Take-home points

  • Cows on patch-burned pastures consistently put on weight, usually the most of other management.
  • No difference among PBG approaches (Spring-burn only vs. Spring + Summer burns)
  • Evidence of annual variability in cattle performance on continuously-grazed rangeland without fire continues to accumulate
  • Animals on twice-over rotational grazing generally under-performed those on continously-grazed pastures, especially those with patch-burning. This is especially apparent in the calf data.

Raw data and distributions

No reason to believe calf gains differed, but it appears that calf gains in the no-fire, continuously-grazed pastures (Refuge) came at the expense of cow performance. In 2017, cows lost weight in continuously-grazed pastures without fire, presumably because they continued to make milk to support their calves while grazing a lower-quality forage base.

Note that the twice-over rotational system was added in 2018. Calves from TOR pastures consistently under-perform calves from continuously-grazed pastures to an even greater degree than the cows do.

Distribution of Average Daily Gains for cows and calves by fire management treatment. Colors denote separate pastures.

Distribution of Average Daily Gains for cows and calves by fire management treatment. Colors denote separate pastures.

Hypothesis testing

There are two different ways to ask questions of these data:

  • Which treatments show the greatest gains/are gains different from zero?
  • The experimental question: Are gains on fire treatments different than the no-fire treatment?

Comparing performance among treatments within years

  • Cows put on weight in both PBG treatments in all four years.
    • Despite higher average gains within patch-burned pastures in 2017, high variability around these means caused the analyses to fail to reject the null hypothesis (i.e., 2017 gains on patch-burned pastures were not statistically different than zero).
    • Gains on patch-burned pastures 2018-2020 were statistically greater than zero.
  • Cows on continuously-grazed pastures without fire (Refuge) tended to lose weight in 2017, put on a bit in 2018 (not significantly-different than 0), and had significant gains in 2019 and even greater gains in 2020.
    • This supports the theoretical expectation that pastures without patch contrast will show greater inter-annual variability.
  • Cows on the twice-over rotational grazing pastures were also variable.
    • 2018-2019: gains not statistically different from zero
    • 2020: Better year for cows on TOR, with gains significantly better than zero.
Differences from zero were tested with LMER models fit to data within each year followed by post-hoc analysis testing the null hypothesis that mean gains per treatment were not different from zero.

Differences from zero were tested with LMER models fit to data within each year followed by post-hoc analysis testing the null hypothesis that mean gains per treatment were not different from zero.

Comparing treatments

4 years, patch-burning vs. no fire

Cow gains on both PBG treatments are significantly different than no-fire, but PBG treatments are not different from each other.

Regression coefficients from a linear mixed-effect regression (LMER) model fit to 2017-2020 data together serve as measures of treatment effect sizes for cattle weight gains throughout the study period so far:

Average daily gain (lbs/day) for three treatments with associated 95% confidence intervals taken from regression coefficient estimates in linear mixed-effect regression.

Average daily gain (lbs/day) for three treatments with associated 95% confidence intervals taken from regression coefficient estimates in linear mixed-effect regression.

Post-hoc pairwise analysis of the LMER used to estimate the above coefficients indicates that, over the full four years:

  • Within the PBG treatment, the different seasonal fire regimes did not differ from each other.
  • When compared to continuous grazing without fire (No fire), Spring + Summer patch-burning had statistically-signficantly greater gains (\(P = 0.02\)), whereas the trend for greater gains on Spring-only burns was only marginally different (\(P = 0.06\))
Results of Tukey pairwise post-hoc analysis of linear mixed-effect regression comparing average daily gain across three treatments.
  Coefficient StandardError z P
Spring + Summer - No fire 0.61 0.23 2.62 0.02
Spring only - No fire 0.52 0.23 2.24 0.06
Spring only - Spring + Summer -0.09 0.23 -0.37 0.93

3 years, all treatments

Average daily gain (lbs/day) for three treatments with associated 95% confidence intervals taken from regression coefficient estimates in linear mixed-effect regression.

Average daily gain (lbs/day) for three treatments with associated 95% confidence intervals taken from regression coefficient estimates in linear mixed-effect regression.

Over the last three years:

  • Cattle on the modified twice-over rotational grazing pastures gained an average of 0.11 lbs/day. This value is significantly different from zero but was lower than all other grazing treatments.
  • Among continuously-grazed pastures, mean gains on unburned pastures (0.38 lbs/day) were not statistically-significantly different than on patch-burned pastures (0.42-0.48 lbs/day)
Results of Tukey pairwise post-hoc analysis of linear mixed-effect regression comparing average daily gain across all four treatments.
  Coefficient StandardError z P
Rotational - No fire -0.54 0.16 -3.3 0.01
Spring + Summer - No fire 0.2 0.16 1.25 0.59
Spring only - No fire 0.09 0.16 0.55 0.95
Spring + Summer - Rotational 0.74 0.16 4.59 0
Spring only - Rotational 0.63 0.16 3.86 0
Spring only - Spring + Summer -0.11 0.16 -0.69 0.9