is not the number of trees we plant but the number of
trees we grow that will restore our forests and protect
grant from the National Fish and Wildlife Foundation Chesapeake Bay
Small Watershed Grants Program. The project is titled: “Failure is Not
an Option: Investigating a Cost-Effective Approach to Reducing Deer
Damage in Reforestation Programs.” Read press release
are a problem for forest health and agricultural viability
throughout the Northeast. The standard approach to deer
exclusion fencing, in areas larger than backyard gardens, uses
permanent fencing installed as a physical barrier to deer. Whether
electric, woven wire or plastic, these fences are quite expensive.
We are testing an approach that does not create a physical barrier.
Rather, it uses inexpensive temporary electric fencing materials to create
an area where it is simply uncomfortable for deer to be.
plantings test sites in areas with high deer density were selected.
Two "official" control and experimental subplots were delineated
within each study site (official is in quotes because the entire
area of each site that is not enclosed in wires will serve as
control). Each test block is 100 feet long; the width varies
between sites (see site pictures below). The experimental subplots
have a double or triple perimeter of single strand, temporary
electric fencing that enclose the area to be protected. The wires
are from 24” to 30” above the ground.
Site 1, near Yellow
Springs, WV. This is a WV Potomac Tributary Strategy
Riparian Buffer Demonstration Project site, described in
detail here. The site
has a multi-year history of failed plantings indicating a severe problem with
deer browsing (it was replanted in April 2005). Test blocks are 100 feet long and ~60
feet wide. There are three perimeter wires, with the
distance outer two being four feet apart and the inner pair
eight feet apart. Fence is charged using a solar
Site 2, a USDA-CREP
(Conservation Reserve and Enhancement Program) site near
Baker, WV. Forest condition in the area and the
landowner indicate a deer problem exists in this area.
This area was planted and a high tensile, electric cattle
fence installed in the spring of 2007. Test blocks are
100 feet long and ~35 feet wide, the standard buffer width
in WV. There are two perimeter wires spaced four feet
apart. Fence is powered via a tap from the high
Data collected includes deer browse and tree vigor on previously planted
trees, grid mapping of tree recruitment by position and species,
proximity effects of electric wires (single and enclosures), signs
of deer within each subplot, and maintenance required to keep the
fence clear and functioning.
If successful, this
method could be used to economically increase success of riparian
plantings in high deer density areas. It would also allow no cost
natural recruitment of trees from seed or roots to occur, and allow
the use of much shorter tubes (for rodent protection) or no tubes at
all, thereby reducing a major cost element of these plantings.
Site 1: The following
graphs present data collected at the
Forest Demonstration Project at Yellow Spring, WV between June 2007
and June 2008.
Three site categories are included: control no wire – no electric
wire along long axis of study area; control w/wire – one electric
wire along long axis of study area; and experimental – 100’ x 60’
study plot enclosed with a triple perimeter of single strand,
temporary electric fencing (spaced 4 and 8 feet apart).
Data for August 2007 shows
significant browse damage in one of the experimental blocks,
with no further damage in later months. This damage was
probably due to heavy weed growth that reduced voltage on
the electric fence to less than 2500V during the night and
early morning hours when the weeds were covered with dew.
This condition was also the case in July, but training deer
to the fence that occurred during the previous months when
the fence voltage was much higher may have been sufficient
to keep them out for a time. The literature on using
electric fencing to dissuade deer is pretty clear on the
need for 4500V or more to be effective. Weeds were cleared
from the fence in August after the problem was detected, and
voltages were restored to >5000 volts at all times
(typically >6000 volts). No browsing damage has been
observed since that time, although some fence repair was
needed in November 2007. No winter browsing was
observed in any block.
The only trees that have
gotten ahead of the deer in the unprotected areas are a few
scraggly crabapples, hawthornes, and one
Graph on left
is a 100% stacked bar graph shows
only living plants with leaves above tube that are
susceptible to browse damage.
Graph at left shows the height of the highest intact leaf
above the top of the tree tube. Growth well
above the tube is occurring in both experimental blocks, and
not occurring in control blocks.
Below is what the above graph
represents in the real world. The photo at the
left is the unfenced area 5 years after the area was
planted. The photo on the right is the NEB
fenced area 3 years after fencing.
