Báo cáo khoa học: "Predicting the yield of Douglas fir from site factors on better quality sites in Scotland"

Chia sẻ: Nguyễn Minh Thắng | Ngày: | Loại File: PDF | Số trang:16

Thêm vào BST

Báo xấu

39
lượt xem 4
download

Download Vui lòng tải xuống để xem tài liệu đầy đủ

Tuyển tập các báo cáo nghiên cứu về lâm nghiệp được đăng trên tạp chí lâm nghiệp quốc tế đề tài:Predicting the yield of Douglas fir from site factors on better quality sites in Scotland...

Chủ đề:

Bình luận(0) Đăng nhập để gửi bình luận!

Lưu

Nội dung Text: Báo cáo khoa học: "Predicting the yield of Douglas fir from site factors on better quality sites in Scotland"

article Original Predicting the yield of Douglas fir from site factors on better quality sites in Scotland AL DC Macmillan J Dutch Tyler 1 Macaulay Land Use Research Institute, Craigiebuckler, Aberdeen AB9 2QJ; 2 Forestry Authority Northern Research Station, Roslin, Midlothian EH25 9SY, UK 14 June (Received 2 January 1994; accepted 1995) Summary — In Scotland, as a result of recent changes in agricultural policy and grant schemes, there is now greater potential for planting a wider range of more productive forestry species on better quality land. In order to permit accurate production forecasting and financial appraisals for any such afforestation, it is necessary to develop predictive yield models. This article describes the development of a multiple linear regression model for the prediction of General Yield Class (GYC) of Douglas fir using readily assessed, or derived, site factors. Climate surfaces developed by spatial analysis of weather data were used to predict temperature and rainfall for 87 sample sites to a resolution of 1 km Estimates . 2 of wind climate were derived from a regression model using geographic location, elevation and topo- graphic exposure. Multivariate analysis of these and other soil and topographic variables indicate that temperature and exposure are most important in determining the productivity of Douglas fir on better quality sites in Scotland. As crop age increases, GYC declines and the possible reasons for this effect are discussed. Other factors are also discussed, such as the genetic variability of Douglas fir, and problems associated with establishment and form. Douglas fir / productivity / yield models / site factors / climate Résumé — Prédire la production du douglas à partir de facteurs stationnels sur des terrains de meilleure qualité en Écosse. Suite aux récents changements de politique agricole et de schémas d’at- tribution des subventions, il existe actuellement en Écosse de nouvelles possibilités pour planter un éven- tail plus large d’espèces forestières plus productives sur des terres de meilleure qualité. Afin de pré- dire de façon précise les productions et les implications financières de tels reboisements, il est nécessaire de développer des modèles de prédiction des productions. Cet article présente le déve- loppement d’un modèle de régression multilinéaire de prédiction des classes générales de production du douglas en utilisant des facteurs stationnels mesurés directement. Des surfaces climatiques, obte- nues par une analyse spatiale des données climatiques, ont été utilisées pour prédire la température et la pluviométrie de 87 sites échantillons à une résolution du km Des estimations du vent ont été obte- . 2 nues en appliquant un modèle de régression linéaire utilisant la localisation géographique, l’altitude et l’exposition. Une analyse multivariée incorporant les 2 précédents aspects plus des variables décrivant
le sol et la topographie montre que la température et l’exposition sont les 2 principales variables expli- quant la productivité du douglas sur des terrains de meilleure qualité en Écosse. On discute ensuite la contribution d’autres facteurs, tels que la variabilité génétique du douglas et les problèmes liés à l’éta- blissement et à la forme. douglas / productivité / modèle de production / facteur stationnel / climat INTRODUCTION later importations has not been as good (Phillips, 1993). Britain’s climate is temperate oceanic, In the European Union, tree planting on agri- and wind is therefore an important factor cultural land is seen as a way to reduce tree growth (Pears, 1967; Grace, limiting agricultural production, diversify farm income 1977; Dixon and Grace, 1984), particularly and provide a range of environmental ben- in exposed situations (Worrell and Malcolm, efits. In the United Kingdom, special grants 1990a). Scotland’s position is at higher lat- to encourage afforestation are available itudes than the extent of Douglas fir on the under the Farm Woodland Premium American continent, and although the cli- Scheme and uptake by farmers has been mate is moderated by the Gulf Stream, high. Although timber production is an impor- mean temperatures are well below the broad tant objective, little is known about the poten- optimum of 20°C that has been recorded tial productivity of species other than Sitka for Douglas fir (Clearly and Waring, 1969). spruce for many agricultural regions of Scot- In addition, a greater proportion of the land. These considerations, and the require- annual rainfall in Scotland occurs during the ment for better strategic forecasts of wood summer months than in the Pacific Coast flows, have given rise to the need for site region (Wood, 1962). yield models for species suitable for better quality land. Douglas fir (Pseudotsuga Existing site yield models are limited in manziesii [F] Mirb) is a potentially high yield- their coverage of Douglas fir, although a ing species that presently provides an alter- model for England and Wales has been native to Sitka spruce for better quality sites, developed recently (Forestry Commission, and was chosen as the subject of this study. 