Mineral lick use by GPS radio collared mountain goats in southeastern British Columbia

Transcription

1 March 2006 Mineral lick use by GPS radio collared mountain goats in southeastern British Columbia Kim G. Poole 1 and Karl D. Bachmann 2 1 Aurora Wildlife Research 2305 Annable Rd. Nelson, BC V1L 6K4 Canada Tele (250) ; kpoole@aurorawildlife.com 2 Bachmann Consulting Services th Av. S. Cranbrook, B.C., Canada V1C 2M3 Tele (250) ; bachmann.consulting@gmail.com EXECUTIVE SUMMARY In most populations of mountain goats (Oreamnos americanus), mineral lick use is an essential part of the ecology of the species. Forestry development has the potential to impact use of licks by altering vegetation along traditional trails, around rocky bluffs near licks, and at licks themselves. Little is known about potential impacts of such development. As part of a larger study, we examined lick use by GPS radio-collared mountain goats in the White River and St. Mary River study areas in southeastern British Columbia. We documented the suspected location, timing and frequency of lick use. Twenty-six and 28 collared goats were monitored during 2004 and 2005, respectively. We observed goat use of 6 verified and 10 suspected mineral licks, 6 in the St. Mary and 10 in the White. Thirteen of the licks were below the forestry operability line. All but 3 of the licks were 600 m from the closest cutblock, and 5 of the licks were <100 m distant. The habitat type immediately surrounding the licks included cutbanks and sloughs along rivers, gullies, road cutbanks, treeline, and alpine areas. Number of visits to licks per season ranged from 0 to 9, and varied between years and between areas. Goats often moved considerable distances prior to visiting licks, up to a maximum of 17.3 km. Most visits by males occurred between early May and late June (median 7 June), and most visits by females occurred between early June and mid-july (median 18 June). Mean time spent at licks on each visit was 1.5 days for females and 1.7 days for males. Two of 5 goat mortalities observed during the 20- month study died at the same lick. Our observations conform to general information about mineral licks from the literature. Distances travelled were often extensive, and the number of licks visited by an individual goat often great. Our technique likely focused on identification of low-elevation licks, many of which were in close proximity to cutblocks and forestry activity. It is unclear how forestry development affects goat use of licks and goat survival. We suggest that field verification of the 10 suspected goat licks identified in this study should be conducted. Additional research on the impact of forestry development and disturbance on licks, traditional trails, and nearby rocky bluffs used for staging areas would enhance goat conservation and management.

2 2 INTRODUCTION Interior populations of mountain goats (Oreamnos americanus) generally make extensive use of natural mineral licks, often travelling to low elevation sites or areas distant from their usual home ranges (Hebert and Cowan 1971, Hopkins et al. 1992, Poole and Heard 2003). While prevalent in most goat populations, lick use is not universal (Varley 1994), and there is a great deal of individual variation among goats in the number of lick visits, timing and duration of visits, and distance traveled to licks (Hebert and Cowan 1971, Singer and Doherty 1985, Hopkins et al. 1992, Klaus and Schmid 1998, Lemke 1999, Poole and Heard 2003). Lick use occurs primarily between April and early autumn, with males generally using licks earlier in the year, and females and family groups beginning to use licks in early June (Hebert and Cowan 1971). Goats generally use traditional trails to access licks. These trails often traverse extensive areas of forest, and goats may stage and rest at rocky bluffs within the timber as they make periodic excursions to the lick (Hebert and Cowan 1971). In areas with limited lick use by goats, mineral sources may be abundantly dispersed throughout the range (Varley 1994). Goat licks are often found in forested valley bottoms where people would be most likely to find them (McCrory et al. 1977, Poole and Heard 2003), but high elevation licks are used (McCrory 1979), and can be identified using collared animals (Poole and Heard 2003). It appears that peak use of licks varies with elevation, with higher elevation licks being used later in the season (Hebert and Cowan 1971, Poole and Heard 2003). Dormaar and Walker (1996) characterized 3 types of licks: dry earth exposures, muck (wet) licks, and rock face licks. While earlier research suggested that sodium was the primary source of attraction to licks for goats (summarized in Hebert and Cowan 1971), elevated levels of magnesium, manganese, iron and copper have also been reported (Dormaar and Walker 1996). Natural mineral springs (wet licks) generally contain more sodium, magnesium, potassium, and bicarbonate, with much higher concentration of sodium and bicarbonate (Bechtold 1996). Bicarbonate may be valuable in neutralizing excess stomach acid produced during ingestion of spring and summer forage (Bechtold 1996). Goats may use licks to maintain sodium balance, since the sodium content of food plants used by goats is low (Hebert and Cowan 1971). A selenium deficiency was suspected in goats in one study, but no link to mineral licks and selenium deficiency was established (Hebert 1967). Forestry development has the potential to impact lick use by altering vegetation along traditional trails, around rocky bluffs near licks, and at the licks themselves. Little is known about the potential impacts of such development. Logged (or burned) areas around a lick site may benefit goats by providing more open sight lines, reducing predation risk, and increasing forage supply. Alternatively, forest cover removal through logging or wildfire may cause abandonment of the site due to higher predation risk. Early seral habitats result in less hiding cover, and may increase the number and adjacency of deer (Odocoileus spp.), elk (Cervus elaphus), and moose (Alces alces), thereby increasing predators such as wolves (Canis lupus), cougars (Puma concolor), and bears (Ursus spp.). No published data are available to support either statement (Côté and Festa-Bianchet 2003). Females during the kidding and post-kidding period appear to be most sensitive to disturbance (Penner 1988). Thus, the most reproductively significant segment of the population may be the greatest affected by reduced mineral intake caused by disturbance. Limited research on lick use by goats has been conducted within the East Kootenay area of southeastern British Columbia. Non-inventory work conducted in the 1960s primarily involved assessment of licks relative to goat distribution and use (Hebert 1967, Hebert and Cowan 1971). In recent years, the Columbia Basin Fish and Wildlife Compensation Program (CBFWCP) has established a database on licks used by goats and other ungulates in the East Kootenay, and has placed trail cameras at various high-use goat licks (L. Ingham, CBFWCP, Invermere, unpublished data). Forestry operations within the area have resulted in alteration of forest cover surrounding traditional travel routes to several

