Amanda R. Goldberg

• Doctor of Philosophy
• Biologist at United States Geological Survey

Hi all, I'm about to start a telemetry project but I'm working at a lab that has no equipment in hand already so I need to start at the beginning. I will need to home in on VHF tags on various sized animals. I have used quite a few different receivers from companies over the years but all that equipment is no longer manufactured. Any ideas on who is making the best VHF receivers right now? I'm buying collars from Lotek this year so I can certainly buy their receiver as well but I wanted to see if there were better options. The last receiver I used was the R1000 which were handy for hiking with compared to the old brick style receivers I used to use that required a trap since they were super boxy. Thanks.

As part of a study I am working on we need a way to count human use in the area we are trapping wildlife. Many of our locations are adjacent to recreational trails so we thought it might be best to estimate use by placing trail count cameras nearby. Has anyone used these and if so, are there any companies that produce them that you like? How difficult is the setup?

Hi, I am analyzing a regression evaluating litter size and a number of covariates for multiple taxa so I need to take phylogeny into account. I have multiple entries for some of the species (for instance latitude is quite variable for some species so I need multiple entries to take that variation into account). However, I realize that may create a problem with my analysis. I am able to run the data without a problem in nlme using the gls function but I'm not sure if that is ok with having multiple entries for the same species.

pglsModel <- gls(littersize ~ lat, correlation = corBrownian(phy = tree, form=~Species),

data = data, method = "ML")

Furthermore, I have attempted to use pgls.SEy in the phytools package but that does not seem to run as I get the following error: Error in na.fail.default(list(vf = c(64488.9418569458, 64488.9409710521, : missing values in object

fit<-pgls.SEy(littersize ~ absLat, data=data, corClass=corBrownian, tree=tree,

se=NULL, method=c("ML"))

I have read that some folks have suggested that with multiple specimens in a species you need to run a MCMCglmm but I don't have multiple values for every species so I don't think I can really run this type of analysis (furthermore I typically only have a few entries per species so I don't know if that is enough either).

I am knew to including phylogenetic trees into data analysis so any suggestions of help is greatly appreciated.

Thanks

For example: I have Response variables (1,2,3) and predictors (A,B,C,D,E,F,G). I need to impute data for missing values and I am running following models

1~A+B+C+D

2~A+C+D+E+G

3~B+D+E+F+G

So I impute data using MICE for missing values with all my data combined or do I impute three times including only the variables listed in each model?

Thank you

I am recently been asked to figure out a way to measure nest concealment using data that was collected prior to my position. They have years of data where photographs were taken of a sphere placed where a nest would be on the ground and then took a photograph of the sphere. The idea is that the sphere is similar to a bird on a nest and they want the percent vegetation cover over that sphere.

Does someone know of a program (preferably free and relatively easy to use (at least once it is set up) so technicians can continue to use it in the future) available to run many of these calculations? We would prefer something that can a) detect the sphere and b) calculate the percent cover. I assume I will need to train the program but I do not know where to begin (I know how to manually do these measurements but we have far too many to count superimposed squares on every photograph).

Thank you.