Evaluating the long-term consequences of forest management and stream cleaning on coarse woody debris in small riparian systems of the central Rocky M

12 Pages·1997·1 MB·English

by  Bragg, Don C

Historic, archived document
Do not assume content reflects current
scientific knowledge, policies, or practices.

2QL627
.F48
Fish Habitat Relationships Technical Bulletin
Number 21 April 1997
4
__
Evaluating the Long-Term Consequences of Forest
Management and Stream Cleaning on
Coarse Woody Debris in Small Riparian Systems
^ - of the Central Rocky Mountains
^
>
r Don C. Bragg^ andJeffrey L. Kershner^
,
- ‘Department ofForest Resources and Ecology Center,
y Utah State University, Logan UT84322-5212
' ' ^ National Fish Habitat Relationships Program, USDA Forest Service,
_ ' Department ofFisheries and Wildlife,
' _ _ ' ' J Utah State University, Logan UT84322-5210
Abstract
Coarse woody debris (CWD), while often providing the prim^ source ofstructure in ^all streams, has
been frequently subjected to intensive removal efforts with uncertain long-term consequences. We present
the integration ofa growth and yield model (FVS) with a mechanistic recruitment model (CWD) to
CWD
simulate long-term effects ofboth harvesting and removal from streams in northwestern Wyoming.
The CWD model predicted a gradual recovery from stream cleaning 80 -110 years later, but both selective
harvesting and clearcutting significantly decreased the volume ofmaterial entering the stream when
compared to uncut controls. The combination ofstream cleaning and clearcutting frequently limited the
CWD
total load to < 10% ofundisturbed levels. Selective harvesting ofriparian forests, while decreasing
CWD inputs, may deliver enough to allow'harvests in some systems when coupled with proactive CWD
management. Proactive management, in this sense, means we may want to deliberately form snags or
CWD CWD
initiate delivery, or at least retain favorable trees in the residual stand to ensure continued
recruitment. a
"-V'
USDA
Forest Service
/
'
Forest Service Fish Habitat Relationships Program Leaders
or Representatives
FHR
Currents Purpose
National FHR Program Region 8
The USDA Forest Service Fish Jeffre&y L. Kershner, Washington Office CindyA. Williams, Southern Region
Fish Wildlife Department Wildlife, Fish 63: Botany Staff
Habitat Relationships Rrogram Utah State University 1720 Peachtree Rd.
was established to further the Logan, UT 84322-5210 N.W. Atlanta, CA 30367
development offisheries Region 1 Region 9
technology and transfer this
Kathy Moynan, Northern Region Bob Hollingsworth, Northeast Region
technology to field biologists. (Anadromous Fish Program) USDA Forest Service
With ever increasing demands Nez Pierce National Forest 310W. WisconsinAve.
for natural resources, protection Route 2, Box 475 Milwaukee, WI 53203
Grangeville, ID 83530
and management of aquatic Region 10
Brian Sanborn, Northern Region
communities requires biologists (Resident Fish Program) Ron Dunlap, Alaska Region
to be knowledgable of current Deerlodge National Forest Wildlife, Fish & BotanyStaff
research findings and state-of- FederalBuilding, Box 400 Federal Office Building, Box 21628
the-art techniques. The purpose Butte, MT 59703 _ Juneau, AK 99802-1628
ofFHR Currents is to provide a Region 2 Intermountain/Northern Region
vehicle to quickly disseminate R. Nick Schmal, RockyMountain Region Kerry Overton, Fish ResearchWorkUnit
information important to field- Wildlife, Fish & Botany Staff Intermountain Research Station
level biologists in the USDA Univ. ofWyoming, College ofAgriculture 316 E. Myrtle St.
Laramie, WY 82071-3354 Boise, ID 83702
^Forest Service.
Associates
Region 3
Bryce Rickel, Southwest Region Glenn Chen
Submissions: Wildlife, Fish & BotanyStaff Eisheries Biologist/MonitoringSpecialist
Federal Building USDA Eorest Service
&
517 GoldAvel, S.W. Eish Wildlife Department
Albuquerque, NM 87102 Utah State University
If you wish to submit a paper Logan, UT 84322-5210 ^
for publication in FHR Currents, Region 4 Ered Mangum — _
please contact the following Seona L. Brown, Intermountain Region Aquatic EcosystemAnalysis Lab
people for information and^ ^ Wildlife, Fish & BotanyStaff 105 Page School ~
?
