P. H. Dunham and D. M. Bourgeois
    Biometrics Center Leader and Biometry Project Assistant, respectively,
          Westvaco Corporation, P.O. Box 1950, Summerville, SC 29484

ABSTRACT:  A study established in nine plantations to quantify the long-term
recovery of plantation-grown loblolly pine from damage by Hurricane Hugo was
installed in early 1990.  After three years, mortality of storm damaged trees
was 8.1%, with 2% of the trees being non-leaners, 8% had an initial lean
between 0 and 25 degrees, 15% were leaning between 25 and 50 degrees, 40% were
leaning between 50 and 75 degrees, and 35% of the dead trees were leaning at
least 75 degrees initially.  The initial average angle of mortality trees was
61 degrees.  Three-year diameter growth gradually decreased as initial average
angle increased.  On average, height growth increased as initial angle
increased up to a given angle, then decreased below the growth of non-leaners.
Initial angle significantly influenced diameter growth for the 3-year period
for all initial ages, but was a significant contributor to 3-year height
growth for initial ages 2 and 5 only.  For all leaning trees, the average
positive change in stem angle generally decreased as the age of the stand
increased, particularly for those stands initially older than age 5. Recovery,
however, did occur along all measured points on the stem, with the greatest
improvement occurring during the first year.


   Tropical storms and hurricanes strike the Gulf and Atlantic Coasts almost
annually (Touliatos and Roth, 1971).  On September 21, 1989, Hurricane Hugo
struck the Coastal Plain of South Carolina with sustained winds of
approximately 135 mph (estimated) and gusts up to 160 mph (Powell, 1990).
Based on aerial and ground surveys, 23 counties were judged to have
substantial forest damage (Sheffield and Thompson, 1992).  Monetary loss, if
the timber could have been salvaged for its highest product use, was estimated
at $10 billion (Spell, 1990).

   Westvaco Corporation, a major pulp, paper and packaging manufacturer,
suffered timber damage on 150,000 acres.  While the salvage of marketable
timber received the most immediate attention, relatively young loblolly
plantations also had extensive damage, generally in the form of stem lean.  A
review of the literature showed most of the published storm-related material
dealt with either damage assessment (e.g.  Van Hooser and Hedlund, 1969;
Trousdell et al., 1965) or changes in ecological succession (see Foster,
1988).  There was little in the literature on the recovery of existing trees
and stands from any storm damage, much less hurricanes.

   McKellar (1942) reported on the short-term recovery of 6- and 7-year-old
plantation-grown loblolly, slash and longleaf pines from ice damage in the
Piedmont of Georgia.  More recently, an unpublished Westvaco Forest Research
Report covered the 8-year response of ice damaged 10-year-old plantation-
grown loblolly in the Piedmont of Virginia (Bell, R. D. & P. H. Dunham.  1987.
Response of 10-year-old plantation-grown loblolly pine subjected to ice
damage.  Westvaco Forest Science Laboratory - Biometrics Group, Res. Report
No. 65.  17 p.).  In both cases, individual pines showed partial to complete
recovery, depending on the amount of damage and the response variable

   Brewer and Linnartz (1973) conducted a 1-year recovery study of 26 11-year-
old plantation-grown loblolly pine trees damaged by Hurricane Camille.  Trees
leaning less than 45 degrees from vertical recovered to some extent while
trees leaning 15 to 25 degrees straightened to the point where the lower
portion of the bole had returned to vertical.  The authors recommended
salvaging any trees leaning more than 45 degrees as these trees are not likely
to recover substantially.

   Their results indicate that at least for one age class, plantation-grown
loblolly pine will recover to some extent, depending on the initial lean after
the storm.  Long-term responses to hurricane damage over a range of ages are
unknown.  Guidelines are needed to determine which plantations will recover
and which to salvage or replant.  Growth and yield of damaged plantations need
to be quantified to determine if and how our models should be modified to
reflect these conditions.

   The study has three specific objectives:  (i) to quantify individual
loblolly tree and stand responses to hurricane damage; (ii) to develop
operational guidelines for determining the future status of damaged loblolly
plantations; and (iii) to publish the results in the forestry literature.
This paper reports the recovery and growth responses three years after
Hurricane Hugo.


