146 
JJ. Haapala et al. 
changes in sea ice drift. Apart from these basic physical 
variables, the length of the ice season and dates of freezing 
and break-up of the ice cover have also been widely used as 
climatological indicators of ice season severity. 
The Baltic Sea ice season usually lasts for six to eight 
months. Typically, coastal shallow areas in the Bothnian 
Bay become ice-covered in October-November. Dining the 
next phase, the ice cover advances towards the central parts 
of the Bothnian Sea, and coastal areas of the other Baltic Sea 
sub-basins also become ice-covered. On average, the ice- 
covered area is greatest in March, when the maximum ice 
extent in the Baltic Sea (MIB) is about 165,000 km 2 , or 
40 % of the total area of the Baltic Sea. The limit of the ice is 
typically located in the northern Baltic Proper, which means 
the Gulf of Bothnia, Gulf of Finland and Gulf of Riga are all 
completely ice-covered. In extremely severe winters, the 
entire Baltic Sea has been ice-covered (420,000 km 2 ). 
During the mildest winters, sea ice is formed only in the 
Bothnian Bay and the eastern Gulf of Finland. The minimum 
MIB to date is 49,000 km 2 . The length of the ice season is 
130-200 days in the Bothnian Bay, 80-100 days in the Gulf 
of Finland and 0-60 days in the southern Baltic Sea. 
Near the shore, sea ice remains stationary, and its thick 
ness is controlled by thermodynamic factors only. That ice is 
termed ‘fast ice’. In those regions, horizontal ice thickness 
variations are small, typically 10-20 cm within a scale of 
kilometres. Most of the monitoring data, in particular for ice 
thickness, have been collected from the fast-ice regions. The 
record for measured fast-ice thickness is 1.22 m from Tor- 
nio, Finland. 
Ice thickness in the drift ice regions is much more vari 
able. Drift ice is a mixture of different ice types, and within a 
scale of kilometres, ice thickness can vary from a few cen 
timetres of thin new ice up to thick pressure ridges. Ice 
ridges, formed by compressive ice motion, are typically 3- 
5 m thick, but freely floating ridges 25 m thick have been 
observed in the Baltic Sea. Ice thickness measurements in 
drift ice areas have been made during several field cam 
paigns, but the observations are too sparse in time and space 
to examine long-term change. 
The state of sea ice depends on the surface energy bal 
ance, momentum flux and water flux. Some of the climate- 
related state variables, such as ice extent or freezing date, are 
mainly driven by the energy balance, and the observed 
variation in these variables very closely reflects large-scale 
variation in air temperature. However, many variables, such 
as ice thickness, ice type and duration of the ice-covered 
period, are also very dependent on winds and ocean currents 
and may reflect changes occurring at a local scale: this 
highlights the importance of a unified analysis of ice 
parameters for quantifying change in sea ice on a climato 
logical scale. An alternative approach for analysing directly 
measured geophysical parameters is to use a particular ‘sea 
ice index’ as a general indicator of ice conditions. For the 
Baltic Sea, two methods have been used. Koslowski and 
Glaser (1995) used an index that describes the mass of sea 
ice by integrating ice thickness over the sea area in question, 
while Sztobryn et al. (2009) developed an index based on the 
normalised length of the ice-covered season of several 
coastal stations. In the Baltic Sea, all these variables display 
a large interannual variability and are very closely related to 
the variability in the large-scale atmospheric circulation 
(Omstedt and Chen 2001; Jaagus 2006; Vihma and Haapala 
2009, see also Chap. 4, Sect. 4.2). 
Recent years have highlighted the magnitude of the 
interannual variability of ice conditions in the Baltic. The 
winters of 2008 and 2009 were very mild; sea ice formed 
only in the Bothnian Bay and coastal areas of the Bothnian 
Sea and Guff of Finland. The MIB was only 49,000 km 2 in 
2008 and 110,000 km 2 in 2009. The winters of 2010 and 
2011, on the other hand, are classed as severe ice winters, 
and the MIBs were correspondingly larger: 244,000 km 2 
(2010) and 309,000 km 2 (2011). More importantly, the ice 
was thick, and severe storms induced pack ice compression 
events that caused major difficulties for winter navigation. In 
the past, there have been similar relatively short periods with 
both extremes of ice conditions, such as 1873-1876, 1938— 
1941 and 1985-1990. It is also evident that there is no 
correlation between consecutive ice seasons, simply because 
the thermal memory of the Baltic Sea is only 2-3 months 
(Leppâranta and Myrberg 2009). 
The first Baltic Sea assessment (BACC Author Team 
2008) concluded that climate warming can be detected from 
the time series of MIB and ice season length. This conclu 
sion was based on studies utilising observations that exten 
ded only until 2000. The present review is an updated 
analysis and includes observations up to 2011. Observed 
trends in sea ice parameters are summarised in Table 8.1. 
8.2 Ice Extent 
Annual MIB is the most widely used indicator of sea ice 
changes because it integrates the winter period weather over 
the entire Baltic Sea basin (Fig. 8.1). The time series is based 
on various observation methods. The early part of the time 
series has considerable uncertainty. It is mainly based on 
correlation between air temperature recordings, observations 
from lighthouses, information from newspapers and records 
of travels on ice (Jurva 1937; Palosuo 1953; Seina and 
Palosuo 1996). According to Jurva (1952), MIB has been 
fairly accurately determined from 1880 onwards from notes 
made onboard ships navigating in the middle parts of the 
Baltic Sea during winter (Vihma and Haapala 2009).