1993). In Scotland, there has been one quantitative study limited to the Perthshire In its natural habitat, Douglas fir covers a region (Dixon, 1971), although general very wide geographic and climatic range guidelines have been produced for eastern from British Columbia to New Mexico. It was areas (Busby, 1974). In Dixon’s study, topex first introducted to Britain in 1826-1827, and score (which is the sum of the angles to the became more widely planted from the 1850s horizon at the 8 cardinal points of the com- onwards (MacDonald et al, 1957). Due to pass) was the single most significant factor the phenotypic variation observed within its affecting productivity, explaining 32-69% natural range (Peace, 1948), and the fact of the variation in yield. In a study in North that UK rainfall and temperature regimes Wales, elevation, soil type and texture, as are similar only to a very small part of the well as indices of topographic position and entire range of Douglas fir, the need for shape, were all significantly related to top attention to seed sources for importation height at 50 years (Page, 1970). was soon realised. Good stands were pro- duced from seed imported in the early 1920s The success of site yield studies aiming from the Lower Fraser Valley in British to elucidate the relationships between yield Colombia, but the form of stands from some and environment for Douglas fir have been
METHODS even within its natural range. Mon- variable, serud et al (1990) attributed part of the cause of poor correlations between site and General Yield Class (GYC) is conventionally used soil factors, and height growth on the wide to estimate siteproductivity for forest crops in the genetic variation of Douglas fir. Decourt et al United Kingdom and measures the mean annual growth rate of timber (m per hectare (ha ), -3 ) -1 (1979) had similar problems with poor cor- per year (yr over the rotation period. It is ), -1 relations in a study in the Massif Central in derived from the relationship between height France, and suggested that the absence of growth and volume and is estimated from the mycorihizal associations could also have mean top height and age of the stand (Edwards contributed. Hill et al (1948) had better suc- and Christie, 1981). cess correlating soils and site index within a Factors known to influence tree growth in Scot- single climatic region in Washington state. land were identified from previous studies and a An investigation of the respective contribu- review of the literature. Eighty-seven temporary sample plots of 0.04 ha were randomly located tions of genotype and environment to site on sites throughout Scotland where site and soil index variation by Monserud and Rehfeldt factors could be accurately assessed. The pro- (1990), again in Washington state, indicated cedure for the collection of field data and the that genotype (as assessed by 3-year derivation of climatic data are described later. (A seedling heights) was a third more impor- full list of all the variables assessed for each site with abbreviations is given in Appendix 1.) tant than the current environment in deter- mining the variation in dominant height in natural stands. Genetic variability is also Sampling evident in the United Kingdom. For example, an investigation of tree growth patterns within Forestry Commission permanent As the study focused on better quality land, sam- sample plots indicated that differences in pling targeted sites below 350 m in both state and private estate ownership. Pure stands between 20 growth rate were not attributable to site fac- and 60 years old were visited at the locations illus- tors (Christie, 1988). trated in figure 1. The lower age restriction avoids The aim of this study was to develop site problems associated with estimating productivity yield models which could predict the poten- accurately for younger stands from published GYC curves and the incomplete expression of site poten- tial productivity of Douglas fir at the stand, tial (Coile, 1952), while 60 years is generally the forest and regional level throughout Scot- maximum rotation length. Plots were randomly land. As end users differ in the information located within compartments, avoiding possible they have available, 2 regression models edge effects, small scale variations in