Field Location

   Nine existing Westvaco Forest Research study sites, all on the coastal
plain, were selected on the range and degree of storm damage present.  The
majority of the studies were age 5 or younger based on the assumption that
stands much older than 5 could be salvaged.  Three older stands (ages 8, 10
and 20) are being followed, primarily for survival information.

   The original objectives of the studies ran the gamut of applied forestry
research.  Three locations were genetics studies, three locations test various
site-preparation treatments, and three locations were fertilizer trials with
one trial also having a thinning treatment.  In addition to the degree of
damage, the plots for the recovery study were treatments which would most
likely be used in future operations.  Due to the vagaries of the storm, no
attempt was made to link the original treatment with the frequency or severity
of storm damage.

   The study was installed in late February and early March of 1990,
approximately 5 months after Hugo.  At seven of the locations, all or portions
of the original measurement plots were used.  The size of the measurement
plots varied, but generally centered around the 0.05 acre.  At two locations
the recovery plots were established in the buffer areas surrounding the
original plots due to study salvage efforts.

Response Variables

   Several objective and subjective attributes were measured.  These included
traditional variables such a dbh and crown class.  They also included angle
measurements for quantifying initial stem lean and subsequent recovery by stem
position.  Three sets of horizontal and vertical distances were measured on
all study trees.  The vertical distances are the height to 8 feet of stem
length, crown base, and tip of the tree.  The horizontal distances were
measured from the stem base to the point perpendicular to each of the vertical
distances.  A carpenter's level was used periodically to insure both
measurements were true.  The horizontal distance defaults to zero for non-
leaning trees.  A subjective adjustment factor to account for badly bending
stems of leaning trees was ocularly estimated to "correct" the calculated
total stem length and length to crown base.

   Because crown base and terminal distances on surviving trees will change
irrespective of recovery, on a subset of the trees ages 2 through 5,
horizontal and vertical distances to 4.5 feet, one-half of original stem
length, and the original tip were also measured and marked on the stem ( Figure 1  ).  The number of trees selected varied by
plot but was the greater of 10% or 6 trees.

   Several subjective codes were used to describe stem form, stem damage,
crown damage and root damage.  The seven stem damage codes reflect the most
severe stem damage inflicted by the hurricane and may be used later in the
study to help select trees for wood quality analyses.  These eight stem form
codes include lean, forking, bole sweep, butt sweep and short crook.  These
codes are expected to change over time as trees recover.  Crown damage
reflects the amount of foliage and/or branches lost in 25% increments.  Root
damage differentiates between visible, broken roots and "wallowing" (the
formation of a hole in the soil at the root collar).

   At every location, selected trees and vistas are being photographed during
each measurement to visually document change.


   The data were analyzed using graphic, correlation, and regression analysis
to determine the relationship between post-storm and pre-storm tree size and
the influence of initial lean on 3-year growth and angle recovery.  For
presentation purposes the data are grouped into five lean categories:  no
lean, 25 to <50, 50 to <75, and 75 or greater degrees, loosely based on Brewer
and Linnartz (1973).  Statistical significance is based on an a priori alpha
level of 0.05.  In most cases, however, the actual probability values are


   Some variables could not be measured over time or were dropped for these
analyses due to confounding.  The original tip could not be located after even
one year and was dropped.  The angle at the base of live crown was also
dropped from the analyses due to ambiguous changes in angle.

Initial Damage

   At establishment there were 1534 living trees across the nine locations.
Initial hurricane mortality found during establishment was 2.7%.  On a
percentage basis, 61.5% of the live trees were classified as leaning.  The
average stem angle (leaning trees only) was almost 36 degrees.  Table 1
presents these data according to age class, along with the occurrence of other

Table 1a.  No. sites, no. plots, initial average dbh (inches) for trees > 4.5
           feet, initial average stem length for all trees as vertical height
           (feet) from ground to terminal, initial % hurricane mortality, and
           initial average degrees lean for leaning trees only.

