New Forests 2:17--30 (1988) © Kluwer Academic Publishers, Dordrecht -- Printed in the Netherlands
Climatology of Acacia mearnsii. 1. Characteristics of natural sites and exotic plantations TREVOR H. BOOTH & TOM JOVANOVIC Division of Forest Research, CS1RO, P.O. Box 4008, Canberra, A. C. T. 2600, Australia Received 29 April 1987; accepted 30 September 1987
Key words: natural distribution, species trials, climate, species selection
Application. This study demonstrates an explicit procedure for investigating a species' climatic adaptability. The method utilizes information from trial sites, as well as locations within the natural distribution. The method could be applied to any species and would help to identify sites where the species could be successfully grown.
Abstract. Climatic variation in the natural distribution of Acacia mearnsii in Australia was analysed. Data from sites where the species has been successfully grown in plantations and trials in Africa, Asia and South America were used to determine the species' climatic adaptability. Annual mean temperatures at these plantations and trial sites ranged from 13.9 to 23,9 °C, whilst annual mean precipitation ranged from 693 to 2263 nun. The coldest month minimum temperature at these sites was - 0 . 6 *C. This information, along with other details of climatic requirements presented in the paper, will assist in selecting sites for future trials at new locations.
Introduction The importance of climate in influencing tree species distributions has been known for many years (Stern and Roche 1974; Woodward 1987). Foresters have given considerable thought to climatic matching when introducing species to new areas. Initially, species which came from areas with similar climates were often selected for trials (see, for example, Prescott and Lane Poole 1947). As information on species' adaptability to conditions outside their natural distribution was gathered, these data have been used to assist species' selection (Golfari 1962, FAO 1979; Webb et al. 1984). Unfortunately, meteorological stations are often remote from forest sites. Educated guesses or crude interpolations have frequently had to be made to estimate climatic conditions at both natural and plantation sites. Modem methods of interpolation (e.g. Wahba and Wendelberger 1980) now allow these conditions to be estimated with great accuracy. Nix, Busby and Hutchinson have used interpolation methods for bio-
18 climatic analysis in the BIOCLIM program. This program can determine the climatic characteristics of a species' natural distribution. Hutchinson and Bischof (1983) have described the type of climatic interpolation methods used by the program, whilst Nix (1986) and Busby (1986) have described applications. Booth (1985) used the program to analyse the climatic characteristics of the natural distribution of Eucalyptus citriodora Hook. in Australia. The "climatic profile" was then used to identify sites in Africa with similar conditions. An analysis of a species' natural distribution can provide a first indication of that species' requirements, but some species are more climatically adaptable than others (Streets 1962; FAO 1979). Booth et al. (1988) studied the distribution of a further twelve eucalypt species using BIOCLIM and then showed how the results of trials in Africa could be used to analyse the climatic adaptability of each species. However, they did not use this information to identify new areas where the species might be tried. In this study, the complete procedure for the climatic analysis of a tree species is demonstrated for the first time. There are three main stages. -- Climatic characteristics of the natural distribution are analysed. -- Climatic conditions are determined at sites where the species is growing satisfactorily outside its natural distribution and the species' climatic requirements are then further defined using this information. -Climatically similar sites (homoclimes) are identified within the region being evaluated for new forests. This paper examines the first two stages of the procedure, whilst the second paper in the series (Booth 1988) demonstrates the third stage. Acacia mearnsii De Wild. (black wattle) was chosen to demonstrate the procedure. The Australian Centre for International Agricultural Research (ACIAR) is sponsoring a collaborative project between the CSIRO Division of Forest Research and the Chinese Academy of Forestry to test the potential of Acacia mearnsii and several related acacias in southern China. The aim is to establish new plantations to supply tannin, as well as industrial and domestic wood products. Early work on this project highlighted the need to identify the broad extent of areas which would be climatically suitable for A. mearnsii.
