USGS - science for a changing world

Northern Prairie Wildlife Research Center

  Home About NPWRC Our Science Staff Employment Contacts Common Questions About the Site

Effects of Grazing and Burning on Densities and
Habitats of Breeding Ducks in North Dakota

Results


We found 2,512 nests of 9 species of ducks (Table 1.), of which the mallard was the most common.

Nest Density

Total Nest Density. -- We found no effects (P >/= 0.28; Fig. 1) of the burning and grazing treatments on total nest density on any of the 7 species we tested. For 5 of the 7 species, the confidence interval of the difference between the treatments with the highest and lowest values of total nest density spanned 0 (Appendix 2). The confidence intervals were wide (1.0 to 5.8 times the estimated difference) for all species except shoveler (0.5 times the estimated difference).

GIF - Figure 1.
Figure 1. Total nest density (no. of nests during 1982-88 divided by the no. of upland ha) on treatment and control fields of 7 species of waterfowl at Lostwood National Wildlife Refuge, North Dakota.

Annual Nest Density. -- Treatment-year interactions for annual nest density of gadwall (P = 0.02), northern shoveler (P = 0.02), blue-winged teal (P = 0.03), and northern pintail (P = 0.01) were significant, which prevented us from interpreting the effects of treatment and year alone.

There were differences among years for mallard (P = 0.0001) and American wigeon (P = 0.0001). The annual nest density of mallard (Table 2.) was higher (LSD test, P < 0.05) in the early years of the study (1980-82) than in the later years (1983-88). The annual nest density of American wigeon was higher (LSD test, P < 0.05) in 1981 than in other years.

For mallard and American wigeon, we found no effect of treatment (mallard P = 0.79; wigeon P = 0.75) or treatment-year interaction (mallard P = 0.70, wigeon P = 0.81) on annual nest density. For lesser scaup we found no year (P = 0.09), treatment (P = 0.99), or treatment- year interaction (P = 0.11) effects. For lesser scaup and American wigeon, the confidence interval of the difference between the treatments with the highest and lowest values of annual nest density spanned 0 (Appendix 2). The confidence intervals were wide (0.7 to 2.6 times the estimated difference) for scaup, wigeon, and mallard.

For annual nest density of gadwall and blue-winged teal, which had treatment-year interactions and adequate samples of nests, we examined year effects within treatments and treatment effects within years (Milliken and Johnson, 1984:322). We examined the 9 differences among years within treatments and 11 differences among treatments within years listed in the Methods. Differences were relative to control fields (see Methods).

For gadwall on the spring burn fields, annual nest density was higher (F = 6.17; 1, 64 df; P = 0.016) in the pre-treatment years than in the post-treatment years, whereas on the spring graze fields, annual nest density was lower (F = 6.45; 1, 64 df; P = 0.014) in the grazing years than in the post-grazing years, and annual nest density increased linearly for 3 years (F = 5.91; 1, 64 df; P = 0.018) beginning in 1984, the last year of spring grazing (Fig. 2). For blue-winged teal on the summer burn/spring graze fields, annual nest density was higher (F = 7.00; 1,64 df; P = 0.010) in the pre-treatment years than in the post-treatment years. On the spring graze fields, annual nest density was higher in the pretreatment years than in the grazing years (F = 5.41; 1,64 df; P = 0.023) and the post-treatment years (F = 4.22; 1, 64 df; P = 0.044) (Fig. 2). Annual nest density among treatments within years did not differ (P > 0.33) for either gadwall or blue-winged teal.

GIF - Figure 2.
Figure 2. Annual nest density by treatment of 3 species of waterfowl at Lostwood National Wildlife Refuge, 1980-88.

Nest Success

Daily nest survival rate varied among years for mallard (P = 0.003) and gadwall (P = 0.02), but not for blue-winged teal (P = 0.40). All 3 species for which daily nest survival rate was analyzed had the lowest rates in 1982 and 1985 (Table 3). Daily nest survival rate was low in these same 2 years for mallard on control fields. Mayfield estimates of nest success of mallard on control fields were 10% and 14% in 1982 and 1985, whereas the range of nest success in the remaining 7 years between 1980 and 1988 was 21 to 70%.

