of which is driven by visits to the KAVHA historic of the timber resource. Beyond the island lie further resources of precinct (Evans, ibid.; K. Anderson, ibid.), the souvenir trade and Norfolk Island pine which might be acquired, milled, treated and the progressive dinners. The Administration has recognised this installed at less expense than the casuarina currently imported by reallocating annual funding of $AUD50 000 from re-roofing from Australia. The research also strongly suggests that treatments work to interpretation work around KAVHA (Jack, ibid.). are available for the pine that extend the service life of the timber without risking permanent damage to the environment or the There are however two arguments against the choice of Norfolk authenticity of the KAVHA precinct. Island pine for re-roofing. The first is that supplies of the wood are very tightly regulated by contemporary legislation, and The fundamental issue identified in this study is that the island secondly the timber is very vulnerable to attack by insects and rot economy depends on tourism driven by history visits, in turn encouraged by inclusion of the island on lists such as the UNESCO World Heritage list. The likelihood of the KAVHA precinct being added to the UNESCO World Heritage list in 2009 deliberations 5 http://www.jameshardie.com.au/NR/rdonlyres/9052665D-4614-4972-83A7- 9E1974B4CCD6/0/Hardiflex_Sheets_Certificate_of_Phys_Prop.pdf
Australian Forestry 2009 Vol. 72 No. 4 pp. 206–211 John Cokley 211 would be much enhanced by the use of authentic Norfolk Island Feist, W.C. and Hon, D.N.-S. (1984) Chemistry of Weathering and pine roofing material at the KAVHA site. Protection. American Chemical Society. http://www.fpl.fs.fed.us/ documnts/pdf1984/feist84a.pdf . Gosselink, R.J.A., Krosse, A.M.A., van der Putten, J.C., van der Kolk, Further research J.C., de Klerk-Engels, D. and van Dam, J.E.G. (2004) Wood preservation by low-temperature carbonisation. Industrial Crops I recommend that further research and testing be conducted and Products 19, 3–12. to reduce the waste incurred in the logging and milling of Gunduz, G., Korkut, S., Aydemir, D. and Bekar, I. (2007) The effects Norfolk Island pine for shingling, and into the sustainable and of heat treatment on physical and technological properties and environmentally benign treatment of the timber prior to use on surface roughness of Camiyani black pine (Pinus nigra Arn. the historic buildings. subsp. pallasiana var. pallasiana) wood. Bioresource Technology 99, 2275–2280. HFIA (Hawaii Forest Industry Association) (2000) Website, accessed Acknowledgements April 2008: http://www.hawaii-forest.org. Horton, C.T., Syme, J.H., Duke, C.R. and Straka, T.J. (1994) Distribution Research assistance and advice by Dr Pauline Heinrich is and marketing of southern yellow pine shakes: potential for gratefully acknowledged. Dr Heinrich is a past recipient of the expansion. Forest Products Journal 44(11–12), 21–26. Institute of Foresters of Australia’s Maxwell Ralph Jacobs Fund ITAD (Industrial, Technology and Design Teachers’ Association of award. Participation as interviewees and fact-checkers by Puss Queensland) (2008) Norfolk pine. Accessed May 2008: www.intad. Anderson, Kane Anderson, Vicky Jack, Gaye Evans, Arthur asn.au/materials/wd_norfolk.asp. Evans, Bevon Nocolai, Howard Christian and Brian Hubber is also Kalnis, M.A. and Detroy, B.F. (1984) Effect of wood preservative gratefully acknowledged, while noting that any remaining errors treatment of beehives on honey bees and hive products. Journal and opinions are the responsibility of the author. I am especially Agriculture, Food and Chemistry 32, 1178–1180. indebted to Gaye Evans, Deputy Clerk of the Norfolk Island KAVHA (Kingston and Arthur’s Vale Historic Area Management Assembly, who arranged introductions and the key interviews. Plan) (2008) Final Draft Attorney General’s Department, accessed May 2008: http://www.ag.gov.au/www/agd/agd.nsf/ Page/TerritoriesofAustralia_NorfolkIsland_DraftKingston References andArthursValeAreaManagementPlan. Loferski, J. (2001) Technologies for wood preservation in historic