topography were developed, 1 incorporating climatic or drainage and areas of windthrow. data developed using trend surface analy- sis and kriging (Matthews et al, 1995), and a second that employs data that can be Field data collection readily collected in the field. Principal com- ponent analysis (PCA) was used to assist For each site, soil drainage, site drainage, major interpretation of the ecological nature of the soil group and rooting depth were assessed from relationships between yield and site fac- a soil pit at the centre of a 0.04 ha plot. The soil drainage classification is based on profile colours, tors. The precision and accuracy of the position in the landscape and the permeability of Douglas fir models were tested with an underlying horizons. It consists of 5 categories: independent data set. These models aid excessive, free, imperfect, poor and very poor the assessment of the economic costs and (Soil Survey of Scotland, 1984). Site drainage benefits associated with planting Douglas consists of 3 categories: shedding, normal and fir. receiving, which were determined by subjective
assessment of the net moisture status of the site plots. For the purposes of analysis, aspect was and its topography. Topex score was used as an transformed using sine and cosine functions into objective measure of geomorphic shelter. It is north-south and east-west components, and grid assessed by summing the angle to the horizon reference was converted to easting and northing at the 8 cardinal points of the compass. Other by replacing the 100-km grid square letters with factors such as elevation, national grid reference, numbers. The precision of easting and northing is slope and aspect were also recorded for the 87 to the nearest 100 m.
in their requirements from the predictions. Climate data Models that predict productivity most accu- rately are often not readily applied in the field, relationships achieved to date for a site The best so a "best fit" model and a model employing yield study in Britain used regression equations to only field measurements will be developed. spatially and altitudinally extrapolate meteoro- logical station data (Worrell and Malcolm, 1990a). Initially, all the independent variables More recently, work by the Climate Change Group Appendix 1 were included in the listed in at the Macaulay Land Use Research Institute has analysis. Forwards stepwise multiple linear taken this approach further. The regional climate regression analysis was used to derive the in Scotland has been modelled to a kilometre grid models as this is one of the best procedures square resolution using a combination of trend surface analysis and kriging for the spatial inter- for deriving regression equations by Draper polation of meteorological station records and Smith (1981).Only variables that were (Matthews et al, 1995). These "climate surfaces" significant at the 5% level or better were are based on data of 30-year means of monthly included in the models. The effects of soil temperature records from 150 stations for the factors were investigated using dummy vari- period 1951-1980, and 1 500 rainfall stations for the period 1941-1970. The kilometre grid cell ables (see Digby et al, 1989). An "average" estimates for each site were extracted from these regression line is used to calculate the dis- surfaces, and adjusted to the specific elevation of placement from this line due to each soil each sample site using standard monthly lapse factor. Confidence intervals for predictions rates. were calculated, and the models validated There are a large number of climate indices using an independent data set of 10% of that can be derived from mean monthly records of the samples collected. temperature and rainfall, so consideration was restricted to those likely to promote or inhibit The mean and range of each variable used growth. The indices investigated were mean spring in the model development are given in table. I temperature (April to June), mean summer tem- The range indicates the intervals within which perature (July to September), mean winter tem- it is generally valid to apply the model. perature (December to February), mean annual accumulated temperature above 5.6°C, mean spring rainfall, mean summer rainfall and mean total annual rainfall. The overall mean annual tem- "Best fit" model perature was divided by mean rainfall to give a measure of the effectiveness of precipitation. Graphical analysis of the trends in individual Cotton "tatter" flags are an established method site variables with GYC did not reveal any for assessing wind climate in upland Britain, with relationships that could be considered non- the rate of attrition of the unhemmed flags depen- dant on mean wind speeds (Rutter, 1968; Jack linear for the range of data. The "best fit" and Savill, 1973). Differences in tatter rates multiple linear regression model was devel- between sites have been related to elevation and oped using all available site, soil and cli- geographic location (Worrell and Malcolm, 1990a) mate data. The resulting model explains and the Stability Project Group of the Forestry 45.5% of the variation in GYC, and its form Commission Northern Research Station have is presented in model 1 and table II. used these relationships to develop a regression model for the prediction of tatter. It is their esti- mates of tatter that are used in this analysis. model1 24.57 + 5.24 * SPRT + 0.04109 GYC = - * TOPEX - 0.1163 * AGE - 2.061 * WINT REGRESSION ANALYSIS None Adjustement for SITEDR (shedding): Adjustment for SITEDR (normal): End users vary in the information they have available for input to such models and differ