Methods
Climatic analysis of natural distribution Information on the natural distribution of A. mearnsii was obtained from specimens lodged at major herbaria located near the natural range of the species. This first data set consisted of 64 points spread across the distribu-
19
/ 4-
÷
(
140"
142 °
144 °
146"
148"
150"
152 °
154 °
Fig. 1. Natural distribution of A. mearnsii. (+) indicates 64 original sites, (×) indicates 223 additional sites.
tion of A. mearnsii (Fig. 1). Observations included a very brief locational description, together with latitude and longitude in degrees and minutes. Elevation, if not available, was estimated from 1 : 100 000 scale maps. The complete data set was analysed using the BIOCLIM program. For each location the program estimated monthly mean conditions of daily maximum temperature, daily minimum temperature and total precipitation. From these 36 values, 18 important climatic factors shown in Table 1 were calculated. These factors were selected on the basis of experience with climatic analysis of agricultural crops (Nix, pers. comm.) and forest species (Booth et al. 1987). Many of the factors are also commonly found in descriptions of tree species climatic requirements (Boland et al. 1984; Webb et al. 1984). The annual temperature range (factor no. 4) for an individual site is simply the hottest month maximum temperature (factor no. 3) minus
2O Table 1. Climatic variation within the natural distribution ofA. mearnsii.
64 sites
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18.
Annual mean temp. Coldest month minimum temp. Hottest month maximum temp. Annual temp. range Wettest quarter mean temp. Driest quarter mean temp. Annual mean precip. Wettest month mean precip. Driest month mean precip. Annual precip, range Wettest quarter mean precip. Driest quarter mean precip. Warmest quarter mean precip. Coldest quarter mean precip. Warmest quarter mean temp. Coldest quarter mean temp. Annual precip, range/(annual mean precip./12) Annual temp. range/annual mean temp.
287 sites
Min.
Max.
Min.
Max.
10.0 -0.7 20.6 17.6 6.2 6.3 440 49 15 15 139 51 51 103 14.1 5.7 0.3 1.1
17.5 6.8 28.1 27.7 20.3 19.4 1601 210 70 149 590 229 420 590 21.8 12.7 1.4 2.2
9.8 -2.6 20.6 16.2 6.2 4.2 440 49 15 15 139 51 51 75 14.1 3.9 0.3 1.1
17.5 7.1 28.6 27.7 20.5 19.4 1601 210 74 149 590 242 420 590 21.8 12.7 1.4 2.6
All temperatures measured in *C and precipitation in mm. temp. = temperature; precip. = precipitation.
the coldest m o n t h m i n i m u m temperature (factor no. 2). Similarly, the annual precipitation range (factor no. 10) for an individual site is the difference between the m e a n precipitation of the wettest and driest months (factors no. 8 and 9). Factors 17 and 18 are measures of the seasonal variation in precipitation and temperature. T h e highest and lowest value for each of the 18 factors was selected f r o m the 64 sites (see Table 1). F o r example, the highest annual m e a n temperature of 17.5 °C was estimated for Port Jackson (33°51'S, 151°16'E, 20 m), whilst the lowest estimate of 10.0 °C was for a site near H a m p s h i r e Hills (41"20'S, 145°55'E, 420 m). Different sites could also represent extremes within the highest and lowest categories. F o r instance, the location near H a m p s h i r e Hills, which had the lowest annual m e a n temperature, was not the location with the lowest coldest m o n t h m i n i m u m temperature. This dubious h o n o u r went to a site near the head of A r a l u e n C r e e k (35*39'S, 149"48 ', 750 m) with a temperature of - 0 . 7 °C. T h e values in Table 1 thus summarized total variation in climatic conditions within the natural distribution.
21 However, though the 64 sites were fairly evenly spread over the distribution of A. mearnsii, no area was intensively sampled. It would be interesting to know if the climatic profile would be much affected by more intensive sampling. Fortunately, a second data set was available to test this effect. Austin, Nicholls, and others (CSIRO Division of Water and Land Resources) have gathered data collected by various researchers recording the presence and absence of plant species at over 6000 sites in an area of south-eastern New South Wales. These data are being used to develop new methods for ecological analysis (e.g. Austin, Cunningham and Fleming 1984). The database was examined for records of A. mearnsii and 223 observations were found to add to the 64 herbarium records (see Fig. 1). The BIOCLIM analysis was rerun and the revised climatic profile is shown in Table 1.
Climatic analysis of plantations and successful trials A. mearnsii is a good subject for climatic analysis as it has been widely tested outside its area of natural occurrence. Sherry (1971) reviewed its growth in 50 countries. He did not attempt a synthesis of all the results in terms of climate, but his descriptions provided a very useful basis for this analysis. The objective here was to gather climatic data for successful trial and plantation sites in the same form as the BIOCLIM analysis described above. "Successful" sites were defined as those where the trees were growing well and climatic conditions were not markedly limiting growth. A more quantitative measure of success would have been desirable, but suitable data were not available. The following sections outline the analysis of major regions.