Australian Department of Environment, Water, Heritage and the Arts preservation. Archives and Museum Informatics 13, 273–290. (2008) Website accessed April 2008: http://www.environment.gov. Lennon, J. (2005) ‘Splendid Spars’: A Norfolk Island Forest History. au/heritage/places/world/nominations/convict-sites/index.html. Australian and New Zealand Environmental History Network Benson, M.L. (1985). Forestry Working Plan, Norfolk Island. Forestry website, accessed September 2009: http://fennerschool-associated. and Timber Bureau, Canberra, cited in Norfolk Island National Park anu.edu.au/environhist/links/publications/anzfh/anzfh2lennon. and Norfolk Island Botanic Garden Management Plan 2008-2018, pdf. Australian Government Director of National Parks, Department of the MacLellan, D.E. and Wolson, G.E. (2004) The National Home Environment website accessed April 2008, http://www.environment. Maintenance Manual. MacLellan Wolfson Associates, http://www. gov.au/parks/publications/norfolk/pubs/management-plan.pdf housefixit.com/images/NHMMCvr_TOC.pdf . [The Forestry and Timber Bureau transferred many functions to Meynink, R. (2006) Feasibility of plantations substituting for timber CSIRO Division of Forest Research in 1975]. currently harvested from Melbourne’s water catchments. Report Berg, B. (2001) Qualitative Research Methods for the Social Sciences. for the Department of Sustainability and Environment, Victoria. Allyn and Bacon, Boston. Accessed September 2009: http://www.dpi.vic.gov.au/CA256F31 Cokley, K. (1995) Preserving our natural heritage: timber. PhD thesis, 0024B628/0/08EEA2DA9366F70DCA25735600260CB1/$File/ Department of Botany, University of Queensland, Brisbane. Timber+Substitution+Part+1.pdf . Accessed April 2009: http://espace.library.uq.edu.au/view/ Moorehead, A. (1966) The Fatal Impact. Hamish Hamilton, London, UQ:9514. p. 143. Cookson, L.J. (2008) Influence of CCA-treated pine structures on the Niemiec, S.S. and Brown, T.D. (1993) Care and Maintenance of arsenic content of soils in preschool playgrounds near Melbourne, Wood Shingle and Shake Roofs. Extension Service, Oregon State Australia. Forest Products Journal 58(11), 94–99. University. Dickey, P. (2003) Guidelines for Selecting Wood Preservatives. The San Nolan, G. (2004) Design consequences of sustainable building and Francisco Department of the Environment. forestry in Australia. Eighth World Conference on Timber De Groot, R.C. (1994) Treatability of western softwood and red alder Engineering. 14–17 June 2004, Lahti, Finland, pp. 95–100. shakes. Forest Products Journal 33, 34–44. Taylor, A.M., Gartner, B.L. and Morrell, J.J. (2002) Heartwood formation De Groot, R. and Nesenson, P. (1995) Performance of preservative-treated and natural durability: a review. Wood and Fibre Science 34, wood shingles and shakes. In: Science and Technology: The Basis 587–611. for Improved Roofing. Proceedings of the 11th Conference on Roof Salaita, G.N., Ma, F.M.S., Parker, T.C. and Hoflund, G.B. (2008) Technology: National Roofing Contractors Association, 11–22 Weathering properties of treated southern yellow pine wood September 1995, Gaithersberg, Md, pp. 22–30. examined by x-ray photoelectron spectroscopy and physical De Groot, R.C. and Stroukoff, M. (1988) Efficacy of Alternative characterization. Applied Surface Science 254, 3925–3934. Preservatives Used in Dip Treatments for Wood Boxes. Research Termiz, A., Alfredson, G., Eikenes, M. and Terziev, N. (2008). Decay Paper FPL-RP-481. Forest Product Laboratory, Madison, United resistance of wood treated with boric acid and tall oil derivates. States Department of Agriculture. Bioresource Technology 99, 2102–2106. Earl, C.J. (ed.) (1998) Araucaria heterophylla. Gymnosperm Database UNESCO (1972), Convention Concerning the Protection of the World Taxon Descriptions. Accessed April 2008: www.conifers.org. Cultural and Natural Heritage. Adopted by the General Conference Evans, P. (2003) Emerging technologies in wood protection. Forest at its seventeenth session, Paris, 16 November 1972. Accessed 2008: Products Journal 53(10), 14–23. http://whc.unesco.org/archive/convention-en.pdf .
