Extension Note

70 Extension Note

December 2004 Windthrow in Partially Cut Lodgepole Pine Forests in West-Central British Columbia

Michaela Waterhouse and Introduction (Ruel 995), but few studies report Harold Armleder B.C. Ministry of Forests windthrow rates within partially cut Southern Interior Forest Region Northern caribou (Rangifer tarandus stands, particularly lodgepole pine 200-640 Borland St. caribou) are a species of concern (Pinus contorta Dougl. ex Loud.) Williams Lake, BC V2G 4T1 (blue-listed) provincially. In southern forest types. Whitehead and Brown British Columbia, they are classified (997) found that uniform commercial as threatened under the federal Species thinning (50–65% basal area reduc- at Risk Act. Under the Cariboo– tion) of mature lodgepole pine ag- Chilcotin Land-Use Plan (cclup), gravated windthrow. Studies from two the Northern Caribou Management other forest types in British Columbia Strategy designated a large portion report small but increased losses of of the wintering area of the Itcha– standing trees in single tree and group Ilgachuz caribou (8 000 ha) to be selection silvicultural systems com- managed through “modified harvest- pared with uncut forest within 2 years ing” as one of the actions necessary of logging (Coates 997; Huggard et to address habitat concerns (Youds et al. 999). al. 2002). The Itcha–Ilgachuz project The “modified” harvesting options tests the effects of various silvicul- being tested in this study are group tural systems on key caribou habitat selection and irregular group shelter- features (particularly terrestrial and wood silvicultural systems. The group arboreal lichen), as well as regenera- selection silvicultural system targets tion, microclimate, long-term site a 33% area removal based on 5 m productivity, commercial mushrooms, diameter openings and is based on and biodiversity. an 80-year cutting cycle. The irregular The focus of this study is to assess group shelterwood system is based whether partial cutting could increase on removing 50% of the area in 20- levels of windthrow in stands to a to 30-m openings. The system is point that would reduce the quality of irregular because the final cut is in 70 caribou habitat. It is also important to years. The size of the harvested open- measure the impact of partial cutting ings and total area removed differ on stand stability because of potential between treatments; therefore, they economic loss and forest health risk. could affect the rate of windthrow. The incidence of windthrow on As the opening size and area cut clearcut edges is well documented increase, wind could influence the

Ministry of Forests Forest Science Program stand, especially the edges. The Orthic Dystric Brunisols developed Within each treatment unit, two point at which gap size and area on loamy glacial till. Throughout all transects (0 × 350 m) were randomly harvested cause the rate of windthrow five blocks, soil textures are sandy selected from many lines previously to become a management concern is loams and the humus types are mors laid out in a grid over each treatment unknown for stands on the Chilcotin (2–6 cm thick). unit. Windthrow (>7.4 cm diameter at Plateau of west-central British The five blocks are located within breast height [dbh]) originally rooted Columbia. stands that were likely initiated after within each transect was measured, The objectives of this study are: stand-destroying wildfires between described, mapped, and marked 60 and 250 years ago. Stands in the annually. The following attributes • to compare the annual, and 5.3- SBPSxc are much more open and less were recorded: dbh, crown class year post-harvest period rates of productive for timber than stands in (Luttmerding et al. 990), direction of windthrow among the partially cut the MSxv due to drier site conditions. fall (°), decay class (Backhouse 993), and no-harvest treatments; and Stand characteristics are summarized and type of wind damage (Stathers et • to describe the characteristics in Table . al. 994). Trees were classified as to of trees most susceptible to whether they were in or out of a 3-m windthrow. Methods forested edge strip around openings. Summaries of decay, damage, edge, Study Area Before harvest, each block was wind direction, and crown closure divided into four treatment units of were based on all trees over 2.4 cm The study area is located about 0 km 5–28 ha and one of four treatment dbh pooled from all blocks and northwest of Alexis Creek, B.C., on options was randomly assigned to treatments. The 2.4-cm value is the a gently rolling, high-elevation pla- each unit (Table 2). The two irregular minimum close utilization standard teau near Satah Mountain (52°28ń, shelterwoods were targeted at 50% and was the minimum size of tree 24°43´w). The five replicate study area removal using 20- to 30-m included in cruise compilations. Trees blocks, which occur in either the openings but differed by harvesting over 2.4 cm dbh represent 57% of Very Dry, Very Cold Montane Spruce system (stem-only or whole tree). As the trees in the data set while those (MSxv) or Very Dry, Cold Sub-Boreal a result of the harvesting, the percent between 7.5 and 2.4 cm dbh make up Pine–Spruce (SBPSxc) biogeoclimatic cut was 0% lower than the target and 43%. Data were summarized for those subzone, are located in northern the treatments resulted in different trees over 7.4 cm dbh but the results caribou habitat and range in eleva- opening sizes. On average the whole were almost identical, except where tion from 260 to 640 m. Lodgepole tree harvesting produced smaller noted in the text. pine is the dominant tree species on openings. Because of the differences Data are reported as number of these mesic, relatively flat blocks. in opening size, the two shelterwood windthrown trees per hectare (>2.4 In the MSxv blocks, the lodgepole treatments were separated rather cm dbh) and as a percentage of the pine–grouseberry–feathermoss site than combined. The group selection standing trees per hectare per year type (MSxv/0) is predominant, while harvesting was closer to the proposed after harvest, for each treatment unit in the SBPSxc blocks, the lodgepole targets of 33% area removal and 5 m (transect lines summed together). pine–kinnikinnick–feathermoss site diameter openings. Harvesting was Non-parametric analyses of vari- type (SBPSxc/0) is predominant done on a snowpack deeper than 50 ance (Friedman test), based on the (Steen and Coupé 997). Soils are cm and was completed by April 996. randomized block design, were used table  Characteristics of the five blocks from cruise compilations (1995) before harvest (minimum dbh = 12.5 cm). Grey attack refers to trees killed by mountain pine beetle.