3 3 known licks (K. Stuart-Smith, personal communication). Large wildfires and in some cases subsequent salvage logging have affected other well-used licks. In this study we examined lick use by global positioning system (GPS) radio-collared mountain goats in 2 study areas. The main study was multi-focused, and included assessment of goat wintering strategies (Poole and Stuart-Smith 2006) and examination of factors affecting goat sightability during aerial surveys (Poole, in press). Our main objectives in the lick portion of this study were to document the location, timing, and frequency of lick use by study animals. We examined whether logging in the vicinity of low-elevation mineral licks will decrease goat use of these areas. This information can then be used to refine forestry development plans and develop guidelines to minimize impacts to low-elevation licks and traditional trails to these licks. While some licks used by the collared goats were known previous to the study, other focal areas of goat concentrations were not verified as mineral licks; thus, our analysis and conclusions should be considered preliminary. STUDY AREAS We selected 2 study areas, the St. Mary River area within the southern Purcell Mountains, and the White River area within the Rocky Mountains (Fig. 1). The study areas were 65 km apart, separated by the Rocky Mountain Trench. In both areas the terrain is steep and rugged, with valley bottoms at approximately 1,000 m, and mountain peaks up to 2,850 m in the St. Mary, and 3,200 m in the White. Most goat habitat occurs in the Engelmann Spruce Subalpine Fir (ESSF) biogeoclimatic zone on valley slopes up to ~1,700 2,200 m elevation (depending upon aspect), and in the Alpine Tundra (AT) zone at the highest elevations (Meidinger and Pojar 1991, Braumandl and Curran 1992). Stands of hybrid white spruce (Picea glauca x engelmannii), subalpine fir (Abies lasiocarpa), and lodgepole pine (Pinus contorta) are common within the ESSF, with a transition into parkland and krummholtz at the highest elevations. Whitebark pine (Pinus albicaulis) and alpine larch (Larix lyalli) are common in the parkland immediately below the AT. Below the ESSF zone, both the ICH and MS zones occur in the St. Mary area, and the MS zone occurs in the White area. The ICH zone includes a wide variety of conifer tree species, with western hemlock (Tsuga heterophylla) and western redcedar (Thuja plicata) predominating. The MS zone commonly has hybrid white spruce, western larch (Larix occidentalis), and Douglas-fir (Pseudotsuga menziesii), with extensive seral stands of lodgepole pine due to past fires. Several deciduous species may occur in both the ICH and MS zones. Compared with St. Mary study area, the White River area has a greater amount of AT, with more extensive open mountainous ridgelines and 2 3 times the density of goats (B.C. Ministry of Environment, unpublished data). Many of the valley bottoms in both areas have been commercially logged over the past 50 years, and large burns of varying age occur in portions of the White area. Goat numbers appear to be stable or increasing in both areas, with approximately 430 goats estimated for the St. Mary area and 930 goats for the White area (B.C. Ministry of Environment, Cranbrook, unpublished data). Average mean daily temperatures at the Cranbrook airport, located at 940 m elevation in the Trench south of both areas, are 7.5ºC and 18.3ºC for January and July, respectively. Precipitation averages 383 mm annually (Environment Canada 2005). Climate varies across the study areas. The St. Mary area is within the moist climatic region and the White area is in the dry climatic region of southeastern British Columbia (Braumandl and Curran 1992). Snow depth/density averages approximately 60% higher in the St. Mary study area compared with the White. Average snow water equivalent (snow pillow data) on 1 March from a site adjacent to the St. Mary area at 2,140 m elevation was 390 mm (36 years of data), while a site within the White area at 2,010 m was 239 mm (34 years of data) ( Snow depths during winter were at or slightly below normal until early April, when an early spring resulted in snowmelt about 3 weeks earlier than normal. Snow depths during winter were on average 25 40% below normal.