Age   #sites   #plots   Avr. dbh   Avr. ht.   % mortality   Avr. lean
---   ------   ------   --------   --------   -----------   ---------
 2       2       12        0.6       4.4          1.7          39.7
 4       2       22        2.6      12.6          1.7          37.4
 5       2        5        4.0      14.8          1.4          39.9
 8       1        4        6.6      31.4          7.3          21.8
10       1        2        6.0      37.6          2.9          18.6
20       1        8        9.3      59.3          6.9          22.4
  MEAN--------------------------------------->    2.7          35.8

Table 1b.  Initial numbers of live trees and number damaged.

Age   #living  #visually damaged  #leaning  #root damage  #blowdown
---   -------  -----------------  --------  ------------  ---------
 2      407           307            305        165           14
 4      510           281            258         69            6
 5      285           231            211         27            1
 8      102            94             81          7            5
10       68            37             25          5            0
20      162           119             60         14            7
 TOTAL 1534          1069            943        287           33

   Across the study, initial average angle (average of all angles collected
for a given tree) was significantly negatively correlated with initial dbh,
plantation age and stem length (analogous to total height prior to the storm -
measured or calculated from horizontal and vertical measurements).  Through
initial age 5, dbh and stem length were significantly correlated with initial
angle although the sign of the coefficients varied by age.  For the three
older stands the correlations were generally not statistically significant
(Table 2).

Table 2.  Linear correlations with average stem angle at study establishment
          across ages and by age.

Variable           No. trees           Correlation         Probability
--------           ---------           -----------         -----------
...............................Across locations..............................
Age                  1534                -0.2425              0.0001
DBH                  1423                -0.2988              0.0001
Stem length          1534                -0.2583              0.0001
...............................Initial age 2.................................
DBH                   297                 0.1918              0.0009
Stem length           407                 0.3662              0.0001
...............................Initial age 4.................................
DBH                   509                -0.3280              0.0001
Stem length           510                -0.2871              0.0001
...............................Initial age 5.................................
DBH                   285                -0.2231              0.0001
Stem length           285                -0.1308              0.0272
...............................Initial age 8.................................
DBH                   102                 0.0211              0.4828
Stem length           102                 0.2123              0.0322
...............................Initial age 10................................
DBH                    68                -0.1321              0.2830
Stem length            68                -0.1882              0.1244
...............................Initial age 20................................
DBH                   162                -0.1035              0.1900
Stem length           162                -0.0921              0.2439

Post-Storm Mortality

   First-year mortality overall was 4.4%.  Mortality of trees judged to be
damaged by the hurricane was 5.7% or 61 trees.  By year three, total mortality
had increased to 6.5% with storm-damaged mortality rising to 8.1% (Table 3).
By lean category, the damage mortality was 2% (2 trees) for non-leaners, 8%
(7 trees) for trees initially leaning between 0 and 25 degrees, 15% (13 trees)
leaning between 25 and 50 degrees, 40% (35 trees) leaning between 50 and 75
degrees, and 35% (30 trees) leaning at least 75 degrees initially.  Initial
average angle of the mortality trees was 61 degrees.  Approximately 70% of the
mortality of storm damaged trees to date occurred in the first year.

Table 3.  Three-year DBH growth (inches), height growth (feet), and mortality
          percent by five categories of initial stem lean (degrees) and
          initial age.  Mortality based on total trees in column-class as
          per Table 1.  A * indicates that initial average angle was
          statistically significant at the 0.05 level in the model:
          growth = b0 + b1(initial size) + b2(initial angle).

Lean             DBH      Height   damaged       Blowdown       Root damage
category   Age   growth   growth   mortality%    mortality%     mortality%
--------   ---   ------   ------   ----------    ----------     -----------
   0         2    2.7*     10.8*       0
  0-25            2.8      11.7        0
 25-50            2.5      11.5        0
 50-75            2.1      12.3        0.1           0               0.3
 75&up            1.5      11.6        0

   0         4    2.4*     13.6        0
  0-25            2.3      13.8        0.1           0               0
 25-50            1.9      13.9        0.1           0               0
 50-75            1.5      12.5        0.5           3.0             1.0
 75&up            0.7       7.4        1.5           6.1             3.5

   0         5    2.3*     13.1*       0
  0-25            2.4      15.0        0
 25-50            1.7      15.6        0.6           0               0.3
 50-75            1.3      13.5        2.3           6.1             2.4
 75&up            0.9       7.2        0.3           3.0             0.7

