openCR and sigma
Cheryl Lohr
Hi folks,
Thank you in advance for any advice you can give me.
I’m trying to estimate the density of golden bandicoots on an island using openCR. The trapping data unfortunately has unequal intervals between primary sessions and an unequal number of secondary sessions. I seem to be getting a lot of movement in sigma which is drastically altering the estimates of density.
model npar rank logLik AIC AICc dAIC AICwt
sigma_withinbetween_JSSAsecrD lambda0~1 phi~1 D~1 sigma~b + t 11 11 -6203.378 12428.76 12429.56 0.000 1
bwithin_primary_sigma_JSSAsecrD lambda0~1 phi~1 D~1 sigma~b 5 5 -6349.819 12709.64 12709.82 280.882 0
Dsigmawithin_JSSAsecrD lambda0~1 phi~1 D~t sigma~b 11 11 -6345.813 12713.63 12714.43 284.871 0
Dsigma_JSSAsecrD lambda0~1 phi~1 D~t sigma~t 16 16 -6370.344 12772.69 12774.37 343.932 0
PhiD_JSSAsecrD lambda0~1 phi~t D~t sigma~1 15 15 -6386.530 12803.06 12804.54 374.304 0
sigma_JSSAsecrD lambda0~1 phi~1 D~1 sigma~t 10 10 -6396.644 12813.29 12813.96 384.532 0
Phi_JSSAsecrD lambda0~1 phi~t D~1 sigma~1 9 9 -6402.413 12822.83 12823.37 394.070 0
z_JSSAsecrD lambda0~1 phi~1 D~1 sigma~1 4 4 -6408.461 12824.92 12825.04 396.166 0
lambda_JSSAsecrD lambda0~1 phi~1 D~1 sigma~1 4 4 -6408.461 12824.92 12825.04 396.166 0
Constant_JSSAsecrD lambda0~1 phi~1 D~1 sigma~1 4 4 -6408.461 12824.92 12825.04 396.166 0
D_JSSAsecrD lambda0~1 phi~1 D~t sigma~1 10 10 -6402.623 12825.25 12825.91 396.489 0
bbetween_primary_sigma_JSSAsecrD lambda0~1 phi~1 D~1 sigma~bsession 5 5 -6408.405 12826.81 12826.99 398.055 0
Could anyone please tell me if I’m on the right track and how I might be able to extract and plot the real estimates of sigma? From the best model the real estimates of sigma are:
sigma
session b estimate SE.estimate lcl ucl
1 0 673.8997 39.16538 601.34754 755.2052
2 0 778.5471 86.74708 625.81009 968.5617
3 0 791.3380 74.58550 657.86115 951.8967
4 0 995.2895 180.40817 697.68427 1419.8417
5 0 196.5815 27.84028 148.93399 259.4727
6 0 152.1714 13.29254 128.22671 180.5874
7 0 727.2378 70.33320 601.66444 879.0195
1 1 496.1397 39.51190 424.43919 579.9527
2 1 573.1835 47.83213 486.69956 675.0352
3 1 582.6004 61.86439 473.13442 717.3928
4 1 732.7540 166.03729 469.98224 1142.4440
5 1 144.7276 27.54590 99.66506 210.1648
6 1 112.0319 10.89286 92.59328 135.5514
7 1 535.4084 67.42479 418.30407 685.2961
I don’t understand how to interpret these values. If b=1 then within primary session sigma is constant?
Distance analyses by other authors on another site found that bandicoots will move into the centre of web array of detectors and density estimates become inflated. Within the mask I’ve set the buffer to include the full extent of the island and spacing is based on sigma.
Code used
Capthist
IA<-read.capthist(captfile="GB_capthist.txt",
trapfile = list("GB_traplocs1.txt",
"GB_traplocs2.txt",
"GB_traplocs3.txt",
"GB_traplocs4.txt",
"GB_traplocs5.txt",
"GB_traplocs6.txt",
"GB_traplocs7.txt"),
detector = "multi",
fmt=c("trapID","XY"))
intervals(IA)<-c(67,46,179,368,705,823)
summary(IA)
Mask
fence = readOGR(dsn=wd,layer="Doole_island_LWM_GDA94MGA50", verbose=FALSE)
##Suggested values for buffer=4signma; and spacing=0.2-1 sigma
##To find out sigma from capthist object use:
RPSV(IA,CC =TRUE)
##In this case sigma ranges from 29 to 39
mask.clipped = make.mask(traps=traps(IA),
buffer=2000,
spacing=35,
poly=fence,
poly.habitat = TRUE)
plot(mask.clipped)
Model
sigma_Session+t_JSSAsecrD<-openCR.fit(IA,mask=mask.clipped,type='JSSAsecrD', ncores=4,model = sigma ~ Session+t,
method="Nelder-Mead", details = list(control = list(maxit = 500)),start=bwithin_primary_sigma_JSSAsecrD)
saveRDS(sigma_Session+t_JSSAsecrD, file = "sigma_Session+t_JSSAsecrD_IA_JSSAsecrD_DOOLE.rds")
Cheryl Lohr
Research Scientist, Animal Science Program
DBCA Biodiversity and Conservation Science
Location: 37 Wildlife Pl, Woodvale, WA 6026
Mail: Woodvale Wildlife Research Centre
Locked Bag 104 Bentley Delivery Centre
WA 6983
Ph: 94055150 (internal ext 5750)
Mob: 0407335004
This message is confidential and is intended for the recipient named above. If you are not the intended recipient, you must not disclose, use or copy the message or any part of it. If you received this message in error, please notify the sender immediately by replying to this message, then delete it from your system.