SPRT and TOPEX were most closely ature has already been accounted for in the correlated with yield, together explaining model, the effect of WINT may actually 29.9% of the variation in GYC; AGE and reflect a statistical relationship between GYC WINT were selected subsequently. The and another site factor not included in the slope (b) coefficient for mean spring tem- final model but which is correlated to WINT. perature is positive, reflecting higher pro- As could be expected, the effect of increas- ductivity of Douglas fir at lower elevations ing geomorphic shelter is to increase pro- and at more southerly latitudes. The effect of ductivity. age in the model is to increase productivity Tests of the effects of qualitative soil vari- either for younger crops, or crops that have ables in the model resulted in the addition of been planted more recently. This could be SITEDR. The 2 drainage categories to which due to a number of factors, such as the model can be applied are shedding sites increased nitrogen deposition or genetic and sites with normal subsurface through improvements, though advances in site . 1 drainage Model 1 predicts that GYC will amelioration techniques are most probable. be greater on sites with "normal" through- The correlation between WINT and GYC is . yr -1 ha 3 m drainage by 1.6 negative. This is unexpected but since SPRT and WINT are highly correlated, and In order to assess the precision of the the variation in GYC due to spring temper- models over a range of sites, the GYC and
respectively, and 18.2 n for the yr -1 ha 3 associated confidence interval (Cl) were predicted from the model for 3, quite differ- typical site. The 95% Cl for the mean GYC ent, hypothetical sites. Two of these are for the site ranged from ±0.7 for the typical extreme sites, and the third is more typical site to ±2.4 m for the high yielding yr -1 ha 3 (table III). The low yielding site is an older site. The range for a single new site was stand on a high, exposed site with low tem- greater and ranged from ±4.8 to peratures during the spring, and the high . yr -1 ha 3 m ±5.3 yielding site is the opposite: a younger stand at low elevation in the bottom of a sheltered valley. Validation Confidence intervals have been calcu- lated for 2 situations; first, the prediction of independent plots were chosen ran- Nine the mean GYC for all cases in the popula- domly from the data set prior to model devel- tion, and second, the estimate of a single opment to test the validity of model 1. One of new site. The intervals for a single new pre- these fell outside the 95% Cl for a single new diction are wider than for the mean as the prediction (fig 2), although overall, the differ- variation of individual variables about their ence between the observed and predicted means (ie residual mean squares) is GYC values was small (-0.2 mA ). yr -1 ha 3 included. The first case is of interest when single sample T test indicated this value was considering the average yield for large areas not significantly different from zero. of land with a particular combination of site factors, such as for regional assessments of productivity. The second case arises A "field" model when predicting GYC for single small blocks of land such as at replanting or prior to land acquisition. The regression model employing only site variables that can readily be assessed in The GYCs predicted for the low and high the field is given in model 2 and table IV. yielding sites are 14.4 and 22.5 m , yr -1 ha 3 proved difficult in practice to find sample sites that were "receiving", as such stands generally 1 It had inadequate survival or suffered windthrow. As there were only 2 "receiving" sites sampled, they were omitted from the data.
Topex and elevation explained 19.5% of the Adjustments for Major Soil Group (podzol): variation in GYC, and age increased the 2 R to 0.271. The addition of northing, and major soil group as a dummy variable, Again the effect of climate on GYC is evi- improved the R 0.413. 2to dent with the inclusion of topex and eleva- tion. The combination of elevation and nor- model 2 thing appears to replace the role of the temperature indices by incorporating both the geographic location and elevation aspects of temperature variation. The slope coefficients for topex and age indicate that for Major Soil Group (brown Adjustments the variables are acting in the same manner earth): None as described for model 1.
