The dendrochronological potential of lime ( Tilia spp.) from trees at Hampton Court Palace, UK

  • Published on
    08-Dec-2016

  • View
    218

  • Download
    6

Transcript

  • This article was downloaded by: [The University Of Melbourne Libraries]On: 12 May 2013, At: 10:59Publisher: Taylor & FrancisInforma Ltd Registered in England and Wales Registered Number: 1072954 Registeredoffice: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

    Arboricultural Journal: TheInternational Journal of Urban ForestryPublication details, including instructions for authors andsubscription information:http://www.tandfonline.com/loi/tarb20

    The dendrochronological potential oflime (Tilia spp.) from trees at HamptonCourt Palace, UKAndy K. Moir a b & Suzanne A.G. Leroy ba Tree-Ring Services, Hungerford, Berkshire, UKb Institute for the Environment, Brunel University, Uxbridge,London, UKPublished online: 13 Apr 2013.

    To cite this article: Andy K. Moir & Suzanne A.G. Leroy (2013): The dendrochronologicalpotential of lime (Tilia spp.) from trees at Hampton Court Palace, UK, Arboricultural Journal: TheInternational Journal of Urban Forestry, DOI:10.1080/03071375.2013.783173

    To link to this article: http://dx.doi.org/10.1080/03071375.2013.783173

    PLEASE SCROLL DOWN FOR ARTICLE

    Full terms and conditions of use: http://www.tandfonline.com/page/terms-and-conditions

    This article may be used for research, teaching, and private study purposes. Anysubstantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,systematic supply, or distribution in any form to anyone is expressly forbidden.

    The publisher does not give any warranty express or implied or make any representationthat the contents will be complete or accurate or up to date. The accuracy of anyinstructions, formulae, and drug doses should be independently verified with primarysources. The publisher shall not be liable for any loss, actions, claims, proceedings,demand, or costs or damages whatsoever or howsoever caused arising directly orindirectly in connection with or arising out of the use of this material.

  • The dendrochronological potential of lime (Tilia spp.) from treesat Hampton Court Palace, UK

    Andy K. Moira,b* and Suzanne A.G. Leroyb

    aTree-Ring Services, Hungerford, Berkshire, UK; bInstitute for the Environment, Brunel University,Uxbridge, London, UK

    Common lime (Tilia europaea L.) and large-leaved lime (Tilia platyphyllos Scop.)are dendrochronologically and dendroclimatologically analysed for the first time. Limeis thought to be sensitive to climate change. Once a dominant species in Europe, it hasbeen in general decline from 3100 BC, but recently it has been found to be increasingthe northern limits of its range. Twenty-five trees from Hampton Court Palace (UK) arecross-matched to form a 138-year chronology spanning from AD 1866 to AD 2003.The relationships with climate were investigated using monthly instrumental records ofprecipitation and temperature from Kew between AD 1872 and AD 1997. The age ofthe lime trees was found to correlate well with girth (r 0.87). The annual resolutionof the chronology is robustly supported by regional cross-dating against established oakand yew chronologies. Summer precipitation (May, June and August) was shown to bea time-stable determinant of annual variation in radial growth. Problems of indistinctboundaries and missing rings, which become more prevalent in trees over 100 years ofage, may limit the dendrochronological potential of lime.

    Keywords: climate change; dendrochronology; dendroclimatology; lime trees; Tilia;Hampton Court

    Introduction

    In arboriculture and urban forestry, the determination of tree age is useful to identify the

    chronology of parks and gardens, to forecast future tree size and threats associated with

    increasing age and to identify individuals of particular conservation value. Dendroclima-

    tology, one of the sub-disciplines of dendrochronology, enables the identification of both

    tree age and relationships between ring width and climatic variables. In areas where a

    particular tree species is suitably responsive and long-lived and/or where past wood is

    recoverable, the discipline may enable climate records to be reconstructed at annual

    resolution, centuries before instrumental data are available. In the British Isles, the climatic

    relationships of few tree species other than oak (Kelly, Leuschner, Briffa, & Harris, 2002),