   0         8    0.9*      6.1        0.1           0               0
  0-25            1.3       9.5        0
 25-50            0.9       7.2        0.2           0               0.3
 50-75            0.4       7.0        0.2           3.0             0
  75&up           All died in 1991     0.4          12.1             1.4

   0        10    0.7*      5.6        0
  0-25            0.7       6.2        0
 25-50            0.1       3.0        0.3           0               0.3
 50-75              None recorded at establishment
 75&up              None recorded at establishment

   0        20    0.5*      1.8        0.1           0               0
  0-25            0.5       1.7        0.6           0               0.3
 25-50            0         8.0        0.1           0               0.3
 50-75            All died in 1991     0.2           0               0
 75&up            All died in 1991     0.7          21.2             2.4

   Root damage and blowdown were also factors in the mortality.  By the end of
the third year, almost 45% of the root damaged trees had died.  This compares
with the study population of approximately 19% of the living trees in 1990
with root damage.  Twenty-one percent of the third-year damage mortality
occurred in trees marked as "blowdown" (where the one entire side of the
crown touched the ground).  This is over one-half of the trees coded initially
as blowdown.  Not surprisingly, 67% of the "blowdown" trees also had visible
root damage.

   Proportionally, mortality of damaged trees was least in the youngest
stands, increased with age through age 5, then decreased slightly through age
10 with a slight increase at age 20.  Three years after the storm, mortality
ranged from 0.3% for the initial age-2 stands to 14.7% for the initial age-4
stands.  The decrease in mortality for the age-10 stand is probably due to the
low number and severity of leaners (Table 1).  Except for the initial age-2
locations, weighted average mortality over the three-year period for hurricane
damaged trees was 11.3% (combined information from Tables 1 & 3).


   After three years, the tips of most of the surviving leaners had
essentially returned to vertical.  No horizontal measure was required, thus
stem length was not estimated.  To keep height values compatible across
measurement periods, height growth refers to changes in vertical height rather
than stem length.

   Because initial average angle was significantly correlated with initial
diameter and stem length, linear regression was used to examine the
relationship of 3-year growth to initial average angle after accounting for
initial size.  Residuals from the simple linear regression of growth on
"initial" size were plotted over initial angle.  The residuals showed a
distinct decreasing trend as average angle increased, at least for the younger
ages ( Figure 2a ).  Including average angle in the regression removed the trend
in the residuals ( Figure 2b ).  For 3-year growth, initial average angle was a
significant addition to the diameter-growth regressions for all initial ages
but only contributed significantly for ages 2 and 5 in the 3-year height-
growth regression.  Average angle did not significantly improve the
regressions on height growth for ages 4, 8, 10, and 20.

   Based on the sign of the regression coefficient for initial average angle,
3-year diameter growth by initial age shows a gradual, though significant,
decrease as average angle increases.  While not total agreement, this trend in
decreased growth as angle increases is evident in Table 3.  The sign of
initial angle for 3-year height growth, however, is not consistent across
ages.  Where initial average angle was significant (ages 2 and 5), the sign is
positive, indicating increasing vertical growth as initial angle increases.
Otherwise, the sign of the coefficient is generally negative, and the
coefficient is not statistically significant.  The mean height-growth values
by age in Table 3 (and scatter plots of the data) indicate a parabolic
relationship with growth increasing as initial average angle increases, then
decreasing after the angle reaches some "critical level."  Including a
quadratic term in the model did not change the results dramatically.

Recovery of Lean

   The greatest improvement in average angle:


always occurred in the 0-25 degree category while the the least improvement
occurred on those trees leaning at least 75 degrees initially, irrespective of
age.  This is true for both the 1- and 3-year changes.  The largest percent
recovery toward vertical occurred after the first year for the two initial
lean categories less than 50 degrees, but this varied somewhat by age (Table

Table 4.  First- and third-year changes in average angle by five categories of
          stem lean and initial age.  Changes are based on all live trees by
          category in the respective years.  A * indicates initial average
          angle was statistically significant at the 0.05 level in the model:
          % change = b0 + b1(initial angle).  The "zero lean" category was
          dropped for this analysis.