extreme sites and a typical site (table V). As with SITEDR, the soil types to which The predicted GYCs were 10.4, 23.6 and model 2 can be applied are restricted. There were not suffificent sites with gley soils for 18.7 m respectively. In 95% of the , yr -1 ha 3 analysis as the majority of sites were either cases, the true mean GYC value will lie brown earths or podzols. Model 2 predicts between ±0.9 and 2.3 m which is , yr -1 ha 3 GYC for brown earth sites, with an adjust- sufficiently precise for practical application to ment of +2.6 m being applied to yr -1 ha 3 large forest areas. The true value for a sin- the regression model for podzolic soils. gle site prediction will lie within a maximum range of ±5.1 to ±5.7 m which is , yr -1 ha 3 As for the "best fit" model, hypothetical too wide a range to provide any improve- site values were used to test the effective- ment over a local forester’s educated guess. ness of "field" model predictions for 2
Validation ents explaining a large proportion of the vari- ation in the data, otherwise interpretation is less straightforward and the purpose is The same independent data set was used somewhat defeated. An advantage of PCA for validation. Again one site fell outside the is the fact that each component is orthogonal, 95% Cl for single predictions (fig 3). and employs some part of all the variables. Although there is a difference between The principal components obtained from observed and predicted values of GYC of analysis were then correlated with GYC. -1.2 m the single sample t test is , yr -1 ha 3 The variables having the greatest effect on not significant (t value =-2.03). Models 1 8df GYC were then determined from signifi- and 2 do not predict accurately the high levels and the standard of the yield class observed for 1 site (shown as &squf; errors cance regression coefficients. The value and sign in figs 2 and 3). This site was located on a of the weights (or loads) of the variables in moderate slope with a very good subsur- each component were used to interpret pro- face water supply. cesses or relationships between variables. PRINCIPAL COMPONENT ANALYSIS Results Principal component analysis (PCA) is a data The fourth principal component 4 (PC[4]) reduction technique which uses weighted was the component most highly correlated linear combinations of each of the original with GYC (table VI). The load values indicate variables to form a new set of independent that it is predominantly an age effect. The variables. The first component will be ori- correlation coefficient is positive, reflecting ented to explain as much of the variation as a decrease in GYC as age or planting year possible in the data by minimising the resid- ual sum of squares, as will the second, and increases. This effect is the same as that so on (Digby et al, 1989). The technique is demonstrated in the multiple linear regres- most effective when there are strong gradi- sion analysis. The load values of PC[2]
PC[1] has been included in table VI trend of decreasing rainfall with describe a because it describes the negative relation- progression towards the northeast of Scot- a ship between elevation and temperature land. This acts in a negative manner, reflect- which features in the regression models, ing a decrease in GYC on more northerly although, contrary to the results of models 1 and easterly sites a a result of lower annual and 2, it is not significantly correlated with rainfalls. The third component is also cor- GYC. It includes a topographic trend with related with GYC, probably due to the high sites at lower elevations, which tend to occur load value for topex. The effect on GYC is in areas with flatter terrain and lower topex the same as in the previous models as both scores. This may help to explain the appar- the load and correlation are negative, giving ently ambiguous load values for slope in a net positive effect of increasing geomor- phic shelter on GYC. PC[8].
DISCUSSION extreme sites. These intervals are reason- able and should provide adequate predic- tions for strategic wood flow forecasts. The From the results of the regression analysis, Cls for predictions made for individual sites it is evident that both temperature and topo- lie between ±4.8 and ±5.3 m , yr -1 ha 3 graphic exposure are 2 of the principal influ- which are probably too wide to be of use as ences determining the productivity of Dou- the entire range of observed GYC was 13.6 glas fir on better quality sites in Scotland. . yr -1 ha 3 m The combination of elevation This concurs with site yield studies on Dou- and northing appears to replace the role of glas fir conducted over smaller areas for the temperature indices by incorporating parts of Britain (Page, 1970; Dixon, 1971). both the geographic location and elevation When climatic data are not available, ele- aspects of temperature variation. The 95% vation performs a similar function to that of confidence intervals for mean predictions temperature without a major loss in predic- are wider than those for the "best fit" model, tive power in model 2. The selection of mean reflecting a loss of precision of ±0.2 m 3 spring temperature over other temperature yr -1 ha for estimates made on a regional indices is not surprising since spring is the scale, and by ±0.3 to 0.6 m for sin- yr -1 ha 3 main period of height extension. gle site estimates. Regression model 1 explained 45.5% of The effect of age on GYC is consistently the variation in GYC. For predictions over negative, so either younger crops, or more large areas such as might be done for recent plantings, are higher yielding. It is regional wood flow, forecasting the 95% confidence intervals for predictions of mean not possible to determine from this study GYC vary between ±0.7 m for yr -1 ha 3 whether the effect arises from the crop age average sites to ±2.4 m for more yr -1 ha 3 or the time at which the crop was planted.
