Model parameter
Multimodel average estimate ± S. E
Z
P-value
Completed model estimate ± S. E
Relative importance
Intercept 0.99 ± 0.01 70.36 *** 0.99 ± 0.01
es 0.14 ± 0.01 11.03 *** 0.13 ± 0.01 100%
o -0.32 ± 0.05 6.63 * -0.31 ± 0.07 98%
li 0.75 ± 0.03 2.24 *** 0.01 ± 0.04 85%
es × o -0.12 ± 0.04 3.41 *** -0.11 ± 0.05 89%
es × li -0.07 ± 0.03 2.68 * -0.006 ± 0.02 8%
lo -0.13 ± 0.02 6.41 *** -0.003 ± 0.019 2%
es × lo -0.06 ± 0.02 3.36 *** -0.001 ± 0.009 2%
p -0.12 ± 0.03 3.55 *** -9.76e-10 ± 1.103e-05 <1%
es × p -0.03 ± 0.01 2.21 *** -7.997e-11 ± 1.676e-06 <1%
LMM: Linear Mixed-effect model, es: Exposure Salinity, o: Origin, lo: Location, p: Population, li: Lineage. Estimates of variables were computed after centering and standardizing the explanatory variables to a mean of 0 and a s.d. of 0.5. Results are presented in both multimodel inference procedure (Burnham & Anderson, 2003) and completed LMM to account for a greater range of alternative, more parsimonious combinations of explanatory variables with respect to their relative weight of statistical support. The relative importance scores were also measured for each variable based on their frequency within parsimonious models. The variables with higher scores are considered more important than those with lower scores, and therefore receive higher support as potential correlates of movement distance. The significant level for the model averaging and relative importance estimations was set at 0.05. *** indicates P < 0.001, ** indicates 0.001 < P < 0.01, * indicates 0.01 < P < 0.05. LMM: Linear Mixed-effect model, es: Exposure Salinity, o: Origin, lo: Location, p: Population, li: Lineage. Estimates of variables were computed after centering and standardizing the explanatory variables to a mean of 0 and a s.d. of 0.5. Results are presented in both multimodel inference procedure (Burnham & Anderson, 2003) and completed LMM to account for a greater range of alternative, more parsimonious combinations of explanatory variables with respect to their relative weight of statistical support. The relative importance scores were also measured for each variable based on their frequency within parsimonious models. The variables with higher scores are considered more important than those with lower scores, and therefore receive higher support as potential correlates of movement distance. The significant level for the model averaging and relative importance estimations was set at 0.05. *** indicates P < 0.001, ** indicates 0.001 < P < 0.01, * indicates 0.01 < P < 0.05. LMM: Linear Mixed-effect model, es: Exposure Salinity, o: Origin, lo: Location, p: Population, li: Lineage. Estimates of variables were computed after centering and standardizing the explanatory variables to a mean of 0 and a s.d. of 0.5. Results are presented in both multimodel inference procedure (Burnham & Anderson, 2003) and completed LMM to account for a greater range of alternative, more parsimonious combinations of explanatory variables with respect to their relative weight of statistical support. The relative importance scores were also measured for each variable based on their frequency within parsimonious models. The variables with higher scores are considered more important than those with lower scores, and therefore receive higher support as potential correlates of movement distance. The significant level for the model averaging and relative importance estimations was set at 0.05. *** indicates P < 0.001, ** indicates 0.001 < P < 0.01, * indicates 0.01 < P < 0.05. LMM: Linear Mixed-effect model, es: Exposure Salinity, o: Origin, lo: Location, p: Population, li: Lineage. Estimates of variables were computed after centering and standardizing the explanatory variables to a mean of 0 and a s.d. of 0.5. Results are presented in both multimodel inference procedure (Burnham & Anderson, 2003) and completed LMM to account for a greater range of alternative, more parsimonious combinations of explanatory variables with respect to their relative weight of statistical support. The relative importance scores were also measured for each variable based on their frequency within parsimonious models. The variables with higher scores are considered more important than those with lower scores, and therefore receive higher support as potential correlates of movement distance. The significant level for the model averaging and relative importance estimations was set at 0.05. *** indicates P < 0.001, ** indicates 0.001 < P < 0.01, * indicates 0.01 < P < 0.05. LMM: Linear Mixed-effect model, es: Exposure Salinity, o: Origin, lo: Location, p: Population, li: Lineage. Estimates of variables were computed after centering and standardizing the explanatory variables to a mean of 0 and a s.d. of 0.5. Results are presented in both multimodel inference procedure (Burnham & Anderson, 2003) and completed LMM to account for a greater range of alternative, more parsimonious combinations of explanatory variables with respect to their relative weight of statistical support. The relative importance scores were also measured for each variable based on their frequency within parsimonious models. The variables with higher scores are considered more important than those with lower scores, and therefore receive higher support as potential correlates of movement distance. The significant level for the model averaging and relative importance estimations was set at 0.05. *** indicates P < 0.001, ** indicates 0.001 < P < 0.01, * indicates 0.01 < P < 0.05. LMM: Linear Mixed-effect model, es: Exposure Salinity, o: Origin, lo: Location, p: Population, li: Lineage. Estimates of variables were computed after centering and standardizing the explanatory variables to a mean of 0 and a s.d. of 0.5. Results are presented in both multimodel inference procedure (Burnham & Anderson, 2003) and completed LMM to account for a greater range of alternative, more parsimonious combinations of explanatory variables with respect to their relative weight of statistical support. The relative importance scores were also measured for each variable based on their frequency within parsimonious models. The variables with higher scores are considered more important than those with lower scores, and therefore receive higher support as potential correlates of movement distance. The significant level for the model averaging and relative importance estimations was set at 0.05. *** indicates P < 0.001, ** indicates 0.001 < P < 0.01, * indicates 0.01 < P < 0.05.