For none of the 3 species tested did we find an effect on daily nest survival rate of treatment (mallard P = 0.66; gadwall P = 0.16; blue-winged teal P = 0.26) or treatment-year interaction (mallard P = 0.93; gadwall P = 0.57; blue-winged teal P = 0.49). For mallard and blue-winged teal, the confidence interval of the difference between the treatments with the highest lowest values of daily nest survival rate spanned 0 (Appendix 2). The confidence intervals were wide (0.9 to 1.6 times the estimated differences) for mallard, blue-winged teal, and gadwall.


Vegetation Structure

Effects of Treatments. -- Vegetation structure in June was influenced by the grazing and burning treatments (Fig. 3). After spring burning, spring grazing, and the summer burning/spring grazing manipulations, the percentage of vegetation <0.5 dm was high, whereas the percentage of vegetation > or = to 2.5 dm was low. In the spring burn fields, the effect was observed in the first year after the burn, because the fields were burned after vegetation sampling in June. Because the spring grazing manipulation occurred between the April and June sampling, the June readings were affected during the grazing year. The summer burn/spring graze treatment spanned 2 years. Because the summer burning for the summer burn/spring graze treatment was conducted the first year after the June readings were taken, an effect was observed during the second year. Spring grazing for the summer burn/spring graze treatment occurred between the 2 readings in the second year, so the June readings were affected during the grazing year.

The length of time it took for the visual obstruction reading profile to reach the profile similar to that on the control fields varied among treatments (Fig. 3). On the spring graze fields during the grazing years, most vegetation was in the shortest visual obstruction category (<0.5 dm). After grazing ended, the fields recovered quickly. Just 1 year after grazing ended, the visual obstruction reading profile was similar to that on the control fields. In the spring burn treatment, the percentage of short vegetation (<0.5 dm) increased 1 year after burning but decreased to a level similar to the control fields by the second year after the burns. The percentage of tall vegetation never was as high as on the control fields during the 2 years after the burns that we monitored the area. The summer burn/spring graze fields showed the longest lasting effects of the treatments. We studied the fields for 2 years following each of 3 burning/ grazing manipulations and found that the proportions of the shortest and tallest categories were never similar to those on the control fields.

GIF - Figure 3.
Figure 3. Distribution of visual obstruction reading classes (dm) in June for 3 burning and grazing treatments and control at Lostwood National Wildlife Refuge, North Dakota, 1980-88. Each treatment is divided into appropriate periods reflecting years before grazing and burning (1982 for control, spring burn, and summer burn/spring graze), and number of years after grazing and burning applications. For the spring burn treatment, data for 1983, 1985 and 1987 were averaged for 1 year after burn and 1984, 1986 and 1988 were averaged for 2 years after burn. For the summer burn/spring graze treatment, 1984 and 1987 were averaged for 1 year after graze, and 1985 and 1988 were averaged for 2 years after graze. For the spring graze treatment, 1985 was 1 year after graze, 1986 was 2 years after graze, 1987 was 3 years after graze, 1986 was 2 years after graze, 1987 was 3 years after graze and 1988 was 4 years after graze.

Use-availabllity Comparisons. -- For mallard there was a difference (X² = 1,348, 5 df, P < 0.0001) between the visual obstruction readings of available vegetation, based on the April transect readings, and the visual obstruction readings of vegetation at nests found before 25 May (Fig. 4). Mallard selected (P < 0.05) vegetation >/=1.00 dm and avoided (P < 0.05) vegetation <1.00 dm. More than 40% of nests were in vegetation >/= 2.5 dm; 3% of habitat was in that class. Fewer than 18% of the nests were in vegetation <1.0 dm; 74% of the habitat was in that class.