Australian Forestry 2009 Vol. 72 No. 4 pp. 206–211 212 Obituary
Obituary
Eric Peter Bachelard
1 March 1931 – 14 September 2009
Emeritus Professor of Forestry, Australian National University
Eric completed an Associate Diploma at the Creswick School of This was a period of great satisfaction to him, as he related recently Forestry in 1952, leading to employment with the then Victorian during interviews for the Emeritus Faculty Oral History Project1. Forests Commission. As an outstanding member of his class he The profession of forestry was undergoing great change during was selected to undertake a degree in forestry at the University this period. Eric recognised and was involved with the beginnings of Melbourne, which he completed in 1958. Not long after he of a marked transition in forestry education and research, a trend went to Yale University where he completed a Master of Forestry now ably carried forward by his successors. degree. After a short stint back in Australia he returned to Yale to complete a PhD in 1962. A year later he was offered a post at Eric was a staunch member of the Institute of Foresters of Harvard University where he remained for a year teaching biology, Australia, of which he was a Fellow. He was editor of Australian but rather than pursue an academic career in the United States he Forestry for five years during the 1970s. He always gave freely returned to Australia to work as a member of a research group of his time to his colleagues with editorial advice on scientific within the Forests Commission. He was subsequently appointed publications arising from their work. He retained his professional Chief Silvicultural Research Officer. I first met and became a interests in the institute throughout his retirement and was friend and colleague of Eric in Melbourne during the period chairman of its Publications Committee until quite recently. between completing his PhD at Yale and returning to the US to The facts detailing a career, like those above, do not tell much work at Harvard. We were good friends for some 47 years. about the person involved and their journey through this career. In 1965 the Australian National University replaced the Australian Eric mentioned his ‘personal skills’ in his recent oral interviews, Forestry School at Yarralumla with a Department of Forestry and these skills are widely recognised and lauded amongst the within the Faculty of Science. Eric was one of the initial group of alumni of the Department of Forestry. His former students now teaching staff assembled by the founding Professor and Head of occupy many responsible positions both here and abroad, bringing Department, Derek Ovington. He later became a senior lecturer. distinction to the university he served. He was greatly pleased By 1978 the department had increased in size and Eric was when one of our own distinguished graduates, Peter Kanowski, appointed as a second professor. The then head of department, succeeded him as professor in 1996. Professor David Griffin, and Eric came to an arrangement that Eric His many graduate students developed particularly close personal would undertake the main role in relation to Australian forestry as well as professional bonds with him and still hold him in great whilst David focused on international forestry. Eric coordinated affection and esteem. This was especially evident during the the administration of post-graduate students. troubled times of his recent illness. His research interests ranged Eric remained with the Department of Forestry for the rest of his widely but gradually tended to concentrate on water relations of professional career. From 1980 to 1983 he was Dean of Science. forest trees, particularly eucalypts — an interest in what plant From 1986 to 1989 he was chairman of the Board of The Faculties, could grow and survive where. His graduate students were The Faculties being responsible for the undergraduate teaching frequently mature aged with a background of field research and and associated staff and post-graduate research activities within investigation. Eric was happy to adapt his interests to further their ANU. He served as Acting Deputy Vice-Chancellor on occasion. chosen field of research, and this led to much productive scientific From 1992 he served as head of the Department of Forestry until endeavour for both student and teacher. his retirement from the university in 1996. Professor Eric Peter Bachelard was a distinguished member of Eric saw the 1980s as the culmination of his career as both scientist the profession of forestry in its broadest sense. He was a loyal and academic manager. It was a period when his personal and colleague and mentor to many. We pay tribute to a life lived well, professional skills came together to considerable effect — for the with integrity and great friendship. Department of Forestry, for the Faculty of Science and for the Ken Shepherd university more generally through his inputs within the Chancery Canberra and on a number of university committees. One wider connection he valued was maintaining significant departmental contact with the CSIRO Division of Forest Research at Yarralumla. During the 1980s he served for some years as the university representative on the CSIRO Forest Research Advisory Committee, part of the 1 A synopsis of this oral history is accessible at www.anu.edu.au/emeritus, oral time as chairman. history tab. Australian Forestry 2009 Vol. 72 No. 4 p. 212 Indexes to Volume 72 213