Block Mean Gross Grey Elev size Aspect Slope dbh volume Total Live attack Block Subzone (m) (ha) (°) (%) (cm) (m3/ha) stems/ha stems/ha stems/ha 1 SBPSxc 1290 113 0 0–1 16.9 115 1033 827 202 2 SBPSxc 1320 105 0 0–4 19.5 100 607 527 74

3 MSxv 1420 96 0 0–6 19.0 228 1316 1217 97

4 MSxv 1495 60 45 5–15 18.7 294 1721 1539 181

5 MSxv 1620 70 90 7–20 21.6 283 1137 894 236

2 to compare rates of windthrow among most stable. The predominant direc- This result is similar to findings in the treatments for all trees in the tion of fall (all trees > 7.4 cm dbh) was central British Columbia where stron- 5.3-year post-harvest period, mean to the northeast (Figure 3), indicating ger wind events prevailed from the annual live tree rate of fall, and mean that wind came from the southwest. southwest (Huggard et al. 999). annual dead tree rate of fall. The live and dead tree rates of fall are calculat- ed as a percentage of the standing live table 2 The average area harvested and opening size by treatment (n = 5) or dead. The 996 data were included Harvested area Opening diameter only in the analyses of variance for (%) (m) the whole 5.3-year period because they were collected in a 3-month post- Treatment Standard Standard harvest period and decay class was Treatment abbreviation Mean deviation Mean deviation not recorded. Results are considered No harvest C – – – – significant at α = 0.05. Irregular group shelterwood – stem IGS-SO 39 10 26.2 7.5 Results and Discussion only harvested Irregular group The majority (80%) of the wind- shelterwood – whole IGS-WT 39 8 22.9 5.6 thrown trees were dead before falling tree harvested (Figure ). The most frequent category of dead trees was class 3 (62%) where Group selection GS 28 4 14.5 3.3 more than 75% bark is on the tree. Trees in more advanced stages of decay may already have fallen or, more likely, the abundance of class 3 �� trees reflects the pine beetle epidemic �� that swept through the area in the figure  Proportion of windthrow �� (>12.4 cm dbh) in each �� early 980s. The most common type �� decay class (n = 172 trees) of damage recorded was root break �� (1997–2001). (53%) (Figure 2), often caused by the �� main, supporting, bracket roots being progressively broken during storms. This pattern is typical for lodgepole pine trees that have small root systems and are rooted in sandy soil (Stathers �� et al. 994). figure 2 Proportion of windthrow �� Most of the fallen trees were in (>12.4 cm dbh) in each �� the codominant (53%) and dominant damage class (n = 248 trees) �� (25%) crown classes; however, when (1996–2001). all trees over 7.4 cm dbh were tal- lied, the distribution shifted to 45% N intermediate, 36% codominant, 5% dominant, and 4% suppressed trees. This result indicates that the smaller dead trees are susceptible to windthrow. The average diameter figure 3 Direction of windthrow (>7.4 of the windthrow (based on 2.5 cm 25 20 15 10 5 cm dbh) based on number of W E minimum dbh) for blocks –4 was stems per 10° interval (n = within .5 cm of the mean reported on 434 trees) (1996–2001). the cruise compilations. The average for block 5 was 3 cm smaller, perhaps indicating that the largest trees are