Two site categories are shown in the graph
below: two control blocks (with one wire along long axis)
and two experimental blocks (with 2 wires around perimeter
as noted above). The following conditions pertain to the
site. This was a new planting when we installed the
fence, and the vast majority of trees were fully contained
within their tree tubes at the start. Trees selected
for this site were a mix of fruit (apple, peach), nut (such
as hazelnut), and flowering trees (such as dogwood). Early results were much less clear than at the Yellow
Spring site, with a certain amount of browse damage to
plants occurring within the treatment area. We were
concerned that the farm border fence that was supplying the
electricity for the study area fence might not be on consistently, as
it was powered through a ground fault interrupter circuit
that shut down on occasion. We installed a solar
charger at the site on July 20, 2007, but a certain
amount of browsing continued in August 2007. We purchased a digital fence
voltage meter and determined that a short in the fence was
dropping the voltage down to near 2200V, which was also the
voltage of the farm’s perimeter fence. Once the short was
found, and weeds cleared, the voltage rose to > 5000V.
Little browsing occurred within the
treatment areas in September 2007 and May 2008, while regular browsing outside the treatment
blocks was widespread. Browsing occurred in both
treatment blocks in June 2008, likely due to excessive weed
growth that was reducing voltage on fence. Vegetation
got ahead of us in June 2008 and again in 2009, fence
voltage down below 3000V. All of the above issues, plus
numerous electrified farm fences in vicinity, led to
“training to the fence” inconsistency. Fully 1/3 of two of
the plots had nearly 100% mortality due to high water
table. CI replanted with species appropriate to conditions
(buttonbush, alder, willow) in July 2010.
1. This graph
includes a "below tube top" category, for trees with
leaves that did not reach to the top of the tree
2. The large number of browse-able trees (leaves
above top of tube) showing early in the control area
were mostly in two foot high tubes in the west
control block. These were all hazelnuts that
sprouted above the tubes quickly - and were just as
quickly browsed to the top. This despite
increasingly heavy thistle growth that made
surveying this area quite unpleasant; the deer did
not seem to care.
Graph above shows the height of the highest
intact leaf above the top of the tree tube.
Growth well above the tube is occurring in
both experimental blocks in both 4' and 2' tubes. Good growth was
observed in fruit trees (particularly apple trees in 4' tubes) in the eastern
Control Block (ECB) as of July 2010. No growth
significantly above tubes was observed in Western
Control Block (WCB); this area was primarily planted with
Hazelnuts in 2' tubes.
The difference between treatment and
control in terms of growth above tubes could not be more
dramatic. See if you can figure out which plants in the
table below are protected:
Here is a summary of what
we saw have observed through 2010:
If the fence is properly
installed, energized, and clear of heavy weed growth,
protection within fenced areas can be ~100%.
In the first few weeks
after installation, you need to check the fence for
damage as the deer get very upset when they first get in
it. The inner wire is more likely to be broken than the
There does not appear to
be much of an edge effect (if any) where trees near the
fenced area are protected.
Failure to maintain the
fence reasonably clear of heavy weed growth results in
lowered voltage on the fence and, over time,
dramatically reduced success.
Just as the literature
says, voltages in the 2000-2500 voltage range are not
The fencing sadly doesn’t
protect against drought or insects.
The following cost comparison is based on retail material
costs, and ballpark labor estimates. It assumes that
either tubes (with stakes) or fence are used to protect the
plantings from deer browse. Everything else about the
planting is assumed to be the same. Not shown are
yearly maintenance costs, for which materials would be
negligible and labor for maintaining fence and cutting weeds
of fence would probably amount to ten hours per acre ($150
at $15/ hour).
The fence design was initially tested
in an informal experiment along the riparian corridor of
Skaggs Run (a tributary of the North River, tributary of the
Cacapon River). This site has a long history of
excessive deer browsing, with little natural recruitment of
trees anywhere on a 54 acre property over the past 20 years.
A variety of fencing configurations were installed, from 4
feet apart to 7 feet apart. After two years, these are
•There have been only two plants slightly
browsed (both recently).
Every tree outside the fence exclosures has
been heavily browsed.
Three non-tubed plants were lost overwinter
2008 to vole damage.
A significant amount of natural recruitment
has taken place, including: black walnut, witch
hazel, hornbeam, hophornbeam, linden, red maple,
oak spp, hickory, dogwood, cherry, ash, box
elder, tulip poplar, locust.
Lack of browse in these areas offers the
potential for proper forest structure and biotic
community formation below the 6 ft browse line.
Planting riparia is not just about the trees,
it's about restoring the forest ecosystem.
Forested riparian buffers are tree lined corridors alongside
streams and rivers that reduce the flow of pollution moving
from the land into the water. People are planting thousands
of miles of buffers to protect the Chesapeake Bay and its
rivers and streams, many of these miles with funding through
the USDA-CREP program. However, it is not the number of
trees planted but the number of trees we grow
that will restore our forests and protect our waters.
