Hydrological impacts of thinning in a deciduous forest ecosystem

1
Hydrological impacts of thinning in a deciduous
forest ecosystem
PRESENTED BY HZAL
SERENGIL, Y., GÖKBULAK, F., ÖZHAN, S., HZAL, A.,
SENGÖNÜL, K., BALCI, A.N., ÖZYUVACI, N
ISTANBUL UNIVERSITY FACULTY OF FORESTRY DEPT.
OF WATERSHED MANAGEMENT
2
Hydrological impacts of thinning in a deciduous
forest ecosystem
NEED FOR THE STUDY Turkey is located on a
geography that includes diverse ecological
conditions.
3
Hydrological impacts of thinning in a deciduous
forest ecosystem
The 72 millions of population is concentrated in
larger cities which led to the implementation of
huge water resources development projects.
Istanbul, the economic capital is one of the
largest cities in Europe with a population of
over eleven million
4
Hydrological impacts of thinning in a deciduous
forest ecosystem
More than 15 water reservoirs are in operation
around the city and new pipeline projects to
carry all the water within a circle of 300 km in
diameter are on the way. The annual supply of
water resources doubled from 1994 to 2002 with
the accelerated investments (ISKI, 2002).
5
Hydrological impacts of thinning in a deciduous
forest ecosystem
BACKGROUND The results of most paired studies
have been summarized and discussed in a number of
reviews starting with Hibbert (1967) and followed
by Bosch and Hewlett (1982), Hornbeck et al.
(1993), Stednick (1996), Sahin and Hall (1996),
Vertessy (1999, 2000) and finally by Brown et al.
(2005).
6
Hydrological impacts of thinning in a deciduous
forest ecosystem

• To highlight some of the conclusions based on
  these reviews
• the response to treatment is highly variable and,
  for the most part, unpredictable (Hibbert, 1967),
  
• streamflow response to the treatment depends on
  both the mean annual precipitation of the
  watershed and on the precipitation for the year
  under treatment (Bosch and Hewlett, 1982),
• infiltration and evapotranspiration (ET) plays a
  key role in determining what happens to the flow
  regime after treatment (Bruijnzeel, 1988),
• changes in annual water yield from forest cover
  reductions of less than 20 of the watershed
  could not be detected by streamflow measurements
  (Bosch and Hewlett, 1982 Stednick, 1996).

7
Hydrological impacts of thinning in a deciduous
forest ecosystem

• In large watersheds or the ones that are managed
  toward multiple uses including water production
  or recreation, intensive or sharp treatments
  might not be applicable.
• Gotmark et al (2005) emphasized the benefits of
  partial cutting on the biological diversity in
  European perspective while Shao et al (2005)
  mentioned the new forestry policy in China
  requiring that forests be harvested with
  selective cutting methods.
• In Turkey, the situation is similar. Clearcutting
  is performed in some situations but out of
  consideration in many cases, and thinning is the
  most preferred harvesting method in water
  producing watersheds and the ones assigned for
  recreation or aesthetics purposes.

