Bayesian and Likelihood-based density estimate results discrepancy

[Carolyn Devens]

Hi there,

I conducted a study determining leopard density estimates using program secr (likelihood based) and SPACECAP (Bayesian). In most papers, the results using both these methods are extremely comparable to each other (while being much lower than traditional non SCR methods). My study however had overall density estimates differing by a factor of 3.37 (secr: 0.35 leopards/100km2; SPACECAP: 1.18 leopards/100km2). 

I'm having trouble determining what this means in terms of our confidence in the extrapolated regional leopard density and abundance estimates? I'm suggesting that our results are a population baseline for comparison with replicate surveys in the future, but i feel like my discussion is lacking a confident explanation as to why there is such a density estimate discrepancy. Also, because my results vary so much, I'm being asked to suggest which method is more suitable for future research, but both methods have a series of pros and cons so I hesitate to name one... I personally found SPACECAP more user friendly and comprehensive, but I also struggle with R which made secr a nightmare. Not really reasons I can mention in a paper haha.

The reason I am posting this question here and not on a secr forum is because I am wondering whether my HRC input file might have made the difference. To create the HRC file, we used a Maxent habitat mask with Worldclim bioclim variables, altitude, vegetation data, etc., as well as sample telemetry data from 2 collared leopards within the study region. Many studies recommend incorporating telemetry data into density estimates and colleagues of mine developed leopard densities estimates based solely off collar data and those results are much more comparable to my SPACECAP results. Could this possibly be what sets my Bayesian analysis apart from my likelihood based one... and consequently, do you think I could make a case for my SPACECAP results being more reliable?

Any thoughts would be tremendously helpful! 

Regards,

Carolyn