Northern pintail also had a visual obstruction reading profile of vegetation at nests that differed (X² = 67.6, 5 df, P < 0.0001) from available vegetation. Pintail avoided (P < 0.05) vegetation <0.5 dm and selected (P < 0.05) vegetation 1.00-1.49 dm. They used other visual obstruction reading classes in proportion to their availability. About 46% of northern pintail nests were in vegetation <1.0 dm, and 17% of nests were in vegetation >/= 2.5 dm.

The peak of nest initiation by the remaining species was later in the year than by mallard and northern pintail; therefore, visual obstruction readings of available vegetation was based on the June sampling (Fig. 4). For all late-nesting species, the visual obstruction readings of available June vegetation was different (P < 0.0001) from the visual obstruction readings of vegetation at nests. Both gadwall and American wigeon selected (P < 0.05) tall vegetation >/= 2.5 dm and avoided (P < 0.05) vegetation <1.5 dm. Both blue-winged teal and northern shovelers avoided (P < 0.05) vegetation <0.5 dm, but they both selected (P < 0.05) vegetation between 1.5 and 2.5 dm. Lesser scaup selected (P < 0.05) vegetation >2.0 dm and avoided (P < 0.05) vegetation <1.0 dm.

GIF - Figure 4.
Figure 4. Comparison of nesting use and availability of visual obstruction reading classes (dm) by 7 species of waterfowl at Lostwood National Wildlife Refuge, North Dakota, 1980-88. Early nesting species are compared with April transect readings; late nesting species are compared with June transect readings.

Vegetation Types

Temporal Changes. -- The relative proportions of available types of upland vegetation changed (MANOVA; Wilks' lambda = 0.076; F = 15.13; 4, 5 df; P = 0.005; Table 4.) during the study. In 1987 there was more grass/brush (F = 35.07; 1, 8 df; P < 0.001) and less brush/grass (F = 39.49; 1, 8 df; P < 0.001) and brush (F = 4.87; 1, 8 df; P = 0.058) than in 1981. The most common vegetation type in 1981 was brush/grass, whereas in 1987 the most common vegetation type was grass/brush. There was no difference (MANOVA; Wilks' lambda = 0.025; F = 2.06; 12, 8.2 df; P = 0.15) among treatments in changes in habitat type and there was no treatment-year interaction. The area of wetlands with water decreased in the study fields between 1981 and 1987; 42 ha were present in 1981 and 14 ha in 1987.

Use-availability Comparisons. -- In 1981 and 1987, mallard preferred brush, used brush/grass in proportion to its availability, and avoided grass/brush, grass/forbs, and shallow marsh emergent vegetation (Table 4). In 1981 and 1987, gadwall preferred brush and avoided grass/brush and grass/forbs (Table 4.). In 1981 they preferred brush/grass also, but in 1987 they used this vegetation in proportion to its availability. In both years gadwall used forbs and shallow marsh emergent in proportion to its availability. In 1981 and 1987, blue-winged teal used grass/ brush, grass/forbs, forbs, and shallow marsh emergent in proportion to their availability (Table 4). In 1987 they avoided brush/grass, but in 1981 they used this vegetation type in proportion to its availability. Blue-winged teal never nested in the brush type.

For the remaining species, there were not enough nests in the years when vegetation type data were collected (1981 and 1987) to allow use-availability comparisons. For these species, we have presented the use of each vegetation type based on the number of nests for all years combined and the available proportion of each vegetation type based on the mean of 1981 and 1987 data (Table5.). For both northern pintail and American wigeon, most (80 and 95%, respectively) nests were found in brush and brush/ grass types. For lesser scaup and Northern shoveler, most nests (85 and 87%, respectively) were in brush/grass, grass/brush, and grass/forb types, and only 1 nest was found in brush.


Previous Section -- Methods
Return to Contents
Next Section -- Discussion

Accessibility FOIA Privacy Policies and Notices

Take Pride in America logo USA.gov logo U.S. Department of the Interior | U.S. Geological Survey
URL: http://www.npwrc.usgs.gov/resource/birds/density/results.htm
Page Contact Information: Webmaster
Page Last Modified: Friday, 01-Feb-2013 19:09:44 EST
Sioux Falls, SD [sdww54]