4 Mineral lick use by GPS-collared mountain goats Figure 1. St. Mary River and White River goat study areas and capture locations, East Kootenay, British Columbia, January and November Green areas denote both provincial parks and conservation areas. 4

5 5 METHODS Mountain goat capture and monitoring We deployed 15 GPS collars on goats in each of the 2 study areas in early January 2004, spreading search effort throughout our areas of interest (Fig. 1). We attempted to roughly equalize sex of collared goats within each area. Four collars were redeployed in early November 2004 to replace animals that had died. Goats were captured by helicopter net-gunning (Barrett et al. 1982). We aged animals by counting the number of distinct horn annuli plus the fainter kid annulus formed at 6 months of age (Smith 1988, Stevens and Houston 1989). We fitted a GPS collar (model G2000, Advanced Telemetry Systems, Isanti, Minnesota, USA) on each goat, programmed to obtain a location every 6:15 hours (3.84 times daily). Capture and handling protocol followed the principles and guidelines of the Canadian Council on Animal Care (1993), and was conducted under British Columbia Ministry of Environment scientific permit DCB0913. The collars were released from the animals in late August and early September 2005 by a remotely fired mechanism (Mech and Gese 1992), recovered, and the data downloaded. These data provided 2 seasons of lick use. Prior to analysis, we deleted locations with a positional dilution of precision (PDOP) >10 to remove major outliers (D Eon and Delparte 2005). Mean GPS location error was assumed to be 11 m (D Eon et al. 2002), likely within the range of accuracy of the habitat data we used. Analysis We examined use of licks by identifying extra-home range movements and use of low elevation habitat. While we were aware of a number of licks within both study areas, and verified new licks during the course of the study, we stress that many of these sites are of suspected licks, and suspected lick sites should be verified with fieldwork. Accordingly, we have rated mineral licks as known or verified licks (V), and suspected licks that require field verification (S). We used 1:20,000 scale terrain resources information management files (TRIM; Geographic Data BC 1992), 1:20,000 scale current forest inventory planning files (FIP; Resources Inventory Branch 1995); ortho photos (Tembec Inc., 1998 for the White area, 2004 for the St. Mary area); and the most recent timber harvesting history with road coverage (Tembec Inc., Cranbrook, British Columbia, Canada). Forestry operability was defined using Tembec's standard operability line layer (last updated in 2004). Analyses were conducted using ArcView geographic information system (ArcView 3.2a and 8.3, Environmental Systems Research Institute, Redlands California, USA). We examined lick use separated by year and individual. To facilitate tracking of licks, we gave licks arbitrary names based on landscape features. We put a 500 m radius buffer around the verified or suspected lick points, and then selected all the points falling within that buffer to calculate arrival, departure, and duration of lick visit by each individual. If the 500 m radius included an obvious rocky staging bluff near the lick, the radius was decreased to exclude that bluff. Given >6 hours between GPS location attempts and some missed locations, the 500 m buffer appeared to a reasonable compromise between detecting most likely visits to licks, and not over-recording animals hanging around the broader area of the lick. Goats could presumably travel 500 m to a lick in minutes on a good trail. We used previously recorded UTM coordinates of verified licks, or the rough centre of a location point cluster in the case of suspected licks. Distance from last core area used was calculated by examining the last broad concentration of goat locations, and following points in sequential date sequence to the lick area using the ruler tool in ArcView. No distance was given if the goat remained in the immediate lick area for longer than ~1 week prior to lick use. Suspected travel through cutblocks en route to licks or in the vicinity of licks was examined using the timber harvest data layer.