^ 324 25th St. BrighamYoungUniversity
guidelines. Ogden, UT 84401 Provo, UT 84602
Jerry Boberg/Karen Kenfield Region 5 Ken Roby
Technical Editors ^ ^ Hydrologist, USDA Eorest Service
, Jerry Boberg, Pacific Southwest Region
[phone] (Anadromous Fish Program) Pacific SouthwestExperimentStation
P.O. Box 245
SixRivers National Forest
Six Rivers National Forest ./A 1330 Bayshore Way Berkeley, CA 94701-0245
Fisheries Department Eureka, CA 95501 Larry Schmidt
1330 Bayshore Way ^ " Stream Systems Technology Center
Jeffery Reiner, Pacific Southwest Region
Eureka, CA 95501 ^ (Resident Fish Program) USDA Eorest Service
Lake Tahoe BasinManagementUnit Rocky Mountain Experiment Station
8So7u0tEhmLearkaeldTaBhaoye,RdC.,ASu9i6t1e501 2Po4r0tWC.ollPirnoss,peCcOt 80526-2098 r
MarkVinson
Region 6
Hydrologist/Monitoring Specialist
Deborah Konnoff, Pacific Northwest Region USDI Bureau ofLand Management
& &
Wildlife, Fish BotanyStaff Pish Wildlife Department
333 S.W. 1stAve., P.O. Box 3623 Utah StateUniversity
Portland, OR 97208 Logan, UT_84322-5210
—
fp
FHR
Currents
Introduction simulate how management actions influence CWD
recruitment and delivery to streams, even though
it is not current federal policy to harvest riparian
For decades we have cultured a rather
forests in this manner. This exercise provided
antagonistic relationship with instream coarse insights as to what information is most important
wcomoidnyddieabmreitser[aCnWdD-1 mpielcoensgo(fMwcoDoaddeatelteaals.t119090, wkihnednomfeplodsit-npgroacgerssoowrt.hTahnedseyiseilmdumloatdieolnswiwtehrethis
Richmond and Fausch 1995)]. For the sake of
more than simply academic, however, as it is also
navigation, mining, or fish migration, humans
conceivable that private or industrial landowners,
CWD
have extracted great quantities of from less constrained by buffer regulation, may be
rivers and streams, often with unanticipated and interested in some level ofharvesting in riparian
detrimental consequences (Swanson and forests. Eventually, some agencies may also decide
Lienkaemper 1978, Beschta 1979, Bilby 1984,
to adopt other buffer policies to address habitat
Harmon et al. 1986, House and Boehne 1987, variegation concerns, which could include limited
Young et al. 1994). Recent research has shed a
entries for harvesting.
considerably more positive light on the influence
of CWD on creating fish habitat, organic carbon Regardless of current riparian forest management,
production, and CWD effects on channel CWD’s greatest applicability may come from the
morphology (e.g., Swanson et al. 1982, Megahan information it can provide on the restoration of
1982, Harmon et al. 1986, Bisson et al. 1987). damaged streams. Until recently, stream clearing
and riparian forest cutting were standard practices
Riparian forests are critical sources of CWD. across much of the western United States (see
Murphy and Koski (1989) and McDade eCtWalD. Froehlich 1973, Bilby 1984, Bisson et al. 1987,
(1990) estimated thatm> 70% ofriparian was Young et al. 1994). These practices have resulted
generated within 20 of the bank. The in many streams depauperate of CWD, and if
preservation of an intact, continuous buffer of simultaneously heavily cut, there may be little
riparian forests has frequently been cited as forest remaining from which to recruit new pieces.
CWD
necessary for sustaining delivery to a stream CWD can be used, in conjunction with historical
(e.g., Murphy and Koski 1989, McDade et al. inventories, to forecast recovery periods (given
1990, Potts and Anderson 1990). Resource
forest type and growth rates) and indicate if
managers in many areas are faced with the CWD
natural recovery shall be sufficient or if
dilemma ofintegrating forest management further human intervention is needed.
