Dead-reckoning facilitates determination of activity and habitat use: a case study with European badgers (Meles meles)

Table 1 Final best-fit general linear (GLM) and general linear mixed models (GLMM) examining variation in area use, nightly travel distances, and activity as determined using GPS-enhanced dead-reckoned data

Model no.	Response	Fixed effects	AICc	Marginal R²	Conditional R²	Final model effects	χ²	df	p
1	Area use	Average distance + Average sum VeDBA	7.557	–	–	Average distance Average sum VeDBA	20.311 4.389	1 1	<0.001 0.036
2	Distance travelled per night	Sum VeDBA + Average speed*Sum tortuosity + Hours + Average temperature	470.932	0.958	0.975	Sum VeDBA Speed Sum tortuosity Hours Average temperature Speed:Sum tortuosity	7.339 52.740 5.404 3.340 2.770 5.161	1 1 1 1 1 1	0.007 <0.001 0.020 0.068 0.096 0.023
3	Sum VeDBA per night	Hours + Average speed + Sum tortuosity	584.064	0.681	0.753	Hours Average speed Average temperature	51.238 0.018 0.125	1 1 1	<0.001 0.893 0.726
4	% Time spent (sqrt transformed)	Habitat	512.235	0.592	0.595	Habitat	162.163	3	<0.001
5	Distance travelled	Habitat	627.361	0.510	0.510	Habitat	71.742	3	<0.001
6	Average VeDBA per second	Habitat	345.158	0.155	0.502	Habitat	20.587	3	<0.001
7	Average speed (log transformed)	Habitat	109.676	0.248	0.580	Habitat	38.949	3	<0.001
8	Average tortuosity per second (sqrt transformed)	Habitat	439.677	0.256	0.354	Habitat	26.361	3	<0.001

Model numbers (1–8) are shown, alongside the dependent ‘response’ variable (and relevant transformations). Independent variables (‘fixed effects’) are shown with an asterisk (*) denoting an interaction between variables and a plus sign (+) denoting main effects. AICc represents the Akaike information criterion values corrected for small sample sizes, which were used to select the final best-fit model. Marginal and conditional R² values are noted. Final model effects are shown, with corresponding Chi-squared statistic (χ²), degrees of freedom (df), and probability value (p). Significant p values are shown in bold. Model 1 was a GLM examining variation in area use during the collared period using Kernel Density Estimate (KDE). Models 2 and 3 were GLMMs examining GPS-enhanced dead-reckoned travel distances and sum VeDBA per night, respectively, with Badger ID included as a random effect to account for repeated measures from individuals. Models 4–8 were also GLMMs with Badger ID included as a random effect, and examined variation in time spent, distance travelled, and activity (VeDBA, speed, and tortuosity) in different habitats

ISSN: 2050-3385