8
Hydrological impacts of thinning in a deciduous
forest ecosystem
HYPOTHESIS The common knowledge that suggests a
20 percent lower limit for cutting treatments to
be able to cause any detectible change on the
streamflow is valid for Belgrad Forest ecological
conditions. Therefore an 11 percent thinning is
not expected to cause an alteration in water
yield and regime.
9
Hydrological impacts of thinning in a deciduous
forest ecosystem
EXPERIMENTAL WATERSHEDS IN BELGRAD FOREST
The experimental watersheds are located in the
Belgrad Forest (41 N, 28 E) which has been
preserved as the only old-growth oak-beech
natural forest near Istanbul (Figure 1). The
climate of the watersheds and surrounding area
according to Thornthwaite classification system
is humid, mesothermal oceanic with a moderate
soil-water deficit in summer. Long term mean
annual precipitation is 1050 mm and mainly fall
from October to March.
Figure 1. Location of experimental watersheds.
10
Hydrological impacts of thinning in a deciduous
forest ecosystem
Table 1. Mean annual precipitation, temperature
and Thorntwaite potential evapotranspiration
(PotET) during the calibration period (1979-1985).
Parent materials mainly consist of carboniferous
clay schists and neogene loamy, gravelly
deposits. The soils are usually shallow to deep,
gravelly, loamy clay in texture, rich in organic
matter with medium to good permeability rates and
high erodibility potentials without carbonate
reaction. The mull type forest floor has an
average depth of 5 cm (Ozhan, 1977). Subsurface
flow is the main mechanism to feed the streams in
the watersheds (Balci et al., 1986).
11
Hydrological impacts of thinning in a deciduous
forest ecosystem
Topography is generally gentle, and mean
elevation is around 140 m. Both watersheds are on
a southern aspect adjacent to the divide which is
about 3-4 km from the Black Sea, under the
influence of prevailing northern maritime winds
during the rainy period. Dominant vegetation
includes oak (Q. frainetto Ten., Q. cerris L.)
and beech (F. orientalis L.) tree species mixed
with varying amounts of Carpinus betulus L.,
Castanea sativa Mill., Populus tremula L., Alnus
glutinosa L., Acer trautvetteri Med., Acer
campestre, Ulmus campestris L., and Sorbus
torminalis Crantz. (Yaltirik, 1966) with a normal
crown closure.
Table 2. Some properties of experimental
watersheds.
12
Hydrological impacts of thinning in a deciduous
forest ecosystem
Table 3. Forest stand properties in the
watersheds before treatment (derived from
Belgrad Forest forest management plan).
Species OOak, HHornbeam, BBeech, MMinor
broadleaved (lt10), Overstorey/Understorey Diamete
r classes (cm) a lt7.9, b 8-19.9, c 20-35.9,
d gt36 Crown closure (0 - 1.0) 1 0.11-0.40, 2
0.41-0.70, 3 gt0.71
13
Hydrological impacts of thinning in a deciduous
forest ecosystem
2.2. Field methods Both streams draining
Watershed-I (control) and Watershed-II
(treatment) were instrumented with 90 and 120
concrete sharp-crested V-notch weirs and
automatic water level recorders. Data collection
has been started in 1979 in an attempt to study
the effects of timber harvesting upon water
quality and yield. The paired watersheds were
calibrated from 1979 to 1985. In February 1986
11 percent of the standing timber volume was
removed from the treatment watershed
(Watershed-II) by employing a standard
individual selective cutting technique. The
felled timber was taken out of the watershed
through horse dragging on forest roads.
14
Hydrological impacts of thinning in a deciduous
forest ecosystem
2.3. Rainfall-Runoff Process Mean annual
precipitation was 1090.5 mm during the
calibration period (1979-1985), while the
following year (March 1986-February 1987) after
treatment it was 1124.9 mm, little more than
average. Runoff coefficient of W-II was mostly
higher than the W-I in the long term (0.27gt0.21)
(Figure 2), both highly variable ranging from
0.02 to 0.63. Coefficients were higher in the
rainy years and after rainy periods of 2-3
subsequent years. The precipitation in the
previous 3 hydrologic years before 1985 was lower
than average.
Figure 2. Runoff coefficients of the
experimental watersheds.
15
Hydrological impacts of thinning in a deciduous
forest ecosystem
Figure 3. Annual hydrographs of the experimental
watersheds during the calibration period
(1979-1985).
The high flow period was from October to May
(Table 1). The monthly precipitation was over 100
mm in the first 6 months of this period,
including a nonsignificant amount of snowfall.
The hydrographs reached to the peak in January,
while the rainiest month was December. The
general shape of the hydrographs reflected the
soil water deficit in summer months, and the soil
water replenishment, continued throughout the
autumn.
16
Hydrological impacts of thinning in a deciduous
forest ecosystem
The history of the rainfall-runoff process can be
determined by correlograms (Muftuoglu, 1991
Salas, 1993). The monthly correlogram shows at