Indexes to Volume 72
Author index Adams, Ron 147 Herbohn, John 184 Hickey, John E. 157 Bartlett, Tony 52 Barton, C. 172 Institute of Foresters of Australia 117 Bishop, Ian D. 157 Joe, Bill 3 Boyton, Steve 3 Johnson, I.G. 172 Brand, Jonathan E. 149 Bryant, Chris 71 Kathuria, A. 172 Kimberley, M.O. 32 Carr, Jenny 29 Kube, Peter 25 Chew, Narelle 29 Cokley, John 206 McCaw, Lachie 51 Cookson, Laurie J. 29 Michael, M. 12 Creffield, Jim W. 29 Morris, A.R. 99 Crous, J.W. 99 Occhipinti, Sharon 39 Dickson, Ross 3 Parekh, J. 172 Ellis, J.C. 32 Ryan, Michael 59, 113 Elms, Stephen 25 Emtage, Nick 184 Sands, Roger 54 Erskine, Wayne D. 61 Sasse, Jo 25 Scholes, M.C. 99 Ferguson, Ian 195 Searle, Suzette 71 Fife, D.N. 12 Shepherd, Ken 212 Ford, Rebecca M. 157 Fox, Julian C. 39 Thomas, Dane 3 Thompson, David 184 Goodwin, Adrian N. 87 Turner, Brian 113 Grant, C.D. 80 Grigg, A.H. 80 Vanclay, Jerry 49 Hamilton, Fiona 39 Wilkinson, Graham 1 Harrison, Steve 184 Williams, Kathryn J.H. 157 Henson, Michael 3
Australian Forestry 2008 Vol. 71 No. 4 pp. 213–214 214 Indexes to Volume 72
Subject index Model, farm forestry, financial 184 Modelling tree taper 87 Acacia acuminata 149 Ash forests: The ANU Westoby Lecture 195 Norfolk Island, shingle roofs 206 Australian Farm Forestry Financial Model 184 Nutrient content of Pinus patula stems at various stem heights 99 Bachelard, Eric Peter: obituary 212 Bauxite mining and growth of rehabilitated jarrah forest 80 Obituary: Eric Peter Bachelard 212 Bushfires Royal Commission, submission to 117 Phosphorus and potassium fertiliser Cubic tree taper model 87 and Pinus patula stems 99 Pine aphid Essigella californica 25 Defoliation of Pinus radiata by the pine aphid Pinus Essigella californica 25 patula 99 radiata and the pine aphid Essigella californica 25 Editorials: Plantation-grown Eucalyptus dunnii 3 Forestry and wood products research and promotion 147 Plantations, Eucalyptus globulus 12 Transparent forestry 1 Pulpwood, a sampling system for scaling 32 Victorian fires: retrospective and prospective 59 Victorian fires: the Royal Commission and the future 113 Referees 2008 55 Essigella californica, pine aphid 25 Reviews Eucalypt forests, Tasmania, silviculture of 157 Forest Wars 49 Eucalyptus Plantation Eucalypts for High-Value Timber 54 dunnii 3The Wild Trees: What if the Last Wilderness globulus plantations 12 is Above our Heads? 51 marginata forest after bauxite mining 80 The Zealous Conservator: A Life of Charles Lane Poole 52 Failure of Eucalyptus globulus plantations 12 Roofs, shingle, on heritage buildings at Norfolk Island 206 Farm forestry, a financial model 184 Royal Commission, Victorian bushfires 113,117 Fertiliser effects on stem form, basic wood density and stem nutrients of Pinus patula 99 Sampling partly segregated pulpwood 32 Finance, farm forestry 184 Sand and gravel in the Letts Creek catchment, NSW 61 Fires in Victoria 59, 113 Sandalwood (Santalum spicatum) 149 Fires, forests and futures 195 Santalum spicatum 149 Forestry and wood products research and promotion 147 Scaling partly segregated pulpwood 32 Forestry degrees: students choice 71 Shingle roofs, Norfolk Island 206 Forestry profession: students choice 71 Silviculture, Tasmanian wet eucalypt forests 157 Form, various species, New South Wales 172 Social acceptability of silvicultural alternatives, Tasmania 157 Species trials in New South Wales 172 Genetic resistance in Pinus radiata to the pine aphid Student choice of a forestry degree 71 Essigella californica 25 Submission to Bushfires Royal Commission 117 Growth Survival, various species, New South Wales 172 Eucalyptus dunnii 3 response to thinning, burning and fertiliser in jarrah 80 Thinning, burning and fertilising jarrah 80 sandalwood (Santalum spicatum) 149 Transparent forestry — seeing is believing 1 various species, New South Wales 172 Tree hollow incidence in Victorian forests 39 Guest editorial — see editorials Tree taper, a cubic model 87
Hardwoods, lyctine susceptibility of 20 Victorian Heritage buildings, Norfolk Island 206 Bushfires Royal Commission 117 Hollows in trees in Victorian forests 39 fires 113 fires: retrospective and prospective 59 Irrigated pasture sites in South Australia, state forests and tree hollows 39 failure of plantations on 12 Westoby Lecture 195 Jarrah (Eucalyptus marginata) 80 Wheatbelt, Western Australia: sandalwood 149 Wood Letts Creek catchment, NSW: sources of sand and gravel 61 density of Pinus patula at various stem heights 99 Logging in the Letts Creek catchment 61 quality of Eucalyptus dunnii 3 Low-rainfall species trials in New South Wales 172 Lyctine susceptibility of selected hardwoods 20