practices with sensitive riparian zones, and
frequently lack the tools to prescribe management
plans that consider both timber and riparian Methods
needs. To address some of these concerns, we
have developed an analysis tool driven by a
l, Study Area Descriptions
growth and yield model [the ForestVegetation
Simulator (FVS)] that predicts CWD delivery to Eield sites were sampled in the Bridger-Teton
the stream through a combination ofstochastic National Eorest (Appendix A) during the summer
and deterministic processes. Originally calibrated of 1995 to provide both model parameterization
for small Engelmann spruce (Picea engelmannii) and riparian stands for future simulation (Eigure
and subalpine fir (Abies lasiocarpa) dominated 1). Small (first- to third-order) streamsCwWeDre
streams on the Bridger-Teton National Forest in selected because of the importance of in
northwestern Wyoming, the objective of the CWD these systems (Swanson et al. 1976), and because
model is to help biologists and silviculturists they are the most likely stream systems in this
predict CWD delivery in riparian zones under a region to possess adjacent, continuous spruce-fir
variety ofmanagement options. ^ forests. Sampled streams tended to be low to
moderate in gradient (< 10%) and low in sinuosity
In designing this model, we had twofold (Appendix A). Table 1 lists the physical
intentions. First, we wanted to hypothetically parameters of the three streams selected for
r
FHR
Currents
simulation in this paper. Climatically, the regioti is
characterized by cold winters and abundant
snowfall, while summers are warm with sporadic
precipitation primarily from thunderstorms.
Spring snowmelt floods are the primary
mechanism influencing CWD transport in many
small streams (Harmon et al. 1986), including
those oh the Bridger-Teton. ^
Model Development r c
^
CWD
Numerous riparian forest and stream
characteristics were sampled, including live tree
attributes (species, height, d.b.h., lean injrelation
CWD
^to the stream), riparian zone properties
(species, orientation relative to stream, end-ohlog
CWD
diameters, length), and stream
characteristics (species, end-of-log diameters,
length, piece adjacency). Some of these data were
then used to help parameterize the CWD model,
which will be detailed in forthcoming work (Bragg
et al. 1996). Once the CWD model predicted
CWD
delivery of to the stream, a simple formula
is applied to determine periodic (10-year interval)
CWD
stream volume: _
^ V -j
CWD - - CW, D^
volume = previous volume +
CWD
production^losses of
Figure 1. Summer 1995 field sites for CWD project, CWD
Bridger-Teton National Forest, Wyoming. Site codes .t Most research on the mechanics of transport
are as follows: A- Sheep Creek; B- Blind Bull Creek; has been confined to steep, unstable streams in the
C- Murphy Creek; D- Willow Creek; E- S. Fork Pacific Northwest affected by debris torrents (see
Gypsum Creek; F- Hoback River; G- Adams Creek; Swanson et al. 1976, Swanson and Lienkaemper
H- Buck Creek; I- Mosquito Creek;]- Moose Gulch. 1978, Bisson et al. 1987) and may not reflect the
dominant processes influencing debris d^amics^
^ ^ ; , , " . - - in the central Rockies (Young 1994, Richmond
and Fausch 1995). In the absence of catastrophic
Table 1. Properties for selectedBrider-Teton National Forest (WY) streams used in this study. All values are for -
m
250 reaches sampleiduring the summer of 1995. _ J '
Mean Riparian
—
^ Drainage bankfull _ - -w ' " Reach forest
basin size ^ width Reach“-^ ^ gradient density
Stream Drder Oia) o ^ (m) - 7 sinuosity (%) (m^/ha)
•'V '
Moose Gulch Creek 1'^ 414 ” 3.4 A '1.12^^ .7^ 72.0 23
. .
Dry Lake Creek 1033^ - 5.5^ 1.05 ; ^ 3.5 33
'
Hoback River 3 9965 14.5 1.02 1.5 31
'
FHR
Currents
CWD
Table 2. parameters used with the model for this study. For an undisturbed old-growth stand (this
example), we assume average cyclic^ CWD delivery equals average cyclic CWD loss (steady-state conditions).
Stream Average
Mean bankfull CWD CWD Cyclic Average cyclic
cyclic CWD CWD
Stream width ^ volume delivery turnover loss
- (m) rate^ (mVlOOm)
(mVlOOm) (mVlOOm)
Moose Gulch Creek C _ 3.4 21.0 2.7 0.13 2.7
Dry Lake Creek 5.5 8.6 2.3 ; k - 0-26 2.3
"
Hoback River 14.5 11.9 3.2 0.27 3.2
^ 'Cyclic'refers to theuseofa 10-yeartimestep toreflectsynchronisitywithFVSvegetationpredictions.
