Seasonal patterns of bison diet across climate gradients in North America

General patterns of dietary quality and composition

Averaged across the months, site-level [CP] ranged from 62.6 to 147.3 mg g⁻¹ and averaged 96.7 ± 3.3 mg g⁻¹. [DOM] ranged from 571.3 to 643.6 mg g⁻¹ and averaged 605.8 ± 2.4 mg g⁻¹ while [DOM]: [CP] ranged from 4.4 to 10.3 with an average of 6.9 ± 0.2.

Examining seasonal patterns averaged across all sites, [CP] averaged 82.5 ± 5.4 mg g⁻¹ in April, peaked at 113.2 ± 3.3 mg g⁻¹ in June, and declined to 80.6 ± 4.5 mg g⁻¹ in September (Supplementary Table S1, Fig. 2). Similarly, [DOM] averaged 583.0 ± 6.1 mg g⁻¹ in April, peaked at 630.2 ± 3.4 mg g⁻¹ in June, and declined to 585.1 ± 4.2 mg g⁻¹ by September (Supplementary Table S1, Fig. 2). [DOM]: [CP] averaged 8.1 ± 0.5 in April, declined to 5.8 ± 0.2 in June, and increased to 8.1 ± 0.4 by September (Supplementary Table S1, Fig. 2).

Examining dietary functional group composition across sites, on average 38.2 ± 2.6% of dietary protein intake came from grasses, ranging from 12.7 to 82.2% across sites. Examining patterns for the two grass functional groups, 26.5 ± 2.9% (1.3–82.0%) of protein intake was derived from cool-season (C₃) graminoids and 11.7 ± 1.6% (0–36.7%) came from warm-season (C₄) grasses. Of the Eudicots, protein intake from legumes averaged 37.8 ± 2.8% (1.3–70.2%), from non-leguminous forbs averaged 21.5 ± 2.0% (0.2–68.0%), and 2.5 ± 0.8% (0–31.6%) from woody species.

Examining monthly patterns across sites, C₃ grass protein intake was highest in May (40.8 ± 5.0%) and lowest in September (15.8 ± 2.8%) while C₄ grass protein intake peaked in September (16.2 ± 2.9%) and was lowest in July (10.2 ± 2.4%) (Supplementary Table S2, Fig. 3). Legume protein intake was highest in August at 55.8 ± 5.3% and was lowest in May (20.0 ± 4.0%) (Supplementary Table S2, Fig. 3). Forb protein intake peaked in June at 27.6 ± 3.5% and was lowest in August (14.2 ± 2.8%) (Supplementary Table S2, Fig. 3).

Climate and dietary quality

Examining patterns of [CP] across sites, [CP] was highest in cool, wet climates (Fig. 4, Supplementary Table S3). For example, bison in June at a site with 1200 mm of rain and 6 °C MAT would have [CP] of 186.3 mg g⁻¹. Bison in June at a site with 400 mm of rain and 18 °C MAT would have [CP] of just 67.8 mg g⁻¹. There was also a statistical interaction between the identity of the month and MAT on [CP] (Fig. 4, Supplementary Table S3). In April, [CP] tended to increase with increasing MAT (2.51 ± 1.46 mg g⁻¹ °C⁻¹), as warm-climate sites tended to have higher [CP] than cold-climate sites (Fig. 4, Supplementary Table S3). By May, colder sites tended to have higher [CP] (− 2.01 ± 1.81 mg g⁻¹ °C⁻¹), which became stronger by June (− 5.34 ± 1.83 mg g⁻¹ °C⁻¹) and lasted through September (Fig. 4, Supplementary Table S3). Examining patterns of functional group composition on top of climate relationships, [CP] was lower when greater amounts of warm-season grasses were in the diet, independent of climate and season (Supplementary Table S4).

Like [CP], [DOM] was also highest in cool, wet climates (Fig. 4, Supplementary Table S3). For example, bison at a site with 1200 mm of rain and 6 °C MAT in June would have [CP] of 670.4 mg g⁻¹ as opposed to bison at a site with 400 mm of rain and 18 °C MAT in June, which would have [CP] of just 580.4 mg g⁻¹. In April, bison in warm-climate sites tended to consume a diet that was higher in [DOM] than bison in colder-climate sites (2.04 ± 1.73 mg g⁻¹ °C⁻¹) (Fig. 4, Supplementary Table S3). By May, [DOM] was invariant across MAT gradients (− 0.08 ± 1.98 mg g⁻¹ °C⁻¹) and by June, cold-climate sites had higher [DOM] (5.51 ± 2.00 mg g⁻¹ °C⁻¹) (Fig. 4, Supplementary Table S3).

Examining patterns of DOM:CP reveals the seasonal and spatial patterns of protein limitation. In general, bison in cool, wet climates had the lowest [DOM]: [CP] (Supplementary Table S3). Yet, the statistical interaction between MAT and month shows that in April, bison in cold climates were more protein limited than in warm climates (− 0.16 ± 0.11 °C⁻¹) (Supplementary Table S3). By May, [DOM]: [CP] tended to become higher in warm climates, with the gradient becoming strongest by August (0.61 ± 0.15 °C⁻¹) (Supplementary Table S3).

