Intro Outline

You mentioned making drought and animal response the focus when I last sent this over. Is that how you would prefer the intro to look also?

Drought/climate change affecting arid and semiarid regions Typical animal performance under drought and cancellation of severe drought studies PBG as an alternative land management strategy that tries to improve drought resiliency HREC drought: gains, fecal count, forage production

Animal Gains

ANCalves <- aov(X2Dday~Year, data=CalfRaw)
summary(ANCalves)  # Fvalue 8.792  #Pr 0.00354
##              Df Sum Sq Mean Sq F value  Pr(>F)   
## Year          1  0.823  0.8229   8.792 0.00354 **
## Residuals   145 13.571  0.0936                   
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
ANCows <- aov(X2Dday~Year, data=CowRaw)
summary(ANCows) #Fvalue 3.363 #Pr 0.0686
##              Df Sum Sq Mean Sq F value Pr(>F)  
## Year          1   1.65  1.6530   3.363 0.0686 .
## Residuals   152  74.72  0.4916                 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
ANSheep <- aov(X2Dday~Year, data=SheepRaw)
summary(ANSheep) #Fvalue 5.349 #Pr 0.0209
##               Df Sum Sq Mean Sq F value Pr(>F)  
## Year           1   0.06 0.06043   5.349 0.0209 *
## Residuals   1092  12.34 0.01130                 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
ANCalves2 <- aov(X2Dday~Year+(1|Pasture), data=CalfRaw)
summary(ANCalves2)

ANCows2 <- aov(X2Dday~Year+(1|Pasture), data=CowRaw)
summary(ANCows2)

ANSheep2 <- aov(X2Dday~Year+(1|Pasture), data=SheepRaw)
summary(ANSheep2)

For comparing the gains, I have been going off of the basic ‘average daily gain ~ year’ model. Calves and sheep gains were significantly different. The sheep have very little to no variation which makes their negligible differences in means (graph) make some sense. Cow gains were not different between years.

I thought I should include the Pasture as a random effect since it might count as a repeated measure, but R gives me an error about it not being meaningul for the factor (year?) when I tried it. An argument against using pasture would be that we don’t put the same animals in the same pasture each year and just looking at gains all together covers that.

ggplot(LivestockRaw, aes(x=Pre2, y=GainProp, color=CalfCow))+
  geom_point(size=2)+
  labs(y="Proportion of Initial BW", x="Initial BW (lbs)")+
  theme_bw(16)+
  theme(legend.position = "top")+
  facet_wrap(~Year)

Fecal Density

Fecal.H1 <- lmer(FecalCount ~ Month + Patch + Treatment + (1|Location), 
                   data=DroughtForage, REML=FALSE)
Fecal.H2 <-lmer(FecalCount ~ Patch + Treatment + (1|Month) + (1|Location), data=DroughtForage, REML=FALSE)
Fecal.NULL <-lmer(FecalCount ~1  + (1|Location), data=DroughtForage, REML=FALSE)
Fecal.NULLM <-lmer(FecalCount ~1 + (1|Month)  + (1|Location), data=DroughtForage, REML=FALSE)