           1990              1991                  1993
Lean       Average     Average  Change %     Average  Change %
category   angle       angle    in angle     angle    in angle
--------   -------     -------  --------     -------  --------
...............................Initial Age 2.................................
   0        ////
  0-25      16.9         2.7      85.0*        0       100*
 25-50      36.1         6.9      80.5         0.2      99.6
 50-75      60.4        17.1      71.5         0.2      99.7
 75&up      83.5        35.9      57.2         8.8      90.3
...............................Initial Age 4.................................
   0        ////
  0-25      15.3         2.4      85.4*        0.1      99.4*
 25-50      36.0        10.3      71.2         0.8      98.3
 50-75      60.2        34.6      43.5        12.5      80.3
 75&up      79.6        69.2      13.0        43.8      45.0
...............................Initial Age 5.................................
   0        ////
  0-25      14.4         2.1      87.2*        0.5      96.8*
 25-50      37.7        14.1      63.7         4.7      88.2
 50-75      60.1        37.7      37.6        11.6      81.1
 75&up      78.5        69.6      13.0        46.2      40.5
...............................Initial Age 8.................................
   0        ////
  0-25       9.5         2.0      87.3*        0.6      96.2*
 25-50      36.3        25.5      29.6        17.4      55.6
 50-75      58.3        42.0      23.0        37.1      31.1
 75&up      88.8        All died in 1991
...............................Initial Age 10................................
   0        ////
  0-25      11.3         6.7      56.3*        3.5      78.8*
 25-50      33.8        31.1       2.6        28.7      14.0
 50-75      None recorded at establishment
 75&up      None recorded at establishment
...............................Initial Age 20................................
   0        ////
  0-25      10.7         9.3      26.6*        3.2      74.7*
 25-50      39.1        28.4       9.3        29.6       9.3
 50-75      55.4        All died in 1991
 75&up      89.6        All died in 1991

   Three years after the storm, the leaning trees on the initial age-2
locations had at least a 90% recovery, regardless of initial angle.  For
initial ages 4 and 5, the recovery was 80% or better for the trees leaning
less than 75 degrees at establishment.  Trees leaning at least 75 degrees at
establishment improved less than 50%.  As the initial age increases, the
percent recovery decreases, particularly for trees leaning more than 25
degrees initially.  Simple linear regressions of percent change to initial
average angle were significant for all ages, for both 1- and 3-year recovery.
The sign of the angle coefficient was consistently negative, indicating a
decreasing percent recovery with increasing initial lean (Table 4).

   By stem position, the greatest change in angle occurred at the tip of the
tree.  Three years after the storm the average improvement at the tip was 97%.
The least recovery occurred at the 8-foot position, which showed a 3-year
change of 77%.  While proportionally a small number, several leaning trees
showed a negative change in angle at the 8-foot position one and three years
after the storm (19 and 11 trees, respectively).  Overall, the initial average
lean improved by 92% after three years although this varied considerably by
age (Table 5).  Leaning trees will return towards vertical over time,
especially the younger trees.

Table 5a.  Percent change in angle from 1990 by stem location based on
           number of living, leaning trees in 1991 and 1993; ages 2-10.
           Number of trees in parentheses; "average" does not include
           crown-base angle.

Stem           Age 2        Age 4        Age 5        Age 8        Age 10
location     1991  1993   1991  1993   1991  1993   1991  1993   1991  1993
--------     ----  ----   ----  ----   ----  ----   ----  ----   ----  ----
Tip      %   79.5  99.4   76.8  96.3   77.6  97.8   78.4  90.7   58.6  76.5
         #  (303) (303)  (246) (234)  (187) (177)   (73)  (73)   (25)  (24)

Crown    %   41.5  95.3   41.9  86.6   31.4  79.1   64.1  78.5   32.3  50.9
         #   (57)  (57)  (232) (220)  (170) (160)   (66)  (66)   (24)  (23)

Original %   61.8  98.3   66.8  92.5   46.7  87.0
1/2 ht.  #   (71)  (71)  (118) (115)   (19)  (19)

DBH      %   66.0  99.1   53.2  89.0   25.6  55.7
         #   (59)  (59)  (118) (115)   (18)  (18)

