Arkansas loblolly pine families growing on a droughty site in southeastern Oklahoma

Forest Ecology and Management 194 (2004) 83–94

Productivity, crown architecture, and gas exchange of North Carolina and Oklahoma/Arkansas loblolly pine families growing on a droughty site in southeastern Oklahoma$ Michael A. Blaziera,*, Thomas C. Hennesseya, Thomas B. Lyncha, Robert F. Wittwera, Mark E. Paytonb a b

Department of Forestry, Oklahoma State University, Stillwater, OK 74078, USA Department of Statistics, Oklahoma State University, Stillwater, OK 74078, USA

Received 7 June 2003; received in revised form 7 September 2003; accepted 4 February 2004

Abstract Biomass production, crown architecture, leaf gas exchange, and specific gravity of North Carolina Coastal (NCC) and local Oklahoma/Arkansas (O/A) families of 15-year-old loblolly pine (Pinus taeda L.) were examined on an excessively droughty site in southeastern Oklahoma. The O/A family produced more branch and foliage biomass per hectare than the NCC family, but the two families produced equivalent amounts of stem biomass per hectare. The O/A family achieved its greater branch and foliage biomass production by virtue of supporting a higher number of live branches per tree. Photosynthetic capacity, needle stomatal conductance of water vapor, transpiration rates, and intrinsic water use efficiency, which were each measured periodically throughout the study, were similar for the two families, as were stand densities and survival. Specific gravity and DBH were similar for the two families, but trees of the NCC family were significantly taller than those of the O/A source. Given the comparable gas exchange characteristics and stem biomass production of the two families on this droughty site as well as the NCC family’s production of wood with fewer knots, we conclude that planting of the NCC family in favor of the local family on excessively drained soils in the northwestern portion of the loblolly pine range may be a justifiable management option. # 2004 Elsevier B.V. All rights reserved. Keywords: Pinus taeda; Biomass production; Family testing; Biomass partitioning

1. Introduction Families from the eastern portion of the natural range of loblolly pine (Pinus taeda L.), particularly $

This work was approved for publication by the director of the Oklahoma Agricultural Experiment Station, and supported in part under project OKLO 2120. * Corresponding author. Present address: Hill Farm Research Station, Louisiana State University Agricultural Center, Homer, LA 71040, USA. Tel.: þ1-318-927-2578; fax: þ1-318-927-9505. E-mail address: [email protected] (M.A. Blazier).

those from North Carolina Coastal (NCC) provenances, outgrow local families when planted in the western portion of the natural range of loblolly pine. Many seed source studies have shown that when planted in common environments in southern Oklahoma and Arkansas, NCC seed sources frequently produce taller, faster-growing trees than those produced by local Oklahoma/Arkansas (O/A) seed sources (Wells and Wakeley, 1966; Wells and Lambeth, 1983; Lambeth et al., 1984; Sprintz et al., 1989; Harrington, 1991; Douglass et al., 1993). The height growth superiority of

0378-1127/$ – see front matter # 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.foreco.2004.02.014

84

M.A. Blazier et al. / Forest Ecology and Management 194 (2004) 83–94

NCC seed sources has been attributed in part to their ability to grow later into the growing season and to produce additional growth flushes (Harrington, 1991; Jayawickrama et al., 1995, 1998). Several authors have suggested that height growth is strongly linked to stem volume production and growth form (Perry et al., 1966; McCutchan, 1983; Sprintz, 1987). The implication of the relationship between height growth, volume, and form is that NCC provenances, given their superior height growth, may yield a greater quantity and quality of wood and fiber than western seed sources. Douglass et al. (1993) have identified several NCC families sufficiently adapted to the western portion of the loblolly pine range that produce trees with higher volumes and straighter stems than trees from Oklahoma and Arkansas families. Talbert and Strub (1987) found that eastern loblolly pine families grown in southwestern Arkansas allocated stand-level volume to fewer trees per acre than western families. This phenomenon was attributable to stands of eastern families growing at a faster rate and consequently reaching a maximum size-density trajectory at a younger age. Thus, stands of eastern seed sources commonly produce larger, more economically valuable trees than those produced by stands of western seed sources (Wells and Lambeth, 1983; Talbert and Strub, 1987; Tauer and Loo-Dinkins, 1990). Coastal North Carolina families are widely planted across the western portion of loblolly pine’s natural range in favor of local seed sources, but some discretion is used when selecting sites for planting non-local sources. Since the early 1980s, forest managers have been planting faster-growing NCC families in Oklahoma and Arkansas to increase stand productivity (Wells, 1983; Wells and Lambeth, 1983; Duba et al., 1984; Lambeth et al., 1984). Stands of NCC trees are expected to yield a 20–30% increase in volume at harvest when compared to stands of local O/A families (Lambeth et al., 1984). However, estimates of wood and fiber gains from NCC sources may be somewhat overestimated since the specific gravity of NCC seed sources is frequently lower than that of O/A sources (Byram and Lowe, 1988; Tauer and Loo-Dinkins, 1990). Another possible factor offsetting volume gains from planting of non-local families may be mortality. When grown in droughty common environments beyond the western extremity of loblolly pine’s natural

