Meteorol. Z. (Contrib. Atm. Sci.)
PrePub Article. 2020
S. Schwegmann & J. Holfort: Baltic sea ice volume 1982-2019
7
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
Table 1: Correlation coefficients between ASIC/ASIV and air temperature trends for selected regions: all - the entire Baltic Sea region,
SBS - Southern Baltic Sea with latitudes from 54° N to 57.5° N, CBS - Central Baltic Sea with latitudes from 58° N to 61.5° N,
NBS - Northern Baltic Sea with latitudes from 62° N to 66° N.
January
February
March
April
May
ASIC
ASIV
ASIC
ASIV
ASIC
ASIV
ASIC
ASIV
ASIC
ASIV
all
-0.635
-0.671
0.157
0.593
0.331
0.659
-0.054
-0.205
-0.674
-0.571
SBS
-0.517
-0.408
-0.186
-0.134
-0.542
-0.332
0.119
0.151
Nan
Nan
CBS
-0.458
-0.363
-0.174
-0.095
0.288
0.362
0.301
0.329
-0.422
-0.307
NBS
-0.249
-0.542
-0.049
0.799
-0.285
0.375
-0.090
-0.581
-0.634
-0.661
In February, trends arc generally lower, and in March,
the entire Bay of Bothnia and also parts of the Gulf of
Finland and the coastal areas in the eastern Baltic Sea
south of the Gulf of Finland show interestingly a slight
decrease in SST. This negative trend has also been ob
served in Bradtke et al. (2010) using the BSH SST data
set for the period 1990 to 2008. In April, most regions
show again a positive trend. Only the coastal area in the
Bay of Bothnia still reveals decreasing SSTs. The gen
erally positive trends in SST arc well correlated with
the decrease in ASIV and ASIC. The situation is dif
ferent considering the negative trend in SST particularly
in the Bay of Bothnia in March, where neither ASIC nor
ASIV trends correlate with the decreasing SSTs. How
ever, in these regions also the correlation between SST
and ASIC and ASIV itself was low, indicating that sea
ice at the eastern coast of the Gulf of Bothnia is influ
enced by other processes, like the predominantly west
erly winds.
In contrast to the negative SST trends the air tem
perature shows a general positive trend. The highest
trends in Tail- (not shown) occur in January in the south
ern Baltic Sea and in February, when nearly the entire
Baltic Sea reveals trends of 0.2 °C to more than 0.5 °C
per decade. In March, trends arc lower but still positive.
So even if the low negative SST trends at the eastern
coast of the Gulf of Bothnia arc real, the observed trend
in ASIC and ASIV arc not inconsistent with the lower
SST as the air temperature is increasing.
Table 1 lists the correlation coefficients between the
regional distributed trends in T Lur and ASIC (ASIV) for
the months January to April. We analysed the trend cor
relation for selected regions: i) the entire Baltic Sea re
gion, ii) the Southern Baltic Sea (54° N-57.5° N), hi) the
Central Baltic Sea (58° N-61.5° N) and iv) the Northern
Baltic Sea (62° N-66° N) in order to account for differ
ent ice conditions. We would expect a negative corre
lation as increasing temperatures arc expected to force
a decreasing amount of ice. In January, changes in air
temperature can explain to a certain amount those of the
sea ice, coefficients vary between -0.25 and -0.67. In
January, usually, a large amount of sea ice forms and
therefore we would expect a good correlation in this
month. However, from February to April, correlation is
weak and occasionally also positive, so that we can ex
pect that air temperature changes arc not the only drivers
for sea ice changes in these months. In May, air temper
ature trends again correlate in a negative way with ASIC
and ASIV. May is, as January, an important month, as
during May, all the remaining ice melts away. Summa
rized: generally, the correlation between trends in T Lur
and ASIC (ASIV) is negative and the highest during the
beginning and the end of the ice season. In between,
other processes may have an additional impact on the
changes in sea ice. j^.
5 Discussio
o
ulated v:
me inste
*
values of sea ice concentration
ead of mean or maximum val-
We used accumulai
and sea ice volume
ues in order to give a good representation of ice winter
strength. Mean values arc dependent on the number of
days, but the length of the season and time of beginning
and end of the ice season can be highly variable in the
Baltic Sea, particularly in the southern areas. Further
more, ice is not necessarily present ah over the winter
but rather may appeal - and disappear frequently. Hence,
mean values would only make sense if we had chosen a
fixed interval for the winter period, e.g. October to June,
regardless when ice occurs in this time. This would have
made numbers very small in regions where ice occurs
only for few days within these nine months. We instead
considered only the period in which ice was present. Do
ing so, comparing mean values is not suitable as it might
happen that there is thick ice with high concentration
for only a few days in the season in a certain region,
which's average over the ice covered time could be as
high or even higher than the average value in another re
gion where ice was present ah over the winter. Similar
problems arise if using just the length of the season, the
number of days, maximum concentration, etc.
Giving an error estimate for the ice data is difficult
(Feistel etal., 2008). Projections, scales and require
ments on the ice charts from which the sea ice data orig
inates from have changed over time. The data used to
draw the maps also changed. In former days, charts were
based on in situ observations from the coast and recon
naissance flights with planes. Today, less in situ data but
information from various satellite data is used.
Another delimiting factor for data accuracy is the
fact that mostly discrete intervals are given in ice charts
rather than absolute values. That means, ice concen
tration is subdivided in six categories from open wa
ter (< 10 %) over very open ice (10-30 %) to very close
