Use of pellet group counts in determining density and habitat use of moose Alces alces in Finland

1. What they did

In Finland, moose numbers have been monitored with questionnaires from hunters (moose sighting cards) and with aerial and ground censuses, but these methods have been criticised. The authors tested pellet group counts (counting the piles of dung moose leave) as another way to estimate how many moose use an area in winter and which habitats they use.

The study took place at Lakomäki, a well-known moose wintering area of managed pine forest in central Finland. In spring and early summer 1997:

• Strips. Three observers walked 45 strips, each 4 metres wide and 300 to 1,440 m long (772 m on average). One walked a compass line along the middle and counted, and the other two each counted a 2-metre band to either side.
• Plots. Along the same strips they counted pellet groups on small circular plots (2 m radius, 12.57 m²) every 20 metres, 1,734 plots in all. They also worked out results using only every second, third or fourth plot (40, 60 and 80 m apart).
• Which pellet groups counted. Only groups lying on dead grass, hay or other dead plant material from late autumn to spring were counted, so that only winter droppings were included. A group on the edge was counted if half or more of its pellets lay inside.
• Density. They calculated moose per km² from the number of pellet groups, the area searched, the number of days the pellets had built up (210 or 240) and the number of times a moose defecates per day (14 or 20.9, from two earlier studies).
• Habitat. They classified every strip and plot by soil (mineral soil or peatland), main tree species, forest site type, stage of stand development and the amount of food moose could reach in winter and in summer, and compared where pellet groups were found with how much of each habitat was available.

2. What they found

Moose density

• Estimates varied a lot with the method and the numbers used: from 4.0 to 6.8 moose per km² on strips, and from 7.2 to 14.3 per km² on plots.
• Plot counts gave about twice the density of strip counts.
• Spacing the plots further apart made little difference.
• The defecation rate used had a large effect (about 30% difference between 14 and 20.9 per day). The text puts the 30% down to the accumulation period, but in the table that changes the estimates by about 12% (see detailed keywords).
• For comparison, aerial surveys gave about 2 moose per km² in winter 1995/96 and about 3 per km² in winter 1998/99. Hunters estimated about 1 per km² after the hunting season in December 1996.

Habitat use

• Moose did not use habitats in proportion to their availability.
• Pellet groups were found more than expected in mineral soil forest, pine-dominated stands, the poorer Calluna site type and young thinning stands, which made up 63% of the area but held 86% of the pellet groups on strips.
• Pellet groups were found less than expected in peatland forest, spruce and birch stands, the more fertile site types, open and young stands, and older thinning and mature stands.
• Habitats with moderate or good winter food held more pellet groups than expected. Habitats with good summer food held fewer.
• Strips and plots gave very similar results for habitat use.

3. What they concluded

• Based on other information about the area, the strip counts were closer to the real number of moose, and the plot counts overestimated it. One possible reason is that observers may include more groups lying on the border of a small plot than along the edge of a strip.
• Both methods may still have large errors, even with three experienced observers and moose pellets being fairly large. The variation in how often moose defecate is a well-known limit on accuracy.
• Strip counting covers more ground and is likely more accurate, but takes longer. Plot counts gave similar results even at long plot intervals.
• The density estimates were much higher than the hunters’ estimate, which could be because moose move into this wintering area from their summer ranges.
• Because pellet counts cover a whole winter rather than the short moment of an aerial survey, the authors believe they can be reliable enough for management, and they are much cheaper than helicopter surveys.
• The highest pellet group densities were in young pine thinning stands, where moose can reach many side twigs and browse heavily. The authors believe the distribution of pellet groups gave a good picture of moose activity and can be used to estimate habitat use.
• Pellet group counts could be added to Finnish wildlife triangle censuses (12 km triangles counted each year) as a third count in spring, soon after snowmelt and before plants grow too dense, and should include as many game species as possible. Permanent marked plots, cleared of old pellet groups in late summer, would give the exact number of days the pellets had built up.
• Overall, pellet group counts can be useful for monitoring the habitat use and density of moose, and likely other deer, but the sampling method and the numbers used in the calculation can strongly affect the results. Comparing trends between years would be better than calculating exact densities.

4. How tracking was or was not used

The whole study rests on finding, identifying and counting one kind of sign: moose pellet groups. Observers counted only groups lying on dead plant material from late autumn to spring, to separate winter droppings from older or newer ones, and followed a rule for groups on the edge of a strip or plot. No trailing was reported. The authors also mention an earlier suggestion by others to combine winter track counts with pellet group counts for moose.

5. Tracker or observer certification

Reported but not validated. The counts were made by “three experienced observers”, but their experience was not described or tested. Pertti Hokkanen, Mira Kooiman and Corina Kremer are thanked for skilful help with the fieldwork. The authors assumed that very few pellet groups were missed on the strips, and suggest that observers may count border groups differently on plots and strips, which matters for the density results.

6. Other technologies, skills or innovations

• An affinity index for each habitat: the share of pellet groups found there divided by the share of the area it covers (1.0 means used just as often as it occurs).
• A food-based habitat rating (poor, low, moderate, good) from the cover of edible trees above 50 cm in winter, and of deciduous seedlings and saplings in summer.

7. Implications for conservation (from the paper)

• Moose management in Finland needs reliable population estimates, and pellet group counts could offer a cheaper alternative or addition to questionnaires and aerial surveys.
• Pellet counts could be added to the existing wildlife triangle censuses, ideally together with other game species, so they make economic sense.
• Because the plot method, widely used so far, appears to overestimate moose density, the choice of method and of defecation rate needs care.
• The pellet distribution links moose to young pine thinning stands, useful for combining forest and wildlife management.