Table Of Content

( document Historic, archived Do not assume content reflects current scientific knowledge, policies, or practices, United States Effects of Thinning on Department ofAgriculture Temperature Dynamics and Forest Service IRnetseeramrocuhntSatiantion l\/lountain Pine Beetle Activity Research Paper a Lodgepole Pin^ Stand INT-RP-479 in December 994 1 The Authors Research Summary Dale L. Bartos is an ecologist with the Mountain Temperature measurements were made to better Pine Beetle Population Dynamics researchwork understand the role ofmicroclimate on mountain unit, Intermountain Research Station, Logan, UT. pine beetle,Dendroctonusponderosae Hopkins Prior tojoining the unit in 1984, he was with the (Coleoptera: Scolytidae), activity as a result of Intermountain Station as a member ofthe aspen thinninglodgepole pine stands. The study area is ecosystem project for 12 years. He holds B.S. and on the north slope ofthe Unita Mountain Range in M.S. degrees from Fort Hays State University, northeastern Utah. Samplingwas done over 61 Hays, KS, and a Ph.D. degree in range science from days startingJune 23, 1987 (174). Eight lodgepole Colorado State University. His principal research pines,Pinus contorta var. latifolia Engelm., were interests are in ecological processes and disturbance sampled for surface and subsurface (phloem) tem- ecology. peratures in a thinned and unthinned stand. Prin- Gordon D. Booth (retired) was leader ofthe Statis- cipal components analysis was applied to all temperature variables. Most ofthe variation was tics and Computer Sciences Group for the Inter- mountainResearch Stationin Ogden, UT. Hereceived attributed to two variables, coolest part ofthe night (1:00 a.m. to 9:00 a.m.) and hottest part ofthe day a B.A. degree in 1960 and a B.S. degree in statistics in 1963, bothfrom BrighamYoungUniversity, Prove, (1:00 p.m. to 6:00 p.m.). These two variables were UT. Then he was granted an M.S. degree in 1967 smoothed using time series analysis that permitted and a Ph.D. degree in 1973 in statistics from Iowa us to see general patterns and small differences State University. From 1963 to 1965, he worked as betweentree temperatures. The thinned stand was approximately 1 °C warmerthantheunthinned consulting statistician with U.S. Steel Corporation and with Philhps Petroleum Co. From 1967 to 1981 stand andthe day temperature was 10 to 11 °C higher than the corresponding night temperature. he worked as consulting statistician with theAgri- Models were developedto predict phloem tempera- cultural Research Service, U.S. Department of Agriculture. From 1981 to 1994 he was with the ture from bark surface temperature. The resvdtant equations had values of0.98 or greater. Intermountain Research Station. The use oftrade orfirm names in thispaper isforreader information only anddoes not implyendorsement by the U.S. DepartmentofAgriculture ofanyproductorservice. IntermountainResearchStation 324 25thStreet Ogden, UT84401 Effects of Thinning on Temperature Dynamics and Mountain Pine Beetle Activity a Lodgepoie Pine Stand in Dale L. Bartos Gordon D. Booth Introduction the stand, and (2) growth ofresidual trees is usually slowto respond to thinning. The objective ofthis Thinning has been usedin the pastto increase study was to determine the surface and subsurface tree vigor (Graham and Knight 1965; Keen 1958), barktemperature variabilitybetween thinned and and tree resistance to attacks by mountain pine unthinned lodgepoie pine forest and to relate these beetle. The removal oflarge-diameter lodgepoie differences to mountain pine beetle activity. pines, which are preferred by mountain pine beetle, can also resultin a reduction in tree loss during Methods and Materials beetle epidemics (Cahill 1978; Cole and others 1983; Hamel 1978; McGregor and others 1987). Reduced tree losses occur before residual trees can express The study site is south ofMountain View, WY, resistance by increased growth (Amman and others on the north slope ofthe Uinta Mountains in north- 1988). This phenomenon suggests that factors other eastern Utah. Mature lodgepoie pine (80-120 years than vigor may be