range, O/A families have demonstrated better survival rates than NCC sources (Long, 1980). To minimize the risk of drought-related mortality, forest managers have restricted planting of NCC stock to soils where mortality of NCC trees would not exceed that of O/A sources, i.e. on soils with adequate moisture-holding capacity. Thus, decisions regarding the planting of NCC families in Oklahoma and Arkansas are guided by a compromise between an anticipated gain in growth rates and a perceived risk of loss to mortality (Lambeth et al., 1984). Planting guidelines for NCC families in the western fringe of the loblolly pine range may be overly cautious. Although guidelines that discourage the planting of NCC families on droughty soils were developed to minimize mortality, Lambeth et al. (1984) have suggested that such planting recommendations are somewhat conservative. They observed that even during the worst drought ever recorded for the southern Oklahoma/Arkansas area some NCC stands on soils with lower moisture storage capacity than planting recommendations allowed exhibited mortality rates similar to stands of local O/A sources. In addition, it is possible that higher mortality rates of fastergrowing eastern sources are due to earlier onset of competition-induced mortality rather than maladaptation to the planting environment (Zobel, 1979; Wells and Lambeth, 1983; Talbert and Strub, 1987; Kung, 1989; Tauer and Loo-Dinkins, 1990). Potential for improving the productivity of excessively drained sites in the western portion of the loblolly pine range may exist. Identification of productive families that are sufficiently adapted to harsh sites is an important pursuit since planting of more productive planting stock increases the profitability of managing such sites (Zobel, 1979). Since NCC families are often more productive than O/A families, planting NCC stock on excessively drained sites in the western edge of the loblolly pine range could allow forestry to be more profitable on what are now considered marginal sites. The need to monitor growth and survival of stands of NCC and O/A seed sources in order to verify the accuracy of the system of matching these seed sources to proper soils in the southern Oklahoma/Arkansas area has been stressed by Lambeth et al. (1984). The objectives of this study were to determine whether on an excessively drained southeastern Oklahoma site: (1) biomass production of trees

M.A. Blazier et al. / Forest Ecology and Management 194 (2004) 83–94

common in the region from June to October. Average monthly precipitation in late summer is substantially less than regional potential evapotranspiration during the same period (Fig. 1), which typically results in rapid depletion of soil moisture (Lambeth et al., 1984). The climate of the region is hot (27–38 8C) and humid through much of the growing season. In all, nine 0.20 ha plots were examined in this study; five of these plots were planted with the NCC family and four were planted with the O/A family. The plots were arranged in a completely randomized design at two adjacent locations within the plantation. Five plots (3 plots with NCC family, 2 with O/A family) were established at one location and four plots (2 plots of NCC family, 2 with O/A family) at the other location; these two locations differed slightly in slope. All trees were planted on a 2:4 m  2:4 m spacing.

from a NCC family is greater than that of trees from an O/A family, (2) crown architecture (growth form) of NCC trees is superior to that of O/A trees, and (3) NCC and O/A families are similarly acclimated to the site in terms of gas exchange characteristics.

2. Materials and methods 2.1. Site description A complete description of the study site is given by Blazier et al. (2002); key characteristics of the site will be discussed here. The site was a 15-year-old loblolly pine plantation established in 1983 with NCC 8-01 and O/A mix 4213 families in McCurtain County, OK (34.067438N, 94.709618W). The site was subsoiled prior to planting to improve seedling survival. The soil is classified as a Goldston–Carnasaw–Sacul association, which is predominately a gravelly silt loam soil (USDA, 1974). Water holding capacity of this soil is 7 cm (NRCS, 2000), which is well below the 11 cm water holding capacity that Lambeth et al. (1984) suggested as a limit to the planting of NCC material in the Oklahoma/Arkansas region. Droughts are

2.2. Climate

18 Avg precip PET

16

16

14

14

12

12

10

10

8

8

6

6

4

4 April

May

June

July

Aug

Sept

Thornthwaite potential evapotranspiration (cm)

Temperature, humidity, precipitation, barometric pressure, wind speed, and solar radiation were monitored throughout the 1998–1999 growing season at a weather station located approximately 8 km from the study area. The station is part of the Oklahoma

18

Average regional precipitation (cm)

85

Oct

Month Fig. 1. Average growing season precipitation and potential evapotranspiration (PET) for southeastern Oklahoma.

86

M.A. Blazier et al. / Forest Ecology and Management 194 (2004) 83–94

Climatological Survey statewide Mesonet network (OCS Mesonet, 1998, 1999). 2.3. Biomass In May 1998, 15 trees per family were destructively sampled to develop regression equations for predicting the aboveground biomass components of the two families. A complete description of the criteria used in sample tree selection to ensure that the 15 sample trees per family were representative of the range of tree diameters present on the site has been described in Blazier et al. (2002). Several traits were measured for regression equation development on each sample tree. First, stem diameter at breast height (DBH), height of every dead branch from the base of the tree, and the diameter of every dead branch were determined. Trees were then felled (with ropes used to slow their descent, thus minimizing loss of branches), and the tree’s total height, height of every live branch from the base of the tree, and diameter of every live branch were measured. The bole was sawn into 1-m sections after all branches had been measured and removed. The bole sections were weighed in the field, and the insideand outside-bark diameters of the top and bottom of each stem section were measured. Disks approximately 2.54 cm in thickness were then cut from the top and bottom of each stem section and weighed in the field.

Biomass and branch relationships based on data collected from the 15 sample trees per family were used to compare biomass accumulation and crown architecture of the two families. The NLIN procedure of the SAS System (SAS Institute Inc., Cary, NC) was used to fit the models to the data. Separate sets of coefficients were estimated for the two families to best isolate family effects on biomass production and crown architecture. The resulting equations were evaluated on the basis of standard fit statistics and graphical examination of the residuals. Regression equations for predicting a tree’s total branch and foliage dry mass were developed for each family. First, the branch-level dry masses of branch and foliage on the sample trees were predicted using equations developed in a study previously conducted on this site (Blazier et al., 2002). The branch and foliage mass predictions were summed for each sample tree to obtain the total branch and foliage mass of the sample trees. The total dry branch mass, height, and DBH data from each sample tree were used to estimate coefficients of a nonlinear model that predicted the total dry branch biomass ðbranchwood þ barkÞ of a tree using DBH and total height as independent variables (Table 1). Similarly, a nonlinear model that predicted the total dry mass of foliage using DBH and total height as independent variables was fitted to total dry foliage mass, height, and DBH data (Table 1). Models for predicting stem dry mass were also developed. First, the disks taken from the 1 m stem

Table 1 Regression coefficients for equations used in prediction of components of aboveground biomass of 15-year-old trees from disparate seed sources grown on a droughty study site in southeastern Oklahoma, 1998a Dependent variableb

NCBRDMS OABRDMS NCFOLDMS OAFOLDMS NCSTMDMS OASTMDMS

Statisticsc

Parameter estimates b0

b1

b2

FI

S.E.