Climate and dietary composition

Examining cross-site relationships between the relative amounts of different functional groups in diet and the predictors of climate and season, the proportion of cool-season grass in the diet was highest in cool sites (decreasing at a rate of 3.8 ± 0.5% °C⁻¹), with no significant influence of MAP (P = 0.16) (Fig. 5, Supplementary Table S5). The proportion of cool-season grass in the diet varied among months, but there was no difference among months in the relationship with MAT (P = 0.54). In contrast, the proportion of warm-season grass in diet was greatest in hot, dry regions (Fig. 5, Supplementary Table S5). For example, bison at a site with 400 mm of rain and 18 °C MAT would have 42.9% of their dietary protein from warm-season grasses as opposed to a site with 1200 mm of rain and 6 °C MAT, where warm-season grasses would provide 0% of their dietary protein.

Among Eudicots, there was no seasonal variation in the percentage of forbs in the diet, but the percentage of dietary protein from forbs was highest in hot, dry regions (Fig. 5, Supplementary Table S5). For example, bison at a site with 400 mm of rain and 18 °C MAT in June would have 54.2% of their dietary protein from forbs as opposed to a site with 1200 mm of rain and 6 °C MAT in June, where it would be just 9.6%. Legumes had strong seasonal variation in its dietary contribution, but little variation with climate, increasing at a rate of 2.0 ± 0.8% per 100 mm MAP (P = 0.02) (Fig. 5, Supplementary Table S5).

Comparison with 2018 data

Comparing [CP] between years, June [CP] was higher on average in 2019 than 2018 (113.2 ± 3.3 vs. 93.6 ± 3.2 mg g⁻¹; P < 0.001). In contrast, September [CP] was not different between the two years (80.6 ± 4.5 mg g⁻¹ in 2019 vs. 80.6 ± 4.5 mg g⁻¹ in 2018; P = 0.98). Like [CP], June [DOM] was higher in 2019 than 2018 (630.2 ± 3.4 mg g⁻¹ vs. 617.6 ± 3.7 mg g⁻¹; P = 0.01) while September [DOM] did not differ between years (585.1 ± 4.2 mg g⁻¹ in 2019 vs. 591.7 + 4.4 mg g⁻¹ in 2018; P = 0.28). June [DOM]: [CP] was lower in 2019 vs. 2018 (5.8 ± 0.2 vs. 6.9 ± 0.2; P < 0.001) while September [DOM]: [CP] did not differ between years (8.1 ± 0.4 vs. 8.7 ± 0.6; P = 0.36).

Comparing monthly functional group composition of bison diet between 2018 and 2019, in almost all cases, patterns in June 2019 were somewhat similar to June 2018 (Table 1). For example, June protein intake from cool-season grasses was approximately 31% in both years (32.5% ± 4.4% in 2019 vs. 30.1% ± 3.9% in 2018; P = 0.68; Table 1). In 2019, cool-season graminoids, warm-season grasses, and legumes differed significantly from 2018. In 2019, 45.6% ± 5.1% of September dietary protein intake came from legumes (Table 1). In 2018, it was 24.6% ± 3.5% (P < 0.001). September dietary protein intake from cool-season graminoids was 10% lower in 2019 than 2018 (15.8 ± 2.7% vs. 26.0 ± 3.5%; P = 0.027). For warm-season grasses, it was 14% lower in 2019 than 2018 (30.2% ± 4.1% vs. 16.2% ± 2.9%; P = 0.009) (Table 1).

Comparing relationships between [CP] with climate between the two years, although [CP] was higher on average in 2019 after standardizing for climate (P < 0.001), there was no difference in the relationships between [CP] and either MAT or MAP between years (P > 0.05 for both) (Table 2). For [DOM], 2019 [DOM] was 12.3 mg g⁻¹ higher after standardizing for climate (P = 0.02) (Table 2). The relationships between [DOM] and MAP were similar between years (P = 0.22), with [DOM] decreasing at a faster rate with increasing MAT in 2019 than 2018 (− 2.6 mg g⁻¹ °C⁻¹ higher in 2019 than 2018; P = 0.05) (Table 2). After controlling for site climate, [DOM]: [CP] was 1.0 ± 0.5 lower in 2019 than 2018 (P = 0.02), but the relationships with climate did not differ between years (P > 0.1 for both) (Table 2).

Controlling for site climate, which would take into account potential differences in the distribution of sites in climate space between the years, there was no difference in the protein contribution of cool-season grasses to bison diet between years (P = 0.77) and no difference in relationships between climate and the protein contribution of cool-season grasses (P > 0.7 for both MAT and MAP) (Table 3). The same was true for warm-season grasses (Table 3). Despite significant differences in legume contribution to diet between years for the different months (P = 0.002), controlling for climate, legume contribution to diet was similar between years (P = 0.34) with no differences in the relationship between legume contribution and climate (P > 0.3 for both MAT and MAP) (Table 3). There were also no significant relationships with climate for forb or woody contribution to diet between years (Table 3). Despite many of the similarities between years, comparing dietary functional group composition in September between years, legumes contributed 20% more protein and warm-season grasses 14% less in 2019 than 2018.

Source: Ecology - nature.com