Unfortunately, where deer are abundant young trees are often
damaged by deer browsing; Cacapon Institute’s (CI) surveys
show more than 90% can be severely impacted by browse. We
are currently testing a relatively low cost defense using
temporary electric fencing materials. Results are
promising. This approach, or other electric fence
approaches such as the 3 wire offset fence, have the
potential to dramatically increase survival of riparian
plantings in high deer-density areas – at a reasonable cost.
We suggest the CREP committee consider adopting the
following tree planting/protection protocol for
implementation in areas with high deer densities, except in
areas where the public has regular access to the site and
liability issues would be a problem. The intent of the
fence is to protect the planting area for a sufficient time
(3-5 years) until the trees have reached a height where deer
are not a major threat.
This practice assumes that riparian plantings are the
standard 35 foot width. The literature indicates that deer
fence success is reduced as exclosures become larger.
Tree tubes are not installed to protect trees from deer
browse. Protection may be necessary to prevent rodent
damage, particularly if the planting area is not routinely
mowed by the landowner. CI suggests the use of 1-2’ tree
tubes, with inexpensive ˝” x 4’ bamboo stakes used to both
support the tube and mark the location of each plant.
Temporary electric fencing should be installed, according to
CI’s 2-wire perimeter design (see below), the New Hampshire
3 wire offset fence, or a similar design. Unless the
landowner is interested in maintaining a permanent fence (as
will be necessary for development of a natural understory
and natural tree recruitment), this fence should be
installed using temporary fencing supplies to reduce costs.
The voltage on the competed fence, clear of weeds, properly
installed without shorts to ground, must exceed 5000 volts.
If it does not, first try installing additional ground rods
per the energizer manufacturer’s instructions. If that does
not work, consider breaking the fence into several sections
that are independently energized.
A low wire (6”-12” high) may be installed for the purpose of
improving the path to ground and increasing the likelihood
that deer will be shocked. However, if installed, this wire
will be a maintenance problem both in terms of weed control
and collecting flood debris.
When the fence is first installed, training the deer to the
fence using peanut butter on aluminum foil wrapped around
the hot wire will certainly not hurt prospects for success.
The fence wires must never be up without being energized.
Long fence installations (500’ and greater) must incorporate
gaps in the fenced exclosure that allow deer shock free
passage across the width of the buffer. These gaps should
be situated at locations where deer trails are observed. If
there is an area with numerous deer trails, the fence
designer may consider initially building a more robust fence
in that area, with two hot wires (at ~20” and 30”), a ground
wire at 6”-12”, and fence baiting to ensure that deer are
quickly conditioned to avoid the new condition on the
If the fence is installed in the vicinity of other electric
fence(s), expect the deer to be more acclimated to being
shocked and more difficult to control. Additional wires
(vertical and/or horizontal) may be needed to accomplish an
acceptable level of control.
Do not use a farm’s perimeter fence as the source of
electricity unless: a) the voltage on that fence always
exceeds 5000 V; b) the fence is known to be well maintained
clear of weeds and shorts. Insufficient voltage will insure
If a source of 120 volt AC power is readily available,
consider buying a plug-in energizer instead of a solar
energizer. They are less expensive, and able to put a
higher voltage on the fence.
Take voltage measurements in the morning when dew is still
on the grass. This is the time of day when any shorts will
have the greatest effect on fence voltage, and the best
indicator of the voltage that will likely be on the fence
during the nighttime. Always use a digital voltage meter to
test the fence.
If the area is known to flood frequently, the fence
energizer should be installed above the typical observed
If the area is known to flood frequently, the landowner
might consider respooling the fence wires before a likely
flood event, and reinstalling the wire after the flood
waters recede. That will reduce the issue of either
cleaning flood debris from the fence, or removing that wire
and installing new wire.
The costs of every planting should include the labor to
check the fence periodically for breaks (especially in the
first few weeks after installation), and to clear the fence
of weeds as needed during the growing season. It is much
easier to clear a fence before the weeds and grass are high
enough to be a problem; one person can weed-eat a 600 foot
long exclosure in about two hours under optimal conditions.
We suggest setting a three month clearing schedule, starting
in middle to late June, and sticking to it. Three visits
should be sufficient in all but the wettest years. Certain
plants, like jewelweed, can grow very quickly in wet areas
and may require additional attention.