7.445 4.411 73.836 43.399 36.629 40.355

3.367 3.392 2.832 3.348 1.825 1.692

1.076 0.839 1.678 2.048 0.819 0.933

0.94 0.92 0.97 0.97 0.99 0.98

0.129 0.184 125.1 167.1 628.8 1401.0

a The model is Y ¼ b0  Db1  H b2 , where Y is the predicted dry mass (g) of branches, foliage, or stem of one tree, D the diameter outside bark (cm) at 1.37 m (4.5 ft), H the total tree height (m), bi the coefficients estimated from data, i¼1, 2. b NCBRDMS: dry mass of branches on coastal North Carolina trees; OABRDMS: dry mass of branches on Oklahoma/Arkansas trees; NCFOLDMS: dry mass of foliage on coastal North Carolina trees; OAFOLDMS: dry mass of foliage on Oklahoma/Arkansas trees; NCSTMDMS: dry mass of stem of coastal North Carolina trees; OASTMDMS: dry mass of stem of Oklahoma/Arkansas trees. c FI: fit index; S.E.: standard error of the estimate.

M.A. Blazier et al. / Forest Ecology and Management 194 (2004) 83–94

sections of sample trees were oven-dried at 80 8C to a constant weight. The disk dry weights were combined with the fresh weights of the disks determined in the field to calculate the dry:fresh weight ratios of the disks at the time of harvest. To estimate the dry:fresh weight ratio for each 1 m stem section, a weighted average of the dry:fresh weight ratios of the two disks collected from each stem section was calculated, with disk area used as the weighting factor. The fresh weights of each 1 m stem section were then multiplied by its weighted average dry:fresh weight ratio to ascertain the dry weights of each stem section. The dry weights of the stem sections were then summed for each sample tree to determine total stem dry mass. Stem dry mass, total height, and DBH measurements for all sample trees were used to fit a nonlinear model that predicted the total dry mass of a stem using the DBH and height of a tree as independent variables (Table 1). In March 1998, 0.04 ha plots were established within each 0.20 ha plot, and the DBH and height of each tree within the 0.04 ha plots were measured. These measurements were used as inputs in the regression equations to predict branch, foliage, and stem dry mass of every tree within the 0.04 ha plots. These treelevel biomass estimates were then summed for each 0.04 ha plot to yield the total branch, foliage and stem masses of the plots. Using the appropriate expansion factors, the total branch, foliage, and stem mass estimates were converted to the per hectare basis. Several additional measurements were taken on each of the destructively harvested trees to provide information on wood quality. The number of live and dead branches was counted on each of the sample trees and expressed as the number of knots per meter of stem length. In addition, wood specific gravity (g dry wood cm3 saturated volume) was measured on all disks removed at 1 m above the base of the stem. The 15 trees per family from which the specific gravity data were gathered is close to a suggested optimum sample size for specific gravity analysis (Lowerts and Zoerb, 1989). The dry weight of each disk was obtained through the procedures previously mentioned. Each disk was then soaked in water until saturation (which was verified using a Delmhorst J-2000 wood moisture meter); disk volume at 100% moisture content was then ascertained by displacement. Specific gravity was then calculated as

87

the oven-dry weight of the wood divided by the weight of the displaced water (Tauer and Loo-Dinkins, 1990; Haygreen and Bowyer, 1996). To augment the predictions of per hectare mass of aboveground biomass components derived from the regression equations, stand density and survival data were collected. Trees per hectare for each plot was calculated by expanding the tree count made in the 0.04 ha plots in March 1998. Survival of trees from 1998 to 1999 was obtained by comparing the tree tally made in the 0.04 ha plots in March 1998 to one made in April 1999. Mortality rates were calculated as the percentage of tagged trees in the 0.04 ha plots that were dead in the 1999 tally. 2.4. Leaf gas exchange and plant water relations Needle gas exchange under optimal conditions and water contents of the two families were measured to supplement biomass, specific gravity, and stand density data. Light-saturated net photosynthesis (Amax), needle conductance to water vapor (gwv), transpiration (Et), and needle water potential (Cn) were measured in July, September, and November 1998 and in February 1999. One tree was sampled from each plot for all measurements. Trees were selected for observation on the basis of representative size, crown form, and absence of insect, disease, drought, or ice damage. On each sampled tree, a subsample of one branch per crown third (upper, middle, and lower) was harvested for measurement. Immediately after branches were detached from the stem, they were placed upright in a sealed container of water and measured within 5 min of branch excision to optimize accuracy of measurements (Ginn et al., 1991; Cregg, Personal communication). Excised branches remained in the water containers during gas exchange measurement. Gas exchange measurements, which were paired by family to reduce variability in data attributable to time of measurement collection, were collected simultaneously using two CO2 gas analyzers (model CI-301 PS, CID Inc.) in an open-system configuration. The gas analyzers were equipped with adjustable light sources (model CI-301 LA, CID Inc.) and CO2 control units. On each sampled branch, three fascicles of needles (9 needles total) from the first flush of growth produced in 1998 were measured. Photosynthetic rates of the needles were measured repeatedly at 40-s