    Scots Pine (Briffa et al., 2001; Moir, Leroy, & Helama, 2011), yew (Moir, 1999) and elm

    (Brett, 1978) have been examined. All these genera have well-defined rings suitable for

    tree-ring analysis. Lime (Tilia spp.) has not been previously dendroclimatologically

    analysed. This is possibly because it is only rarely found in old buildings and it has a

    tendency to decay rapidly if damp. Furthermore, lime trees over 350400 years of age are

    typically hollow (Pigott, 1989).

    At the northern limits of its range in theLakeDistrict ofEngland, limehasbeenconsidered

    a relict species, due to its limited production of fertile seeds and pollen evidence of decline

    after 3100 year BC (Pigott &Huntley, 1980). The small-leaved lime (Tilia cordataMill.) and

    q 2013 Taylor & Francis and Aboricultural Association

    *Corresponding author. Email: akmoir@tree-ring.co.uk

    Arboricultural Journal: The International Journal of Urban Forestry, 2013

    http://dx.doi.org/10.1080/03071375.2013.783173

    Dow

    nloa

    ded

    by [T

    he U

    nivers

    ity O

    f Melb

    ourn

    e Libr

    aries

    ] at 1

    0:59 1

    2 May

    2013

  • large-leaved lime (Tilia platyphyllos Scop.) species are both native to the UK but are near the

    northern limits of their European ranges. Small-leaved lime is essentially a continental species

    known to require summer temperatures$208C at flowering for several consecutive days toproduce viable seed (Pigott & Huntley, 1981). It has had difficulty growing from seed in the

    UK under contemporary climatic conditions, but suitable conditions to permit fertilization

    may have been more prevalent during the medieval warm period (Pigott, 1989). Recently,

    Gray andGrist (2000) reported the natural regeneration of lime as far north as Perth, Scotland.

    Large-leaved lime is nationally a scarce tree (Newlands, 1999).Where both small- and large-

    leaved limes occur they can hybridise, resulting in common lime (Tilia europaeaL.),whichis also a widely planted ornamental tree.

    Limes (Tilia spp.) are one of the tallest broad-leaved trees inmost areas of Great Britain.

    They can achieve a height of 3540m, a diameter of 100300 cm and live up to 1000 years

    (Mayer, 1977). Multi-stemmed self-coppicing limes can live much longer (Pigott, 1993).

    Lime trees have played a major role as an architectural element in gardens in many

    European countries since the late seventeenth century. Owing to their aesthetic value, lime

    trees have become increasingly important in urban and open landscape in recent decades.

    The avenues of a double row of lime trees on both sides of the Long Water at Hampton

    Court Palace, UK, form the central feature of a great baroque patte doie layout (Figures 1

    and 2). This was originally commissioned by King Charles II soon after his restoration in

    AD1660. It is the finest surviving example of its kind inGreat Britain. The following history

    of the lime trees at Hampton Court is summarised from a report by Gough (2000). Adrian

    May, a royal gardener, purchased a consignment of 758 common limes fromHolland, which

    were planted in AD 1661. The average life expectancy of a European lime tree is 200250

    years, so the LongWater avenuewas in its prime at the beginning ofQueenVictorias reign.

    A policy of gapping up those trees in the avenue, which had perished, was largely

    unsuccessful due to competition with older trees. Many trees used in gaps were also a

    different species of lime, or planted in the wrong positions, which resulted in a gap-toothed

    canopy. The 1987 hurricane in Southern England affected the LongWater avenue badly. By

    2000, the original population of 544 trees had dwindled to 300, with only 7 original tree

    specimens remaining. The original trees were also a cause of problems as disease

    (particularly several types of bracket fungus) and old age had left them in a dangerous

    condition.