3 Barring extraordinary events, the table 3 Annual windthrow rate (stems per hectare) for all trees (>12.4 cm dbh). The 1996 first few years following harvesting data, collected in a 3-month period, were standardized to 1 year for comparison in are considered the most susceptible to this table. windthrow (Stathers et al. 994). The Block 1996 1997 1998 1999 2000 2001 number of windthrown trees per hectare in all blocks, particularly block , 1 74.3 1.1 0.4 2.5 5.0 5.4 was greatest in the 3 months follow- 2 7.1 0.4 1.1 1.4 1.1 2.9 ing harvesting in 996 (April–June) 3 5.7 1.4 2.5 4.3 2.1 2.5 (Table 3), then was fairly steady in subsequent years. The large amount 4 5.7 2.1 2.5 5.4 4.3 2.9 of windthrow in block  was about 5 15.7 0.7 1.4 3.9 2.5 1.8 equally divided among treatments. Perhaps this block experienced an iso- lated wind event that was exacerbated by the high density of mountain pine table 4 Analysis of variance (Friedman test) of windthrow rates (%/ha) for: all trees in a beetle–killed trees (about 20% of the 5.3-year period, live trees (annual), and dead trees (annual) (n = 5 yr) (C = no standing trees). harvest, IGS-SO = irregular group shelterwood – stem-only, IGS-WT = irregular It is more meaningful to examine group shelterwood – whole tree, GS = group selection) rates of windthrow that have been ad- Treatment (mean) justed by the number of standing trees in each treatment unit. Over a 5.3-year Type period, there were no significant IGS- IGS- III differences in windthrow of all trees Analysis C SO WT GS SS F3,12 p (dead and alive combined, >2.4 cm All trees (% ha–1 · 5.3-yr–1) 1.76 2.58 2.14 2.68 1.0 .17 .92 dbh; Table 4) among treatments. The Live trees (% ha–1 · yr–1) 0.04 0.18 0.04 0.04 1.2 .22 .88 mean annual rate of live windthrow –1 –1 was also non-significant among treat- Dead trees (% ha · yr ) 1.43 1.54 1.93 2.33 1.8 .31 .82 ments and ranged from 0.04 to 0.8%/ Note: 996 included only in 5.3-year period. ha/yr (Table 4). When dead trees over 2.4 cm dbh were considered, the rate of windthrow was substantially higher openings more wind could penetrate trees are more likely to come down (.4–2.3%/ha/yr) but not significantly the stand. during wind events because their different among treatments. Similar to other studies in British root systems are slowly decaying. It The stem-only shelterwood Columbia (Coates 997; Huggard et remains to be seen if the rate of tree treatment tended to have more live al. 999), the rate of live windthrow fall will remain steady, increase, or windthrow than the other partially in the partially cut units in this trial is explode during one big storm event. cut treatments but the opposite was low (0.04–0.8% stems per hectare per true for the dead trees. When all year). Even the cumulative amount Management Implications windthrow was summed in the of both dead and alive over the past 5.3-year period, the control tended 5.3 years is well below the suggested The annual rate of windthrow for to have less windthrow. This result is 0% loss for operational significance live stems has been extremely low, probably due to the dead tree com- (Coates 997). The rate of loss may and therefore of low risk to caribou, ponent being more protected in the also be mitigated by ingress of new timber supply, and forest health. On uncut forest. Fifty percent of all fallen trees into the 2.4 cm dbh and larger the other hand, in stands where an trees in the harvesting treatments size class over time. abundant dead tree component is fall- were within 3 m of a forest/opening This trial also considered the fate of ing out, there could be several effects edge, indicating a strong susceptibility the standing dead component, which on caribou habitat: of freshly exposed edges to wind. This is up to 20% in some stands. These effect is lost when averaged over the trees were mostly killed in the early . decreased substrate for arboreal whole treatment area. For a set level 980s by mountain pine beetle. In the lichen (forage); of harvest, increasing opening size 5–20 years since this major mortality 2. changed stand microclimate, which should decrease the amount of edge. event, these trees have slowly and could influence terrestrial and However, by increasing the size of steadily fallen (≤2.3%/ha/yr). These arboreal lichen health and survival;

