Table 1.

Estimated rates of range expansion (km/year) derived from different range expansion models. The estimates using survey data are based only on the three periods when the area around the invasion point was comprehensively surveyed to the range margin (1981; 1989–1990; 2008–2010), and include 95% CIs in parentheses. Sampling accuracy shows the proportion of occupied sites (i.e., 5 × 5 km grid squares) that need to be sampled to be confident that the range statistic (e.g., mean) is within 5 (or 10)% of the true value. The estimates using yearly data include all years where records exist (1981–2012), regardless of how large an area was surveyed in that year; thus these estimates include years with highly biased data

Range expansion model	Estimate [km/year] using survey data (95% CIs)	Sampling accuracy within 5 (or 10)%	Estimate [km/year] using yearly data
Grid occupancy (linear)	1.11 (0.55–1.66)	0.95 (0.90)	0.31
Grid occupancy (nonlinear)	1.58 (1.02–2.21)	0.95 (0.90)	0.69
Mean	1.52 (0.97–2.07)	0.66 (0.33)	1.28
Median	1.50 (1.05–1.95)	0.97 (0.83)	1.24
Gamma	3.03 (1.62–4.45)	0.51 (0.20)	3.21
Marginal mean	3.12 (1.85–4.41)	0.58 (0.38)	1.97
Maximum	3.09 (1.49–4.69)	0.36 (0.17)	2.59
Conditional Maximum1	3.09 (1.49–4.69)	–	1.45

¹Sampling accuracy is not given for conditional maximum because its value is conditional on previous years' values that are not included in the data simulation.