    In 2003, the lime trees of the royal palace park in Hampton Court were felled (Figure 2)

    to allow the replanting and restoration of the avenues, which presented a useful opportunity

    for dendroclimatological and dendrochronological analysis. Indeed, on one hand, lime

    should be considered if it has the potential to become a new proxy for estimating global

    change; on the other hand, the dating of old lime trees will contribute to the corpus of

    knowledge on gardens of a historically significant place. The aims of this research were (1)

    to establish the ages of the trees sampled, (2) to establish relationships between the radial

    growth and meteorological climate records, (3) to identify whether lime might be a useful

    species in dendrochronology and (4) to interpret the ages of the sampled trees into the

    cultural landscape of Hampton Court.

    Materials and methods

    Sampling and chronology building

    Full trunk sections were sawn by operators at Hampton Court, typically from 20 to 30 cm

    above the ground level. For practical reasons, V-section samples were cut across the widest

    diameter to provide two radii for measurement (Figure 3). Standard dendrochronological

    2 A.K. Moir and S.A.G. Leroy

    Dow

    nloa

    ded

    by [T

    he U

    nivers

    ity O

    f Melb

    ourn

    e Libr

    aries

    ] at 1

    0:59 1

    2 May

    2013

  • techniques were then utilised for sample preparation, measurement, cross-matching and

    dating (Stokes & Smiley, 1968). Cross-matches are reported using raw ring-width data and

    the standard Students t-value statistic. Those t-values in excess of 3.5 are accepted as

    significant where supported by visual comparison.

    Growth rates

    Tree-ring series commonly contain age trend, caused by the general reduction in the ring

    width as trees get progressively older and pass through the formative, mature and senescent

    phases of growth (White, 1998). For useful visual comparison between tree growth rates,

    cumulative plots were produced to help emphasise the underlying biological age growth

    trend (Figure 4).

    Figure 1. Map of Western Europe showing the location of Hampton Court.

    Arboricultural Journal: The International Journal of Urban Forestry 3

    Dow

    nloa

    ded

    by [T

    he U

    nivers

    ity O

    f Melb

    ourn

    e Libr

    aries

    ] at 1

    0:59 1

    2 May

    2013

  • Dendroclimatic analysis

    The series were standardised (a process to remove age trends) using ARSTAN software

    (Cook, Briffa, Shiyatov,&Mazepa, 1990), andwere detrended using a negative exponential

    curve or linear regression with power transformation (Cook & Peters, 1997) to reduce

    potential end-effect inflation of resultant indices. The chronology statistics generated from

    the standardised series are described in Table 1. Mean sensitivity is a measure of the mean

    relative change between adjacent ring widths (Fritts, 1976). Values over 0.30 are high and

    indicate that the tree-ring series are highly responsive to environmental factors, while low

    Figure 3. A V-section sample cut from a full section.

    Figure 2. A standing and felled lime tree along the Long Water at Hampton Court Palace, UK.

    4 A.K. Moir and S.A.G. Leroy

    Dow

    nloa

    ded

    by [T

    he U

    nivers

    ity O

    f Melb

    ourn

    e Libr

    aries

    ] at 1

    0:59 1

    2 May

    2013

  • values indicate weak inter-annual variance. The expressed population signal (EPS)

    (Wigley, Briffa, & Jones, 1984) measures the degree to which the chronology correlates (or

    agrees) with a theoretical population chronology. The value of EPS ranges from 0 to 1, with

    1 being the best possible value (the hypothetically perfect chronology).

    Growthclimate relationships were examined using correlation functions as a

    statistical model to compute coefficients between tree-ring chronologies and monthly

    climatic variables (Blasing, Solomon, & Duvick, 1984). These coefficients are univariate

    estimates of Pearsons product moment correlation. Correlation function analyses and

    moving interval correlation function analysis were carried out using DENDROCLIM2002

    software (Biondi & Waikul, 2004), which tests significance at the 0.05% level. A 14-

    month analysis period extending from September in the year before growth to October of

    the year of growth was selected. Residual tree-ring chronologies (which have proved to

    Figure 4. Cumulative plot tree rings.