responsible forreductions in old) occupy the site, which is on the Wasatch-Cache mountain pine beetle infestations. National Forest at an elevation of2,865 m. Part of Thinningforests can cause subtle changes in tree the site was thinned in the early 1970's. Replicating physiology (Nebeker and Hodges 1983) and microcli- the thinningtreatment was not possible because of mate, especially temperature (Bartos andAmman the length oftime since treatment and the prohibi- 1989; Schmid and others 1992). In western North tive cost ofequipment to monitor numerous sites. America, changes brought about bythinninglodge- We decided that useful information could still be pole pine (Pinus contorta var. latifolia Engelm.) for- obtained by selecting adjacent stands (thinned and ests have had profound effects on mountain pine unthinned). These two stands were selected so that beetle (Dendroctonusponderosae Hopkins [Coleop- physical characteristics, such as slope, aspect, and tera: Scolytidae]) behavior(Schmitz andothers 1989), elevation were uniform. resultinginreducedtree mortalityinthinned stands The stands were sampled for 61 days duringthe (McGregor and others 1987). summer of1987 to determine bole's surface and sub- In addition to increasing growing space, thinning surface (phloem) temperature. During 1986 the also affects parameters ofthe physical environment, area had a very active mountain pine beetle infesta- such as temperature, light, and wind speed. Tem- tion andwasused foranin-depthstudyofnumerous perature is an important factor in the ecology ofin- microclimatic differences including; temperature, incident solar radiation, windspeed, wind direction, sects as it affects the physical conditions ofhabitat and the physiology ofinsects themselves (Safranyik and stand temperatures (Bartos and Amman 1989). 1978; Wellington 1950). In the case ofmountain Beetles were captured during peak flight time in pine beetle, observations were made on the effects both thinned and unthinned stands in 1986; only ofextremely high (Patterson 1930) and low (Somme 5 percent ofthe total beetles caught were in the 1964; Yuill 1941) temperatures. Between the ex- thinned stand. The epidemic beetle population tremes is an optimum zone oftemperature that may was not active during the summer of1987. be modified by other microclimaticfactors (Rudinsky Stand Characteristics 1962). We explored the effects oftemperature on altered Characteristics ofthe thinned and unthinned lodgepoie pine stands because (1) temperature lodgepoie pine stands were determined through changes are immediate followingtree removal from variable plot (10 BAF) cruising. Plots were placed 1 m 50 apartin a grid patternin each stand. Lodge- These means were labeled: 1:00 a.m., 4:00 a.m., pole pine trees on each plotwere tallied as to live or 7:00 a.m., 10:00 a.m., 1:00 p.m., 4:00 p.m., 7:00 p.m., dead by cause ofdeath and were measured for diam- and 10:00 p.m. The mean labeled 1:00 a.m. was the eter at breast height (d.b.h.). The dominant or co- average ofthree temperature measurements taken dominant tree closest to plot center was measured at 1:00 a.m., 2:00 a.m., and 3:00 a.m. Other means for height and crown length. Stand density was ex- are defined similarly. The measurements were pressed in terms ofbasal area andnumber oftrees taken each day for 61 days. per hectare. Afurther reduction inthe number ofvariables that needed to be dealt with was attempted by using Monitoring Temperature a modificationofprincipal component analysis, some- what similar to that used by Jassby and Powell An automatic recording device (2IX micrologger, (1990). The objective was to identify fewer than Campbell Scientific) was used to measure the tem- eight daily means that would capture all the essen- peratures for a 61-day period startingJune 23, 1987 tial information on dailytemperature patterns. (174). This time period encompassed the "peak" Suchvariables would also have the additional prop- mountain pine beetle flight period, which usually erty ofbeingindependent and thus capable ofbeing occurs in this area the lastweek in July and the studied one at a time. first week