Abstract

The Baltic Sea is one of the fastest-warming semi-enclosed seas in the world over the last decades, yielding critical consequences on physical and biogeochemical conditions and on marine ecosystems. Although long-term trends in sea surface temperature (SST) have long been attributed to trends in air temperature, there are however, strong seasonal and sub-basin scale heterogeneities of similar magnitude than the average trend which are not fully explained. Here, using reconstructed atmospheric forcing fields for the period 1850–2008, oceanic climate simulations were performed and analyzed to identify areas of homogenous SST trends using spatial clustering. Our results show that the Baltic Sea can be divided into five different areas of homogeneous SST trends: the Bothnian Bay, the Bothnian Sea, the eastern and western Baltic proper, and the southwestern Baltic Sea. A classification tree and sensitivity experiments were carried out to analyze the main drivers behind the trends. While ice cover explains the seasonal north/south warming contrast, the changes in surface winds and air-sea temperature anomalies (along with changes in upwelling frequencies and heat fluxes) explain the SST trends differences between the sub-basins of the southern part of the Baltic Sea. To investigate future warming trends climate simulations were performed for the period 1976–2099 using two RCP scenarios. It was found that the seasonal north/south gradient of SST trends should be reduced in the future due to the vanishing of sea ice, while changes in the frequency of upwelling and heat fluxes explained the lower future east/west gradient of SST trend in fall. Finally, an ensemble of 48 climate change simulations has revealed that for a given RCP scenario the atmospheric forcing is the main source of uncertainty. Our results are useful to better understand the historical and future changes of SST in the Baltic Sea, but also in terms of marine ecosystem and public management, and could thus be used for planning sustainable coastal development.

Highlights

  • The Baltic Sea exhibits outstanding sea surface temperature (SST) changes over the last decades with for instance an increase of 1.35 °C in 1982–2006, corresponding to seven times the global rate (Belkin et al 2009)

  • In agreement with the Kniebusch et al (2019)’s assumption, our results show that differences in SST trends between the northern and southern Baltic Sea are explained by summer and spring ice cover

  • Several other sources of variability may modify our results for the future period. To assess their effect on SST trends we used a pool of climate change simulations performed for four atmospheric forcings in three sea level rise scenarios and two nutrient load scenarios, i.e. twenty four climate change simulations for each Representative Concentration Pathways (RCPs) scenarios

Read more

Summary

Introduction

The Baltic Sea exhibits outstanding SST changes over the last decades with for instance an increase of 1.35 °C in 1982–2006, corresponding to seven times the global rate (Belkin et al 2009). Results demonstrated that seasonal SST changes are very heterogeneous among the sub-basins of the Baltic Sea with for instance twice as much warming in the Baltic proper than in the Bothnian Sea in winter, and the opposite in summer (Kniebusch et al 2019) These differences are partly related to the heterogenous climatic and geomorphological characteristics of the Baltic Sea. The exchange of water masses with the open ocean is constrained by the narrow and shallow Danish straits. The Baltic Sea has a low average depth (~ 54 m, Fig. 1) and a strongly variable bathymetry that further limits the exchange between the Baltic Sea sub-basins (Fig. 1) This is reflected in long surface water renewal rates of around 30 years in the southern part and 40 years in the northern part (Meier 2007; Snoeijs-Leijonmalm and Andrén, 2017). Its water masses can be understood as a mixture of saline water from the North Sea and freshwater due to the presence of a very large number of rivers on its shores

Objectives
Methods
Findings
Conclusion
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call