South Africa
A. mearnsii has been an important crop in South Africa with a maximum total plantation area of about 325 000 ha in the nineteen sixties. Since then, a fall in demand for tannin has led to a considerable reduction in area, but the species remains a significant crop. Location (latitude, longitude and elevation) and mean annual rainfall data for 33 plantation sites in South Africa were selected from SchSnau (1969) (see Fig. 2). These locations formed a representative sample of sites when A. mearnsii plantations were at their largest in South Africa. Mean monthly precipitation data for each site were calculated using proportional data from nearby meteorological stations. Mean monthly data for average daily maximum and minimum temperatures were calculated using interpolation surfaces, which had been calculated for Africa (Booth et al. in press). These surfaces were developed using data for 617 African meteorological stations, which
Fig. 2 Location of selected successful A. mearnsii trial sites and plantations.
23 came from the GLOCLIMEANMTH (global climate mean monthly) data base (McMahon 1986). The same 18 climatic parameters used by the BIOCLIM program were calculated for the 33 South African plantation sites, but only four factors of major importance are shown in Table 2. These factors, which were originally chosen on the basis of agricultural studies (Nix, pers. comm.), had proved to be particularly useful in the bioclimatic analysis of E. citriodora (Booth 1985). They represented mean temperature and precipitation conditions, as well as the stress of cold and drought. Table 2. Climatic variation within A. mearnsii trials and plantations.
Location(s)
Annual mean temp. (°C)
Coldest month min. temp. (°C)
Annual mean precip, (mm)
Driest q. precip. (ram)
33 South African sites
13.9/19.1
--0.6/8.5
693/1608
39/155
21 African sites (ex. S.A.)
16.7/23.9
Taquari, Brazil
19.7
10.1
1458
297
Bandoeng, Indo.
17.7
13.8
1799
173
Belgaum, India
24.0
13.9
1551
10
Nilgiri, India
14.3
5.1
1630
86
Palni, India
14.2
8.0
1672
159
Fukuoka/Kumanoto, Japan
16.1
2.2
1918
237
5.3/16.7
536/2263
0/235
Africa (excluding South Africa) Twenty-one locations mentioned by Sherry (1971) were chosen to represent successful A. mearnsii plantation and trial sites outside South Africa (see Fig. 2). Mean annual rainfall data were available for all these sites, so that monthly rainfall data could be estimated. Where necessary, temperature conditions were estimated using the temperature surfaces described above. A. mearnsii has been an important crop in the highlands of eastern Africa. Five sites in Zimbabwe (Mr Selinda, Melsetter, Chipinga, Inyanga and Umtali) represented plantations in the eastern border mountains. A. mearnsii was described by Streets (1962) as a %pecies of major importance" in Zimbabwe with over 27 000 ha planted, but Luyt, Mullin and Gwaze (1987) estimated the total area to be now about 14 000 ha. Eldoret, Turbo and a site
24 east of the Rift Valley were chosen as sites in Kenya. In 1965 the total wattle plantation area in Kenya was estimated to be about 52 000 ha (Sherry 1971). Voll (1980) has described the development of the 12000 ha wattle plantation in the Njombe district of Tanzania. Njombe, along with Iringa and the Usambaras region, was chosen to represent sites in Tanzania. Other African sites described by Sherry (1971) which were chosen included; Antsirabe (Madagascar), Antananarivo (Madagascar), Rabat (Morocco), Larache (Morocco), Gilo (Sudan), Gebel Mara (Sudan), Kigezi (Uganda) and north-east region (Zaire).