8-foot   %   65.7  100.   63.9  91.1   44.9  76.6   53.1  70.4   -4.3  32.7
         #   (11)  (11)  (244) (232)  (183) (173)   (59)  (59)   (20)  (19)

Average  %   76.9  99.3   67.0  92.7   63.1  88.3   79.2  85.5   43.4  65.3
         #  (303) (303)  (247) (235)  (187) (177)   (73)  (73)   (25)  (24)

Table 5b.  Percent change in angle from 1990 by stem location based on
           number of living, leaning trees in 1991 and 1993; all locations
           and age 20.  Number of trees in parentheses; "average" does not
           include crown-base angle.

Stem        All locations         Age 20
location     1991   1993        1991   1993
--------     ----   ----        ----   ----
Tip      %   75.2   96.8        33.4   97.9
         #  (881)  (854)        (47)   (43)

Crown    %   39.3   78.4        20.2   25.4
         #  (595)  (568)        (46)   (42)

Original %   63.3   94.0
1/2 ht.  #  (208)  (205)

DBH      %   54.5   89.0
         #  (195)  (192)

8-foot   %   49.1   77.3       -13.0   10.7
         #  (552)  (525)        (35)   (31)

Average  %   67.2   91.8        26.3   73.1
         #  (882)  (855)        (47)   (43)

   While the results do show a dramatic improvement in initial angle by stem
position, that does not necessarily mean the trees have returned to pre-storm
straightness.  Three years after Hugo, 808 of the 855 surviving leaning trees
had shifted from "leaners" to some other stem form condition.  Of the 808,
31% were classified as straight, 21% had bole sweep, 27% and butt sweep, and
14% were recorded as having some other stem form condition.  Five percent of
the surviving leaning trees were still considereed leaning (Table 6).

Table 6.  Change in number of trees from leaning in 1990 to other stem forms
          in 1991 and 1993 (cumulative) across and by ages.  Values represent
          the total change between 1990 and the measurement year.

Measurement  # live    Total #   --------------Changed to:--------------------
year         leaners   changed   Straight  Bole sweep  Butt sweep  Short crook
-----------  -------   -------   --------  ----------  ----------  -----------
...............................Across locations...............................
   1991        882       532        223        213          68          12
   1993        855       808        250        170         216          57
...............................Age 2..........................................
   1991        303       225         60        114          40           3
   1993        303       300         61         21         150          46
...............................Age 4..........................................
   1991        247       145         75         44          22           1
   1993        235       221         98         49          50           5
...............................Age 5..........................................
   1991        187        81         26         45           0           5
   1993        177       170         42         74          15           1
...............................Age 8..........................................
   1991         73        60         45          7           6           2
   1993         73        63         48         22           1           5
...............................Age 10.........................................
   1991         25         7          6          1           0           0
   1993         24        12          8          2           0           0
...............................Age 20.........................................
   1991         47        14         11          2           0           1
   1993         43        42         13          2           0           0


   These results apply only to the initial conditions of this study.  The
study areas were selected to represent a range in damage conditions and may or
may not apply to the storm path as a whole.  Similarly, while several existing
research studies were utilized to take possible advantage of past stand
history, no attempt was made to link the severity or types of damage with
study treatments.  The vagaries of the storm preclude such an exercise.

   The correlations between initial tree size and initial average angle are
interesting.  Preliminary observations immediately following the storm
indicated the tallest, largest trees took the brunt of the damage.  Across all
study locations this appears to be true with negative correlations between
tree size, age and initial angle.  By age, however, initial angle was
significantly positively correlated with tree size for the age-2 stands,
negatively correlated with tree size for ages 4 and 5, and generally not
significantly correlated for trees age 8 and older.

   Reasons for the change for the younger stands from a positive to a negative
correlation between tree size and initial angle are not obvious and probably
involve factors such as soil rooting depth which were not part of the sample-
selection criteria.  As conjecture, however, the positive correlation for the
age-2 locations may be related to a possible sheltering effect of the smaller
trees by the taller trees and slight competition on the sites.  By age 4 the
sheltering effect becomes negligible while the larger trees, because they are
better established, are better able to withstand the wind.