88

M.A. Blazier et al. / Forest Ecology and Management 194 (2004) 83–94

intervals with the gas analyzers until values became stable. Amax measurements were conducted between 800 and 1700 h Eastern Standard Time at saturating light (photosynthetically active radiation of 1800 mmol m2 s1) and constant CO2 concentrations (355 ppm) (Whitehead and Teskey, 1995; Liu and Teskey, 1995; Zhang et al., 1997). No measurements of Amax were collected once ambient temperatures exceeded 27 8C, as operational experience revealed that photosynthetic rates were reduced when temperatures reached 29–32 8C. Intrinsic water use efficiency (WUE) (mmol CO2 mmol1 H2O) was then calculated by dividing Amax by gwv (Cregg et al., 2000). Needle water potential (Cn) was measured concurrently with gas exchange measurements using a pressure chamber as described by Ritchie and Hinckley (1975). Needles adjacent to those measured for Amax and gwv were excised for Cn measurement. 2.5. Statistical analysis Analyses of all treatment effects were conducted by analysis of variance (ANOVA) using the MIXED procedure of the SAS System (SAS Institute Inc., Cary, NC). Various models were used in the analyses depending on the treatments being assessed, and they are discussed in detail below. When an ANOVA indicated significant treatment effects, least-square means were calculated and separated by the LSMEANS procedure and DIFF option in SAS. The DIFF option separated means by invoking a t-test to evaluate significant differences between the least-square treatment means, thus allowing multiple comparisons of the treatment means. Dry branch mass ha1, dry foliage mass ha1, dry stem mass ha1, total aboveground dry mass ha1, stem:foliage ratio, branch:stem ratio, trees ha1, stand basal area, percent mortality from 1998 to 1999, specific gravity, knots per stem, and the heights and DBHs of trees within the 0.04 ha plots were each analyzed as treatments arranged in a completely randomized design. Family and location within the plantation were the treatments, and plots were the experimental units receiving the treatments. Analysis of variance procedures were performed on a mixed model with family, location, and the interaction between family and location as fixed effects and plot as a random effect.

For analyses of family differences in Amax, gwv, Cn, Et, and WUE, the experiment was treated as a randomized complete block design. Since gas exchange observations were paired by family, paired plots were considered as blocks. Family was the treatment, and blocks were the experimental units that received the treatment. Since gas exchange measurements were taken for each crown third, crown levels were treated as subsampling units. Separate analyses of Amax, gwv, Cn, Et, and WUE were conducted for each month in which measurements were taken to determine whether family differences occurred during the growing season. Analysis of variance procedures were performed on a mixed model with the following effects: (1) family, (2) block, (3) crown level, and (4) all possible interactions between family, block, and crown level. Family and crown level were considered fixed effects; block and all interactions with block were considered random effects. Although Cn data collection was similar to that of the gas exchange measurements, only one pressure chamber was used, preventing pairing of the Cn measurements. Instead, family and location treatments were considered arranged in a completely randomized design for the analysis of Cn. Family and location within the plantation were the treatments, and plots were the experimental units receiving the treatments. Crown levels were considered subsampling units. Analysis of variance procedures were performed on a mixed model with the following effects: (1) family, (2) location, (3) crown level, (4) all possible interactions of family, location, and crown level, and (4) plot. Family, location, crown level, and all their possible interactions were considered fixed effects; plot was considered a random effect.

3. Results 3.1. Climatic The climate during the study was marked by a pronounced drought during July and August. For the 2 months combined, there was only 7.7 cm of precipitation, which was 65% lower than average (OCS Mesonet, 1998, 2001). Temperatures were high during these 2 months; there were 11 consecutive days in which mid-day air temperatures exceeded 38 8C.

M.A. Blazier et al. / Forest Ecology and Management 194 (2004) 83–94

For July and August combined, temperatures exceeded 38 8C for 21 days. However, the hot, dry summer was followed by a wet fall. From September to November, a total of 56.5 cm of rain fell, which was 32% greater than average (OCS Mesonet, 1998, 2001). The total precipitation for the February 1998–February 1999 period was 124 cm, which was 10% lower than average (OCS Mesonet, 1998, 1999, 2001). 3.2. Biomass Crown architecture of the NCC and O/A families differed significantly (Table 2). Branch and foliage dry masses ha1 were significantly greater ðP ¼ 0:01Þ for the O/A family than the NCC family. However, the families did not significantly differ in dry stem mass ha1 ðP ¼ 0:68Þ and total aboveground dry mass ha1 ðP ¼ 0:32Þ. The NCC family produced more stem biomass per unit of foliage, as evidenced by its significantly ðP ¼ 0:03Þ higher stem:foliage ratio. The O/A family was associated with significantly ðP < 0:0001Þ higher branch biomass per unit of stem biomass. The two families differed somewhat in wood quality and growth form (Table 2). Specific gravity did not differ for the two families ðP ¼ 0:62Þ, but the O/A family had a significantly greater ðP ¼ 0:03Þ number of knots per stem (Table 2). There was a significant difference between the families in total tree height; NCC trees were significantly taller ðP ¼ 0:01) than O/A trees. There was no significant difference ðP ¼ 0:46Þ in average DBH of the two families. The NCC and O/A families did not significantly differ in stand density or survival at age 15. The differences in trees ha1 and stand basal area for the two