Rest of the world Estimating climatic conditions for A. mearnsii plantations outside Africa was comparatively difficult as surfaces were not available to allow estimation of temperature conditions. However, a small number of sites were chosen to represent major plantation areas. Brazil has one of the most important A. mearnsii plantation areas outside Africa. It is centred in the Rio Grande do Sul State and Sherry (1971) estimated there were around 60 000 ha in this region. Taquari is in this area and its climate is summarized in Table 2. Black wattle has been successfully grown in plantations at high elevations in Indonesia. A site at an altitude of about 1500 m in an area south-east of Jakarta was chosen to represent these locations. Samraj and Chinnamani (1978) described plantations of about 16 000 ha in the Nilgiri Hills in southern India; the Palni Hills a little further south also have significant plantations. Sherry (1971) reported successful small-scale trials from the Belgaum area in Bombay State. Successful A. mearnsii trials have also been established in the warm south-western districts of Japan on the islands of Shikoku Kyushu and Honshu (e.g. Tadaki 1978). The main commercial plantations are in the Fukuoka/Kumanoto prefectures; their climatic conditions are given in Table 2. The data shown in Tables 1 and 2 indicate the limits of variation in climatic factors within the natural range and for many successful plantation or trial sites in southern Africa and elsewhere in the world. To provide some indication of climatic variation within and between these groups, histograms of annual mean temperature and precipitation have been plotted in Figs. 3 and 4. As the numbers of observations within the Australian, South African and rest of the world groups were so different, the data in the figures are shown in percentage terms, so observations for any one group total 100. It is thus easy to see which classes were the most common within any group and how these compare with the other two groups.
25 60
50
40
3O
X X
IJJ 20
10
×
/
×
/
×i
,,
x,
f
i <10
'
10-12 12-14
Pl
14-16 16-18 18-20
N
20-22 22-24 <24
MEAN ANNUALTEMPERATURE(°C) [ ] AUSTRALIAN SITES
[ ] SOUTHAFRICAN SITES
[ ] RESTOF THE WORLD SITES
Fig. 3. Variation in mean annual temperature for Acacia mearnsii sites.
Discussion
The climatic profiles for the natural distribution of A. mearnsii (Table 1) show that the range of climatic conditions is considerably greater than previously thought. For example, the data in Table 1 can be compared with the following figures presented by Boland et al. 1984): Coldest month minimum temperature Hottest month maximum temperature Annual mean precipitation
0--5 °C 25--28 °C 625-- 1000 mm
Not all the climatic factors shown in Table 1 are as important in influencing growth as these measures are. More work is needed to determine the relative importance of climatic factors in determining the suitability of individual sites. Nevertheless, our knowledge of climatic factors at individual sites is already assisting in the selection of provenances for trial. We shall describe this work in detail elsewhere, but one of the most interesting features of the enlarged data set, summarized here in Table 1, was the lowering of the
26 601
50
40
~3 3O
20
10
<6
6-8
8-10
10-12
1 2 - 1 4 14-16 16-18 18-20
>20
PRECIPITATION(mm/lO0~ [ ] AUSTRALIAN SITES
[ ] SOUTH AFRICAN SITES
[ ] REST OF THE WORLD SITES
Fig. 4. Variationin meanannualrainfallfor Acacia mearnsii sites.
coldest month minimum temperature from -0.7 to -2.6 °C. These cold locations would be worth sampling to test if the trees growing there were particularly frost resistant. However, injury from unseasonal frost and slow growth at low temperatures may still be problems when the species is introduced into new areas. Probably the largest source of error in the analysis was the estimation of climatic conditions at sites outside Australia. Data were available from only 617 meteorological stations to estimate temperature conditions throughout Africa. This was in comparison to 901 stations used by Nix, Busby and Hutchinson (pets. comm.) to develop temperature surfaces for Australia, an area approximately one quarter the size of Africa. The number of African locations from the GLOCLIMEANMTH database was insufficient to allow accurate estimates of precipitation to be made. This severely reduced the number of plantation and trial sites which could be included, as an estimate of mean annual rainfall was required, in addition to accurate locational information. If, through cooperation with researchers in Africa, more climatic