    Table 1. General statistics of lime chronologies from the arstan standard chronology.

    File name MS AR1 R(bt) SNR EPS

    HPLIME 0.33 0.29 0.67 15.99 0.94

    Note: Common interval 19101990. MS, mean sensitivity; AR1, first-order autocorrelation; R(bt), betweenseries correlation; SNR, signal to noise ratio; EPS, expressed population signal.

    Arboricultural Journal: The International Journal of Urban Forestry 5

    Dow

    nloa

    ded

    by [T

    he U

    nivers

    ity O

    f Melb

    ourn

    e Libr

    aries

    ] at 1

    0:59 1

    2 May

    2013

  • yield more climatic information and minimise autocorrelation) were used with monthly

    maximum temperature, minimum temperature and precipitation as predictors. Monthly

    temperature and rainfall series for Kew (a meteorological station 9.5 km south west of

    Hampton Court) were used in this analysis (Wales-Smith, 1980). Temperature and rainfall

    can be intercorrelated causing an apparent negative association between temperature and

    ring width when using correlation to examine climategrowth relationships, as

    highlighted by Fritts (1976). To help resolve this problem, response function analysis

    (which transforms the predictor variables into uncorrelated principal components) was

    also carried out using DENDROCLIM2002. However, as response coefficients tend to be

    lower than correlation coefficients the results are only summarised.

    Results

    Chronology

    The results of the cross-matching between 25 samples against both oak and yew reference

    chronologies are described below. The trees sections were generally quite circular in form

    and showed no signs of hollowing. Pith was recovered in all cases. Twenty-five out of the

    30 samples (83%) were successfully measured and cross-matched. Nineteen series were

    from common lime and four from large-leaved lime. Two samples were labelled with the

    same number, and therefore, could only be established as Tilia spp. The 25 cross-matched

    together were used to form a chronology called HPLIME, which spans 18662003. The

    annual resolution of this tree-ring series is confirmed by cross-matching against both oak

    and yew existing reference chronologies (Table 2). The rings in years 1949/1950, 1964

    and 1985 were the narrowest rings and most commonly missing. Instances of missing rings

    were more frequent in older trees, i.e. after the first 80 years of growth. In five series where

    the rings to bark could not be reliably measured, it was calculated that a missing ring

    occurred on average once every 12 years, suggesting a 12% underestimation of tree age

    from ring counts in lime trees over 50 years of age.

    Growth rates and age

    The ages and girths of the 25 cross-matched trees (Figure 4), together with five ring

    counted trees, are plotted in Figure 5, and the following regression equation is calculated

    as follows:

    AGE 105:77 GIRTH in m2 116:72 Standard error 31:25:Prior to c. 1898, only common lime hybrids appear to have been planted. However,

    after c. 1898, both common limes and large-leaved limes are shown to have been planted.

    The three oldest trees used in this equation could not be reliably measured, but were ring

    counted to be 312, 269 and 250 years of age. The oldest tree is therefore estimated to have

    been growing since at least 1691. However, assuming that it was one of the original limes

    planted in 1661, this suggests an underestimation by ring counting of around 10%, adding

    further evidence of a 1012% underestimation of age of lime trees from ring counts. Plots

    of cumulative ring width show no common differences in the radial growth rates of the two

    species of lime in this study. The mean formative growth rate is 3.84mmyear21, and

    the transition between formative and mature growth occurs after 50 years of growth

    (Figure 4). A mature growth rate of 3.25mmyear21 is applicable to trees between 50 and

    100 years of age. These results confirm that very few of the trees originally planted in 1661

    survived.