6 6 We obtained topographic and forest overstory information for each lick from digital data sources, and calculated the distance to nearest cutblock. We also examined goat response to experimental harvesting adjacent to a well-used lick site (Redding 4 km lick). A cutting permit 17.1 ha in size 190 m northeast and upslope of the Redding 4 km lick was proposed for harvesting in early winter of We compared goat use of this lick in 2004 (pre-harvest) and 2005 (post-harvest) by examining use of the lick area by our collared goats. RESULTS Capture and GPS collars We placed GPS collars on 15 goats (8 F, 7 M) in the St. Mary River study area and 15 goats (5 F, 10 M) in the White River study area in January One collar on a female in the St. Mary failed on deployment, and was removed from further analysis. Average age of collared goats was 5.6 years (± 0.47 SE; range years), and did not differ between study areas (t = 0.56, P = 0.58). Collared goats were well distributed throughout the study areas on a number of individual ranges. Two females died during winter 2004 in the St. Mary, giving 10 females and 17 males (St. Mary: 5 F, 7 M; White River: 5 F, 10 M) monitored during at least a portion of lick season Two additional goats died during June July 2004 in the White. In early November 2004 we captured and fitted GPS collars on 4 more female mountain goats (2 in each area), ranging in age from years. Thus, during lick season 2005 we monitored 13 females and 16 males (St. Mary: 7 F, 7 M; White River: 6 F, 9 M). One male in each lick season in the White was in areas not covered by ortho photos, and was therefore removed from analysis. Twenty-three goats (9 F, 14 M) were monitored over both the 2004 and 2005 lick use seasons. A fifth male goat died in early August 2005 in the White. GPS collar data quality was high, with 87.8% location success and 83.1% 3-dimensional fixes (n = 53,675 GPS locations). Mineral licks Our analysis indicated goat use of 6 verified and 10 suspected mineral licks, 6 in the St. Mary and 10 in the White (Figs. 2, 3). One goat location (from goat 410M) also was located within 380 m of the verified Redding 10 km lick, but since no other collared goat locations were located at the lick, we did not consider this lick further in our analysis. We suspect all licks were of the dry type because all 8 goat licks we have visited to date throughout the East Kootenay have been dry licks. Almost all of the licks have been under large or veteran trees (primarily if not solely Douglas-fir trees), and nothing on the ortho photos suggested that moisture was present at any site. Licks were located at an average elevation of 1,550 m (±77; range 980 2,260 m, n = 16), on an average of 66% slope (±9.5; range 3 135%). Nine of the licks were on a south aspect, 1 on west aspect, and 3 each on north and east aspects. The habitat type immediately surrounding the licks ranged from cutbanks and sluffs along rivers (4 licks), to gully walls (6), road cutbanks (3), treeline (1), and unknown (2)(Appendix 1). Half of the licks were in stands with no overstory data in the FIP database (mostly alpine, rock and clearing descriptors). The other 8 sites with overstory data were in lodgepole pine (n = 4), Douglas-fir (n = 2), and spruce-subalpine fir (n = 2) stands, with an average age of 142 years (±17; range years) and 44% crown closure (±7.3; range 10 70%). Thirteen of the licks were below the forestry operability line. Distance from licks to the closest cutblock ranged from 45 7,700 m; however, all but 3 of the licks were 600 m from a cutblock, and 5 of the licks were <100 m distant. Exposed mineral soils, either natural or anthropomorphic in origin, appeared to be a common factor for many licks.

7 7 Figure 2. St. Mary River study area showing the location of verified (red star-burst) and suspected (purple question mark) mineral licks used by mountain goats, southern British Columbia, One collared goat may have used the Redding 10 km lick in June A suspected lick also was also found in the headwaters of Dewar Creek in the Purcell Wilderness Conservancy, 32 km north of the Redding 10 km lick. One goat used the Redding 4 km lick in 2004, making 3 visits between May and June. The adjacent block was harvested 7 January 1 February Additionally, another cutting permit was harvested during this same time period approximately 1,400 m to the northwest. No use of this lick by collared goats was detected in the 2005 lick season, although the same goat was located once 350 m from the lick along the creek. Although it was impossible to determine where trails to licks were located using digital mapping, no goat locations occurred in cutblocks above 14 of the licks, in some cases because no blocks occurred above the licks, and in other cases because no goat locations occurred in blocks located above licks. In 2 cases (Hellroaring Creek, 25 locations in a 4-year old cutblock; and Lower Blackfoot, 2 locations in a 3- year old cutblock) goats appeared to make some use of cutblocks on suspected access routes to mineral