Australian Forestry 2008 Vol. 72 No. 4 pp. 213–214 Australian Forestry
Notice to contributors Scope Australian Forestry provides a vehicle for formal publications in all areas of interest to the Institute of Foresters of Australia. The editors welcome original scientific and technical contributions, reviews, suitably argued opinions, short communications and reviews of books on topics related to forest research, forest management and forest policy in Australia and the South Pacific region. Papers relating to any aspect of knowledge or management of forest species and land use matters of interest to the region will be particularly valued. It is a condition of submission of original papers that they have not been accepted for publication elsewhere. The Editorial Panel evaluates the suitability of material for publication. All papers submitted to the journal will be reviewed by referees nominated by the panel. Criteria for the assessment of papers centre on the suitability of the topic for publication in Australian Forestry, technical and scientific quality of the work, clarity and conciseness of presentation, appropriateness of the conclusions on the basis of the evidence presented, and suitability of the material for printing. Manuscripts Contributions should be printed with double spacing on one side of the paper only. Three copies of each paper should be submitted. Following review, please submit the revised manuscript as one hard copy and an electronic file or files. The latter should be in Word or RTF, and should include tables and illustrations. The latter should also be supplied in their original files (e.g. in Excel) or as photographs. Low-resolution files are not acceptable. Title The title should be brief and where possible begin with a word suitable for indexing. Summary The summary should follow immediately after the author’s address and should concisely describe the aim and nature of the work, the principal results and conclusions. It should normally not exceed 5% of the length of the paper and should be suitable for reproduction by abstracting journals. If possible provide several keywords drawn from the CAB thesaurus; if this is not available it will be helpful to look at the keywords assigned to similar articles in Forestry Abstracts or in previous issues of this journal. Presentation Authors should follow the pattern of presentation used in this issue. Figures and tables should be numbered with arabic numerals. Photographs should be identified as figures, and numbered in a single sequence with line drawings. Photographic prints should be of high quality on gloss paper, with a full range of tones and good contrast. Tables should be constructed in Excel or in the ‘Table’ facility of Word — not tabbed, nor as pictures. The preferred placement of tables and figures in the text should be marked with for example ‘Figure 1 near here’. Actual tables and figures should be gathered at the end of the paper. Illustrations Graphical material may be submitted as computer-generated graphics or line drawings. Diagrams should be provided in postscript (ps) or encapsulated postscript (eps) format if possible; low resolution jpg files are not satisfactory. If diagrams are supplied only as hard copy, they should be camera ready, with labels applied at an appropriate size, printed by high quality (e.g. laser) printer. Where possible, figures will be printed within a single column width. References References should be cited in the text by author and year of publication. Three or more authors should be referred to by the senior author et al. The reference list should cite the full title of the publication. Please follow the style used in recent issues of this journal. Submissions Please forward your submission to: Executive Director, The Institute of Foresters of Australia, PO Box 7002, Yarralumla ACT 2600, Australia. Copyright protection For the protection of authors and the IFA, a simple form of copyright agreement has been developed and is available on request from the Executive Director. A copy of the agreement will be sent to the senior author or his or her agent with the typeset proofs; it should be signed by the senior author or agent on behalf of all co-authors and returned to the office of the IFA. Under the terms of this agreement the IFA will obtain an interest in copyright only if the article is published by the institute. Offprints A copy of the final paper in portable document format (pdf), and guidelines for the distribution of copies, will be provided to authors after publication of their paper. Correspondence on editorial matters should be addressed to the Managing Editor, Dr Colin Matheson, C/- The Institute of Foresters of Australia.