27 information could be collected, better relationships could be developed to estimate mean climatic conditions for sites anywhere on the continent. Despite these limitations, the analysis indicated the remarkable adaptability of A. mearnsii to a wide range of climatic conditions (see Tables 1 and 2). This versatility is well illustrated in Figs. 3 and 4, and is, perhaps, surprising as seed used in overseas plantations are thought to have come from a very limited part of the natural range (Boland, pers. comm.). The resulting trees might therefore be expected to have relatively limited adaptability. The data from plantations and successful trials presented here can be compared with previous efforts to define the climatic requirements of A. mearnsii. The two editions of "A Guide to Species Selection for Tropical and Sub-tropical Plantations" (Webb et al. 1980, Webb et al. 1984) were compiled using published data and expert opinion. They included the following information for A. mearnsii climatic requirements:
Annual mean temperature Coldest month minimum temperature Hottest month maximum temperature Annual mean precipitation
1st ed. 12--18 °C 2--8 °C 18--24 °C 900--1600mm
2nd ed. 12--20 °C 2--8 °C 18--24 °C 700--2000 mm
The changes between editions, increasing the upper limit for annual mean temperature and increasing the rainfall range, approach the ranges shown in Table 2. The limit of 12 °C for annual mean temperature is a little low in comparison to the lowest value (13.9 °C) in Table 2. The data in Table 2 also indicate that A. mearnsii can withstand lower (and much higher) coldest month minimum temperatures than the Species Selection Guides suggest. The detailed data presented here are complementary to the useful, but necessarily brief, summaries in the Guides. For example, Fig. 3 shows that 20 °C is a reasonable upper limit for annual mean temperature, as a high proportion of plantations is found below this level. But Fig. 3 also shows that a few successful plantations have been established at higher temperatures. If A. mearnsii plantations are being considered for similarly high temperatures, these successful sites can be identified and conditions compared with the potential sites. As the method described here identifies all the sites used to derive the climatic information it provides a means for individuals with experience in particular areas to contribute to the development of more accurate climatic profiles. They may suggest the removal of specific "successful" sites or suggest the addition of other sites. These revised data sets could be used to produce improved climatic profiles for publications such as Webb et al. (1984).
28 In addition to the overall limits shown in Table 2 it is also possible to compare the limits for South African plantations shown in the same table. The South African profile is of particular interest as it represents the largest area of A. mearnsii plantations in the world. Schrnau and Schultz (1984) summarized "optimum" conditions for healthy growth of A. mearnsii based on South African experience. They indicated a requirement for annual mean precipitation in the range 850--1200mm and annual mean temperature greater than 16"C. This corresponds to the main South African groups shown in Figs. 3 and 4. The values are somewhat different to the ranges shown in Table 2, as the aim here was to identify sites which would be considered successful in a worldwide context. It was not intended to describe only the good sites within the most important plantation region. Nevertheless, it is interesting that the "optimum" conditions identified by Schrnau and Schultz (1984) are intermediate between the generally cool dry natural distribution and the broadly hotter wetter conditions prevailing where plantations occur in other parts of the world (see Figs. 3 and 4). Climate is only one consideration in site selection. Other factors, such as socio-economic conditions, soil conditions and pest/disease risks must also be evaluated. Other researchers are developing methods to assess these factors (Davidson and Nuruzzaman Khan, in press; Hackett, in press). However, these methods require detailed information both to assess a species' requirements and to evaluate a potential site's suitability. The much simpler method described here can identify areas worthy of these more detailed analyses.
Acknowledgements This study was supported by the Australian Centre for International Agricultural Research (ACIAR). We thank Henry Nix (Centre for Resource and Environmental Studies, ANU, Canberra), Michael Hutchinson, June McMahon (CSIRO Division of Water and Land Resources) and John Busby (Bureau of Flora and Fauna) for the use of programs, map plotting routines and data bases. Doug Boland, Peter Martensz (CSIRO Division of Forest Research) and A. Mc Geary Brown (Tasmanian Herbarium) provided information from herbaria on the natural distribution of A. mearnsii. Mike Austin and his colleagues (CSIRO Division of Water and Land Resources) supplied information from their south-east New South Wales data set. A. P. G. Schrnau (Inst. Commercial Forest Res.) provided information on the location of black wattle plantations in South Africa. Maria Valesca Popp Borges (Tanac) supplied climatic information for Taquari, Brazil. We also thank Alan Brown, Doug Boland and John Turnbull (CSIRO Division of
