University of California Press

Figure 4

$Regression of Arctic minimum sea ice extent against observed crossing dates over two latitudinal bands. Shown are correlation results (a–d) and comparisons (e) of observed summer minimum sea ice extent with those forecast using different thresholds for fractional coverage of specular returns (FSR, threshold x) and surface albedo (α¯, threshold y). In (a), ΦSR.L(x) and ΦSR.H(x) refer to the dates when the area-averaged FSR (F—SR) for 2011–2017 climbs above threshold x (see Equation 1) in the lower (L, 70–80°N) and higher (H, 80–90°N) latitudinal bands. Squared correlation/adjusted squared correlation (first row) and standard error (second row) are given for three thresholds; underlined values are the statistical significance (p values) of selected regression analyses. In b), same, except that ΛL(y) and ΛH(y) indicate the dates when surface albedo (α¯) falls below threshold y (see Equation 2) in latitudinal bands L and H, for two thresholds. In c) and d), same as a) and b), respectively, except using only crossings from 2011–2016. Additional row (in italics) shows the difference between the forecast and observed ice extent (IE, September minimum) in 2017 (of 4.636 × 106 km2). In upper e), day of year for the different threshold crossings in 2011–2017 is shown. Blue lines/triangles show ΛL(0.6) and ΛH(0.6), red lines/circles show ΦSR.L(0.5) and ΦSR.H(0.5), and thin and thick lines are the crossing dates at the lower (L, 70–80°N) and higher (H, 80–90°N) latitudinal bands, respectively. In lower e), observed ice extent (black line) is compared with those estimated using regression coefficients from crossing dates of 2011–2017 (colored lines, corresponding to information in a) and b) in the colored line boxes) and from only 2011–2016 (colored dashed lines, corresponding to information in c) and d) in colored dashed-line boxes). DOI: https://doi.org/10.1525/elementa.311.f4$

Regression of Arctic minimum sea ice extent against observed crossing dates over two latitudinal bands. Shown are correlation results (a–d) and comparisons (e) of observed summer minimum sea ice extent with those forecast using different thresholds for fractional coverage of specular returns (F_SR, threshold x) and surface albedo (⁠ $\bar{α}$ ⁠, threshold y). In (a), Φ_SR.L(x) and Φ_SR.H(x) refer to the dates when the area-averaged F_SR (⁠ ${\overset{—}{F}}_{SR}$ ⁠) for 2011–2017 climbs above threshold x (see Equation 1) in the lower (L, 70–80°N) and higher (H, 80–90°N) latitudinal bands. Squared correlation/adjusted squared correlation (first row) and standard error (second row) are given for three thresholds; underlined values are the statistical significance (p values) of selected regression analyses. In b), same, except that Λ_L(y) and Λ_H(y) indicate the dates when surface albedo (⁠ $\bar{α}$ ⁠) falls below threshold y (see Equation 2) in latitudinal bands L and H, for two thresholds. In c) and d), same as a) and b), respectively, except using only crossings from 2011–2016. Additional row (in italics) shows the difference between the forecast and observed ice extent (IE, September minimum) in 2017 (of 4.636 × 10⁶ km²). In upper e), day of year for the different threshold crossings in 2011–2017 is shown. Blue lines/triangles show Λ_L(0.6) and Λ_H(0.6), red lines/circles show Φ_SR.L(0.5) and Φ_SR.H(0.5), and thin and thick lines are the crossing dates at the lower (L, 70–80°N) and higher (H, 80–90°N) latitudinal bands, respectively. In lower e), observed ice extent (black line) is compared with those estimated using regression coefficients from crossing dates of 2011–2017 (colored lines, corresponding to information in a) and b) in the colored line boxes) and from only 2011–2016 (colored dashed lines, corresponding to information in c) and d) in colored dashed-line boxes). DOI: https://doi.org/10.1525/elementa.311.f4

This Feature Is Available To Subscribers Only