least 2 months of strong, 2 months of weak linear
history (r00.648, r10.644, r20.533, r30.234,
r40.058) which means that the runoff amount of
the current month is significantly affected not
only with the precipitation of this month but
also the previous 4 months.
Table 4. The 0-8 lag cross correlations between
monthly precipitation and streamflow.
17
Hydrological impacts of thinning in a deciduous
forest ecosystem
Figure 4. The monthly variation of precipitation,
streamflow (W1) and the change in the
correlation coefficient between the two
experimental watersheds.
18
Hydrological impacts of thinning in a deciduous
forest ecosystem
STATISTICAL METHODS
2 REGRESSION APPROACHES WERE APPLIED. DAILY AND
MONTHLY. We considered the monthly flows that run
over the prediction interval as statistically
significant. Because the frequency distribution
of the daily flows, as mentioned in the flow
regime section below, was far from being
Gaussian, application of a t test procedure would
not be adequate. It is a representation of the
range of values that an individual y might take
on for a given xi. It incorporates the parameter
uncertainty as well as the unexplained
variability of y. The (1-?) 100 prediction
interval for a single response, given xi, is yi ?
t? s 1 1/n (xi-?)2 /SSx1/2 where s
(SSy-b1Sxy) / n-21/2 yi is the best estimation
of Y according to Xxi, while t? is the t value
for n-2 degree of freedom and exceedence
probability of 0.05. ? is the mean value of X, s
is the standard error of the regression (Bayazit,
1996) and Hirsch et al., 1993). The prediction
interval bands are further from the best-fit line
than the confidence bands, a lot further if you
have many data points. The 95 prediction
interval is the area in which you expect 95 of
all data points to fall (Motulsky and
Christopoulos, 2003).
19
Hydrological impacts of thinning in a deciduous
forest ecosystem
Figure 5. The predictions of both regression
approaches for the post treatment year.
The post treatment year water surplus calculated
by regression equations (observed-predicted)
based on monthly and daily data were 79.28 mm and
106.11 mm, respectively.
20
Hydrological impacts of thinning in a deciduous
forest ecosystem
Figure 6. The observed and predicted monthly
streamflow of W-II with prediction intervals (p
0.05).
January was the only month that observed flow
exceeded the prediction interval (plt0.05). The
difference was 7.76 mm which means that at least
this amount of water was achieved by thinning,
statistically. It is 2.92 percent of the annual
post treatment flow.
21
Hydrological impacts of thinning in a deciduous
forest ecosystem
TREATMENT NOT EFFECTIVE ON FLOW REGIME
FLOW DURATION CURVES
Figure 8. FDC for the dry months
(April-August) after the treatment.
Figure 9. FDC for the rainy months
(October-February) after the treatment.
22
Hydrological impacts of thinning in a deciduous
forest ecosystem
DISCUSSION and CONCLUSIONS The climatic condition
of a region is a significant factor to determine
the results of a forestry treatment. In this
study the hydrologic response of the treatment
watershed to the thinning was mostly concealed in
the following summer months because of the high
potential ET rate in the region and also possibly
stimulated growth of the thinned stands. In
published studies based on the works in USA, the
largest relative changes in streamflow were
observed in summer months after removal of
eastern deciduous forest and western conifer
forest. However, the largest absolute streamflow
increases occurred during wet winter months
(Jones and Post, 2004). In this study the
situation was quite different, as the summer
flows were very low due to limited summer
precipitation in the region. The largest absolute
streamflow increases were observed in the winter
months.
23
Hydrological impacts of thinning in a deciduous
forest ecosystem
The 11 percent thinning was not effective enough
to change the flow regime of the streams draining
watersheds covered with deciduous forests, but
caused a statistically significant increase on
the streamflow of January in the following year.
The mechanism to explain the water surplus in
the winter months in this situation might be the
quicker replenishment of soil water beneath the
thinned stands in fall. Owing to the 4 months
long rainfall runoff history of the watersheds,
the difference in the recharging time of the soil
moisture might have caused the difference on the
water yield in the dormant winter months.
24
Hydrological impacts of thinning in a deciduous
forest ecosystem
The additional water gained by the thinning
treatment was around 3 percent of the annual flow
in the first year while the annual precipitation
was over the average. Consequently, it is not
possible to suggest such a slight treatment as a
tool to be used in water resources development
works. However, it can be considered as a
baseline which suggests that less than 20 percent
thinning might also cause a statistically
detectible change in water yield.