Seasonal Growth Pattern of Cypress Pine

Cypres6 Plnr Intcrh Scalonal Rcport Grorrth proJect odelltng XLLbrr Pattern J.I(. I, 1985 of Clnrrcs! plne Vanclay Ab!tract Analysis totaling of nonthly height and dendroneter nore than 600 tree years reveals ttrat: o Height growth to Septenber; exhibits a period of dornancy o Diameter increnent may accrue wtrenever conditions are experiencedr relative hulidity be a najor controlling factor; and tlrat neasurellents during July favourabfe appears to 6 Natural reversi.ble changes j.n sten size due to noisture status [ay attain a nagrnitude equivalent to ttle annuat increlent of the stem. An 6published report for Lnternal Not to be cited lrithout perllission circulation onLy. of tl'e author. 1. Introduction Investigations and applications using grol'th tree infornation generally require the anmral- increment. This can be readily deterDined when a tree (or plot) l.s reneasured on the anniversary of hhe previous leasure. Horrever, rJhen tlte tiDe interval is not an exact Eultiple of one year, soltre i.nterpolation is necessary to derive annual increttrent. In soutltern g:rolrth nay be continuous tree oueensland, througtrout year, the or nay exhi.bit a period of dor[ancy. To illustrate the effect of these differing qrolrth patterns on estidates of annual increnent, it is expedient to consider lrro sinple eralples. Figure I illustrates a tree 9,hich qrolrs at a constant rate throughout tl.e year. Tlte tree represented in Figure 2 grohrs at a constant rate for nine nonths and is dor[ant for three nonths during July to August, Itre assunption of constant grotrth lrhen not dornant is unrealistic, but is convenient for this exa.nple. If the assueption of constant grosth ls applied to the cree of Flgure I, an unbiased estiDate of annual increuent is obtained. ff hoeever, the tree had a dorr0ant peri.od, considerable bias could result, depending upon holJ Duch of the dornant period occurred between neasures. Fiqures 3 and 4 indicate the extent of the possible bias reiultinq fron the incorrect assunptlon of grorrtb pattern. This bias is signif icant, particularly stten reneasurenent occurred af ter a short interval, when bias nay exceed IOO*. It is not only the length of the dornant period lJhich influences the estinates of annual increnent. Fi$tre 5 ilLustrates the effect of faiLure to correctly detect tlte onset of doruancy by only one tronth. fn order to utilize data derived frou annual Eeasure[ents, sode infornation pattern of the species is required. other ttran strictlv concernincr the crowtir Ibta (153 Dalby) rras established to observe fn 1974 an elperinent the short groeth tern pattern of Cypress pine_ This experiment recorded: ( 5-15 of sixteen trees Western Creek (Figure 5) 1974 to January 1979. 6 Monthly observations cD dbh) Iocated during July at on heiqttts SF 154 Fqow 7-. tF rti Ag 3 e s rd Frguee ?. :SF t\[ t r Lou sr<g^tt a fra-Fr t {o-rr t} r -x D s Ftrlr^J.t Tbre.--moa$ Dor-.o7 rrs 3 t 3 c-r, D s Ft{l ar ! }'t o-D sFfarl',rs5A a Ace.r.g Cori,rnnr fo..rrr srrr Ddrrrrry 5 e T 3 a. F,gn 4. I ? a. 1.O t'3' d a h r.t' d o H Ir t 6 c a.\ l'o 4 { G .b "; b to 4 tL d. t t l.o .t J. 4 o l | r r O l t t 1 fcrloadcc 6 , l *rtr ,l- !&i.r fitri*r drtttl ' - 3 o ltonthly observations different locations T'hese incLuded on diantreter of lG8 trees at five during July 1974 to January 1979. - len trees each in 0-10, LO-20. 20-30 classes at SF 10 Hillsittet - ten trees each in 5-15, L5-25, 25-35 classes at SF 55 Pony Hillst - ten trees each in 5-15, 15-25, 25-35 classes at SF 328 Yuleba; - ten trees each in 5-15, L5-25. 25-35 classes at SF 302 Barakula; and - si.xteen trees each in 5-15, 15-25. 