89

families were not significant ðP ¼ 0:46  0:18Þ (Table 2). The two families also did not differ significantly ðP ¼ 0:24Þ in mortality rates from 1998 to 1999. The NCC family experienced a 2% mortality rate, and the O/A family had a 0.5% mortality rate. 3.3. Gas exchange and plant water relations In the analyses of Amax, no significant differences attributable to family were found for any month. Both families showed similar trends in Amax over the growing season (Table 3). During the hot and dry month of July, Amax values were at their lowest; in the much cooler and wetter September, Amax values were at their highest. Amax values dropped from September to November, but were slightly greater in February 1999. In November, a significant ðP ¼ 0:02Þ crown effect was found; for both families, the lower crown had significantly ðP ¼ 0:03  0:02Þ lower Amax than both the middle and upper crown levels (Table 3). No significant differences attributable to family were found in the analyses of Et, gwv, or Cn. Et values followed a trend similar to that exhibited by Amax values for most measurement periods (Table 3); however, there was a decrease in Et in February instead of a slight increase as in the Amax trends. As with the analysis of Amax, a significant ðP ¼ 0:01Þ crown effect was found in only November for Et (Table 3). The lower crown had significantly lower Et than both the middle ðP ¼ 0:002Þ and upper ðP ¼ 0:02Þ crown levels in November. Similar to the analyses of Amax and Et, a significant crown effect on gwv ðP ¼ 0:01Þ was found only in November (Table 3). Comparison of gwv least-square means showed that there were differences within the crown identical to those found

Table 2 Effect of family upon biomass production, crown architecture, wood quality, and survival of 15-year-old plantation-grown loblolly pine in southeastern Oklahoma Familya

NCC O/A

Aboveground dry mass (Mg ha1)

Total height (m)

DBH (cm)

12.3 a 11.6 b

17.7 a 11.8 b 5.3 b 16.1 a 15.9 a 7.0 a

Branch Foliage Stem

74.0 a 77.3 a

Total

Stem: Branch: Specific Knots Stand Stand foliage stem gravity per stem density basal area mass ratio mass ratio (g cm3) (#branches m1) (#trees ha1) (m2 ha1)

91.1 a 14.2 a 100.2 a 11.2 b

0.15 b 0.20 a

Means within a column followed by different letters differ significantly at P < 0:05. a NCC: North Carolina Coastal loblolly pine; O/A: Oklahoma/Arkansas loblolly pine.

0.47 a 0.46 a

52 b 65 a

1307 a 1412 a

33.6 a 37.2 a

90

Measurementa

July 1998 Low

September 1998 Middle

Upper

Low

Middle

November 1998 Upper

Low

Middle

February 1999 Upper

Low

Middle

Upper

NCC family Amax Et gwv Cn WUE

1.57 a 0.15 a 0.004 a 0.93 a 405.8 a

2.11 a 0.18 a 0.004 a 0.96 b 395.3 a

1.73 a 0.21 a 0.005 a 1.00 c 357.3 a

6.56 a 0.70 a 0.04 a 0.62 a 190.3 a

6.53 a 0.90 a 0.04 a 0.70 b 158.9 a

7.96 a 1.12 a 0.04 a 0.75 b 244.7 a

3.09 b 0.51 b 0.04 b 0.62 a 98.0 a

5.56 a 0.69 a 0.06 a 0.74 b 100.9 a

4.41 a 0.65 a 0.05 a 0.75 b 92.2 a

4.94 a 4.31 a 6.31 a 0a 0.20 a 0.20 a 0.002 a 0.001 a 0.02 a 1.17 b 0.98 a 1.14 b 2375.5 a 3099.5 a 1145.2 a

O/A family Amax Et gwv Cn WUE

2.67 a 0.11 a 0.004 a 0.88 a 802.2 a

3.02 a 0.18 a 0.006 a 0.96 b 909.5 a

3.59 a 0.12 a 0.005 a 1.00 c 1494.5 a

7.22 a 0.98 a 0.05 a 0.64 a 237.9 a

7.37 a 0.96 a 0.04 a 0.70 b 205.8 a

6.11 a 1.21 a 0.05 a 0.58 a 140.9 a

1.82 b 0.46 b 0.04 b 0.77 a 58.1 a

4.90 a 0.84 a 0.06 a 0.72 a 83.4 a

3.93 a 0.68 a 0.05 a 0.72 a 79.9 a

5.39 a 0a 0.01 a 1.05 a 1996.1 a

4.69 a 0.07 a 0.01 a 0.92 a 1751.2 a

5.03 a 0.13 a 0.01 a 1.10 b 823.5 a

No significant differences ðP < 0:05Þ attributable to family were found. Within each quarter, means followed by a different letter are significantly ðP < 0:05Þ different. a Amax: light-saturated net photosynthetic rate (mmol m2 s1); Et: transpiration rate (mmol m2 s-1); gwv: needle conductance to water vapor (m2 s mol1); Cn: needle water potential (MPa); WUE: intrinsic water use efficiency (mmol mmol1).

M.A. Blazier et al. / Forest Ecology and Management 194 (2004) 83–94

Table 3 Quarterly gas exchange measurements taken in the upper, middle, and lower crown levels of North Carolina Coastal (NCC) and Oklahoma/Arkansas (O/A) families of loblolly pine grown on an excessively drained site in southeastern Oklahoma

M.A. Blazier et al. / Forest Ecology and Management 194 (2004) 83–94

for both Amax and Et in November, with the lower crown having significantly lower gwv than the middle ðP ¼ 0:001Þ and upper ðP ¼ 0:04Þ crown. Cn values were relatively low in July, plateaued at a higher level from September to November, then declined again by February 1999 (Table 3). Interestingly, the Cn values were lower in February than in July 1998 even though February had more precipitation than July. Analyses of Cn values by month revealed no significant family effects on Cn; however, there were significant crown or crown  family effects for every month ðP ¼ 0:01  0:01Þ in which measurements were taken. Generally, Cn of the lower crown was greater than that of the middle and upper crown levels. In the analyses of WUE, no significant differences attributable to family were found for any month, although the difference in WUE was marginally non-significant ðP ¼ 0:12Þ in July (Table 3). During months with relatively higher precipitation, WUE values were lower. In months with lower precipitation, WUE values were higher. Although the difference was marginally non-significant, the average WUE of the NCC family was lower than that of the O/A source in the dry month of July. No significant crown level and family  crown level effects were detected.

