Hello,
Being rather new to statistics and to this type of models, I find myself struggling to be sure about my model results and how to interpret them correctly. I hope you will forgive the elementary nature of my questions…
I am studying the temporal trends of several bird species. Here is an example of the model used for one species :
### unmarked frame
> summary(umf_sp1)
unmarkedFrame Object
25 sites
Maximum number of observations per site: 21
Mean number of observations per site: 19.12
Number of primary survey periods: 7
Number of secondary survey periods: 3
Sites with at least one detection: 25
Tabulation of y observations:
0 1 2 3 4 5 6 <NA>
156 163 111 30 14 3 1 47
Site-level covariates:
w_area w_height w_dens
Min. :-1.28928 Min. :-1.5452 Min. :-1.94598
1st Qu.:-0.88181 1st Qu.:-0.8470 1st Qu.:-0.49685
Median :-0.01456 Median : 0.2234 Median :-0.08281
Mean : 0.00000 Mean : 0.0000 Mean : 0.00000
3rd Qu.: 0.90019 3rd Qu.: 0.7819 3rd Qu.: 0.33123
Max. : 2.14665 Max. : 1.8989 Max. : 2.40143
Observation-level covariates:
dAs
Min. :-1.5017
1st Qu.:-0.9049
Median : 0.0009
Mean : 0.0000
3rd Qu.: 0.9733
Max. : 1.7983
NAs :47
Yearly-site-level covariates:
dOs rod year
Min. :-2.2336 Min. :-1.0154 Min. :-1.1514
1st Qu.:-0.7556 1st Qu.:-0.8070 1st Qu.:-0.8635
Median : 0.4289 Median :-0.1817 Median :-0.2878
Mean : 0.0000 Mean : 0.0000 Mean : 0.0000
3rd Qu.: 0.8048 3rd Qu.: 0.6519 3rd Qu.: 0.5757
Max. : 1.1796 Max. : 3.3611 Max. : 2.0149
NAs :3
### model
mod_sp1 <- pcountOpen(~1 + w_area + w_height, ~1 + rod + dOs, ~1, ~1 + w_area + w_dens + w_height + dAs + year, umf_sp1, mixture="P", K=40, dynamics="trend")
### model results
> summary(mod_sp1)
Abundance (log-scale):
Estimate SE z P(>|z|)
(Intercept) 3.1471 0.207 15.190 4.09e-52
w_area -0.0395 0.200 -0.198 8.43e-01
w_height -0.0949 0.151 -0.630 5.29e-01
Growth Rate (log-scale):
Estimate SE z P(>|z|)
(Intercept) -0.0521 0.0392 -1.327 0.184
rod 0.0134 0.0204 0.656 0.512
dOs -0.0199 0.0228 -0.873 0.383
Detection (logit-scale):
Estimate SE z P(>|z|)
(Intercept) -2.7830 0.1448 -19.225 2.30e-82
w_area 0.1300 0.2097 0.620 5.35e-01
w_dens 0.0377 0.0709 0.532 5.95e-01
w_height 0.1318 0.1699 0.776 4.38e-01
dAs 0.2756 0.0663 4.157 3.22e-05
year 0.3018 0.1305 2.312 2.08e-02
After that, I backtransformed my estimates with the linear combination method, for the intercept and the covariates, such as :
### backtransformation
lambda_int <- backTransform(linearComb(mod_sp1, type="lambda", coefficients=c(1,0,0)))
lambda_w_area <- backTransform(linearComb(mod_sp1, type="lambda", coefficients=c(0,1,0)))
lambda_w_height <- backTransform(linearComb(mod_sp1, type="lambda", coefficients=c(0,0,1)))
And so on for all the estimates, to obtain this table :
### backtransformed estimates
lambda_int lambda_w_area lambda_w_height
23.26807342 0.96122565 0.90950410
gamma_int gamma_rod gamma_dOs
0.94927353 1.01349871 0.98030617
det_int det_w_area det_w_dens det_w_height det_dAs det_year
0.05825069 0.53244820 0.50942617 0.53290999 0.56847207 0.57488290
So now I have a few questions/clarifications to be sure that I understood the meaning of these estimates properly :
1) lambda intercept gives the mean abundance for the first year of study for the covariates fixed at their mean value (because they are scaled). So it is an average value for one hypothetical site, not over the whole "metapopulation"?
2) gamma intercept gives the finite rate of increase. Is it as an average over all sites? and over the whole study period?
3) the detection estimate seems to be very low (here p=5.8%, and it is one of the highest of all my species). Is it the mean probability of detection of an individual rather than the global p for the entire species? Is it also a mean value over the whole study period?
4) Probably a very basic algebra question : Is there a way to calculate the effect of a covariate on the detection estimate when they are backtransformed? It is easy for lambda and gamma, because their exp() backtransformation has a simple rule of exp(a+b) = exp(a) * exp(b), so what was additive now has to be simply multiplied to see the backtransformed effects on the parameter. But for detection, due to the more complex logit() / logistic() backtransformation, I cannot find a way to combine directly those backtransformed covariates…
Thanks in advance for any help!
Clem