    6 A.K. Moir and S.A.G. Leroy

    Dow

    nloa

    ded

    by [T

    he U

    nivers

    ity O

    f Melb

    ourn

    e Libr

    aries

    ] at 1

    0:59 1

    2 May

    2013

  • Growthclimate relationships

    The statistical parameters of the standardised chronologies developed are shown in

    Table 2. Mean sensitivity, the relative change in ring widths from 1 year to the next (high-

    frequency signal), is 0.33, which is high in comparison to values for oak, Scots Pine and

    yew in the UK. First-order autocorrelation, a measure of the influence of the previous

    years growth on the current year (Fritts, 1976), is low (0.29) indicating little persistence

    from one years growth to the next. The EPS value is 0.94, which is above a 0.85 value

    suggested by Wigley et al. (1984) as reasonably strong and suitable for climatic studies.

    Rainfall in the summer months of May, June and August was shown by correlation

    analysis to be the strongest determinant of ring width in lime (Figure 6). A positive

    relationship between lime growth and rainfall also occurred in winter relating to the

    previous November. Moving correlation analysis shows that the relationships with

    precipitation in May, June and the previous November are time stable (Figure 7).

    However, precipitation in April and July ceased to be a significant factor in growth from

    around the 1980s onwards, while precipitation in August became significant. Correlations

    between ring growth and temperature are not time stable (Figure 7).

    Discussion

    The dendrochronology of lime

    This study identifies that lime series cross-matched together and usefully against existing

    chronologies of oak and yew. It is important not only for estimating the ages of lime trees that

    could not successfully be cross-matched but also for estimating the age of standing trees that

    Table 2. Cross-matches of HPLIME chronology with oak and yew reference chronologies.

    File nameStartdate

    Enddate t-Value

    Overlap(years) Species

    Chronology andshort reference

    SEYEW11 AD1719 AD2009 7.02 138 Yew Churchyards of SE England(Moir, in preparation)

    YATLY-WW AD1829 AD2003 6.95 138 Oak Wych Wood Yateley Hampshire (Moir, unpublished)

    HPYEW92 AD1690 AD1992 6.22 127 Yew Hampton Court Palance GTLondon (Moir, 1999)

    SLG AD1764 AD1993 6.21 128 Oak Scarles Grove SotterleyEstate Suffolk (Moir, 1996)

    EVSLY-BR AD1815 AD2003 6.06 138 Oak Brick House Eversley Hampshire (Moir, unpublished)

    SWW AD1806 AD1992 5.99 127 Oak Southwell Lane SotterleyEstate Suffolk (Moir, 1996)

    HERWOR2 AD1729 AD1969 5.91 104 Oak Hereford and Cumberland (Sie-benlist-Kerner, 1978)

    BRIT002 AD1754 AD1979 5.72 114 Oak Bath Avon (Pilcher, unpub-lished)

    MSC AD1820 AD1995 5.50 130 Oak Mendhams Corner ScotterleyEstate Suffolk (Moir, 1996)

    HVYEW00 AD1814 AD2000 5.38 135 Yew Happy Valley Coulsdon London (North, 2000)

    HVOAK00 AD1814 AD2000 5.38 135 Oak Coulsdon London (North,2000)

    COBHAM AD1770 AD2001 5.13 136 Oak Cobham Kent (Arnold et al.,2003)

    Arboricultural Journal: The International Journal of Urban Forestry 7

    Dow

    nloa

    ded

    by [T

    he U

    nivers

    ity O

    f Melb

    ourn

    e Libr

    aries

    ] at 1

    0:59 1

    2 May

    2013

  • can only be girthmeasured (Figures 4 and 5). The likelymaximumandminimum limits of age

    for a given radius of a standing tree can be identified from Figure 4. This study confirms a

    relatively short lifespan of maiden lime trees in formal avenues. Only 1 of the 30 lime trees

    sampled could be a survivor from the original planting in 1661. Although not shown here, the

    earliest campaign of re-planting probably occurred c. 1898 (Moir, 1996), which suggests that

    trees in the avenue had started to die off around 230 years after the original planting. It is also

    of interest that only after c. 1898 large-leaved limes were used in the avenues. Whether it

    might be possible to extend the 138-year long lime chronology established, back further in

    time is not clear, but the population of ancient lime trees in the north of England (Pigott, 1989)

    could hold the potential to produce millennium-long chronologies.

