ni, 
m 
ADVANCING EARTH 
AND SPACE SCIENCES 
Journal of Geophysical Research: Oceans 
10.1029/2023JC019937 
Kinematic approach 
65°'N 
Volume = 1.5x10'* m? 
27.1 
Volume = 1.8x10'4 m? 
27.2 
Volume = 2x10'* m? 
27.5 
400 
350 
DD X 
= 60'N 
= 55 N. 
9 | 
O 50N 
45°N' 
65°N 
En 
60°'N 
N 
5 
SQ 50°N 
Volume = 1.8x10'* m? 
27 
Volume = 2x10'* m? 
27.2 
% 
Volume = 1.5x10'* m® 
27.5 
300 
{250 E 
N 
8 
(200 & 
me 
150 £ 
— 
© 
i00 & 
— 
50 
0 
45°N 
Thermodynamic approach 
65°N 
D 
E 60N 
” 
& ] 
 55°N 
E 
9 l 
© 50°N 
45°N — 
65°’N 
A 
N 
60°N' 
S5'N/ 
& 
CC) 
50°N- 
45°’N L 
Volume = 1.5x10'* m? 
> 
Volume = 1.8x10'* m? 
»”z x 
Volume = 2x10'* m? 
27.5. 
Volume = 1.8x10'4 m* 
> 
Volume = 2x10'* m® 
27.3 
Volume = 1.5x10'* m? 
27.5 
— 
40°W 30W 20W 10°W 0° 40W 30W 20W 10°W 0° 40W 30°W 20°W 10°W 
0 
„0.01 
‚0.008 
9.006 
0.004 
> 
| © 
0.002 '— 
© 
© 
10 I 
= 
0.002 © 
0.002 © 
E 
db 
-0.004 © 
Lc 
0.006 
-0.008 
„0.01 
Figure 9. Composite maps of the SPMW layer thickness [m] (filled contour with gray color scheme). The color bar is displayed on the top right side, and formation rates 
Sv] from the kinematic approach (a) to (f) and thermodynamic approach (g) to (1) for the isopycnal bins at 09 = 27.1 kg m” (first column a, b, g, and h), 27.3 kg m” 
(second column c, d, i, and j), and 27.5 kg m” (third column e, f, k, and 1). The formation rate is depicted with contour lines using a purple/green color scheme. The 
purple contour lines are positive values and green contour lines are negative values (color bar on the bottom right side). The composite 1 (first row a, c, and e and third 
row g, 1, and k) is calculated as the mean of the years 1995, 2000, 2009, 2015 and 2018. The composite 2 (second row b, d and f and fourth row h, j, and 1) is calculated as 
the mean of the years 1998, 2004, 2010, 2013, and 2017. Bathymetry at 1,000 and 2,000 m is also displayed with gray contour lines. The mean volume of SPMW for 
each isopycnal bin is given on top of each panel. 
About the mean state of SPMW, we observed that its spatial distribution within each isopycnal layer aligns well 
with the corresponding formation rate. However, the formation rate derived from the kinematic approach is much 
noisier than the thermodynamic one. This can be ascribed to the limits in the characterization of mesoscale and 
sub-mesoscale velocity components at the spatial and temporal resolution considered. 
STENDARTI ET 
AL 
+ 
at 2