^Cyclic CWDturnoverrateistherationofaveragecyclicCWDlosstostreamCWDvolumeatt=0.
disturbance, several authors have hypothesized growth-and-yield model with the capacity to
CWD
stable long-term loads (Froehlich 1973, simulate both forest growth patterns and the
,
Swanson and Lienkaemper 1978, Likens and Bilby influence ofmanagement. This model uses data
1982). We also assumed that the CWD volumes taken from thestand ofinterest and projects the
existed at a steady-state condition in which the _ change in physical characteristics (e.g., height,
CWD
long-term loss of (transport off-site decay) diameters, stocking, density) through time. FVS
-i-
is approximately equivalent to inputs from can be customized to produce realistic forest f
adjacent forests and deposition ofmaterials fromc dynamics, including regeneration, growth rate,
upstream. Table 2 lists the CWD characteristics ^ and mortality patterns for many different
applied to simulated streams. ^ locations. For this exercise, we haveThtilized the
CWD Teton Variant (version 6.2) ofFVS.
Generally, losses ofstream arise from
/ decomposition and transport (Swanson et al. Since FVS is capable ofsimulating multiple
1982, Ruediger and Ward 1996). With the harvests over long periods of time (up to 400
exception offlood events, small streams are . years), it is possible to evaluate the long-term
considered capable of transporting only small consequences ofboth past and present
CWD CWD
pieces of (Harmon et al. 1986, Bisson et al. management, and estimate recovery periods
1987), and wTthin-stream decomposition is for streams influenced by timber harvests and/or
CWD
negligible compared to annual inputs (Bisson et al. removals. Six different treatments were
1987). Small streams, because of their smaller tested on a first, second, and third order s^eam
dimensions and lower flow rates, are able to retain (Tabled). ' -- .
CWD
higher loads than larger streams (Swanson
In low intensity stream treatments (i.e., no debris
eral. 1982, Lienkaemper and Swanson 1987, removal), we assume that the streams move a -
Ruediger and Ward 1996), which are capable of CWD
constant ratio of during each cycle (equal to
moving whole trees (especially during peak
theamount on average delivered ever)-" cycle in
flow^s). _ CWD
- . undisturbed stands), regardless of volume_
CWD
'' The Forest Vegetation Simulator [FVS, also known actually delivered after treatment. The
' as Prognosis (Stage 1973, Wykoff et al. 1982)] is a attrition rates reported in Table 2 are similar to
I . ..
CWD
Table 3. Permutations ofriparian forestharvests and stream cleaning treatments used in this paper (each
,
-was applied to all three stands).
No harvest, Clearcut, Selective harvest,
no stream cleaning (TMT#1) no stream cleaning (TMT#3) no stream cleaning (TMT#5)
^ No harvest,^: " ^ Clearcut, Selective harvest,
_ stream cleaning (TMT#2) stream cleaning (TMT#4) strearn cleaning (TMT#6)
Page 3
FHR
Currents
those used by others (e.g., Murphy and Koski ' harvest treatments experienced an initial harvest
1989). For high intensity treatments (i.e., debris at year 50, and then a target stockingi of 195 m3/
cleaning), we eliminate all CWD in the stream at ha was established as the threshold for harvest
We CWD
the first treatment, but not during subsequent scheduling by FVS. used delivery to the
CWD
harvests (simulating current conditions, in which stream, net change in stream volume, and:
streams were cleared in the past but not with differences in merchantable harvest volumes to
present or future harvests). We also assume a evaluate the ecological and silvicultural effects of
neutral-impact harvest system in those stands each treatment.