inAugust (Bartos andAmman 1989). The principal components analysis indicatedthat The time priorto flight was included to monitor the onlythe first two principal components accounted temperature that existed duringthe final stages of for virtually all the variability within a given day. beetle maturation. Eight trees were monitored in Further, the first principal component was essen- each ofthe two treatments. Initially, the instru- tiallythe mean ofthe mean temperatures labeled menttower site was randomly selected within each 1:00 a.m., 4:00 a.m., and 7:00 a.m. We chose to use stand. The eight sample trees were located around a simple unweighted mean ofthese three tempera- this tower and the distance fi'om the tower to each tures and referto it as the nighttime mean tempera- sample tree was limited by the temperature sensor ture. Likewise, the second principal componentwas "leads." All sample trees were similar in height, essentially the mean ofthe mean temperatures at crown length, and d.b.h. (approximately 22 cm). 1:00 p.m. and 4:00 p.m. Again, we chose to use the Temperature sensors were connected to the micro- simple mean ofthese two temperatiires to represent logger to measure temperature attwo points on the second principal component and refer to it as each sample tree. These temperature sensors were the daytime mean temperature. Because ofthe m placed at breast height (b.h.) 1.3 aboveground, close relationships between these two variables on the bark surface, and immediately below the sur- and the first two principal components, the two face, on the south side ofthe trees. Temperatures variables are very nearly independent at any given measured in these stands in 1986 (Bartos 1988) time. Ofcourse, they are not serially independent showed that a strongrelationship existed between overtime (Booth 1993). the surface temperature on the north and south The nighttime and da3i;ime mean temperatures side ofthe tree. Therefore, only the south side was were analyzed as time series. We knew in advance measured to maximize the number ofsample trees. thatthere was variability from one measurement to The below-outer bark surface sensor was posi- the next duringthe course ofthe 61-day study. To tioned in the phloem because phloem is the substra- study the underlyingtemperature patterns, we ap- tum in which mountain pine beetle adults mine and plied the Daniell (1946) window, a smoothing pro- lay eggs and the developinglarvae use it as food. cess, to each ofthe time series. This sensor was placedjust below the thinmbmark of The application ofthis smoothing method elimi- the tree and was gently forced a couple of into nated all high frequency oscillations. Residuals the phloem. fi'om these smoothed values had all major, low- frequency patterns removed, and a fourier analysis Data Analysis ofthem produced no detectable periodicities in the data. Therefore, graphs ofthe smoothed values Characteristics (density, basal area, d.b.h., tree height, and crown length) ofthinned and unthinned themselves can be expected to convey all pertinent stands obtained during stand surveys were sub- information relatingto temperature patterns inthe jected to analysis ofvariance to test for significant data. differences between treatments (Bartos andAmman Schmid and others (1992) found a strong correla- 1989). tion between bark temperature and ambient air Ourbasic daily data consisted ofeightvalues, temperature. Ifpossible, we were interested in es- each being the mean ofthree hourly measurements. tablishing a relationship between bark temperature 2 m and phloem temperature. Therefore, we performed 15.1 high, withUvecrownbeing53 percentoftotal model fitting on 61 nighttime means and on 61 day- height (Bartos andAmman 1989). Ofthese stand time means. We further dividedthe data into those characteristics, only the basal area andtrees per obtained on the thinned stand and those obtained hectare were significantly different between stands — on the unthinned stand. Thus, we obtained atotal (P < 0.05) at this point in time approximately offour fitted models: both a nighttime and daytime 20 years after thinning. When the stands