Northern Prairie Wildlife Research Center
We reran the preceding analysis using numbers of verified complaints on livestock only (pets excluded) for 1976-86 and found a similar relation (r = -0.701, P = 0.016). Generally, the winters of lowest WSI values (i.e., mild) were followed by summers of highest verified complaints and vice versa (Fig. 3). Likewise, the year with the most verified complaints (1981) was preceded by the second mildest winter in the period. The severest winter, 1978-79, was followed by the second lowest number of verified complaints. Thus, although cause and effect cannot be confirmed, severity of the previous winter seemed to be the best indicator of the level of losses to be expected during a given depredation season.
A question of considerable interest to us and other agencies concerned with wolf management was whether the depredation problem was increasing. We believed the best index of depredation severity to be the total number of complaints verified annually during 1976-86; annual numbers of farms having verified losses was possibly the second best indicator (Fig. 3). (Complaint data from 1975 were not included in analyses because of low public awareness of the program in the start-up year.) Simple linear regression on time for the 11-year period did not indicate a significant positive trend in verified complaints (r = 0.433, P = 0.092), although an increase (average, 1.55 per year) was noted in the number of farms having verified complaints (r = 0.578, P = 0.032). The number of complaints increased significantly (average, 3.48 per year) from 1976 to 1986 (r = 0.628, P = 0.020), at least partly because of increased awareness of our program and the state's compensation program. Taken together, however, these results suggest a long-term, increasing trend in the depredation problem.
To evaluate the relative effects of WSI and year on annual numbers of verified complaints, we used multiple linear regression. This analysis supported the relation between WSI and verified complaints (F1,8 = 8.25, P = 0.021). However, when the effect of WSI was accounted for, there was little evidence of a time trend in verified complaints (F1,8 = 2.18, P = 0.178). When the effect of WSI was accounted for, the addition of the time variable accounted for an additional 18.1% of the total variation explained. However, when the effect of time was accounted for, the addition of WSI accounted for an additional 68.7% of the total variation explained.
After taking into account WSI, related specific variables that we suspect contributed to the unexplained variation in depredation rates were date of snowmelt in spring (the earlier the snowmelt the earlier the livestock and their young are exposed to wolves), weather in summer, time of first snowfall (thus restricting movements or causing confinement of cattle and sheep and shortening the annual periods of exposure), number of wolves captured locally the previous year, degree of local illegal killing of wolves, and number of livestock available (related to market fluctuations). Availability of natural prey was also a potential factor but did not explain the increase in depredations in Roseau and Kittson Counties. Estimates of prefawning deer density and moose density in that area indicated no decreasing trend from 1982 to 1986 (Joselyn et al. 1988).
Annual variation in numbers killed (Table 1) was higher for sheep than for cattle (F7,7 = 25.6, P < 0.005), and higher for turkeys than for sheep (F7,7 = 23.7, P < 0.005). This finding reflects the difference in the depredation problem in the northwestern counties where more sheep and turkeys are available (55% of sheep were in Roseau, Kittson, and Marshall Counties in northwestern Minnesota), compared with the major part of the wolf range where cattle are the primary domestic prey. Because the depredation problem was more variable in the northwestern counties, the events there played a disproportionate role in determining whether any given year in Minnesota was characterized by relatively high or low losses. Moreover, often a few wolves had a disproportionate affect on the state's compensation program. A vivid example occurred in 1981 when four packs or pairs of wolves were responsible for $23,276 of the total $38,605 paid in compensation; that year the compensation paid in Roseau and Kittson Counties was 65% of the total.