4. Discussion The two families differed in aboveground biomass production and crown architecture. The well-documented (Wells and Wakeley, 1966; Wells and Lambeth, 1983; Lambeth et al., 1984; Harrington, 1991) height growth superiority of the NCC family relative to the O/A family was maintained on this droughty site. However, the O/A family produced more foliage and branch biomass ha1 than the NCC family. This finding was somewhat unexpected, since a study conducted previously on this site determined that on an individual-branch basis NCC trees supported 29% more foliage per branch than O/A trees (Blazier et al., 2002). It appears that the greater branch and foliage biomass production of the O/A family is attributable to its branch production per tree rather than its foliage production per branch. The O/A family had approximately 20% more branches per tree than the NCC family; the O/A family in turn had 26% more branch mass ha1 and 24% more foliage mass ha1 than the

91

NCC family. Campbell et al. (1995) found that selection for small-crown, fast-growth characteristics in loblolly pine breeding programs had been effective in improving juvenile log quality of North Carolina families. These growth characteristics were exhibited even when North Carolina stock was planted in Mississippi. They found that relative to local Mississippi sources, the NCC stock had a lower number of branches per tree. The results of our study are consistent with their findings in that NCC trees had fewer branches per tree than the local O/A family. The two families did not differ, however, in their production of stem biomass ha1. As such, the NCC family partitioned more biomass into its stem component relative to branches, since it produced an amount of stem biomass comparable to that produced by the O/A family with significantly less branch biomass. The NCC family was also more efficient in producing stem biomass per unit of foliage biomass. In summary, the analyses of the aboveground biomass components suggest that the superior growth form (in terms of greater height growth and fewer branches) of NCC trees relative to O/A trees is maintained even on a soil with very low water-holding capacity. Since specific gravity measurements were equivalent for both families, wood properties (such as strength and stiffness) would likely be similar for the two families. In addition, pulp yields per volume would likely be similar for the two families (Byram and Lowe, 1988; Haygreen and Bowyer, 1996). Our finding that specific gravity did not differ between the two families on this excessively droughty site contrasts with studies that found eastern families produced lower specific gravities than local sources when planted in Arkansas and Oklahoma (Byram and Lowe, 1988; Tauer and Loo-Dinkins, 1990). It is possible that these contradictory results are attributable to edaphic and climatic attributes of our site, which may have been stressful enough to mask genetic effects on specific gravity. It has been demonstrated that differences in the latitude of seed source origin determine the timing of spring growth initiation, with southern sources initiating growth earlier in the spring than northern sources (Byram and Lowe, 1988). NCC and O/A provenances originate from areas that lie along virtually the same latitude and have similar spring rainfall patterns, so it is likely that earlywood production initiates at about the same time for the two families.

92

M.A. Blazier et al. / Forest Ecology and Management 194 (2004) 83–94

It has also been shown that summer drought initiates latewood production (Byram and Lowe, 1988; Cregg et al., 1988). The onset of drought (which is common in our test area) may be especially severe on this site since low rainfall and high temperature effects are accentuated by the poor moisture-holding ability of the soil. As a result, both families likely make the transition to latewood production at virtually the same time, resulting in similar specific gravities. The lack of significant differences in light-saturated photosynthetic rates (Amax) suggests that the photosynthetic apparatus of the NCC and O/A families likely had similar characteristics throughout the growing season (Schaberg et al., 1998). Needle physiology of the two families was possibly similar, so foliage of the two families may have had the same potential for producing photosynthate and thereby, by inference, for producing biomass. However, biomass growth potential of the two families would also be influenced by foliar biomass and the photosynthetic rates per unit of foliage in addition to Amax (Teskey et al., 1987). In this study, the O/A family had more foliage biomass than the NCC family, and thus potentially greater photosynthetic capacity. However, it has been demonstrated that trees with greater leaf biomass have greater mutual leaf shading, which decreases the average amount of photosynthesis per unit leaf area (Burkhalter et al., 1967; Boltz et al., 1986). Thus, although the O/A family had greater foliar biomass, its photosynthetic rates per unit leaf area were likely moderated at the crown level by increased mutual leaf shading. The equivalent survival (as inferred by stand density, basal area, and 1998–1999 mortality rate similarities) of the two families on this site contradicts some studies that suggest eastern families planted on highly droughty sites in the western edge of the loblolly pine range would have greater mortality than local families (Long, 1980; Lambeth et al., 1984). The lack of any significant family differences in plant water relations (Cn, Et, gwv, WUE) suggests that the two families are similar in their abilities to take up water on this excessively drained site. It has been demonstrated that trees with greater drought resistance will exhibit a more rapid decline in gwv than less resistant trees as soil moisture is depleted (Van Buijtenen et al., 1976; Bongarten and Teskey, 1986). The lack of gwv differences throughout the growing season in our study,

particularly in response to the drought that occurred in July, suggests that the NCC and O/A families are similarly drought resistant when both are grown in an area of recurring drought. Bongarten and Teskey (1986) tested drought resistance of seedlings from diverse origins (which included seedlings from both NCC and O/A families) and found no differences in gwv among the various seed lots when seedlings were exposed to recurring drought. They inferred that trees from disparate origins experience comparable drought conditioning when grown in areas with recurring drought. The gwv and Et responses of the two families in our study are consistent with their findings. Due to their higher foliage biomass, O/A trees likely had higher canopy-level stomatal conductance rates than the NCC trees. However, increased mutual shading in the larger crowns of O/A trees would possibly moderate differences in canopy-level conductance between the O/A and NCC families (Ewers et al., 2001). In addition to genetic factors, the subsoiling operation done at stand establishment likely improved survival of both families. Subsoiling improves the condition of shallow, rocky soils by breaking up soil layers, concentrating organic matter near seedlings, and improving soil aeration and water-holding capacity (Wittwer et al., 1986; Fallis and Duzan, 1995). Wittwer et al. (1986) proposed that drought survival could be improved by the increases in soil moisture gained by subsoiling. Fallis and Duzan (1995) found that subsoiling was particularly effective in improving early survival and height growth up to age 19 on the harsher, drier sites of southeastern Oklahoma.

