Can waterfowl buffer the mortality risk induced by GPS tags? A cautionary tale for applied inference across species

Setash, Casey M.; Behney, Adam C.; Gammonley, James H.; Overton, Cory T.; Casazza, Michael L.; Letourneux, Frédéric; Buderman, Frances E.; Schummer, Michael L.; Luukkonen, Benjamin Z.; Huck, Nathaniel R.; Beatty, Karen E.; Legagneux, Pierre; Koons, David N.

doi:10.1186/s40317-024-00381-2

Animal Biotelemetry

Table 1 Results from the Bayesian band recovery model for 13 taxa of waterfowl fitted with only a metal band or also with a GPS tag

From: Can waterfowl buffer the mortality risk induced by GPS tags? A cautionary tale for applied inference across species

Species	S_band (SD)	S_tag (SD)	HR_avg (f)	Years	No. bands recovered (No. released)	No. tags recovered (No. released)
LSGO^b¥	0.87 (0.02)	0.86 (0.10)	1.13 (0.42)	2013–2022	2876 (41,406)	6 (101)
GSGO^a	0.80 (0.03), 0.79 (0.04)	0.41 (0.13), 0.61 (0.15)	4.29 (0.99), 2.21 (0.88)	2006–2010, 2019–2022	136 (1771), 43 (762)	19 (87), 24 (116)
GWFG^b	0.86 (0.02)	0.75 (0.12)	1.98 (0.80)	2012–2022	967 (7157)	12 (131)
BLBR^a	0.91 (0.02)	0.79 (0.07)	2.68 (0.96)	2006–2022	283 (14,076)	10 (122)
CANG^b	0.76 (0.01)	0.56 (0.13)	2.26 (0.93)	2008–2022	20,467 (89,748)	27 (341)
WODU^a	0.55 (0.02)	0.26 (0.19)	2.71 (0.91)	2019–2022	780 (8976)	5 (33)
CITE^a	0.56 (0.03)	0.25 (0.11)	2.61 (0.99)	2017–2022	40 (1604)	15 (119)
GADW^a	0.62 (0.03)	0.48 (0.15)	1.62 (0.82)	2015–2020	41 (434)	12 (105)
AMWI^a	0.56 (0.04)	0.48 (0.24)	1.58 (0.62)	2019–2022	25 (676)	3 (63)
MALL^a	0.60 (0.01)	0.48 (0.06)	1.45 (0.97)	2013–2022	10,018 (108,829)	90 (1140)
ABDU^a	0.60 (0.02)	0.45 (0.19)	1.77 (0.78)	2007–2022	894 (14,054)	5 (116)
NOPI^a	0.59 (0.04)	0.46 (0.22)	1.77 (0.71)	2017–2022	7 (127)	5 (95)
LESC^c	0.56 (0.03)	0.29 (0.13)	2.34 (0.97)	2006–2022	312 (7813)	8 (111)

Hazard ratios were calculated as the ratio of the mortality hazard for individuals wearing a GPS tag and a band to the hazard for individuals wearing only a metal band. Species and sub-species included lesser snow goose (LSGO), greater snow goose (GSGO), greater white-fronted goose (GWFG), black brant (BLBR), Canada goose (CANG), wood duck (WODU), cinnamon teal (CITE), gadwall (GADW), American wigeon (AMWI), mallard (MALL), American black duck (ABDU), northern pintail (NOPI), and lesser scaup (LESC). We used time-averaged hazard rates calculated from a log-linear model that accounted for random time effects (for the band-only group) or time periods (for the band plus GPS tag group) to compute the hazard ratios and survival probabilities. All numbers are rounded to the second decimal place. SD indicates the standard deviation (i.e. sampling uncertainty) of the posterior and f is the proportion of the hazard ratio posterior greater than 1, where hazard ratios greater than 1 are indicative of higher mortality associated with GPS-tagged birds than band-only birds. Point estimates for species with temporal or geographic variation in the model structure were computed using averaged hazard rates
HR hazard ratio
^aBackpack attachment style comprised majority of sample
^bNeck collar attachment style comprised majority of sample
^cImplant attachment style comprised majority of sample
^¥Geographic variation in LSGO survival necessitated assigning Canadian provinces to geographic Flyways. All Flyway assignments are noted in Table S1

Back to article page

ISSN: 2050-3385

Contact us

Submission enquiries: journalsubmissions@springernature.com