V. Maurer et al.: Evaluation of coupled and uncoupled simulations
(a) SW net surface
(b) pr
50°
10}
50°N
CC
40°)
10°
x
30°1
Knall
50°
Ya
ii
MAM
56.
„an
SON
W°N
&
A
/c) LW net surface
sO°h.
Ö°h
NS.
3°:
AQ°N
;o°h
60°N -
-A°M
40°N
30° N
KON
NN
A0°N
AM
cr
‚or
40°
30°
(d) pr
S0°N gu
S0°N
40°N'
US
E 30°
60°N '
105
Z0°N
30°)
650°M
30°N
40°]
Anh
4
RI
a
FO
Figure A1. Seasonal mean biases of surface shortwave net radiation for ROAM-NBS against CERES (a), of precipitation against E-OBS (b),
of surface longwave net radiation against CERES (c), and of precipitation against IMERG (d), all averaged over 2001-2020.
is much larger than the differences between the simulations
(Fig. 6a). Especially in winter, the coupling slightly increases
the negative precipitation bias, related to the negative SST
bias of NEMO-NBS and ROAM-NBS and, with that, decreased
evaporation in ROAM-NBS (see Sect. 3.2.2), but
che effect is small. The tuning of evaporation by laminar resistances
over the ocean can increase precipitation and was
found to be beneficial for most biases over land, but it was
disregarded as it resulted in a strong overestimation of evaporation
over the ocean compared to satellite observations. In
the future, an increase of precipitation efficiency over the
ocean should be tested.
https:/doi.org/10.5194/egmd-19-543-2026
Complementing Fig. 5, the seasonal bias maps for ROAM-NBS
against E-OBS are shown in Fig. A2 for tasmin and
tasmax. They confirm that the tasmin bias is small and nonsystematic
compared to the more systematic cold bias for
tasmax. One slightly outstanding pattern in the tasmin bias
is red pattern in Scandinavia in winter (DJF). However, it 1s
unclear if this also might be over-emphasized by the comparison
against E-OBS as it does not show up that clearly when
comparing against ERA5 (not shown).
To give an insight into the temporal evolution of temperatures
within the evaluation period, time series of yearly averaged
tas are given in Fig. A3 for different countries. In
Geosci. Model Dev... 19. 543-578, 2026
