Baltic Sea Ice Climate Workshop 2005 
80 
(10 km ) 
Reconstruction of Annual Maximum Ice Extent 
in the Baltic Sea 1660-2005 
Lars Axel! and Katarina Lindquist 
Swedish Meteorological and Hydrological Institute. Sweden 
Figure 1. Annual maximum ice extent based on proxy data and ice charts. 
Data used in the reconstruction: 
•Swedish and Finnish ice charts: 
1961-2005 
•coastal observations: 1871—1990 
•Stockholm winter (DJF) air temperature: 
1756-1990 
•Uppsala winter (DJF) air temperature: 
1722-1990 
•Central England winter (DJF) air 
temperature: 1660—1990 
Consolidation method: 
•all proxy-based ice extents weighted as 
Aj = W t Aj, + W 2 A j2 + ... 
where w f is the weight corresponding to 
the nth reconstruction 
•weights calculated according to 
w- l = Y E i /E j 
j 
where E'¡ 2 are the mean-square errors 
Regression methods: 
•calibration period: 1961—1990 
•ice charts used as reference 
•regression of winter air temperature T w 
against ice charts: Aj=c 0 + c,T„ 
•c 0 and c, calculated for calibration period 
for each temperature data set 
•Aj calculated for period 1660—1960 
•regression of coastal observations s yi 
against ice charts: A t = S p a, 
where j denotes year, i denotes the 
number of the coastal station, and a i is the 
ice area associated with coastal station i 
•a t is calculated with a least-squares 
technique for the calibration period 
•Aj calculated for period 1871—1960 
Figure 2. Calculated trends over moving 30- 
year windows. Mann-KendaFs test was 
applied on test time series B J =A J -aAj_ l , where 
a is the autocorrelation for time lag 1 year 
[i.e. con-(A r A jA )\. Significant trends (95% 
C.I.) coloured red. 
SWEDISH MARITIME 
ADMINISTRATION 
http://www.sjofartsverket.se/ 
Sponsored by the Swedish Maritime Administration: www.sjofartsverket.se 
Contact person: Lars Axell e-mail: Lars.Axell@smhi.se 
http:// www.smhi.se/