Sampling moose browse in GMU 19D, February 2000

Tom Paragi, ADF&G, Div. Wildlife Conservation, Fairbanks (25 February 2000)

During 15-18 February 2000, Al Root (ADF&G, McGrath) and I conducted browse surveys on sites along the Kuskokwim River within about 15 miles of McGrath, Alaska (Game Management Unit 19D).  Previous surveys were conducted on these sites by Jack Whitman in late May or early June (1989, 1990, 1991, and 1994) in selected patches of riparian willows to assess overwinter browsing by moose.  Ice scouring obliterated flagging that had marked the start and end of survey transects, so location of transects had to be surmised from 1:63,360 scale topographic maps.  A scaled photograph on slide film was taken in a representative section of each transect for an archive.

Twelve surveys (see attached map for locations) were completed by sampling every five paces along 250 m transects, each a total of 50 data points, to be consistent with previous surveys.  At every fifth footfall, the closest tree or shrub available to moose browsing (defined as <3.1 m [10 ft]) was categorized to species when possible.  Height and distance to nearest available stem of the same species were estimated by use of a graduated meter stick.  Browsing intensity class was assigned to the stem based on the count proportion of current annual growth (CAG) leaders that were browsed by moose:  none, low (1-25% of leaders), moderate (26-75%), or high (76-100%).  Notes were taken on the proportion of stand that had a hedged growth form (browsing intensity from previous years), visible proportion of decadent stems (largely dead), and other criteria indicative of browsing history and stand condition.   

An importance index to gauge browsing intensity on a species for a given year was calculated when data for all transects were combined, similar to previous years (J. Whitman memos, 1989 and 1994, ADF&G files, McGrath).  Midpoint of range for browse intensity classes (e.g., midpoint of low [1-25%] = 0.125) was multiplied by the grand mean percent frequency of samples for each browse intensity class, by species.  The sum of the three products (L, M, and H classes) was then multiplied by the frequency of occurrence of the species along all transects to derive the importance index (Table 1).

Results

The vast majority of riparian willows were of the same growth form: a stout leader up to 2 m on a single stem or small number of stems from a common base, with CAG being smooth bark of olive or carmel color and height sometimes exceeding 5 m.  Although the winter bark on its CAG did not have the wooly hairs characteristics of feltleaf willow (Salix alaxensis) in other parts of Alaska, the species was confirmed to be feltleaf (R. Skinner, Innoko National Wildlife Refuge, McGrath, pers. comm.).  A few specimens of redstem willow (S. arbusculoides), a shrubby form to 2.5 m of multiple stems with delicate reddish leaders, were noted, along with a few possible S. bebbiana having a right-angle branch pattern.   Other species noted occasionally on previous surveys (S. glauca and S. planifolia var. pulchra) were beyond our taxonomic abilities for winter willows.  Detection of shorter willow species would have been hindered by snow depth of 80-110 cm among all transects, with variation in snow accumulation attributed to shrub height and density and wind direction.  Most of this snow fell in three storms between 20 December and the end of January (A. Root, pers. comm.).  Balsam poplar (Populus tremuloides) and American green alder (Alnus crispa) composed <5 % of most transects.

Although the number and location of browse transects sampled has varied among years, feltleaf willow continues to be the predominant species in riparian willow stands near McGrath (Table 2).   The post-winter assessments that occurred during 1989-94 showed a consistently high percentage of individual stems browsed but substantial variation in the distribution of browsing extent on those stems (importance index; Table 2).   The browsing extent on feltleaf willow was higher during winters of deep snow compared to winters of normal snow depth, and in severe winters the moose sometimes browsed beyond the extent of CAG (J. Whitman, ADF&G, pers. comm.).  Although the sampling in February 2000 occurred at least two months prior to the end of snow cover and beginning of new willow growth, 95% of feltleaf stems had already been browsed (Table 2), with 84% of the browsing in the high intensity class (76-100% of stems browsed per individual). 

Most willow bars surveyed in past years have now grown to a height extent becoming marginally feasible for moose browsing.  We observed browsing to 5 m where the moose obviously had “walked down” the stems between their front legs to reach the smaller-diameter CAG on feltleaf willow.  The predictable pattern of increased height and self-thinning of stems (greater nearest-neighbor distance) over time following disturbance was reasonably evident on sites we visited (Table 3).  We skipped transect 89-G (“old town slough”) because most of the plot appeared to be tall alders with very little available feltleaf or redstem willow.   Three other sites near town (89-E, 89-F, and 94-H) were not sampled because of time constraints.  Two of the sites were adjacent to the airport runway and had mechanical disturbance that is maintaining shorter willow.  Undisturbed areas in the three sites had willows >3 m tall but were still being used by moose (A. Root, in litt.). 

Overall, the hedged or broomed growth form of willows that have endured heavy, persistent browsing beyond CAG was not common in the riparian feltleaf willows we observed near McGrath.  Dead stems composed <10% of areas within stands we visited, which had stem breakage only in pockets of heavy browse use.  Because of relatively infrequent browsing pressure and the time lag since widespread ice scouring, substantial parts of some riparian terraces are maturing into willow forest and becoming dominated by alder thickets >4 m tall.