were model forthe thinned steuid and two analogous monitored, the effects ofthe treatmentwere still models forthe unthinned stand. The models were pronounced. fitusing data fi-om five trees in the unthinned stand and fi-om fourtrees in the thinned stand. These Temperature trees were selected because the remainder ofthe The three main questions ofinterest regarding eighttrees produced incomplete data sets. temperature dynamics are: (1) "Whatis the difference intemperature pattern betweenthe thinned Results and the unthinned stands?", (2) "What is the difference in pattern betweenthe air temperature and We looked atthe results interms ofboth stand the phloem temperature?", and (3) "What is the dif- characteristics andtemperature. ference in patternbetween the temperatures atthe warmesttime ofthe day andthose at the coolest Stand Characteristics part ofthe night?" Data relatingto these questions appearinfigures 1 to 3, respectively. Thethinned stand had an average basal area of To facilitate answeringthese three questions, the 22.1 m^/ha, a density of708 trees/ha, and an aver- datawerebroken into four groups: measurements age d.b.h. of20.2 cm.mDominant and codominant taken in the airin the thinned stand (AT), inthe trees averaged 15.1 in height, with live crown phloem inthe thinned stand (PT), in the air in the being 52 percent oftotal height. The adjacentun- unthinned stand (AU), andinthe phloem in the thinned stand had a basal area of37.0 m^/ha, a unthinned stand (PU). The measurements inAT density of1,090 trees/ha, and an average d.b.h. of and PT are fi-om the sametrees in the thinned 18.6 cm. Dominant and codominanttrees averaged Unthinned Stand Thinned Stand M M M 1'7l4IIrIIIIIIIIIII II1IIIIIIIII210I4I1IIIIIIII IIII II 23I4 Days — — Days — AU Day PUDay AUNight PUNight ATDay PTDay ATNight PTNight — Figure 1 Smoothed curves showing treatmentdifferences for61-daytime period between June 23 and August22, 1987. Sets of curves representthe hottestand coldestpartofthe day. First letterofthe code refersto position oftemperature probe (A = airand P = phloem) and second letter refers totreatment (T= thinned and U = unthinned). 3 Air Temperatures Phloem Temperatures 174 204 234 174 204 234 — Days — — Days — AT Day AU Day ATNight AUNight PTDay PU Day PTNight PU Night — Figure2 Smoothed curves showing temperature differencesfor61-daytime period between June 23 and August22, 1987. Sets ofcurves representthe hottestand coldest part ofthe day. First letterofthe code refersto position oftemperature probe (A = airand P = phloem) and second letterrefersto treatment (T=thinned and U = unthinned). Night Temperatures Afternoon Temperatures Days — Days — AT PT AU PU PT PT AU - PU — Figure3 Smoothed curves showingtemperaturedifferences ofthe hottestand coldest part ofthe dayfora 61-daytime period between June 23 andAugust22, 1987. First letterofthe code refers to position oftemperature probe (A = airand P =phloem) and second letterrefers totreatment (T= thinned and U = unthinned). 4 stand and, thus, are paired. Likewise thoseinAU Table 1—Percentagesoftotal day-to-dayvariability and PU are from the same trees inthe \inthinned explained byeigenvalues in principal component stand and are paired. These groupings are main- analysis. tainedthroughoutthe analyses and resulting Percentofday-to-dayvariability graphs. Datagroup Night Day Cumulative A fourth question ofinterestregards the tempera- AT 66 25 91 ture increase from the coolest part ofthe nightto (thinned stand the warmest part ofthe afternoon. These increases on surface) are presented in figure 4. Two variables (principal components) accounted PT 68 25 93 for high percentages ofthe variability in tempera- (thinned stand ture from day to day (see table 1). Therefore, only in phloem) these two variables were studied. The firstis the AU 69 24 93 warmestpartofthe afternoon(1:00 p.m. to6:00 p.m.), (unthinned stand whichis referred to as the "day"variable, andthe on surface) secondis the coolest partofthe night (1:00 a.m. to PU 69 24 93 9:00 a.m.), which is the "night" variable. The high (unthinned stand percentages shown intable 1 indicate that ifthe ac- in