25-35 classes at SF 154 l{estern Creek. cn d. b.h. cn d. b. h. cr d. b. h. cD d.b.h. cn d.b.h, li Daily observations on dianeter of ten stens located near (SF 302) canp during August 1975 to July ttle Barakula L976. Dianeters sere [easured using.a graduated stainless steel dendroneter band {Lining 1957), pernanently fitted to each tree. t'hich ena-bted diameger to be recorded to the nearesr 0,1 nilliuetre. Heights were neasured to nearest centinetre using height sticks and a peruanent reference peg driven flr[ly into the ground near the base of each tree. Ilaily dianeter observatLons sere made at 9:OO au (a1so at 3:00 pu). Uontbly diaueter and height observations lrere nade near the start of tlre nonth, at various tines of the day ' Rain gauges installed at each locaij-on lrere read lonthly at the tiDe of diaBeter/hei.gltt measurement, and paired g,auges at the Barakula Office reaa dally and nonth!.! enabled- a '.true,' conparison betseen nonthly readings and tlte ,'nonthly' ( cumulati.ve daily) gauge at Barakula rainfall. Ttre !'as initial-lzed to 20 nilliDetres each uontlt to allorJ deternination of ewaporation in tl'e absence of rainfall. 3. Analylt! rnd Relult! Tlis investigation attenpls pattern of cypress pine determine this, el.o options to reveal gro['th the typical throughout the year. In order to are available: o It can be assu!tred ttEt the rainfall pattern during the years L974 to 1979 was typical of tlte long terd average pattern, or 9"1"f' : -"f+ 1 I m'l) d P -i- 5 l s-t\ o -3tJl^ : l I o ol I lrr J o a l rvt l = t = o z Y E = 6 i Et frp9 FfiE ---\ o FIGURE 6. : < a\ LOCATIONS OF MEASURED STEMS. - 4 - o The actual observed rainfall nay be used in the analysis as a reqressor (or so-called "independent.,) variable to descrl.be the gror"'th lrattern, and the grolrth typi.caL pattern lay be suruised using lonq tern averagre rainfall observations. The nature of the rainfall observations in this experi.nent are not suita.ble for the second of these options, ai.rd i.t is necessary to nake the forner assunption. It is expedient to uake a further assuuption: 6 Itlat tree height and diameter increase lonotonically. That is, height and diarieter increnenls (adjusted f;r reversible sten srrelLing and shrinking) can never be neqative, and uust alrrays be positive ot zeto. Linear regression cannot be used for this analysis. Ttle serial nature of these d.ata ( repeated neasureneits of the same stens) violates the assurptlon of lndependent residuals ilrplicit in linear regression. Thus non-llnear reqression nust be used. 3.I nalnfaU Regression analysis indicated that a straiqht line relationship existed between observations on ihe paired gal.lges; that ttle slope of this line nas not sigmificantly different fror unityt and that the intercept rras strongly influenced by uonth of year. the Itris inplies that the uonthly gauge faithfully recorded the rainfall, but Lost lrater at a rate qiven by ( nonth+Z. ZS, t 6a]^t z Evap.Loss = t5.383+I.7s9sinfPi^ Tl'rus loss fron a nonthly rain gauge nay vary froD 13 (July) ( Janr.rary) nillifietres to 5l per nont]r. As thls Loss is ;f tlte sare order of nagmitude as the rainfall in this reg.ion (Fignire 7), the presence or absence of water in the gauge is not consLdered to be a retiable indicator of rainfall. For exanple, sone water in the gauge Day indicate heavlr rain early in ttle lonth fol,1olred by fine weatlter, or a dry [onth eith a llght sholrer the previous day. tlrus the voluhe of sater in the gauge is useful only as an general indicator of soil [oisture content and ste[ hydration. 3.2 tletgbt crorrth The height observations conforD !re]l to the assuDption of nonotonic lncreasing height, r"rith only one erceltion. In ltovenber L976, t}):e height of tree 257 suddenly decreased to 11,7 netres, and S. LI.4 Detres after several [ontbs at g. contlnued to reDain a! this lesser hei;ht. To enable a FSor<- J- MtLes /ldir'r ld.ifau trrDA^ -t3o 7o 6" 50 tro ,o ?o lo o 5q. F6 trlr Ay lial'$n llro re tlq&.r. 8"r*u {F .) Stl \ 9P el fldt D4 ID-yer qoon lEilrorr 'fgl|l- t{..