deficits have been demonstrated in the The former case implicates the form of the same area for Sitka spruce (Jarvis and GYC curves, and it should be possible to Mullins, 1987), the rainfall indices were not investigate this further through stem analysis significantly related to GYC in the multiple of individual tree growth rates. Planting date linear regression analysis in this study. Site is, however, a more likely cause since drainage was significant in regression model improvements in site amelioration techniques 1, and provides a crude reflection of soil and seed provenance are likely to have moisture supply. The influence of topex and raised the productivity of Douglas fir con- temperature or elevation, indicated by the siderably over the past 40 years. Environ- regression analyses, are not apparent in mental pollution could be both an age and the PCA result. The effect of topex appeared planting year effect because the deposition to be confounded by gross regional differ- could change continuously both during crop ences in topography. Macmillan (1991) also rotation periods, and from one rotation to found no single factor to have an overrid- the next. The same effect has been reported ing influence on the yield of Sitka spruce on in recent years for Sitka spruce (Worrell and better quality sites in Scotland. Malcolm, 1990b; Macmillan, 1991) and other species (Forestry Commission, 1993). A large proportion of the variation in GYC remained unexplained. Genetic differences Model 2 predicts that on sites with podzol due to provenance and seed origin were soils, the average GYC will be higher than identified from the literature as likely causes. brown earths by ±2.6 m This con- . yr -1 ha 3 According to Forestry Commission annual trasts with Dixon (1971), who proposed that records, there were several main North Douglas fir could tolerate exposure better American seed exporters who supplied seed fertile sites. This could be related to on more during the period 1943-1960. They were general differences in soil texture, as the largely based in northwest Washington best root penetration of Douglas fir occurs state, but, even within this region, there are on fine, well-drained, dry, podzolic brown extremes of climate and considerable vari- earths (Kupiec and Coutts, 1992). This ations in the growth of Douglas fir. agrees with Murray and Harrington (1990), who found that fertility did not appear to be Total rooting depth was not important for a factor limiting growth on former farmland the sites sampled in this study, despite a sites in western Washington state. Models 1 number of North American studies demon- and 2 cannot be applied to peaty or gley strating total effective soil depth to be an soils, or moisture-receiving sites. Very few important factor affecting the productivity of stands on such sites visited during this study Douglas fir through its influence on water were stocked to an acceptable level, as they and nutrient supply, root respiration and had either suffered high mortality or physical space and stability (Hill et al, 1948; windthrow. Lemmon, 1955; Steinbrenner, 1965). Related variables such as soil texture, den- The principal component analysis result sity and aeration can also be important. Sim- was not in complete agreement with that of ilarly, observations have been made relating the regression analysis. The component adverse rooting conditions to a decrease in describing age (PC[4]) was most highly cor- height increment and crown density in Britain related with GYC, although this could be (Day, 1963). to the fact that it acts GYC due on more The ability of a soil to maintain a mois- directly. The second component (PC[2]) described a limitation in moisture supply in ture supply to the roots during the summer months is important for high yields (Hill et the northeast of Scotland. Although the lim- al, 1948; Contreras and Peters, 1982; Mur- itation on growth imposed by moisture