5. Conclusions The rising demand for wood products and the shrinking land base from which wood can be harvested necessitate planting the most productive planting stock on lands being managed for wood products. This study has provided some evidence that forest managers may have more flexibility selecting sites to be planted with NCC loblolly pine in the southeast Oklahoma–southwest Arkansas region than has been traditionally considered. However, it must be stressed that before broad claims of the productivity of NCC loblolly pine relative to O/A loblolly pine on droughty

M.A. Blazier et al. / Forest Ecology and Management 194 (2004) 83–94

sites can be made with greater certainty, a wider variety of NCC and O/A families must be tested throughout a rotation on a larger number of excessively drained sites. Given the greater total aboveground biomass production of the O/A family, it could be advantageous to plant O/A stock on droughty sites when fiber yield is the management objective. If sawtimber production on droughty sites is desired, it may be preferable to plant NCC stock since its survival, specific gravity, and stem production ha1 was comparable to that of the local O/ A family and its stems tended to have significantly fewer knots.

Acknowledgements The authors are grateful to Ed Lorenzi, Bob Heinemann, Randy Holeman, Dennis Wilson, Walt Sanders, and Keith Anderson for their assistance in data collection. The authors also wish to thank Bert Cregg for his critical reading of this manuscript.

References Blazier, M.A., Hennessey, T.C., Lynch, T.B., Wittwer, R.F., 2002. Comparison of branch biomass relationships for North Carolina and Oklahoma/Arkansas loblolly pine seed sources growing in southeastern Oklahoma. For. Ecol. Manage. 159, 241–248. Boltz, B.A., Bongarten, B.C., Teskey, R.O., 1986. Seasonal patterns of net photosynthesis of loblolly pine from diverse origins. Can. J. For. Res. 16, 1063–1068. Bongarten, B.C., Teskey, R.O., 1986. Water relations of loblolly pine seedlings from diverse geographic origins. Tree Physiol. 1, 265–276. Burkhalter, A.P., Robertson, C.F., Reines, M., 1967. Variation in photosynthesis and respiration in southern pines. Ga. For. Res. Pap. 46, 1–5. Byram, T.D., Lowe, W.J., 1988. Specific gravity variation in a loblolly pine seed source study in the Western Gulf Region. For. Sci. 34, 798–803. Campbell, J.B., Land Jr., S.B., Duzan Jr., H.W., 1995. Influence of family and spacing on juvenile log quality of loblolly pine. In: Proceedings of the 23rd Southern Forest Tree Improvement Conference, June 20–22, 1995, Asheville, NC, pp. 24–31. Cregg, B.M., Personal communication. Assistant Professor of Horticulture, Michigan State University, Lansing, MI. Cregg, B.M., Dougherty, P.M., Hennessey, T.C., 1988. Growth and wood quality of young loblolly pine trees in relation to stand density and climatic factors. Can. J. For. Res. 18, 851–858.

93

Cregg, B.M., Olivas-Garcia, J.M., Hennessey, T.C., 2000. Provenance variation in carbon isotope discrimination of mature ponderosa pine trees at two locations in the Great Plains. Can. J. For. Res. 30, 428–439. Douglass, S.D., Williams, C.G., Lambeth, C.C., Huber, D.A., Burris, L.C., 1993. Family  environment interaction for sweep in local and nonlocal seed sources of loblolly pine. In: Proceedings of the 22nd Southern Forest Tree Improvement Conference, June 14–17, 1993, Atlanta, GA, pp. 318–326. Duba, S.E., Goggans, J.F., Patterson, R.M., 1984. Seed source testing of Alabama loblolly pine: implications for seed movements and tree improvement programs. South. J. Appl. For. 8, 189–193. Ewers, B.E., Oren, R., Phillips, N., Stro¨ mgren, M., Linder, S., 2001. Mean canopy stomatal conductance responses to water and nutrient availabilities in Picea abies and Pinus taeda. Tree Physiol. 21, 841–850. Fallis, F.G., Duzan Jr., H.W., 1995. Effects of ripping (deep subsoiling) on loblolly pine plantation establishment and growth—nineteen years later. In: Proceedings of the Eighth Biennial Southern Silvicultural Research Conference, November 1–3, 1994, Auburn, AL, pp. 525–529. Ginn, S.E., Seiler, J.R., Cazell, B.H., Kreh, R.E., 1991. Physiological and growth responses of eight-year-old loblolly pine stands to thinning. For. Sci. 37, 1030–1040. Harrington, C.A., 1991. Retrospective shoot growth analysis for three seed sources of loblolly pine. Can. J. For. Res. 21, 306– 317. Haygreen, J.G., Bowyer, J.L., 1996. Specific gravity and density. In: Forest Products and Wood Science: An Introduction, 3rd ed. Iowa State University Press, Ames, IA, pp. 191–212. Jayawickrama, K.J.S., McKeand, S.E., Jett, J.B., 1995. Phenological variation in height and diameter growth in provenances and families of loblolly pine. In: Proceedings of the 23rd Southern Forest Tree Improvement Conference, June 20–22, 1995, Asheville, NC, pp. 33–36. Jayawickrama, K.J.S., McKeand, S.E., Jett, J.B., 1998. Phenological variation in height and diameter growth in provenances and families of loblolly pine. New For. 16, 11–25. Kung, F.H., 1989. Optimal age for selecting loblolly pine seed sources. In: Proceedings of the 20th Southern Forest Tree Improvement Conference, June 26–30, 1989, Charleston, SC, pp. 302–308. Lambeth, C.C., Dougherty, P.M., Gladstone, W.T., McCullough, R.B., Wells, O.O., 1984. Large-scale planting of North Carolina loblolly pine in Arkansas and Oklahoma: a case of gain versus risk. J. For. 82, 736–741. Liu, S., Teskey, R.O., 1995. Responses of foliar gas exchange to long-term elevated CO2 concentrations in mature loblolly pine trees. Tree Physiol. 15, 351–359. Long, E.M., 1980. Texas and Louisiana loblolly pine study confirms importance of local seed sources. South. J. Appl. For. 4, 127–132. Lowerts, G.A., Zoerb, M.H., 1989. Variation of slash pine wood specific gravity in different progeny test environments. In: Proceedings of the 20th Southern Forest Tree Improvement Conference, June 26–30, 1989, Charleston, SC, pp. 253–241.