\ tu-rrttr' rb) 9o ts go 7o h LO fo fr o 5o -5 4o -b 3o -15 ao 6o 40 30 2D ao lo o o I^, ft n r - 5 - feasible nodel. to be applied that the tip had broken. to these data, 1t lras assuDed Ttre non-linear nodel. proposed I'as H(i,j) = H(i,j-l) .r r(i)*n(j,) - b shere H(i,j) iE the height of tree i at the end of ronth j, H(i,o) is the initial helght of tree I, r(i) is the averaqe annual heigtrt increDent for tree i. [(j) is the proportion of ttre ann(EL increnent occurrlng during Donth j,, J, is the nonth of the year corresponding to Donth j, and b is the breakage and is zero for all treeg and nonths eacept for tree 257 in I{oveDber 1975. This uodel inplies the follosing assunptions: ( increuents O height increases nonotonically positive ot zero, but never negative); nay be o each tree ttas its oIJn unique annual increnent rate, and ,nai.ntains this increuent during each of tlte five years of observation, and o the proportion of tlle nonLh i6 the sane for annual increDent aII trees. accruinq in anv I'he esti'?ted g. Donth]y incre[ents g.lven in are Figure Annual helgttt increDents varied frol 13.2 c! for a t;ee lt.4 n high to 40.I cn for a 7.5 n tree. lte estiDated breakage eas 35 cn. and the standard error of height estinates lras u.J cmThe results sine curve, of this analysis erhibited sone regeDblance Eo a and a refined uodel lras proposed: P( j ) =tc+Sinf ( J+d,tt 2r.pi / 36571I 355 Qhere P(J) is the proportion of the anmral incre[ent j, d deterDines when during tbe year the occurringr on day greatest and least increments occur, and c deteruines the aDotlnt of increlent occurring on the day of teast increnenE. This- analysis yielded sinilar estinates for initial heights, parameter helght increnents and the breakage. c was estinated as exactly 1.0, and paraneter d as 52.5. Ttris predicts a zero heig:trt increnent on 9 August, and the naraibrrD increment on I February. Ttre standatd error of height estinages $as 9.1 cb. The predicted qrowtlt pattern suggests that height grolrth eay cease during the [onths JuIy to Septenber. Era[ination of cliDatic data for the reqion (Anon. 1975) revealed that the nonthly uinidun teDperature folloded a sinilar trend. The lreather i.n late sinter is nore varlable tlran in early FtSote8. lletyt lx*narlr DerAltlzred.: 3 * S"o".t lrcm"'r.* Ht"t *z o, -\\ "1 tib. f-Fartl.r) o.6 0s c.4, o.1 c4 ol. Fqtrc lO- Dnmerat.lff*$anpr0c?seranrv - -, i6 i \r"ol Itrt,nrrt xrt'r3 re ,tAo'rrlr , tl tl r+ a lo g 6 4 z \-*.. - 6 - winter, and the chance of frost is correspondingly -lig, greater. Ttre ten-year nininum teroperature+ reflecti thi; and corresponds closely lJith the heiqht grolrth pattern. 3.3 lblly Dluetcr Obrcrvttlon! A nore conplex nodel is required for the dianeter observations, as tlEse do not Lncrease nonotonically as with height, but show considera-ble variation, Ficrure 9 illustrates the fluctuating nature of dianeter, and the dranatic response to rainiaLl. (f97S p.84) Johnston attributes tlEse revergible changes in steo dianeter to bark snelling and shrinkinq eith changing noisture content. M.R. N-ester (pers. cou[. ) is conducting an analysis of this ptrenonenon, and proposes ttrat: o Following rain, (by an the bark innediately seells anount linearly related to rainfall. up to its naxinur swelllng), then graduall.y shrinks as aD exponential decay as it dries. 6 frreversible dia-ueter grosth is also linearly related to rainfall, but occurs nore gradually over the fotlolring thirty day perlod. The rapid increase in sten dianeter followinq rain nay be (or both) xyle[ or bark thickness. due to cltanqes in eitlter odin (1972) found that eettlng of Scots pine and Norsay spruce bark had rel-atlvely l-ittle effect on sten size. BotL seasonal and diurnal ste[ shrinkage of forest trees are lrel.l docunented (Kozlolrski 1982). Worrall (l.gGG) deDonslraEes a linear re1ationship betlreen sten dianeter and lrater stress in Callttris cupressiforuis, and Kozlowski and flinget (1964) shotJed substantial slrell.ing of drougrtrted poptar st;us eithin 30 ninutes of rainfall. i this is conputed as = Hean - 2.52* ,lean-I4Decile) Io-yr-uin and represents the loeest tedperature likely to be encormtered during that Donth over a ten year period. Hean and l4DeciLe observations rrere derived fron Anon (1975). 