meaning of 'Total' in stratified dht2

[Samantha Ball]

Dear List, 

  I was wondering if anyone could tell me what the value of 'Total' is in the dht2 output (highlighted below) as I am having issues finding an answer to this. I have replicate point surveys, where only half the point is visible (0.5) to a single species throughout the year and am wanting to look at abundance according to season. Therefore, I have fitted a detection function with no covariates and stratified according to season. Here, I have not pooled surveys according to season, but get very similar results when taking that approach. I thought that the total value would be the mean of the estimates, but it doesn't quite work out that way. In the output below, the mean you would expect for the total from the estimates would be 88.5. The result I have gotten is close, but not the same. 

Could anyone please help to shed some light as to what total is estimating here? 

Many Thanks,

Sam

[Eric Rexstad]

Good morning Sam. Sharp eye in examining your output. If your interest rests in the seasonal estimates, then the estimate labelled "Total" is of little interest to you, but happy that you are curious.

The simplest explanation may be rounding. Abundance estimates from report="abundance"​ are rounded to whole numbers (because they are abundances, which should be "whole" animals). We don't see the decimal values of the estimated abundances, if we could average those, they might equate to a value, when rounded, equals 90.

The mystery of the "total" value is caused by your choice of effort_sum​ as the stratification​ argument in your call to dht2​. Even though you didn't provide the code showing your call to the dht2​ function, the output echo shows that effort_sum​ was used.

From what you describe, I wouldn't think effort_sum​ is the option to use. Situations in which it is used are fairly exotic. I gather that your need for dht2​ was to apply your sampling fraction. In the situation you describe, I would think that stratification="geographic"​ would be the better option for the argument to dht2​.

An unsolicited comment unrelated to your question: the number of detections per season is relatively small, particularly for point transect surveys where detection function fitting is more difficult than lines. This could be a situation wherein you perhaps use season as a covariate in the detection function. That model (with 4 or 5 parameters) may be difficult to fit given the 163 detections, but might provide a way to assess if there are seasonal differences in detecability.

---

From: distance...@googlegroups.com <distance...@googlegroups.com> on behalf of Samantha Ball <samant...@ucc.ie>
Sent: 17 November 2022 22:49
To: distance-sampling <distance...@googlegroups.com>
Subject: [distance-sampling] meaning of 'Total' in stratified dht2

 

--
You received this message because you are subscribed to the Google Groups "distance-sampling" group.
To unsubscribe from this group and stop receiving emails from it, send an email to distance-sampl...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/distance-sampling/3a92fb30-67cf-4053-96f6-3c78c9facbdfn%40googlegroups.com.