Discussion

Data from these browse transects and general observations over time suggest that the intensity of browsing on riparian willows by moose along the upper Kuskokwim near McGrath seems proportional to depth and timing of snowfall.  In winters of normal snowfall, moose in this area generally remain on upland or non-riparian lowland range dominated by S. planifolia var. pulchra and paper birch (Betula papyrifera) (R. Skinner, pers. comm.).  Deep snowfall, particularly early in the winter, reduces availability of shorter willows and forces moose to heavily utilize riparian stands dominated by feltleaf willow.  Because of this range use pattern, inferences on browsing intensity or browse quality made solely from riparian sites may not be indicative of moose abundance relative to long-term carrying capacity of their range.

Riparian willow stands can be critical winter forage for moose in years of prolonged deep snowfall in this area (R. Skinner, pers. comm.).  Despite observed browsing by moose to 5 m this year, riparian feltleaf stands on the Kuskokwim River near McGrath appear to be growing beyond the utilization range of moose.  The last flood that produced widespread ice scouring was in 1991 (A. Root, pers. comm.).  Current recruitment of feltleaf willow appears to be primary succession on sandbars below the main river terrace, a relatively small proportion of riparian willow stands in this area.  The risk of substantial mortality by moose from malnutrition in a severe winter would be reduced by the occurrence of a scouring spring flood to knock down the larger riparian willows.  With heightened concern over abundance of moose in GMU 19D, mechanical crushing of willows near towns on the Kuskokwim River would be a definitive approach to rejuvenating winter forage for severe winters in a small portion of the winter range.  The cost of such a limited spatial approach should be weighed against the potential benefits of using prescribed fire to rejuvenate a much large upland range, provided that upland willows are showing substantive evidence of reduced productivity associated with age decadence or heavy browsing by moose. Prescribed fire is a proactive approach relative to waiting for chance ignition to occur, even on areas already designated for limited suppression of wildland fire.

Methods of assessing a population size relative to its carrying capacity are challenging, both conceptually and technically.  Surveys of range condition over a more inclusive cross-section of upland, lowland, and riparian willow stands would be instructive of browsing intensity and possibly predictive of population-level effects driven by nutrition.  However, difficult access to large, sometimes poorly-defined seasonal ranges would require costly techniques of remote sensing and image processing, or costly access and labor for ground sampling.  Rump fat or other condition indices for moose subsume the effect of range quality and weather, but they require capture and handling of individuals, which is costly and not without danger of injury or mortality to free-roaming wild animals.  An annual assessment of reproductive fitness would track the response of a population to range and weather conditions without the need to capture and handle individuals.  For example, neonatal twinning rate could be estimated by observing about 30 parturient cow moose from fixed-wing aircraft.  Periodic range assessment, when and where feasible, would provide a complementary index to population health and possibly identify areas and objectives for habitat enhancement.

Table 1.  Raw counts and calculations for importance index of feltleaf willow in February 2000 along the Kuskokwim River (GMU 19D) near McGrath.

   ^aTwo transects were less than 50 points because at some point along transect all willows were >3 m tall and offset continuation of the transect was not feasible.

Proportion of all transects points that were feltleaf willow = 506/581 = 0.871

Calculation of importance index for feltleaf willow:

Browse            Midpoint          Feltleaf use

class                 of class frequency         Product

Low                 0.125               0.042               0.005

Moderate         0.500               0.111               0.056               Proportion

Heavy              0.875               0.794               0.695               feltleaf

                                                                        -------

                                                            Sum     0.756      x       0.871     =  0.658

Table 2.  Statistics on feltleaf willow stems along the Kuskokwim River near McGrath.  Data were collected in late May or early June except in 2000 (February).  See text and Table 1 for importance description and calculation methods.

Year         # transects                % feltleaf                      % browsed                   Importance      

1989                8                           68                                 89                              0.414

1990                3                           39                                 86                              0.214

1991                6                           87                               100                              0.650

1994              14                           86                                 85                              0.307

2000              12                           87                                 95                              0.658          

Table 3.  Change in mean height (m) and mean nearest-neighbor distance (m) for feltleaf willow stems <3.1 m tall on selected transects distributed along the Kuskokwim River near McGrath.  Data were collected in late May or early June except in 2000 (February).  Sample size can be inferred from Table 2.  Precision of measurements was 0.1 m.  A flood with extensive ice scouring occurred just prior to the 1991 survey. 

Transect   Metric          1989                1990                1991                1994                2000   

89-A        Height          2.59                 --                     1.82                 --                     2.07

89-A        Distance       0.16                 --                     0.31                 --                     0.34

89-B        Height          1.65                 --                     1.46^a                --                     2.10

89-B        Distance       0.41                 --                     0.40^a                     --                     0.34

89-D        Height          1.11                --                     --                     --                     1.66 89-D            Distance       0.48                 --                     --                     --                     0.37

89-H        Height          1.62                 1.78                 1.41^b                --                     2.68

89-H        Distance       0.31                 0.27                 0.39^b                --                     0.59

^aTwo transects were done at this site in 1991, so this is a grand mean of the two means.