6 F d L t l s = s. dt e 4 i s T Y f @ ffi-2 f.{ l1-2 -9- | E - J' C c 4 oa ".. -! Ar _j : -t I t^ tuE-: ? ? o J F ! Y(t .E z t c e! ET,. J u- *-=a?.- ! 4 ter a FlEt t s r d 9 r t tL G8 a -{ tE:EEEnrIlcEr ml t > . 6 r l c d zh .rit t 6' 3 .s: jrl JssFlt!.?_F v r + J i 2 6 6 5 6 : o 3.4 t{ontbly Dluetcr Obrcrv.tlonr Thedonthly dianeter observatj.ons exhibited considerable fluct{ration, pregunably a result of seasonal and diurnal steD shrinkage. RegretCably, no account can be taken of diurnal shrinkage as tile of neasuredent lras not recorded. S-one account nay be taken of seasonal shrlnkage by eEpLoying the rain qaugre data. The [odel proposed I'as D(i,j) = D(i,j-1) i r(i)*u(j,) + b(k) = D(i,j) ^ (I + s^Minfp(j),pDaxl) D'(i.j) nhere D(i,j) is the "alry dianeter" of tiei i at the end of j, nonth D' ( i, j ) is the corresponding' observed diareter { includi.ng an allorrance for any steD srrelllng.), r(i) and n(J') are as above, b(k) allorrs for re-adjustnent of the dendroneter band after dauage by cattle. etc, s is the anormt of sten swelling corresponding to a unit of [Jater in ttle 9au9e, p(j) is the arount of water in the g|aug.e at the end of nonth j, and Pnax represents the a[ount of nater in the gauqe corresponding to sten saturation. This nodel ilrplies the sane assunptions heigtrt lncrenent, and also inplie! tlrat: as o the presence of lJater in the rain deqree of sten srrellinq correlated lrater in the gauge. The nodel is nore corpatible wilh ttre swelling tltan of seasonal shrinkage of efflcient nay of e[ploying the available account for sone variation in the data. atte[pt to nodel ryleE s]trinkage lrould soil [oisture content. the nodel for gauge indicates a lriah the aDount of assuuption of bark the xylen, but is an rain gauge data to A nore sophisticated require esiiDates of Cllanges in stem diameter conforned weII to these assumptions, except for one snall steu Fhlctr violated the assuDption of a constant average annual increnent. Ttris (tree steu 509, dbh 4.3 cn) at Western Creek exhibited a high increnent for the firsb eig'trteen [onths follolred by a very loI' increDenl, and lras excluded frod further analysls. The resuLts of this analysis are sunuarized in Tables l-, z and 3. Tabte I inalicates the precision of tbe estinates, and the apparent reLationshlp betlreen sten swelling and the "Saturation nontltly 9au9e reading. llle entry labelled Level.' refers to the aDount of lrater ln the gauge (at tiue of reading ) lrlrich corresponds to uaxinurtr sten swelling, and. not to the actual rainfalL. - B - TAALE 1. Details Sten Slre1l ing ( ?/nn ) Location I SF 10 Hillside I SF 55 Pony HllLs l SF 328 Yuleba I SF 302 Barakuta I SF 154 t{estern Ckl o.026 0.003 0. 008 0-005 o.024 of Analysis Saturation I (Dn) 90 | I Standard Error (uD dbh ) Deg.rees I o f l Freedonl I 0.815 o.716 0.794 I.L21 L212 977 1364 1 40 4 2459 Ttre estinated saturation level and rate of stem swellinq varies considerabty l.ith location (Table l). This uay be du6 in part, to ttle place[ent of t]re rain gauge Lrit]t respect to surrounding trees. This coulil influence both the iccuracy ltith which the gaug:e records ralnfall, and the rate oi evaporation fron the gauge. Table 2 indicates the proportion of lhe annual lncredent ( standardized accruing each nontlt to 30-day Eonttls), and Table 3 indicates llre estinated averagfe atrnual increnents. Dianeter groerth does not exhibit the sane regular pattern of helght grorth, but nay occur at any ti[e crtren conditions are favourable. The averaged figures (Table 2) indicate that the najority of dl.aneter increuent occurs durinq the first eigbt nonths of the year ( Jamrary to August), and that very litale incretrent occurs during ttre last four Dontlts. Hoqrever. grolrth uay cease at any tine during the earl-y part, or reconnence at any time during the later part of the year as conditions dictate. It is notelrorthy that the average nonthly increments correspond closely to the average 9:OO an hunidity (Fi$rre l0). 