8 8 licks, but in only 1 case (Hellroaring Creek) were more than a few locations detected. Two to 3 goat locations were detected in cutblocks below or adjacent to 5 of the licks. Figure 3. White River study area showing the location of verified (red star-burst) and suspected (purple question mark) mineral licks used by mountain goats, southern British Columbia, Of the 26 goats monitored in 2004, we detected lick use by 19 animals. Of the 28 goats monitored in 2005, we detected lick use by 19 animals. Number of visits to licks per season ranged from 0 to 9, and varied between years and between areas (Table 1). Small sample sizes resulted in poor power for t-tests between pairs of means. However, there was a tendency for goats in the White to have on average more lick visits than goats in the St. Mary, and for males to have more visits than females (Table 1). Goats often moved considerable distances prior to visiting licks, up to a maximum of 17.3 km (Table 2). There were no significant differences in the movement distances between areas (Table 2; t = 0.82, 83 df, P = 0.34) or sexes (t = 0.30, 83 df, P = 0.77).

9 9 Table 1. Frequency of mineral lick visits by collared mountain goats in the St. Mary River and White River areas, southern British Columbia, Area Year Sex n Mean SE Min Max St. Mary 2004 F M F M White 2004 F M F M St. Mary 2004 Both Both White 2004 Both Both St. Mary Both Both White Both Both F M Table 2. Distance traveled (m) by collared goats from previous core areas to mineral licks in the St. Mary River and White River areas, southern British Columbia, Area Sex n Mean SE Min Max St. Mary White M F

10 10 We detected visits to licks between early February and mid-august (Fig. 4). Most visits by males occurred between early May and late June (median 7 June), and most visits by females occurred between early June and mid-july (median 18 June). The same female made 3 visits to the previously known Redding 4 km lick in February and March While outside of our defined period of lick use, visits to the lick area were readily detected. When these winter lick visits were removed, the median date of lick use by females shifted to 21 June. Mean time spent at licks on each visit was 1.5 days for females (±0.19, range 1 8, n = 40) and 1.7 days for males (±0.12, range 1 6, n = 91; t = 0.96, 129 df, P = 0.34) Number of lick visits Feb 24 Feb 9 Mar 23 Mar 6 Apr 20 Apr 4 May 18 May 1 Jun 15 Jun 29 Jun 13 Jul 27 Jul 10 Aug 24 Aug F 2004 M 2004 F 2005 M 2005 Figure 4. Visits to verified or suspected mineral licks by female (n = 40) and male (n = 91) mountain goats by 2-week periods in the St. Mary River and White River areas, southern British Columbia, Dates indicate the start of each 2-week period. Two goats, a male and a female, died at the Middle East Thunder lick (Fig. 5) nearly 1 year apart. Both animals had made multiple, extensive movements from mountain ranges to the west to visit licks within the Thunder Creek drainage. Both mortalities occurred in late July and early August. Although predation was suspected in each case, the actual cause of death was not known.