   The lack of correlation for trees in the older stands is in part due to the
relatively low initial average angle.  This, in turn is partially due to stem
length and stocking.  Many of the leaning trees were caught in their
neighbors, artificially limiting the initial angle.  Almost 11% of the trees
in the ages 8-, 10- and 20-year stands contained trees leaning into them.
This compares with 2.3% for ages 4 and 5, and 0% for age 2.

   The growing season following the storm was relatively dry, based on the
average total precipitation and departure from normal (period 1951-1980) for
all recording stations in the five counties where the study is located
(Anonymous 1991).  Yet first-year mortality of those trees noted as visibly
damaged at establishment was fairly light.  Only 5.7% of the damaged trees
died during the first year.  Of the 5.7%, 77% were leaning at least 50 degrees
immediately following the storm.  Forty-four percent of the trees that died
the first year had visible root damage.  For ages 8 and 20, all trees
initially leaning at least 75 degrees died during the first year (no trees in
that category at the age-10 location).  These results would suggest that
salvage efforts should concentrate first in stands where a majority of the
trees are leaning at least 50 degrees and have a strong occurrence of root
damage.  This roughly agrees with recommendations by Brewer and Linnartz

   Three years after the storm, cumulative mortality to visibly damaged trees
increased to 8.1% with 75% of the mortality again occurring in the two most
severe-lean categories.  By age, the greatest mortality for the 2-year period
1991-1993 occurred for ages 4 and 5 (22 trees).  This represents almost 85% of
the increase from the first-year mortality and is due to the vast majority of
the most severely leaning trees in the older ages having died during the first

   Three-year diameter growth showed a consistent, gradual and statistically
significant decrease as initial average angle increased.  This was true for
all ages.  While not shown, a similar trend was evident for the first-year
diameter growth.

   Three-year height growth was not consistently related to initial average
angle although the mean growth values tended to increase as initial angle
increased, up to some critical angle, then declined.  The height-growth/
initial angle relationship was only positively statistically significant for
ages 2 and 5.  The lack of significant relationship for the other ages sampled
may be due to the effects of initial height on 3-year growth.  For ages 2 and
5 initial height tended to increase as initial average angle increased and
significantly influenced height growth.  The other ages did not show such a
trend.  Mean initial heights by lean category were very similar.  Also, an
examination of the signs of the initial angle coefficient were generally
negative and the initial-height coefficients were not significant for either
the linear or quadratic models.

   The response of a 10-year-old plantation loblolly pine growing in the
Piedmont of Virginia to ice-damage showed an increase in height growth and
a decrease in diameter growth for the damaged trees compared to the undamaged
trees (Bell and Dunham unpublished report cited earlier).  The authors
suggested the growth pattern was a response by the trees to regain lost canopy
(sunlight) position.  It would seem reasonable that the leaning trees in the
present study would exhibit a similar response.  Three-year diameter growth
for hurricane damaged trees does show a similar response, while the 3-year
height growth suggests such a response but is not statistically consistent.

   Three years after the storm the tip of the tree, following a phototropic
response, showed the greatest percent recovery, 97%.  This varied by age but
was generally over 90% for all ages sampled.  Measurements at other stem
locations also indicate at least partial recovery occurring all along the
stem.  This likewise varies by age, with the initial age-2 stands showing a
99% recovery after three years while the ages 10 and 20 stands showed a 65.3%
and 73.1% recovery respectively.  For all ages except age 20, the greatest
percent recovery in average stem lean (roughly 66%) occurred during the first
growing season after the hurricane.

   Stem recovery is also shown in the number of initially leaning trees which
have been reclassified to some other stem form including "straight."  These
classifications are dynamic and may change again as the trees age, and
possibly exhibit additional recovery.  As an example, the number of trees with
bole sweep decreased from 213 after the first year to 170 three years after
the storm.  This was particularly evident in the age-2 stands where the
decrease was over 90 trees.  Most of the shift was into butt sweep which
became more noticeable as the trees increased in height.