4. Dlscullion The results of this analysis need to be interpreted eitlt sone caution. It is rmreasonable to expect treights to be correctly (!{ester leasured lo the nearest centi[etre lgBI), and it possible is not to conpletely distinguish lrreversible qrorJth fron seasonal and diurnal steu shrinkage and- _swelling. T'his constant shrinking and seelling l;d (1941) to conclude that they Fielding and Uillet could- not deternine by use of dendroneters srhen canbial activitv in Pinus radiata began or ended. However, the analysis i.oes yield sone usefut results. o st t zo c o :-- t - - - - i-t c rl, n^ t-l t l J- --1 I E d l;--;; t-l I I E--- - t rr .'l E l l t l c4 t E -t-l t----- O F tt t d c C iDEqI l*l t'-"' t l I J-J t----- Et -J. u q ^g l-l;-l -l tl l t----- E F4l O E t -ll -lt----I H C ) c a E c E {}0r q t-..-'t -tt - l t--;;: l= -.1 l t----_-.1l ' l t-----t--t lt.- r'r F (o F. ! E t;;:; a ;d l-rI d o € l _t J - { o d f r, lr q.>.q 3 .i(rH I lrq t& rt r& f{ F c F .{ I t t*l C <d n^ t l t lJ g E l I .q^ - { C d E d { , E lo. I l.t Jll d t{ t'l t l c E d t o l tt - - --- - -_t_J t---_- o .q .q^ lJC t @ l E 4 o 4 F O F _l-J t t t t |I r I l td I tl "! Il _{ I'I gl d-i +I.d L) -{ c J ! o r r o f d o :EO.FEIB !.] - r 0 - . . Height increnents appear to exhibit a regnlar pattern tJith a period dornant of up to three nonths du;ing July to Seplenber. The regular pattern su99est9 internal (hornonal ) control of groLrth, and the lEttern appears suited to protect the grohti.ng shoot froD frost dauage {Fiqure 8). This pattern is characteristic of sh,oot growth ln tenperate zones lKozlolrski 1982 ) . The slJelling and sttrinking of the sted (Talle I and Figure 9) nay attain a haqnitude equivalent to the annual incremenE poses difficulties {Ta-ble 3), and this i.n detectinq -nodel and neasuring irreversible increnent. Itre adopted expresses sten slrelling as a sinple linear frmction of the erater Ievel in tlre gauge. t'hich nay provide a reasonable estimate of moisture content of the bark, but is wlikefy to reflect the soil noisture content. Lassole (1979) found that sten shrinkage lJas Iinearly growth, related to recent and argued that ttlese newly forned ti.ssues forned najor water storage areas githin the plant. The current noael expressed steu swelling as a proportion of steB size, noE as a proportion of recent increnent. Hoeever. the correlation (Table between stern swelling 1) and average annual (Table increnents (1979) 2) suggests tlrat Lassoie,s contention nay be nore appiopriate. Diaueter increnent is dore varialle than heiqht increnent and appears to occur rthenever favouralle con6.itiorr" o..rr.. There is no suggestion of dornancy in rrinter (Figure IO); presulrably there is no advantage in such dornancy as the carbiun is lrell insutated against the nild frosts experienced in this region. Tllis is conslstent lrith that apical grolrth is rmder stronger -research findings (Kozlolrski horDonaL control than growth caubial l-97La p.372t. The siuilar patterns of average increnent and averaqe 9:OO an humidity sugqest that relatj.ve hunidity Day be a causal. ag,ent in deternining the grorirlh pattern of cypress pine. It ( and is likely that during periods of 1ow uorning huBidity high teuperatures anal strong, winds) the rate of eater depletion fron the fotiage exceeds