11 11 DISCUSSION We detected significant lick use by collared goats in the St. Mary River and White River study areas, and we were able to detect lick use by the majority of collared goats. Our observations generally conform to available information about goat use of licks from the literature. We observed individual variation among goats in the number of lick visits, timing and duration of visits, and distance traveled to licks (Hebert and Cowan 1971, Singer and Doherty 1985, Hopkins et al. 1992, Klaus and Schmid 1998, Lemke 1999, Poole and Heard 2003). We observed males generally using licks earlier than females, and what appeared to be use of rocky bluffs near licks for staging areas for visits to licks. Although we detected suspected lick use by one female goat in the St. Mary in February and March, the bulk of lick use occurred between April and early August. Our analysis should be considered somewhat conservative given our GPS location frequency and 500 m radius buffer used for identifying lick use. For example, goat 414F in the White made km trips from the Quinn Range in July 2004 around the head of Thunder Creek to licks on the east side of Thunder Creek that were detected using our 500 m buffer criteria. However, this goat also made an 11 km trip during June 2004 to areas above the Middle East and Upper East Thunder licks that was not detected by our analysis because no locations occurred within the 500 m radius buffer. We suspect there is a high probability that some lick use occurred during the June visit. Our analysis is biased towards detecting low elevation lick sites outside of the normal high elevation ranges used by mountain goats. At our relatively coarse temporal scale (~ 4 locations per day), long distance and low-elevation movements stood out from core range locations, and were easier to identify as potential lick use. High elevation lick use occurs in some populations (e.g., Poole and Heard 2003), but it is difficult to determine why a goat utilizes a specific area at high elevation or within core home ranges using remote, digital data. Again, given our coarse temporal scale, we acknowledge that our analyses likely missed use of some low-elevation licks by the collared animals. Thus, this analysis should be considered preliminary and somewhat conservative, and not as an inventory of all licks used by the collared goats within the study areas. Distances travelled were often extensive, and some individuals visited a number of licks. For example, 6 collared goats in the upper White River area travelled up to 17 km to visit the Barr Creek lick, inhabiting the mountain block immediately to the north of the lick from the beginning of May to midsummer each year. In late July to mid-august each year, all of these collared goats abandoned this block, and returned to ranges to the northeast and east. One male goat (412M) moved extensively over the Van Nostrand, Quinn, and Mt. Folk ranges to visit all 5 verified and suspected licks in the Thunder Creek drainage. Initial movements to licks in the spring were often long. Male 426M moved 17 km over a 7- day period down the entire length of a ridge to go directly to the suspected Mt. Patrick lick. Use of licks is probably traditional, and subject to a number of factors determining whether they are used by a particular group of goats. CBFWCP identified 4 licks, including the Maiyuk lick (not used in our study), used by goats in Height of the Rockies Provincial Park. These licks were <3.5 4 km from 6 collared goats in our study, the same goats that made extensive movements of up to km to the Barr Creek and Mt. Patrick licks. None of these collared goats appeared to use the 4 CBFWCP licks because no locations were detected even remotely in the direction of these licks. Many licks were in close proximity to cutblocks and forestry activity. It is unclear how proximity to forestry development affects goat access to and use of low elevation sites, and their survival. We found limited evidence of goats either avoiding or using cutblocks en route to licks. On 1 trail system and lick complex in the Ospika Valley in northern British Columbia where a 150 m buffer was left around the main trail to a lick, goat use of the trail decreased and use of the adjacent clearcuts increased (to about half of total trips), with no apparent decline in overall use of the lick (M. Wood, Peace-Williston Fish and Wildlife Compensation Program [PWFWCP], personal communication). A logging road also bisecting the trail resulted in reduced movements for only about 1 year. Two of 5 mortalities reported during our

12 12 collaring study were at the Middle East Thunder lick (Fig. 5), <100 m from a cutblock. Goats have an increased risk of predation at and below treeline (Festa-Bianchet et al. 1994, Côté and Beaudoin 1997); this risk may be compounded if cutblocks alter the prey and predator community at these low elevation sites. Increases in early seral habitats may increase populations of deer, elk, and moose such that potential predators of goats wolves, cougars and bears (Côté and Festa-Bianchet 2003) may become more numerous adjacent to licks, resulting in goats being taken more often as secondary prey. Researchers in the Ospika study noted increased trail camera detections for moose, deer, and elk, as well as bears and wolves subsequent to logging in the area (M. Wood, PWFWCP, personal communication). A similar situation has been proposed for woodland caribou (Rangifer tarandus caribou) in central and southern British Columbia (Wittmer et al. 2005). Trees may concentrate minerals sought by goats within their root systems. At several licks, including the Middle East Thunder lick (Fig. 5), goats appear to focus their soil eating in and around the base of trees. Hebert and Cowan (1971) noted that preferred sites at licks in southeastern British Columbia were comprised of finer textured soils usually located among the roots of Douglas-fir trees. These authors suggested that preferred sites contained a large proportion of moist brown clays (Fig. 6). Although the reasons why goats appear to prefer sites under tress is unclear, harvesting of these sites would not only affect the physical dynamics of the site, but may alter the quality of the soil being ingested. Figure 5. Karl Bachmann at the Middle East Thunder lick, White River study area, southern British Columbia, September Note how goats have excavated soil under the Douglas-fir trees.