94

M.A. Blazier et al. / Forest Ecology and Management 194 (2004) 83–94

McCutchan, B.G., 1983. Estimation of gain in form, quality and volumetric traits of American sycamore in response to the selection for tree dry weight. In: Proceedings of the 17th Southern Forest Tree Improvement Conference, June 7–9, 1983, Athens, GA, pp. 209–219. Natural Resources Conservation Service (NRCS), 2000. Soils descriptions, nontechnical. OK-FOTG Notice 450-110. Oklahoma Climatological Survey Mesonet (OCS Mesonet), 2001. Normal (1971–2000) monthly precipitation in inches. Retrieved February 6, 2003 from http://climate.ocs.ou.edu/ normals_extremes.html. Oklahoma Climatological Survey Mesonet (OCS Mesonet), 1999. University of Oklahoma, Norman. Retrieved March 30, 1999 from http://okmesonet.ocs.ou.edu/ Oklahoma Climatological Survey Mesonet (OCS Mesonet), 1998. University of Oklahoma, Norman. Retrieved February to November 1998 from http://okmesonet.ocs.ou.edu/. Perry, T.O., Chi-Wu, W., Schmitt, D., 1966. Height growth for loblolly pine provenances in relation to photoperiod and growing season. Silvae Genet. 15, 61–64. Ritchie, G.A., Hinckley, T.M., 1975. The pressure chamber as an instrument for ecological research. Adv. Ecol. Res. 9, 165–254. Schaberg, P.G., Shane, J.B., Cali, P.F., Donnelley, J.R., Strimbeck, G.R., 1998. Photosynthetic capacity of red spruce during winter. Tree Physiol. 18, 271–276. Sprintz, P.T., 1987. Effects of genetically improved stands on growth and yield principles. In: Proceedings of the 19th Southern Forest Tree Improvement Conference, June 16–18, 1987, College Station, TX, pp. 338–348. Sprintz, P.T., Talbert, C.B., Strub, M.R., 1989. Height–age trends from an Arkansas seed source study. For. Sci. 35, 677–691. Talbert, C.B., Strub, M.R., 1987. Dynamics of stand growth and yield over 29 years in a loblolly pine source trial in Arkansas. In: Proceedings of the 19th Southern Forest Tree Improvement Conference, June 16–18, 1987, College Station, TX, pp. 30–38.

Tauer, C.G., Loo-Dinkins, J.A., 1990. Seed source variation in specific gravity of loblolly pine grown in a common environment in Arkansas. For. Sci. 36, 1133–1145. Teskey, R.O., Bongarten, B.C., Cregg, B.M., Dougherty, P.M., Hennessey, T.C., 1987. Physiology and genetics of tree growth response to moisture and temperature stress: an examination of the characteristics of loblolly pine (Pinus taeda L.). Tree Physiol. 3, 41–61. United States Department of Agriculture (USDA), 1974. Soil survey of McCurtain County, Oklahoma. Soil Conservation Service, USDA, 99 pp. Van Buijtenen, J.P., Bilan, M.V., Zimmerman, R.H., 1976. Morphophysiological characteristics related to drought resistance in Pinus taeda. In: Cannell, M.G.R., Last, F.T. (Eds.), Tree Physiology and Yield Improvement, pp. 349–359. Wells, O.O., 1983. Southwide pine seed source study—loblolly pine at 25 years. South. J. Appl. For. 7, 63–70. Wells, O.O., Lambeth, C.C., 1983. Loblolly pine provenance test in southern Arkansas—25th year results. South. J. Appl. For. 7, 71–75. Wells, O.O., Wakeley, P.C., 1966. Geographic variation in survival, growth, and fusiform-rust infection of planted loblolly pine. For. Sci. Mongr. 11, 40. Whitehead, D., Teskey, R.O., 1995. Dynamic response of stomata to changing irradiance in loblolly pine (Pinus taeda L.). Tree Physiol. 15, 245–251. Wittwer, R.F., Dougherty, P.M., Cosby, D., 1986. Effects of ripping and herbicide site preparation treatments on loblolly pine seedling growth and survival. South. J. Appl. For. 10, 253–257. Zhang, S., Hennessey, T.C., Heinemann, R.A., 1997. Acclimation of loblolly pine (Pinus taeda) foliage to light intensity as related to leaf nitrogen availability. Can. J. For. Res. 27, 1032–1040. Zobel, B.J., 1979. Trends in forest management as influenced by tree improvement. In: Proceedings of the 15th Southern Forest Tree Improvement Conference, June 19–21, 1979, Mississippi State University, pp. 73–77.