    Insufficient samples of large-leaved limes were available to establish a useful

    chronology to identify possible differences in correlations with climate between this

    species and hybrids. However, as cross-matching and growth rates showed no clear

    differences, all the samples were combined in this analysis. A similar approach is typically

    Figure 5. Plot of girth against tree age.

    Figure 6. Correlation functions of the residual chronology with monthly maximum temperature(T-max), minimum temperature (T-min) and precipitation (Prec). Correlations are for an 80-yearperiod (19111990). Note: none of the relationships shown were found to be significant by responsefunction analysis.

    8 A.K. Moir and S.A.G. Leroy

    Dow

    nloa

    ded

    by [T

    he U

    nivers

    ity O

    f Melb

    ourn

    e Libr

    aries

    ] at 1

    0:59 1

    2 May

    2013

  • applied in the dendrochronological analysis of British oaks (Quercus robur L.) and Sessile

    Oaks (Quercus petraea) where samples from these two species are usually combined.

    Some problems in the analysis of lime were encountered in ring boundary

    identification and the occurrence of missing rings in older trees. Lime appears to be an

    intermediate between diffuse-porous and ring-porous tree species. The start of each ring is

    defined by a ring of xylem parenchyma (Pigott, 1989), but these boundaries are often

    poorly distinguished. In particular narrow rings, the boundary was not always present,

    which can make the identification of an annual ring difficult. The width of rings also varied

    considerably, narrow rings tended to be locally missing (i.e. visible around only part of

    the circumference), but false ring boundaries could also occur within a ring (parallel to

    the ring boundary) and are difficult to distinguish from true ring boundaries. The first 10

    20 rings of growth were found to be particularly prone to very narrow rings, and so they

    could not be reliably measured and were only counted.

    Physiological relationships with climate

    Correlation between ring growth and rainfall in May, June and August rainfall is positive,

    which indicates that higher rainfall tends to lead to the development of wider rings and

    conversely lower rainfall leads to narrower rings (probably through water stress). Low

    precipitation limiting the radial growth during the growing season is a relationship that

    lime shares with yew (Moir et al., 2011) and oak

    Maximum and minimum temperatures in January indicate that lower mean temperatures

    in January are unfavourable for the radial growth in lime. Maximum temperatures in March

    are unexpectedly shown to have negative relationships with ring width, and the absence of

    a corresponding correlationwithminimum temperature inMarch suggests that above average

    early spring temperature is unfavourable for radial growth in lime. A physiological

    mechanism for this relationship may be that premature loss of winter hardiness followed by

    freezing kills or injures expanding tender tissues such as buds, flowers, leaves and shoots.

    Moving correlation analysis shows that these relationships start from 1980 for January and

    1962 forMarch temperatures. Taking into account the 80-year base length of the analysis, this

    indicates that they have become significant since the 1940s and 1920s, respectively. Losses of

    Figure 7. Contour map showing moving correlation values of climatic variables against HPLIMEring-width indices. A moving 80-year base length over the period 18721997 is used; only the lastyear of the interval, coefficients significant at a level of (p , 0.05) and months that show $9consecutive years of values in the HPLIME chronology are shown. Months in CAPS identify those inthe year prior to ring growth.

    Arboricultural Journal: The International Journal of Urban Forestry 9

    Dow

    nloa

    ded

    by [T

    he U

    nivers

    ity O

    f Melb

    ourn

    e Libr

    aries

    ] at 1

    0:59 1

    2 May

    2013

  • correlations between ring width and maximum temperatures in April and precipitation in

    March and April are also shown to occur around this period from the 1930s. While these

    observed changes in correlations could relate to the effects of climate change, they might

    equally be responses to changed age and species composition in the chronology. Other

    environmental factors such as competition between trees might also be involved.