phloem) tual principal component had been used, complete independence would have resulted. However, this would have requiredusing a different weighted mean for each ofthe four datagroups. Because the comparable. Figures 1 to 4 contain the smoothed weights (coefficients from the principal components values ofthe day and night means. analysis) were quite similar for the hours specified Afinal point ofinterestwas the development of in"day" and also in "night," unweighted means for models to predict phloem temperature from bark the appropriate hours were used in all analyses. surfacetemperature. The estimates ofvariance for This makes the variables from the fourdata groups the fitted models were based on means ofseveral trees. Because we wishedto predict phloem temperature for anindividual tree and because the vari- ance is larger for an individual tree than for a mean Temperature Increase During Day ofseveral trees, we inflated these variances to re- flect proper prediction intervals on an individual treebasis. These intervals are presentedinfigure 5. The resultingmodels (table 2) show there is a strongcorrelation (r^> 0.98) between bark surface and phloem temperatures. — Thinned vs. Unthinned Temperature curves for the 61-day period were very similarbetween the thinned and unthinned stands (fig. 1). Similar patterns were obtained forboth the hottest and coldest part ofthe day forboth the thinned andunthinned stands. Subsurface (phloem) temperatures were similarto those onthe bark surface (air). Inthe unthinned stand the airtemperature was higher than the phloem temperatiire duringthe day 234 with a reversal occurring at night. Inthe thinned stand, the same general condition held with the exceptions that for most ofthe month — ofJulythe phloem temperature duringthe day was Figure4 Smoothed curves showing warmerthanthe airtemperature. This phenom- temperature increasesfrom the coldestto enon probably canbe attributed, in part, to overcast the hottest partofthedayfora 61-daytime skies duringthis period. The thinned stand was period between June 23 and August22, more open; therefore, atmospheric conditions could 1987. First letterofthe code refersto position oftemperature probe (A = airand have amore subtle effect on the temperatures, such P = phloem) and second letter refers to as increase in radiant warming. Another condition treatment (T= thinned and U = unthinned). in the thinned stand was the larger difference 5 Predicted Phloem Temperature Predicted Phloem Temperature Thinned Stand during Day Thinned Stand at Night 30 30 T- c: 25 25 - ro 20 2 QCC)L 0Q).20 c E <D 15 15 - \- 1- E(oU 10 E(oU 10 - sz 5 5 D. Q. - 10 15 20 25 30 10 15 20 25 30 AirTemperature AirTemperature Predicted Phloem Temperature Predicted Phloem Temperature Unthinned Stand during Day Unthinned Stand at Night 30 30 T- 25 2 20 Q. E a> 15 h- E 10 o JZ 5 Q. - 10 15 20 30 10 15 20 25 30 AirTemperature AirTemperature — Figure5 Plots ofvarianceassociated with modelsdeveloped to predictphloem temperature from barksurface temperature. Treatment/time represented are (A) thinned/daytime, (B) unthinned/ daytime, (C) thinned/nighttime, and (D) unthinned/nighttime. — Table2 Equationsthatpredictphloem temperaturefrom between phloem and airtemperatures at night. barksurfacetemperature forfourtreatment/time This was due to a consistently higher night-time situations. phloem temperature inthe thinned stand. Overall, the measuredtemperature inthe thinned stands Treatment Predictiveequation was warmerby about 1 °C. This difference does Thinned y= 0.81324 + 0.96285*ear/cSt;r 0.9871 not appearlarge; however, ifaccumulated over a night period oftime itcould have an effect on beetle Thinned 9= 0.81383 + 0.9534Q*BarkSur 0.9798 development. day — Airvs. Phloem Graphs oftemperature inthe Unthinned y= 0.28213 + 0.90000*ear/cStyr 0.9986 air displayed similar patterns in both day and night night measiu'ements (fig. 2). The curves for nighttem- Unthinned y==0.4105 + 0.96250*ea/-/cSur 0.9973 perature contain smoother and less volatile pat- day terns. Nevertheless, the same general patterns pre- vail in both sets ofmeasurements. While the night 6