ray and Harrington, 1990), although too course of the study. The crop age/planting much summer rain in Scotland could be the year effect has been demonstrated consis- cause of the deterioration in form from north- tently in recent studies and is an aspect of east to southwest in the Great Glen. If growth site yield studies in Britain that requires continues through summer when moisture investigation if the future application of the is available, new shoots would be vulnerable models is to be valid. Additional unac- to deformation by wind (Fletcher, personal counted variation in GYC could have arisen communication). This sensitivity to wind has in this study from the lack of a variable accu- been observed as dieback in new planta- rately reflecting plant available moisture. tions and as damage to leading shoots in Rainfall indices are not good measures of mature stands when tops extend above sur- moisture supply for forest conditions rounding canopies (Darrah et al, 1965). because major losses occur to interception, A general problem with site yield mod- particularly during the summer months els and GYC system is that no account is (Jarvis and Mullins, 1987). The synthesis taken of the quality of the timber, which can of rainfall, interception losses, potential evap- have important economic consequences in oration and soil water holding capacity on some instances. A number of stands were a regional scale may improve model pre- encountered during sampling which had dictions. above average height growth but tree form was poor. Typically, such trees were coarsely branched with waves or spirals evi- CONCLUSION dent in the trunks between nodes. Lines (1987) suggested that the summer rainfall, 1) Temperature, topex and crop age are the higher wind speeds or the long summer day main factors determining the productivity of lengths could be the cause of stem sinuos- Douglas fir on better quality sites in Scot- ity. In addition, basal sweep occurs fre- land and can be used to predict GYC for quently in Scottish stands, although it is not brown earths and podzols on sites below always associated with other defects. The 350 m in Scotland. root spread of Douglas fir in Britain is very limited during the first 5 years of growth 2) The level of precision of the predictions for (Kupiec and Coutts, 1992), and this pattern GYC from regression model 1 are adequate of initial allocation of biomass to the crown for strategic modelling of wood flow. In 95% at the expense of the root system could be of the cases, the true value for the mean a factor contributing to the basal sweep. GYC will lie within 0.7 m . yr -1 ha 3 Instability is evident in young stands in open 3) When estimates of temperature indices field situations where good soil fertility could are not available, elevation performs a sim- be promoting canopy development without ilar function to temperature indices with a corresponding development in the root sys- loss of precision of GYC estimates on the tem. Exposure to winds, poor planting tech- order of 0.2 m. yr -1 ha 3 niques and a root restricting layer could all 4) Neither the "best fit" model nor the "field" aggravate the problems. Similar problems model was sufficiently precise to be of prac- with form have occurred in British Columbia tical value in estimating GYC for an individ- with plantings on ex-arable fields, although ual site. the cause is not known (Nixon, personal communication). development of The a methodology for 5) Two principal requiring further available water or soil moisture estimating areas identified during the investigation deficit for trees that can be applied on a were
ACKNOWLEDGMENTS regional scale may improve model predic- tions. 6) The &dquo;age effect&dquo;, which consistently This project was funded by the Scottish Forestry appears in recent site yield studies in Britain, Trust. We would like to thank all those who requires specific investigation before the assisted with the many aspects of field work, be determined. especially J Davidson of the Forestry Authority. cause can
Thanks also to K Matthews of the Macaulay Land Jack WH, Savill PS (1973) The causes of tattering of flags under natural conditions. IntJ Biometerol 17, Use Research Institute for the provision of cli- 185-192 mate data, and C Quine of the Forestry Authority MP (1992) Spatial disposition and Kupiec LC, Coutts for the tatter data. extension of the structural root system of Douglas fir. For Ecol Manage 47, 111-125 Lemmon PE (1955) Factors affecting productivity of REFERENCES some lands in the Willamette Basin of Oregon for Douglas-fir timber. J Forestry 53, 323-330 Lines R (1987) Choice of seed origins for the main for- Busby RJN (1974) Forest site yield guide to upland species in