    Future research

    Apremise of tree-ring studies has long been held that treesmore sensitive to temperature tend

    to be found in the high latitudes and/or altitudes, near their climatically determined limits of

    distribution (Fritts, 1976). Therefore, the sampling of lime trees planted further north is

    important to establish a relationship between radial growth and climate near the limits of

    limes range. Lime trees have potential to become a useful indicator species for global

    warming at high latitudes (Chen, Hill, Ohlemuller, & Thomas, 2011). Radoglou et al. (2008)

    showed that lime trees grow faster in the first 50 years of life than beech, but by the age of 100

    years beech stands yield about 30% more than lime. Additional research on Tilia might be

    considered useful to help predict their future yields and effects under a climate-warming

    scenario.

    The potential to acquire samples useful for dendrochronological analysis, by the

    cutting of V-sections from the stumps of previously felled trees, is highlighted. V-

    sectioning from where the radii were widest helped overcome the problem of missing rings

    in this study, and this could be a useful method to gain material for dendrochronological

    studies from the stumps left of old trees of all species.

    Acknowledgements

    This research was funded by Hampton Court Palace, UK. We are grateful to Graham Dillamore forthe collection and transportation of samples. Donald Pigott and Rikard Andersson made usefulcomments that helped improve this paper.

    Notes on contributors

    Andy K. Moir is Director of Tree-Ring Services and a Post-Doctoral Research Fellow in the Institutefor the Environment at Brunel University. He has worked on the tree-ring analysis of trees andtimber-framed buildings for over 20 years.

    Suzanne A.G. Leroy is Professor of Geography and Earth Sciences at Brunel University. Herresearch focuses on palaeoclimates, palaeoecology and the reconstruction of past natural hazards.

    References

    Arnold, A. J., Howard, R. E., Laxton, R. R., & Litton, C. D. (2003). Tree-ring analysis of timbersfrom Cobham Hall, Cobham, Kent. English Heritage, Centre for Archaeol Rep, 50/2003.

    Biondi, F., & Waikul, K. (2004). Dendroclime2002: A C program for statistical calibration ofclimate signals in tree-ring chronologies. Computer & Geosciences, 30, 303311.

    Blasing, T. J., Solomon, A. M., & Duvick, D. N. (1984). Response functions revisited. Tree-ringBulletin, 44, 115.

    Brett, D. W. (1978). Elm tree rings as a rainfall record. Weather, 33, 8794.Briffa, K. R., Osborn, T. J., Schweingruber, F. H., Harris, I. C., Jones, P. D., Shiyatov, S. G., &

    Vaganov, E. A. (2001). Low-frequency temperature variations from a northern tree ring densitynetwork. Journal of Geophysical Research-Atmospheres, 106(D3), 29292941.

    Chen, I-C., Hill, J. K., Ohlemuller, R., & Thomas, C. D. (2011). Rapid range shifts of speciesassociated with high levels of climate warming. Science, 333, 10241026.

    10 A.K. Moir and S.A.G. Leroy

    Dow

    nloa

    ded

    by [T

    he U

    nivers

    ity O

    f Melb

    ourn

    e Libr

    aries

    ] at 1

    0:59 1

    2 May

    2013

  • Cook, E. R., Briffa, K. R., Shiyatov, S. G., &Mazepa, V. (1990). Tree-ring standardization and growth-trend estimation. In E. R. Cook & L. A. Kairiukstis (Eds.), Methods of dendrochronology.Applications in the environmental science. Dordrecht: Kluwer Academic.

    Cook, E. R., & Peters, K. (1997). Calculating unbiased tree-ring indices for the study of climatic andenvironmental change. The Holocene, 7(3), 361370.