the ability of the roots to replace the Bater, resutting in cessation of photoslmthesis because of internal eater deficits (Kozlosski 197Ib p.13I). Lassoie (1979) concluded that sunDer grolrth of Douglas fir Eas controlled by evaporative denand, rattler tlran soil. noisture content pg! Eg. (1982) also Kozlowski stresses that plant rater deficits depend on relative rates of absorption and transpiration, and not on absorption a_Lone. - 11 Johnston (1975 p.51) observes that cypress pine stens in the understorey -(of r$even-aged stands) ltroe very efficiently, aPParently rtith nininal inpact on overstorey iircrenent. t'tre a-bove hypothesis offers sone feasibte exFlanation of this pltenonenon. Hu[idity, te[perature and eindspeed are likely to be nore favourable in the understorey, and could altow continuing photosynthesis in tlte understorey when sater deficits bave becone lirhiting in the overstorey. 5. Conclullon ltrree inportant conclusions nay be drain! fron this study: o There is some eviilence to suggest growtt! that height erhibits a period of doruancy during JuIy to Septerber. 6 Diameter grorrth appears not to enter a dordant state, (soil- noisture status and but occurs wtleneve; conditions rate of lrater loss fron foliage) are suitable to sustaj.n qroIJttr. 6 IlIe natural variation noisture status) increnent. The inplications are: in sten diareter nay erceed tlte of these conclusions (due to annual for model o Dianeter increnents should be based on a of not Less than two years. cbang.ing alianeter d.evelopnent tine interval 6 !,lhen calculated fron deasurenents correspondinq to tlne intervals not an exact nultiple of one year, aliameter increnent should be assuded to be constant throuqhout tlte year. 6. Acknosledgcrcnt! Experinent 153 Dalby lras initiated by T.Ninvolved nunerous officers of Forest Speclal recogmition is nade of the [Jork of the daily leasurenent of steDs. Dr. c. valuable criticisn of an earlier draft. Jobnston, and has Research Branch. J. llacalister in Bacon conbributed - 1 2 - . . 7. Referencer Anon. (1975) Cliuatic Averages oueensland. Metric klition. Bureau of Meteorologry, Departuent of Science and Consuner Affairs. AGPS. 76 pp. Fielding, J.M. and Uitlet, H.R.0. (1941) Sone studies on the (Pinus radiata). growth of ltonterey pine I. qroLrth. Dlaneter Forestry and Tidber Bureau Bulletin 27zL-33. Johnston. T.N. (1975) Thinning studies in pine cypress in D e p . F o r . R e s . 7 . P a p . l l o . 87 Queensland. Qld, pp. Johnston, T.N. (1978) Ttre effect of silvicultural breatnent on qrolrth and nutrient status of a cypress pine forest. Unpublished thesis, Haster of Natural R e s o u r c e s , U , l { .E . 1 3 6 p p . Kozlowski, T,T. (1971a) crol'th and Development of Trees. Volune I. Seed Gernination, ontogeny and Sttoot cro\rth. Acatlenic Press, l{.V. 443 pp. Kozlolrski, T.T. (I971b) crorrth and Developnent of Trees. Voluue II. CaDbial crorrth, hoot Grolrth and Reproductlve Gro!'th. Acade[lc Press, N.Y. 514 pp. Kozlowski, T.T. (1982) Iilater supply and tree qrowth, Part I. t{ater deficits. Reviee Article in Forestrv Abstracts 43(2r.57 -95. Kozlolrski, T.T. and Winget, C.H. (1964) Diurnal and seasonal variation in radii of tree stens. Ecoloqy 45:149155. Lassoie, (19791 Sten dinensional J.P. fl.uctr.rations in Douglas-fir of different crorm classes. For. Sci. 25(1\ t132-144. Liuing, F . G . ( 1 9 5 7 ) H o D e l r a d ed e n d r o n e t e r s . J . 577. Nester, !{.R. (1981} AssessDent and heasureuent errors in sLash pine research plots. Dep. For. Tech. QId Pap- No. 25. 10 pp. odin, H. (1972) Studies of the increnent pine and Norway spruce plants. Suecica 97. 3l pp. For. 55:575- rhythn of Scots Studia Forestalia - 1 3 - . . Worrall , ( L966 ) A nethod of J. correcting d.endro[eter neasures of tree dianeters for variations induced by troisture stress clranqe. For. Sci. L22427,429-