13 13 Figure 6. Goat teeth marks in brown soils of the Redding 10 km lick, St. Mary study area, southern British Columbia, June Several issues surrounding goat lick use in this area remain unresolved. We noted some anecdotal evidence that forestry operations disrupted goat lick use at the Redding 4 km lick. However, the digital data did not facilitate this kind of analysis, especially since trail systems were un-mapped at all licks. A rigorous experimental study focused on forestry operations disturbance of licks could employ more frequent GPS collar fixes, trail cameras, and data loggers to detect fine scale movements associated with lick use. Also, certain licks seem to be especially important to goat ecology, as suggested by the long distance movements some individuals made to access them, bypassing other, closer licks. Soil and chemical analysis of licks could elucidate if and/or why some licks are critical, and could identify the most important in management plans. MANAGEMENT IMPLICATIONS Licks are important to mountain goat ecology. Disturbance at lick sites may affect goat reproduction and survival. Therefore, resource managers should be aware of and plan around goat lick locations in the landscape. Location information can be used to refine forestry development plans and develop guidelines to minimize impacts to low-elevation licks and traditional trails to these licks. Field verification of the 10 suspected goat licks identified in this study should be conducted. Field visits during May and June would likely identify lick location, traditional trails, and may verify current use of licks by goats. One aspect of goat ecology that is unclear is whether goats prefer open or forested habitat along access routes to low elevation licks, and the resulting impacts on goat use of sites and goat survival. For example, there are limited data on the implications to goats if the forest surrounding a traditional trail to a lick or the lick itself is harvested. It is unclear whether the lick would be abandoned, or another forested trail would be developed, or if the goats would continue to access the lick via the open cutblock. The implications for predation risk and fitness from these options are unknown. Similarly, the implications of disturbing the forest around a rocky bluff traditionally used by goats as a staging area prior to lick access are unclear. Finally, no studies have examined whether licks close to cutblocks have higher goat mortality rates. Given the current concern surrounding goat populations in general in B.C., additional

14 14 research on these questions would give managers valuable information for mitigating impacts from forestry on goat licks. Acknowledging that we sampled lick use by a comparatively small number of collared goats, several areas stood out as of higher importance. The Thunder Creek drainage holds 5 licks that were used by 4 collared goats, with 1 male visiting all 5 licks. This drainage has seen extensive forestry development, but the impacts on goat use of these licks are poorly understood. Similarly, the Barr Creek lick was used by 6 collared goats that travelled long distances. This lick is undoubtedly of importance to the local population. These examples clearly indicate why managers need consistent and effective mitigation strategies they can implement at these and other licks identified in this study. ACKNOWLEDGEMENTS Tembec Industries through their Forestry Investment Account (FIA), BC Ministry of Environment (MoE), and the Habitat Conservation Trust Fund (HCTF) provided funding for this project, and we thank them for their generous support. This project was developed in collaboration with K. Stuart-Smith, Tembec, Inc., and I. Teske, BC MoE, and we thank them for their insights and discussions. M. Belcher and T. Pachara, Tembec Industries, and I. Teske, BC MoE, provided administrative support. We greatly appreciate the capture efforts of L. Ingham and D. Lewis, CBFWCP, and for sharing with us their mineral lick database. C. Wilson, T. Vanderbrink, T. Macdonald, and B. Stewart, Bighorn Helicopters, ably captured goats under trying winter conditions. D. Pritchard provided GIS outputs of lick sites. K. Stuart-Smith provided valuable comments on an earlier draft of this manuscript. LITERATURE CITED Bechtold, J.-P Chemical characteristics of natural mineral springs in northern British Columbia, Canada. Wildlife Society Bulletin 24: Barrett, M. W., J. W. Nolan, and L. D. Roy Evaluation of a hand-held net gun to capture large mammals. Wildlife Society Bulletin 10: Braumandl, T. F., and M. P. Curran A field guide for site identification for the Nelson Forest Region. B.C. Ministry of Forests, Victoria, British Columbia, Canada. Canadian Council on Animal Care Guide to the care and use of experimental animals. Volume 1, 2nd edition. Canadian Council on Animal Care, Ottawa, Ontario, Canada. Côté, S. D., and C. Beaudoin Grizzly bear (Ursus arctos) attacks and nanny-kid separation on mountain goats (Oreamnos americanus). Mammalia 61: Côté, S. D., and M. Festa-Bianchet Mountain goat. Pages in Wild mammals of North America: biology, management, and conservation. G. A. Feldhamer, B. Thompson, and J. Chapman, editors. The John Hopkins University Press, Baltimore, Maryland, USA. D Eon, R. G., R. Serrouya, G. Smith, and C. O. Kochanny GPS radiotelemetry error and bias in mountainous terrain. Wildlife Society Bulletin 30: D Eon, R., and D. Delparte Effects of radio-collar position and orientation on GPS radio-collar performance, and the implications of PDOP in data screening. Journal of Applied Ecology 42: Dormaar, J. F., and B. D. Walker Elemental content of animal licks along the eastern slopes of the Rocky Mountains in southern Alberta, Canada. Canadian Journal of Soil Science 76:

15 15 Festa-Bianchet, M., M. Urquhart, and K. G. Smith Mountain goat recruitment: kid production and survival to breeding age. Canadian Journal of Zoology 72: Geographic Data BC British Columbia specifications and guidelines for geomatics. Content series volume 3: digital baseline mapping at 1:20,000. Release 2.0. British Columbia Ministry of Environment, Lands and Parks, Victoria, British Columbia, Canada. Hebert, D. M Natural salt licks as a part of the ecology of the mountain goat. M.S. thesis, University of British Columbia, Vancouver, British Columbia, Canada. Hebert, D., and I. M. Cowan Natural salt licks as a part of the ecology of the mountain goat. Canadian Journal of Zoology 49: Hopkins, A., J. P. Fitzgerald, A. Chappell, and G. Byrne Population dynamics and behavior of mountain goats using Elliot Ridge, Gore Range, Colorado. Proceedings of the Northern Wild Sheep and Goat Council 8: Klaus, G., and B. Schmid Geophagy at natural licks and mammal ecology: a review. Mammalia 62: Lemke, S. L Cayoosh Range mountain goat study results and recommendations. Unpublished report for BC Ministry of Environment, Lands and Parks, Kamloops, British Columbia, Canada. McCrory, W. P, D. A. Blood, D. Portman, and D. Harwood Mountain goat surveys in Yoho National Park, B.C. Pages in W. Samuel and W. G. Macgregor, editors. Proceedings of the First International Mountain Goat Symposium, Kalispell, Montana. B.C. Fish and Wildlife Branch, Victoria, British Columbia, Canada. McCrory, W. P An inventory of the mountain goats of Glacier and Mount Revelstoke National Parks, British Columbia. Unpublished report. Parks Canada, Western Region, Glacier National Park, Revelstoke, British Columbia, Canada. Mech, L. D., and E. M. Gese Field testing the Wildlink capture collar on wolves. Wildlife Society Bulletin 20: Meidinger, D., and J. Pojar Ecosystems of British Columbia. Special Report Series 6, B.C. Ministry of Forests, Research Branch, Victoria, British Columbia, Canada. Penner, D. F Behavioral response and habituation of mountain goats in relation to petroleum exploration at Pinto Creek, Alberta. Biennial Symposium Northern Wild Sheep and Goat Council 6: Poole, K. G. Does survey effort influence sightability of mountain goats during aerial surveys? Wildlife Biology. In Press. Poole, K. G., and D. C. Heard Seasonal habitat use and movements of mountain goats, Oreamnos americanus, in east-central British Columbia. Canadian Field-Naturalist 117: Poole, K. G., and K. Stuart-Smith Wintering strategies by mountain goats in interior mountains. Unpublished report prepared for Tembec Inc., Cranbrook, and British Columbia Ministry of Environment, Cranbrook, British Columbia, Canada. Resources Inventory Branch Relational data dictionary (RDD) 2.0. British Columbia Ministry of Forests, Victoria, British Columbia, Canada. Singer, F. J., and J. L. Doherty Movements and habitat use in an unhunted population of mountain goats, Oreamnos americanus. Canadian Field-Naturalist 99: Varley, N. C Summer-fall habitat use and fall diets of mountain goats and bighorn sheep in the Absaroka Range, Montana. Proceedings of the Northern Wild Sheep and Goat Council 9:

16 16 Wittmer, H. U., A. R. E. Sinclair, and B. N. McLellan The role of predation in the decline and extirpation of woodland caribou. Oecologia 144:

17 17 Appendix 1 Mineral licks (V = verified; S = suspected) identified using GPS-collared mountain goats, in the St. Mary River and White River areas, southern British Columbia, Area Lick Name Lick Type In CB database Dist. to Cutblock (m) Elev. (m) Aspect (º) Slope (%) UTME UTMN Habitat Around Lick St. Mary Angus Creek V N Road cutbank St. Mary Dewer Ck S N Unknown St. Mary Hellroaring Creek S N Road cutbank St. Mary Lower St. Mary S Possibly Road St. Mary Redding 4 km V Y Stream bank St. Mary Upper Meachen S N Stream bank White Barr Creek V Y Bottom of dry, forested gully White Elk Creek S N Steep-walled gully White Lower Blackfoot S N Cutbanks and sluffs above river White Lower East Thunder V N Sluff on gully edge White Lower West Thunder V Y Treeline area White Middle East Thunder V N Steep-walled gully (ephemeral) White Middle East White S N Cutbanks and sluffs above river White Middle West Thunder S N Sluff on gully edge White Mt Patrick S N Unknown White Upper East Thunder S N Steep-walled gully (ephemeral)