    Fritts, H. C. (1976). Tree rings and climate. New York, NY: Academic Press.Gough, T. (2000). A restoration plan for the Long Water Avenue Hampton Court Place. London:

    Internal Report for the Gardens & Estate Managers Hampton Court Palace.Gray, R. K. S., & Grist, N. R. (2000). Natural regeneration of limes (Tilia spp.) in Scotland: Locally

    widespread and more numerous in 1999. Glasgow Naturalist, 23, 1316.Kelly, P. M., Leuschner, H-H., Briffa, K. R., & Harris, I. C. (2002). The climatic interpretation of

    pan-European signature years in oak ring-width series. The Holocene, 12, 689695.Mayer, H. (1977). Waldbau auf soziologisch-okologischer Grundlage. Stuttgart: Gustav Fisher

    Verlag.Moir, A. K. (1996). A dendrochronological analysis of 9 oak compartments from the Sotterley

    Estate, Suffolk England: An interim report. Hungerford: Tree-Ring Services.Moir, A. K. (1999). The dendrochronological potential of modern yew (Taxus baccata) with special

    reference to yew from Hampton Court Palace, UK. New Phytologist, 144(3), 479488.Moir, A. K., Leroy, S. A. G., & Helama, S. (2011). Role of substrate on the dendroclimatic response

    of Scots pine from varying elevations in Northern Scotland. Canadian Journal of ForestResearch, 41, 822838.

    Newlands, C. D. (1999). Habitats and rare plants of the Durham and Southern Magnesian Limestonenatural areas, with reference to the SSSI series. Naturalist, 124, 2327.

    North, D. (2000). A dendrochronological investigation into yew (Taxus baccata) trees on the NorthDowns, Greater London (Unpublished MSc thesis). University of Greenwich, Greenwich.

    Pigott, C. D. (1989). Estimation of the age of lime trees (Tilia spp.) in parklands from stem diameterand ring counts. Arboricultural Journal, 13, 289302.

    Pigott, C. D. (1993). The history and ecology of Ancient woodlands. In P. Beswick, I. D. Rotherham,& J. Parsons (Eds.), Ancient woodlands: Their archaeology and ecology A coincidence ofinterest. Also published in Landscape Archaeology and Ecology, 1, 111.

    Pigott, C.D.,&Huntley, J. P. (1980). Factors controlling the distributionofTilia cordata at the northernlimits of its geographical range. II. History in north-west England.New Phytologist, 84, 145164.

    Pigott, C. D., & Huntley, J. P. (1981). Factors controlling the distribution of Tilia cordata at thenorthern limits of its geographical range. III. Nature and causes of seed sterility. NewPhytologist, 87, 817839.

    Radoglou, K., Dobrowolska, D., Spyroglou, G., & Nicolecu, V. N. (2008). A review on the ecologyand silviculture of limes. (Tilia cordata Mill., Tilia platyphyllos Scop. and Tilia tomentosaMoench.) in Europe (pp. 129). Retrieved from: http:///www.valbro.uni-freiburg.de/

    Siebenlist-Kerner, V. (1978). The chronology, 13411636, for certain hillside oaks from WesternEngland and Wales. In J. M. Fletcher (Ed.), Dendrochronology in Europe, Vol. 51, BAR Int Ser(pp. 295301). Oxford: Archeopress.

    Stokes, M. A., & Smiley, T. L. (1968). An introduction to tree ring dating. Chicago: University ofChicago Press.

    Wales-Smith, B. G. (1980). Revised monthly and annual totals of rainfall representative of Kew,Surrey and an updated analysis for 16971976. Meteorological Office, HydrologicalMemorandum, 43.

    Wigley, T. M. L., Briffa, K. R., & Jones, P. D. (1984). On the average value of correlated time-series,with applications in Dendroclimatology and Hydrometeorology. Journal of Climate and AppliedMeteorology, 23(2), 201213.

    White, J. E. J. (1998). Estimating the age of large and veteran trees in Britain, Forestry Commission,Information Note, 250.

    Arboricultural Journal: The International Journal of Urban Forestry 11

    Dow

    nloa

    ded

    by [T

    he U

    nivers

    ity O

    f Melb

    ourn

    e Libr

    aries

    ] at 1

    0:59 1

    2 May

    2013