The cost of this fence maintenance for 5 years should be
considered a component cost of installation. Maintenance
could be performed either by the landowner as part of their
match or by the contractor. The planting is not complete
until the trees grow, and maintenance is necessary for that
riparian buffers are wide strips of trees located along river and
stream corridors. They provide many important benefits, including
shade to keep river water cool and wildlife habitat. They also
dramatically reduce the flow of pollution from the land into our
rivers by filtering nutrients, sediments and other pollutants from
runoff as well as removing nutrients from groundwater, allowing
cleaner water to flow through to the stream.
buffers are an important component of every Bay state's Tributary
Strategy. According to the Chesapeake Bay Program, between 1996 and
August 2005 Bay Program partners planted 4,606 miles of riparian
forest buffers throughout the watershed, with a short-term goal of
10,000 miles restored by 2010. (http://www.chesapeakebay.net/newsriparian121205.htm
expense of riparian plantings (over $1000 per acre) is justified by
the goals of improving water quality and restoring ecosystem
functions. Planting the trees is not enough. The planted trees must
actually survive if the goal of improving water quality will be
achieved. In addition, planting failures may in fact do actual harm
by demoralizing the volunteer community that plants many sites, and
may serve as a negative demonstration for farmers who witness
plantings on their property fail.
The challenges to
survival are many, and include drought, weed competition, insects,
rodents, deer, and mechanical damage from mowers or vandalism.
Tree tubes from two
to four feet tall are used to reduce damage from deer and rodents.
Weed mats reduce weed competition and newer designs help conserve
moisture at ground level. However much these technologies help,
they do not represent a silver bullet. Rodents may hide under weed
mats and use them as cover while they gird a sapling. And, if small
seedlings are planted in tree tubes the trees that finally emerge
from the tubes are very spindly and weak. Maryland recommends
leaving the tubes on for 5-8 years to give the resulting tree a
chance to strengthen. But the stakes that hold the tubes upright do
not last that long, and must be replaced.
The state of
Maryland evaluated 130 buffer planting sites in 2000. The sites
were planted between 1997 and 1999. They reported that: 82% met
acceptable stocking levels; survival of planting seedlings averaged
60%; and volunteer seedlings made up a third (36%) of the seedlings.
They identified weed competition as the "most common stressor for
the young seedlings, with drought, deer, machinery such as mowers,
and insects also being noted with lesser frequency." (Pannill et
As a member of the
West Virginia Potomac Tributary Strategy Implementation Team (WVPTS),
CI is assessing survival and browse damage at three forested
riparian buffer demonstration projects, one on the Cacapon, one on
the South Branch, and one on Tuscarora Creek near Martinsburg.
The first WVPTS
demonstration site planted was along the banks of the Cacapon River
near Yellow Spring, in April 2005. Each tree was planted with a
weed mat and in a tree tube for protection from deer browsing.
Overall survival, as indicated by living leaves seen on a plant, was
down to 68% by September 2006. Living trees and shrubs that rose
above the tops of tubes (trees in 4' tubes and shrubs and smaller
trees in 2' tubes) were assessed for signs of recent deer browse.
56% exhibited browse damage three months after being planted in
2005, and 91% in September 2006. Many of the living plants that
were not recently browsed were stunted, many severely, and many of
these showed signs of browse and dieback from the previous year.
demonstration planting site, along the South Branch of the Potomac
River, had a mixture of tube and no-tube plantings. The tube
plantings fared about the same as at Yellow Spring. Plants without
tubes (weed mats only) fared more poorly, with 48% survival three
months after being planted in 2005 and 41% survival (of those that
could be found) in June 2006. Browse damage was nearly universal
and severe on surviving plants in 2005, with the few existing leaves
very small in comparison to the same plants in tubes. As of June
2006, the browse rate was not as severe, with 61% of available
The Tuscarora Creek
planting is doing well, with no apparent deer browsing. However,
deer browsing at the other two sites is a serious problem. In
addition, unlike the Maryland report above, the WV sites with high
deer pressure have had no significant recruitment of native trees
despite the lack of any livestock pressure for more than a decade.
That doesn't mean that no native tree seedlings germinate; they do
in abundance. It means they do not survive the deer browse.
CI’s deer exclusion
fencing experiment is designed to test an innovative and relatively
low cost method to protect riparian forest plantings from
destructive or even catastrophic damage from deer browsing
activities. If successful, this method could be used to
economically increase success of riparian plantings in high deer
density areas. It would also allow no cost natural recruitment of
trees from seed or roots to occur, and allow the use of much shorter
tubes (for rodent protection) or no tubes at all, thereby reducing a
major cost element of these plantings.
The first year of this project
was funded by the WV Division of Environmental Protection.
Institute - From the Cacapon to the Potomac to the Chesapeake Bay,
we protect rivers and watersheds using science and education.
#10 Rock Ford Road
Great Cacapon, WV 25422
possible by funding from The Norcross Wildlife Foundation, the
National Fish and Wildlife Foundation, Virginia Environmental
Endowment, NOAA-BWET, USEPA, The MARPAT Foundation, and our generous