being treated (i.e., no slash was deposited in the
stream). This final assumption, while convenient
Results^and Discussion
for modeling purposes, may not hold in some
harvested areas especially after clearcutting
CWD
(Froehlich 1973, Potts and Anderson 1990). Changes in stream loads through time can
CWD
be found in Figures 2-4. The initial drop in
Thirty replicates of the CWD model were generated delivery before the first harvest (years 0 - 49) was
CWD
on each treatment to calculate average an artifact ofstand dynamics produced by FVS
delivery over a simulation period of300 years. reflecting the flux between recruitment ofnew
m
Three subject stands were modeled as 15 strips mature trees and mortality settings. The no
offorest immediately adjacent to the stream harvest, no stream clearing treatment (TMT#1)
CWD
(Bragg et al. 1996). delivery results are vacillates somewhat over the length of the
m
reported as cubic meters per 100 ofstream simulation period but generally follows the
reach. After allowing FVS a period ofadjustment original steady-state stream CWD volumes. While
to incorporate both regeneration and mortality, all there is little long-term data on stream CWD ^
Page 4
" FHR Currents
FHR
Currents
retention, this trend seems to pauern natural amount of CWD entering the stream because
riparian forest fluctuations in stand density weakened trees were the most likely candidates to
(Froehlich 1973, Likens and Bilby 1982). As succumb under our assumptions. The removal of
CWD
expected, stream loads changed drastically most ofthe largest size classes (common practice
with stream cleaning, even under a no-harvest with selection systems) also decreased the volume
system. Complete stream CWD removal under of CWD delivered by removing the biggest pieces^
unmodified canopies (TMT#2) resulted in (which represent a large share of total CWD).
CWD
decades-long lags in volume for all Although not addressed by the CWD model, this
simulations (Figures 2-4). These streams were change may further affect CWD storage as large
predicted to eventually re-establish uncleared pieces help to store smaller, more transient pieces
stream CWD loads. Depending on the steady-state of CWD (Bisson et al. 1987). Selection systems
CWD loads, the productivity of the stand, and resulted in higher post-harvest CWD levels than
CWD
cyclic turnover rates (Table 2), the lag in clearcutting because there^were residual trees > -
CWD
stocking for undisturbed riparian forests 10 cm d.b.h., whereas cl^rcutting removed all
ranged from approximately 80 years at Dry Lake merchantable (high CWD potential) stems. Both
Creek and the Hoback River to 110 years for treatments that did not involve stream clearings
Moose Gulch Creek. While these lags are less than (TMT#3 and TMT#5) led to significant depletion
what Murphy and Koski (1989) predicted for of total CWD storage in the streams, but were still
CWD
some Alaskan streams, the overall trends are able to retain some CWD, probably as large, pre-
quite similar, with most ofthe differences likely harvest logs and rootwads [see Andrus et al.
resulting from different methodologies, model (1988) and Ralph et al. (1994)]. Both clearcut
assumptions, and species growth rates. and selective harvests combined with stream
However, while the gross volumes may be similar, cleaning (TMT#4 and TMT#6) resulted in long-
CWD
it is difficult to say if these two conditions should_ term loss of and foreCstWs Dless able to
be considered functionally equivalent. The replenish depleted stream loads. Stream
CWD c^^ning coupled with repeated clearcutting
dimensions of the stored by the treatments
(TMT#1 versus TMT#2) is probably very different, largely eliminated all CWD from the streams for -
almost a century after treatment and only
as cleared streams tend to lack the largest pieces of
permitted minimal restocking. Under these
wood. Even 250 years after cleaning, it is unlikely
that a similar volume ofvery large trunks and conditions, the CWD model indicates that even
CWD
rootwads could be replaced, as these structures are 10% ofpre-logging volumes will be difficult
slow to accumulate and may persist for centuries — to maintain (Figures 2-4).
in undisturbed systems (Swanson et al. 1976, Selective harvest systems are capable ofdelivering
Bisson et al. 1987, Murphy and Koski 1989). some CWD consistently from the riparian forest.
As suggested by the model results, applying Table 4 presents the differences in both timber
CWD
and production under the management
standard harvesting systems in riparian forests,
even without repeated stream clearings, can systems applied during this simulation. As this
CWD table shows, both harvest systems reduce
significantly depress total volume within a CWD
stream (also see Murphy and Koski 1989). cumulative production significantly from
Rainville et al. (1985) found similar responses in no harvest. However, according to FVS growth
and yield projections on the subject stands, a
their simulation ofharvest treatments ofnorthern
Idaho forests. Selective harvesting (here simulated selection system can produce anywhere from 68 - —
by periodic thinnings of the largest size classes 104% of the harvest volume ofa clearcut over a - ^
when the stand volume exceeded 195 m3/ha) 300 year period while retaining 38 - 65% ofno
CWD
.harvest volumes. Selection systems appear
instituted a gradual but significant decline in
stream CWD volume as both mortality rates and to offer a compromise solution by preserving
the number oflarge trees declined. Increased tree some harvest while recruiting and retaining a
substantial portion ofstream CWD.
vigor, generally thought to be a favorable aspect of
managed forests, did serve to limit the potential
Page 6 I