4 3. increased wind, which could References Steen, O.A. and R.A. Coupé. 997. A result in wind scouring of arboreal field guide to forest site identifica- lichen; and Backhouse, F. 993. Wildlife tree tion and interpretation for the 4. increased obstacles to animal management in British Columbia. Cariboo Forest Region. B.C. Min. movement. Workers’ Compensation Board, For., Victoria, B.C. Land Manage. B.C. Min. For. and Canada-British Handb. 39. So far, the rate of tree fall in the Columbia Partnership Agreement Whitehead, R.J. and B.N. Brown. study area is not a concern for cari- on Forest Resource Development: 997. Windthrow after commercial bou. If the dead component of the FRDA II, Victoria, B.C. thinning in mature lodgepole pine. stands (20%) suddenly comes down in Coates, K.D. 997. Windthrow damage Can. For. Serv. File Rep. p5800- the next few years, it still may not be 2 years after partial cutting at the 97-0. an issue. Of bigger concern is the cur- Date Creek silvicultural systems Youds, J., J. Young, H. Armleder, M. rent mountain pine beetle epidemic study in the Interior Cedar– Folkema, M. Pelchat, R. Hoffos, in caribou habitat. In areas of high Hemlock forests of northwestern C. Bauditz, and M. Lloyd. 2002. tree mortality, where harvesting is British Columbia. Can. J. For. Res. Cariboo–Chilcotin Land-Use Plan: not planned, eventual windthrow of 27:695–70. Northern caribou strategy. Cariboo most stems is possible and will likely Huggard, D.J., W. Klenner, and A. Mid-Coast Inter Agency Manage- be problematic for caribou (Youds et Vyse. 999. Windthrow follow- ment Committee, Williams Lake, al. 2002). ing four harvest treatments in an B.C. Spec. Rep. Results of this study strongly reject Engelmann spruce – subalpine fir the notion that harvesting associ- forest in southern interior British ated with group selection or group Columbia, Canada. Can. J. For. shelterwood silvicultural systems will Res. 29:547–56. aggravate windthrow in pine forests Luttmerding, H.A., D.A. Demarchi, on the Chilcotin Plateau. At the levels E.C. Lea, D.V. Meidinger, and T. of removal and opening sizes used in Vold. 990. Describing ecosystems the trial, the rate of fall in the uncut in the field. 2nd ed. B.C. Min. forests and treatments were similarly Environ., Victoria, B.C. MOE low, even in stands that had up to 20% Manual . dead trees. This finding should give Ruel, J.-C. 995. Understanding foresters confidence to apply these windthrow: Silvicultural implica- silvicultural systems in caribou habitat tions. For. Chron. 7:434–45. as required under the cclup. There Stathers, R.J., T.P. Rollerson, and S.J. is also potential to implement these Mitchell. 994. Windthrow hand- systems in areas outside of caribou book for British Columbia forests. habitat (with low levels of beetle in- B.C. Min. For., Victoria, B.C. Work. festation) to meet other management Pap. 940. objectives such as biodiversity, visuals, or water quality.

Citation Waterhouse M.J. and H.M. Armleder. 2004. Windthrow in partially cut lodgepole pine forests in west-central British Columbia. B.C. Min. For., Res. Br., Victoria, B.C. Exten. Note 70.

The use of trade, ﬁrm, or corporation names in this publication is for the information and convenience of the reader. Such use does not constitute an oﬃcial endorsement or approval by the Government of British Columbia of any product or service to the exclusion of others that may also be suitable. This Extension Note should be regarded as technical background only.