   While the study indicates at least partial stem recovery to vertical, it
does not yet address the wood quality of the storm damaged (i.e., leaning)
trees.  In a series of experiments on tilting seedlings of various conifers,
Kennedy and Farrar (1964) found the initiation of compression wood formation
at the cellular level within a matter of a few days.  Tentatively, following
the 11-year measurements, selected trees representing a range of damage
conditions will be destructively sampled to look at wood quality issues
(Dunham, P. H. and W. J. Hammond.  1990.  Long-term recovery of plantation-
grown loblolly pine from Hurricane Hugo.  Westvaco Forest Science Laboratory,
Revised Study Plan, unpublished.  9 p.).


   In general, mortality after three years was heaviest in the most severely
leaning trees and, proportionally, the oldest trees being followed.  Other
factors being equal, stands of marketable age should be salvaged first if a
majority of the trees are leaning over 50 degrees and have extensive visible
root damage.

   Height growth after three years increases as initial average lean increases
up to 50 degrees, depending on the age of the plantation.  This is offset by a
corresponding decrease in diameter growth for trees leaning more than 25
degrees initially, regardless of age.

   After three years, recovery in angle towards vertical was noted at all
points on the stem.  The greatest recovery occurred at the tip.  The least
recovery occurred at the lower stem positions, dbh and 8 feet.  Almost 66% of
the recovery noted after three years occurred after the first year.  Still,
average stem angle returned to within 92% of vertical for all initially
leaning study trees surviving after three years.  As expected, angle recovery
varied by age.

   Approximately 94% of the initially leaning trees surviving after three
years had changed their subjective stem form code from leaning to some other
condition such as bole sweep or butt sweep.  These codes will continue to be
applied for the duration of the study and are expected to change further as
the trees mature and possible additional recovery occurs.  They will also be
used later as part of the selection process for determining which trees to
destructively sample for wood quality analysis.

                               LITERATURE CITED

Anonymous.  1991.  Climatological data South Carolina summary.  National
   Oceanic and Atmospheric Administration, South Carolina Vol. 93, 13:1-7.

Brewer, C. W. and N. E. Linnartz.  1973.  The recovery of hurricane-bent
   loblolly pine.  LSU Forestry Note #104.  Louisiana State Univ., Baton
   Rouge, LA.  2 pp.

Foster, D. R.  1988.  Species and stand responses to catastrophic wind in
   Central New England, USA.  Journal of Ecology 76:135-151.

Kennedy, R. W. and J. L. Farrar.  1964.  Tracheid development in tilted
   seedlings.  In:  Cellular Ultrastructure of Woody Plants.  W. A. Cote, Jr.,
   Ed.  Proceedings of the Advanced Science Seminar, Pinebrook Conference
   Center, Upper Saranac Lake, NY.  Syracuse University Press, 1965.  Pp.

Powell, Mark D.  1990.  Meteorological aspects of Hurricane Hugo.  In:
   Hurricane Hugo One Year Later.  Benjamin J. Sill and Peter R. Sparks, Eds.
   Proceedings of a Symposium and Public Forum in Charleston, South Carolina,
   September 13-15, 1990.  Sponsored by the American Society of Civil
   Engineers.  Pp. 11-40.

McKellar, B. F.  1942.  Ice damage to slash pine, longleaf pine, and loblolly
   pine plantations in the Piedmont section of Georgia.  Journal of Forestry

Sheffield, Raymond M. and Michael T. Thompson.  1992.  Hurricane Hugo effects
   on South Carolina's forest resource.  Research Paper SE-284.  Asheville,
   NC:  USDA Forest Service Southeasternn Forest Experiment Station.  51 p.

Spell, N. E., Jr.  1990.  Hurricane Hugo one year later.  Forest Farmer,
   September 1990.  Pp. 23-25.

Touliatos, P. and E. Roth.  1971.  Hurricane and trees:  ten lessons from
   Camille.  Journal of Forestry 69(5):285-289.

Trousdell, K. B., W. C. Williams and T. C. Nelson.  1965.  Damage to recently
   thinned loblolly pine stands.  Journal of Forestry 63(2):96-100.

Van Hooser, D. D. and A. Hedlund.  1969.  Timber damaged by Hurricane Camille
   in Mississippi.  Res. Note SO-96.  New Orleans, LA:  USDA Forest Service
   Southern Forest Experiment Station.  5 p.