Britain. Forestry Commission Bulletin est Britain. Forestry Commission Forest Record no 97, no 66, HMSO, London, UK HMSO, London, UK Macmillan DC (1991) Predicting the Yield Class of Sitka Christie JM (1988) Levels of production class of Dou- spruce on better quality land in Scotland. Forestry 64, glas fir. Scottish Forestry 42, 21-32 359-372 Cleary BD, Waring RH (1969) Temperature collection MacDonald J, Wood RF, Edwards MV, Aldhous JR (eds) of data and its analysis for the interpretation of plant (1957) Exotic forest trees in Great Britain. Forestry growth and distribution. Can J Bot 47, 167-173 Commission Bulletin no 30, HMSO, London, UK, Coile TS (1952) Soil and the growth of forests. Adv 167 p Agron 4, 329-398 Matthews KB, Aspinall RJ, Hudson G, Law ANR, Birnie Contreras CC, Peters NR (1982) Induces de sitio para RV (1995) Mapping climate change: an approach to pino oregon en la Provincia de Valdivia y sus rela- the mapping of regional temperature regimes under ciones con sitios para pino insigne. Proceedings of future climate scenarios. Agric For Meteorol (in press) a workshop on evaluating the productivity of forest Monserud RA, Moody U, Beuer DWA (1990) A soil-site sites, 1982, Valdivia, Chile (compiled by Schlatter, study for Inland Douglas fir. Can J For Res 20, 686- JE), 98-109 695 Darrah GV, Dodds JW, Penistan MJ (1965) Douglas fir Monserud RA, Rehdfeldt GE (1990) Genetic and envi- in Wessex. Forestry 38, 183-200 ronmental components of variation of site index in Day WR (1963) The development of Douglas fir plan- Inland Douglas fir. For Sci 36, 1-9 tations in relation to site conditions. Forestry Com- Murray MD, Harrington CA (1990) Yield comparison of mission Forest Record 49, HMSO, London, UK no 3 Douglas fir plantations on former farmland in west- Decourt N, Tacon F le, Nys C (1979) The influence of ern Washington. Western J Appl For5, 123-126 environmental factors on production of Douglas fir in Page G (1970) Quantitative site assessment; some prac- the north-east of the Massif central. Rev For Fr 31, tical applications in British forestry. Forestry 43, 45-56 20-27 Peace TR (1948) The variation of Douglas fir in its native Lane P (1989) Genstat 5: a second Digby P, Galwey N, habitat. Forestry 22, 45-61 Oxford University Press, Oxford, UK course. Pears NV (1967) Wind as a factor in mountain ecology: Dixon JR (1971) An attempt at site assessment for Dou- some data for the Cairngorm mountains. Scottish glas fir in Perthshire. Scottish Forestry 25, 26-33 Geographical Magazine 83, 118-124 Dixon M, Grace J (1984) Effect of wind on the transpi- Phillips MTT (1993) A short history of an early Douglas ration of young trees. Ann Bot 53, 811-819 fir seed import to Britain and its values. Scottish Smith H (1981) Applied regression analy- Draper NR, Forestry 47, 24-25 sis, 2nd ed. J Wiley and Sons, New York, NY, USA Ruffer N (1968) Tattering of flags at different sites in rela- Edwards MV, Christie JM (1981) Yield models for forest tion to wind and weather. Agric Meteorol 5, 163-181 management. Forestry Commission Booklet no 48, Steinbrenner EC (1965) The influence of individual soil Edinburgh, UK and physiographic factors on the site index of Dou- Forestry Commission (1993) Site classification and yield glas-fir in Western Washington. In: Forest-soil rela- prediction for lowland sites in England and Wales. tions (CT Youngberg, ed), Oregon State University Unpublished Mensuration Branch Report, Forestry Press, Corvallis, OR, USA Commission and Ministry of Agriculture, Food and Fisheries Joint Contract CSA 1563. Forestry Author- Wood RF (1962) Studies of North-west American forests ity Research Division, Farnham, Surrey, UK in relation to their silviculture in Great Britain. Forestry Commission Bulletin no 25, HMSO, London, UK Grace J (1977) Plant responses to wind. Academic Press, London, UK, 204 p Worrell R, Malcolm DC (1990a) Productivity of Sitka spruce in Northern Britain. 1. The effects of elevation Hill WW, Arnst A, Bond RM (1948) Method of correlating and climate. Forestry 63, 105-118 Douglas-fir to site quality. J Forestry 46, 835-841 Worrell R, Malcolm DC (1990b) Productivity of Sitka Jarvis NJ, Mullins CE (1987) Modelling the effects of spruce in Northern Britain. 2. Prediction from site drought on the growth of Sitka spruce. Forestry 60, factors. Forestry 63, 119-128 13-30