anova(Fecal.H1, Fecal.NULL)
## Data: DroughtForage
## Models:
## Fecal.NULL: FecalCount ~ 1 + (1 | Location)
## Fecal.H1: FecalCount ~ Month + Patch + Treatment + (1 | Location)
##            Df    AIC    BIC  logLik deviance  Chisq Chi Df Pr(>Chisq)    
## Fecal.NULL  3 1989.0 2000.9 -991.52   1983.0                             
## Fecal.H1    7 1907.6 1935.3 -946.78   1893.6 89.464      4  < 2.2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
anova(Fecal.H2, Fecal.NULL)
## Data: DroughtForage
## Models:
## Fecal.NULL: FecalCount ~ 1 + (1 | Location)
## Fecal.H2: FecalCount ~ Patch + Treatment + (1 | Month) + (1 | Location)
##            Df    AIC    BIC  logLik deviance  Chisq Chi Df Pr(>Chisq)    
## Fecal.NULL  3 1989.0 2000.9 -991.52   1983.0                             
## Fecal.H2    6 1915.7 1939.5 -951.86   1903.7 79.305      3  < 2.2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
anova(Fecal.H1, Fecal.NULLM)
## Data: DroughtForage
## Models:
## Fecal.NULLM: FecalCount ~ 1 + (1 | Month) + (1 | Location)
## Fecal.H1: FecalCount ~ Month + Patch + Treatment + (1 | Location)
##             Df    AIC    BIC  logLik deviance  Chisq Chi Df Pr(>Chisq)    
## Fecal.NULLM  4 1963.3 1979.2 -977.67   1955.3                             
## Fecal.H1     7 1907.6 1935.3 -946.78   1893.6 61.774      3  2.456e-13 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
anova(Fecal.H2, Fecal.NULLM)
## Data: DroughtForage
## Models:
## Fecal.NULLM: FecalCount ~ 1 + (1 | Month) + (1 | Location)
## Fecal.H2: FecalCount ~ Patch + Treatment + (1 | Month) + (1 | Location)
##             Df    AIC    BIC  logLik deviance  Chisq Chi Df Pr(>Chisq)    
## Fecal.NULLM  4 1963.3 1979.2 -977.67   1955.3                             
## Fecal.H2     6 1915.7 1939.5 -951.86   1903.7 51.615      2  6.195e-12 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
anova(Fecal.H1,Fecal.H2)
## Data: DroughtForage
## Models:
## Fecal.H2: FecalCount ~ Patch + Treatment + (1 | Month) + (1 | Location)
## Fecal.H1: FecalCount ~ Month + Patch + Treatment + (1 | Location)
##          Df    AIC    BIC  logLik deviance  Chisq Chi Df Pr(>Chisq)   
## Fecal.H2  6 1915.7 1939.5 -951.86   1903.7                            
## Fecal.H1  7 1907.6 1935.3 -946.78   1893.6 10.159      1   0.001436 **
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Anova(Fecal.H2)
## Analysis of Deviance Table (Type II Wald chisquare tests)
## 
## Response: FecalCount
##            Chisq Df Pr(>Chisq)    
## Patch     45.813  1  1.301e-11 ***
## Treatment 24.048  1  9.395e-07 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Biomass

Bio.null <- lmer(LbAc ~ 1 + (1|Month) + (1|Location), data=DroughtForage, REML=FALSE)
Bio.H1 <- lmer(LbAc ~ Patch + (1|Month) + (1|Location), 
                 data=DroughtForage, REML=FALSE)

Bio.H2 <- lmer(LbAc ~ Patch + Treatment + (1|Month) + (1|Location), 
                 data=DroughtForage, REML=FALSE)

anova(Bio.H1, Bio.null) #
## Data: DroughtForage
## Models:
## Bio.null: LbAc ~ 1 + (1 | Month) + (1 | Location)
## Bio.H1: LbAc ~ Patch + (1 | Month) + (1 | Location)
##          Df    AIC    BIC  logLik deviance  Chisq Chi Df Pr(>Chisq)    
## Bio.null  4 6844.2 6860.0 -3418.1   6836.2                             
## Bio.H1    5 6809.0 6828.8 -3399.5   6799.0 37.242      1  1.044e-09 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
anova(Bio.H2, Bio.null) #
## Data: DroughtForage
## Models:
## Bio.null: LbAc ~ 1 + (1 | Month) + (1 | Location)
## Bio.H2: LbAc ~ Patch + Treatment + (1 | Month) + (1 | Location)
##          Df    AIC    BIC  logLik deviance  Chisq Chi Df Pr(>Chisq)    
## Bio.null  4 6844.2 6860.0 -3418.1   6836.2                             
## Bio.H2    6 6810.7 6834.4 -3399.3   6798.7 37.555      2      7e-09 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
anova(Bio.H1, Bio.H2)
## Data: DroughtForage
## Models:
## Bio.H1: LbAc ~ Patch + (1 | Month) + (1 | Location)
## Bio.H2: LbAc ~ Patch + Treatment + (1 | Month) + (1 | Location)
##        Df    AIC    BIC  logLik deviance Chisq Chi Df Pr(>Chisq)
## Bio.H1  5 6809.0 6828.8 -3399.5   6799.0                        
## Bio.H2  6 6810.7 6834.4 -3399.3   6798.7 0.313      1     0.5758
Anova(Bio.H1)
## Analysis of Deviance Table (Type II Wald chisquare tests)
## 
## Response: LbAc
##        Chisq Df Pr(>Chisq)    
## Patch 39.124  1  3.978e-10 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Anova(Bio.H2)
## Analysis of Deviance Table (Type II Wald chisquare tests)
## 
## Response: LbAc
##             Chisq Df Pr(>Chisq)    
## Patch     39.1279  1  3.969e-10 ***
## Treatment  0.3209  1     0.5711    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Grazer attraction over time

dropping font sizes down on annotations for Rpub