29 Forest Research), A. P. G. Sch6nau and the anonymous referees for comments on the manuscript.
References Austin, M. P., Cunningham, R. B. and Fleming, P. M. 1984. New approaches to direct gradient analysis using environmental scalars and statistical curve-fitting procedures. Vegetatio 55: 11--27. Boland, D. J., Brooker, M. I. H., Chippendale, G. M., Hall, N., Hyland, B. P. M., Johnston, R. D., Kleinig, D. A. and Turner, J. D. 1984. Forest Trees of Australia. Nelson and CSIRO, Melbourne. Booth, T. H. 1985. A new method to assist species selection. Commonw. For. Rev. 64: 241-250. Booth, T. H. 1988. Climatology of Acacia mearnsii: 2. Homoclime analysis of potential trial sites in China. New Forests. Booth, T. H., Nix, H. A., Hutchinson, M. F. and Busby, J. R. 1987. Grid matching: a new method for homoclime analysis. Agric. For. Meteorol. 39:241--255. Booth, T. H., Nix, H. A., Hutchinson, M. F. and Jovanovic, T. 1988. Niche analysis and tree species introduction. For. Ecol. Manage. 23: 47--59. Busby, J. R. 1986. A bioclimatic analysis of Nothofagus cunninghamii (Hook.) Oerst. in southeastern Australia. Aust. J. Ecol. 11: 1--7. Davidson, J. and Nuruzzaman Khan. in press. Computerized matching of tree species to sites. F A O / U N D P Project BGD/79/017 Assistance to the Forestry Sector, Working Paper, FAO, Rome. FAO 1979. Eucalypts for Planting. FAO Forestry Series No. 11 FAO, Rome. Golfari, L. 1963. Climatic requirements of tropical and subtropical conifers. Unasylva, 17: 33--42. Hackett, C. in press. Matching Plants and Land. CSIRO Division of Water Resources Research, Divisional Report. CSIRO, Canberra. Hutchinson, M. H. and Bischof, R. J. 1983. A new method for estimating the spatial distribution of mean seasonal and annual rainfall applied to the Hunter Valley. Aust. Meteorol. Mag, 31: 179--184. Luyt, I. E., Mullin, L. J. and Gwaze, D. P. 1987. Black wattle (Acacia mearnsii De Wild.) in Zimbabwe. Proceedings of the A C I A R / N F T A workshop on Australian Acacias in Developing Countries, Gympie, Queensland, 4--7 August 1986. ACIAR, Canberra. McMahon, J. P, 1986. Climatic databases held in the Division. CSIRO Division of Water and Land Resources, Research Report 1983--85:60--63. Nix, H. A. 1986. A biogeographic analysis of Australian elapid snakes, pp. 4--15 In: Longmore, R (ed) Atlas of Elapid Snakes of Australia. Bureau of Flora and Fauna, Canberra. Prescott, J. A. and Lane Poole, C. E. 1947. The climatology of the introduction of pines of the Mediterranean environment to Australia. Royal Soc. South Aust. Trans., 71: 67--90. Samraj, P. and Chinnamani, S. 1978. Black wattle cultivation in Nilgiris. Indian Farming 28: 15--18. Sch6nau, A. P. G. 1969. A site evaluation study in black wattle (Acacia mearnsii De Wild). Annale Universiteit van Stellenbosch 44A, No 2. Schtnau, A. P. G. and Schulze, R. E. 1984. Climatic and altitudinal criteria for commercial afforestation with special reference to Natal. Proceedings of the IUFRO Symposium on site and productivity of fast growing plantations, Pretoria and Pietermaritzburg: 551--63.
30 Sherry, S. P. 1971. The Black Wattle (Acacia mearnsii De Wild.). University of Natal Press, Pietermaritzburg. Stern, K. and Roche, L. 1974. Genetics of Forest Ecosystems. Chapman and Hall, London. Streets, R. J. 1962. Exotic Forest Trees in the British Commonwealth. Clarendon Press, Oxford. Tadaki, Y. 1978. The primary productivity and stand density control in an Acacia mollissima stand. Bull. For. Exp. Sta., Meguro, Tokyo No. 216:99--125. Voll, S. P. 1980. Projects of the colonial era I. Tanzania (Tanganyika): Tanganyika wattle scheme, pp, 169--171. In: A Plough in Field Arable: Western agribusiness in third world agriculture. University Press of New England. Wahha, G. and Wendelberger, J. 1980. Some new mathematical methods for variational objective analysis using splines and cross validation. Monthly Weather Review 108:1122-1143. Webb, D. B., Wood, P. J. and Smith, J. P. 1980. A Guide to Species Selection for Tropical and Sub-tropical Plantations. (1 st edn.) Commonwealth Forestry Institute, Oxford. Webb, D. B., Wood, P. J., Smith, J. P. and Henman, G. S. 1984. A Guide to Species Selection for Tropical and Sub-tropical Plantations. (2nd edn.) Commonwealth Forestry Institute, Oxford. Woodward, F. I. 1987. Climate and Plant Distribution. Cambridge University Press, Cambridge.