week11 – Brian D. Gerber

Week 11

The authors state they use a “process model” and an “observation model.” In your own words, what does this mean, and what is it representing here? Can you think of a situation in your own work where you might want to use a similar structure?
What are some of the key model design choices the authors used? Did they face limitations in measurement related to their instruments or staffing, for example, and how did those show up in modeling? Did you feel that they justified their decisions convincingly? Are there components that you have questions about? Under what conditions (if any) do you feel that these decisions could lead to model bias or error?
Does the model answer the conservation question or address the gap that the authors set out to fill? Knowing what you do about the study system, are there any other factors you think should have been considered? Did you find the model satisfactory to predict human-wildlife conflict, and if not, what would improve it?

Optional:

How should we interpret disagreement between different sources of information? What can we learn when farmers report conflict but camera traps do not detect wild boar?
Would you use this model to make management decisions? What additional evidence would you want before acting on its predictions?

Are there inherent biases in the data used? Are lek count data appropriate to answer the BACI study in Analysis 1? Is the sample size sufficient?
Is the distance-based approach to the BACI design providing information about cause-and-effect relationships, as claimed in the paper?
In this class, we emphasized the importance of having a priori hypothesis for robust research. In the paper, authors don’t clearly state their hypothesis. Do you agree with this approach of setting research?
What is the hierarchy in the SSM model this paper used to estimate rate of change in population abundance and lek absence?
In the Discussion, the study blames noise pollution, light pollution, and increased vehicle activity for decreased survival, perhaps because of the opposite relationship between topographic impedance and survival, but predation is the highest source of mortality for nests and females. Why make inference about the unmeasured indirect mechanism rather than the directly measured mechanism of raven abundance?
If you were approached by a federal agency to improve this study, as a researcher what changes would you make in the research design, what parameters/variables would you add and what statistical models would you use to make inference?

# CRAN packages
install.packages(c("terra","raster","amt","MASS","sp"))