How to compare the impact of one predictor variable (native_biomass) across 4 different response variables of different scales? I am trying to assess the effect of different amounts of native biomass and non native biomass on four reproductive metrics:
Lay date
Clutch size
Chick size
Fledgling success

I've run models for these - 2 Linear Models (lay date, chick size) and 2 GLMs (Clutch size, fledgling success). Included in these models are relevant confounding variables.
Models:
lm5 <- lm(First_egg ~ Native_biomass + Non_native_biomass + Species + 
          No_tree_species + Distance_to_light + Anthropogenic_cover, 
          data=datav2)

glm2_native <- glm(No_of_eggs ~ Native_biomass + Non_native_biomass + 
                   Species + Distance_to_light +  No_tree_species + 
                   Anthropogenic_cover + First_egg, 
                   data=clutch_size_df, family=poisson(link=log))

chick_size_native_lm <- lm(Chick_size ~ Native_biomass + 
    Non_native_biomass + Distance_to_light + Anthropogenic_cover + 
    No_tree_species + Species + No_of_hatched, data=chick_size_df)

model_v2 <- glm(cbind(No_fledged, non_fledged) ~ Native_biomass + 
                  Non_native_biomass +
                  Distance_to_light + Anthropogenic_cover + 
                  No_tree_species + Species, 
                  data=fledgling_success_df, 
                  family=binomial(link=logit)) 

For each model, I've received my model outputs. Now, I want to standardise my response variables in order for me to be able to directly compare the effect of native biomass across the responses variables. I.e. to help me ascertain what stage of the reproductive cycle is most affected by native biomass increasing/decreasing.
My goal is to have 4 plots (one each for each response variable - lay date, clutch size, chick size and fledgling success) each showing how increasing native biomass affects the response. And I want to standardise them so I can compare across plots and make direct comparisons re the scale of effect of native biomass on each response.
Does anyone know how I might do this?
(See conversation re this on stackoverflow for more background if interested)
 A: 
My goal is to have 4 plots (one each for each response variable - lay date, clutch size, chick size and fledgling success) each showing how increasing native biomass affects the response. And I want to standardise them so I can compare across plots and make direct comparisons re the scale of effect of native biomass on each response.

I don't know that there's a useful way to do this standardization, as the outcomes differ in fundamental characteristics. You would have to apply your understanding of the subject matter to make that type of judgment.
The lay date is a continuous outcome, similar to a survival model with the "event" being the first egg.* Its Native_biomass coefficient is in units of "change in days per unit change in Native_biomass."
The number of eggs is a count variable. Its Native_biomass coefficient in the Poisson GLM with log link is in units of "change in log of mean egg number per unit change in Native_biomass."
The chick size is a continuous variable. Its Native_biomass coefficient is (perhaps) in units of "change in grams of chick weight per unit change in Native_biomass."
The fledgling success is a binomial outcome. Its Native_biomass coefficient is in units of "change in log-odds of success per unit change in Native_biomass."
Even for the two continuous outcomes, however, how do you compare a 1-day change in lay date against a 1-gram change in chick size? How do you compare either against a 1-egg change in clutch size? Or against a 10 percentage point change in fledgling success?
Then there is a further problem in that the associations of predictors with First_egg end up being re-incorporated into the model for No_of_eggs via its First_egg predictor.
I suppose you could scale your continuous regression coefficients into something like the percentage change in each outcome per unit change of Native_biomass. You could divide each coefficient by typical values of its corresponding outcome--lay date or chick size, respectively--to get all coefficients expressed in units of fractional (or percentage) change in outcome per unit change of Native_biomass. That would depend on your choice of "typical" value. (If you could model those outcomes on natural-log scales instead, the coefficients with respect to Native_biomass would already be in units of "fractional outcome change per unit change of Native_biomass.")
In the Poisson GLM, the original regression coefficient is already in units of "log of egg numbers per unit change of Native_biomass. That's inherently a fractional change.
In the binomial logistic GLM, you would have to make a choice about how to normalize, as the original coefficient in units of "change in log-odds of success per unit change of Native_biomass." It might be simpler to explain if you converted from log-odds to probabilities, but to do that you would have to choose specific values for all of the other covariates in the model first.
But does all that really make sense from a biological perspective? This seems more like a question of biological significance of lay date versus clutch size versus chick size versus fledgling success, once you know how each is associated with native biomass, rather than a statistical scaling problem.

*Did all individuals eventually lay eggs? If not, you would have to take that into account in your modeling.
