Aanes R., Sæther B.-E., Smith F.M., Cooper E.J., Wookey P.A. & Øritsland, N.A. (2002). The Arctic Oscillation predicts effects of climate change in two trophic levels in a high-arctic ecosystem. Ecology Letters, 5, 445–453.

Due to an error in the written syntax when specifying the autoregressive model, the coefficients presented in Table 1 of Aanes et al. (2002) are erroneous. We have recalculated the coefficients using the r software (R-Development Core Team 2004). The true coefficients are given in Table 1 (Corrected), below.

Table 1.  (Corrected). Estimated autoregressive coefficients from eqn  1 shown for both 1- and 2-order models, explaining the variation in population growth rate (Rt) as influenced by regional climate (i.e. AO) during the period 1978–1999. AR models including the year 1993 are given in (a), and models excluding 1993 are given in (b). In the second column AOS refers to the AO summer index, while AOW refers to the AO winter index (see Methods)
Dependent variableClimate variableNAR(1)-coeff.Clim.-coeff.AICcAR(2)-coeff.Clim.-coeff.AICc
β 1 α 1 § β 1 β 2 α 1 §
  1. †Two-tailed t-test, H0:β1 = 0.

  2. ‡Two-tailed t-test, H0:β2 = 0.

  3. §Two-tailed t-test, H0:α1 = 0.

(a) Including 1993
  RtAOS200.054 −0.18229.930.0330.163 −0.11232.50
  RtAOW200.048 −0.19327.780.0590.241 −0.22329.70
(b) Excluding 1993
  RtAOS190.799 −0.282 −8.270.5220.352 −0.284 −6.94
  RtAOW190.805 −0.124 −6.770.6430.200 −0.120 −4.10

The error did not alter the major conclusion of the paper. Our reanalysis is consistent with the original paper showing that the most parsimonious model for the complete period (1978–1998) was a 1-order model including winter climate (AOW). The only discrepancy was that the lagged effect of climate during winter (AOW) disappeared when excluding the extreme year 1993. The reanalysis then suggests that a 1-order model including the AO during the snowfree months (AOS) is the most parsimonious model.