Question about oSCR minimum recaptures

Antonio Sampedro

unread,

Oct 12, 2025, 6:03:22 PM10/12/25

to oscr_p...@googlegroups.com

Hi everyone.

I would like ask for a bit of your wisdom here.

I applied oSCR to fit this multisession model to calculate population density of a very elusive felid. Set the encounter dataframe, trap dataframe, data function (data2oscr()), calculated sigma from 1/2 of the mmdm and set a buffer (3.5*sigma) for the Space State Dataframe.

(m1 <- oSCR.fit(model = list(D~1, p0~session, sig~1), sl.sf, sl.ss)

Data comes from camera trap deployment. 83 location, 2 sessions, 17 captures, and 9 recaptures (One individual even 4 recaptures). Occasion length is 5 days.

As expected the population estimates are not very robust. But my question is, Is there any specific lower limit for the number of recaptures below which SECR estimates become unreliable? The final estimates relate quite nicely to the actual population size.

Here are my estimates after backtransforming.

estimate se lwr upr Session

1 0.00919712 0.00437194 0.003622576 0.02334996 1

per Km2

The standard error is quite high, but does that automatically invalidate the results, or can they still be considered informative?

Also, am I missing any alternative framework to handle such low-recapture multi-session data?

Thank you very much in advance for any guidance!

Antonio Sampedro

Wildlife biologist | Tibetan Language Translator Research Officer at Himalayan Wolves Project

Mobile: +34 639843617 |

anton.s...@gmail.com

https://www.himalayanwolvesproject.org/

Daniel Linden

unread,

Oct 13, 2025, 2:37:34 PM10/13/25

to oSCR

Hi Antonio, it's definitely frustrating when the data are so sparse. How many spatial recaptures did you have?

There are not many solutions for wildlife monitoring when the data are lacking, aside from changing your design in some way. Still, your estimates are certainly not the worst I've seen and the fact that the algorithm converged and gave you a result is a victory itself. Sometimes a 30% CV is used as a general threshold for whether a population estimate is "robust", but that is certainly a gray area. And again, this threshold would suggest a design change is needed for more robust monitoring but it would not stop folks from using the population estimate to make inferences (especially in the absence of alternatives).

Antonio Sampedro

unread,

Oct 16, 2025, 3:56:55 PM10/16/25

to oscr_p...@googlegroups.com

Hi Daniel, thank you very much for your message.

Yes indeed it is frustrating, working with big cats. Little alternatives here to improve data.

We will try to explain it right. I have tried to make some simulations that would measure the uncertainty direction. At the end, it is what we got.

Thank you very much, your message encouraged me.

Best regards!

Antonio Sampedro

Wildlife biologist | Tibetan Language Translator Research Officer

Mobile: +34 639843617 |

anton.s...@gmail.com

https://www.himalayanwolvesproject.org/

El 13 oct 2025, a las 20:37, Daniel Linden <danl...@gmail.com> escribió:

--
You received this message because you are subscribed to the Google Groups "oSCR" group.
To unsubscribe from this group and stop receiving emails from it, send an email to oscr_package...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/oscr_package/82e01614-9db6-4b7f-aa8d-2cc29c193721n%40googlegroups.com.