1. Freshwater transformations in the Beaufort Gyre and model intercomparison results Andrey Proshutinsky, Rick Krishfield, John Toole Woods Hole Oceanographic Institution, Woods Hole, USA Mary-Louise Timmermans Yale University,Yale, USA Elena Golubeva Institute of Computational Mathematics and Mathematical Geophysics, Novosibirsk, Russia Eiji Watanabe International Arctic Research Center, Fairbanks, USA Jinlun Zhang International Arctic Research Center, Fairbanks, USA Andrey Proshutinsky, Rick Krishfield, John Toole Woods Hole Oceanographic Institution, Woods Hole, USA Mary-Louise Timmermans Yale University,Yale, USA Elena Golubeva Institute of Computational Mathematics and Mathematical Geophysics, Novosibirsk, Russia Eiji Watanabe International Arctic Research Center, Fairbanks, USA Jinlun Zhang International Arctic Research Center, Fairbanks, USA 1 14 th AOMIP workshop Woods Hole Oceanographic Institution Woods Hole, MA, USA

2. Arctic Ocean - largest freshwater reservoir 45,000 km 3 17,300 km 3 2,800 km 3 3,800 km 3 12,200 km 3 2,900 km 3 3,400 km 3 5,300 km 3 4,000 km 3 77 km 3 Aagaard and Carmack, 1989 And the oceanic Beaufort Gyre (BG) of the Canadian Basin is the largest freshwater reservoir in the Arctic Ocean (Aagaard and Carmack, 1989). Freshwater content: calculated relative to salinity 34.80 according to Aagaard K. and E. Carmack, JGR, vol. 94, C10, 14,495-14,498,1989

3. Arctic Ocean = 73,588 km^3 Canada Basin = 44,461 km^3 Freshwater content, meters: Summer 1950s

4. Arctic Ocean = 64,991 km^3 Canada Basin = 43,092 km^3 Freshwater content, meters: Winter 1950s

5. Arctic Ocean = 76,121 km^3 Canada Basin = 46,010 km^3 Freshwater content, meters: Summer 1960s

6. Arctic Ocean = 60,628 km^3 Canada Basin = 40,446 km^3 Freshwater content, meters: Winter 1960s

7. Arctic Ocean = 74,713 km^3 Canada Basin = 44,797 km^3 Freshwater content, meters: Summer 1970s

8. Arctic Ocean = 61,463 km^3 Canada Basin = 41,516 km^3 Freshwater content, meters: Winter 1970s

9. Arctic Ocean = 75,349 km^3 Canada Basin = 45,171 km^3 Freshwater content, meters: Summer 1980s

10. Arctic Ocean = 65,166 km^3 Canada Basin = 42,110 km^3 Freshwater content, meters: Winter 1980s

11. Scientific questions 1. What is the origin of the salinity minimum in the Beaufort Gyre? 2. How does this salinity or freshwater content change in time? 3. What are the driving forces of the Beaufort Gyre circulation? 4. What is the current state of the BG system? 5. What is the role of the BG system in Arctic climate change?

12. The Great Salinity Anomaly, a large, near-surface pool of fresher-than-usual water, was tracked as it traveled in the sub-polar gyre currents from 1968 to 1982. Arctic (north of 55N) air temperature anomalies relative to 1961-1990. University East Anglia data archive. (GSA’70s; Dickson et al., 1988)

13. The Great Salinity Anomaly, a large, near-surface pool of fresher-than-usual water, was tracked as it traveled in the sub-polar gyre currents from 1968 to 1982. This surface freshening of the North Atlantic coincided very well with Arctic cooling of the 1970s. At this time warm cyclone trajectories were shifted south and heat advection to the Arctic by atmosphere was shutdown. Cold air Surface Warm deep water 1000 m

14. Each component of the system stores and exchanges mass and energy differently during different climate regimes. Quantifying and describing the state and variability of these components and their coupling is essential to understand the state and fate of present day Arctic climate. Proshutinsky et al., 2002, GRL Proshutinsky et al., 2009, JGR

15. 1950 1960 1970 1980 1990 2000 Year Circulation regimes and freshwater dynamics GSA of 1970sGSA of 1980sGSA of 1990s

16. Motivation for coordinated idealized FW and circulation experiments Theory, observations and model results allow us to conclude that both thermohaline and wind-driven forcing are important to the Arctic Ocean’s dynamics and thermodynamics (hydrography and circulation). But unfortunately, the role of individual factors in the circulation and hydrographic fields cannot be easily evaluated because observed temperature and salinity distributions reflect the combined effects of wind, baroclinicity, and topographic interaction. We also know that there is insufficient observational information for clearly separating the roles of atmospheric and thermohaline forcing in the Arctic Ocean. Through numerical modeling, however, the relative strengths of the circulations and major features of hydrographic fields arising from atmospheric driving and thermohaline driving can be assessed and compared Theory, observations and model results allow us to conclude that both thermohaline and wind-driven forcing are important to the Arctic Ocean’s dynamics and thermodynamics (hydrography and circulation). But unfortunately, the role of individual factors in the circulation and hydrographic fields cannot be easily evaluated because observed temperature and salinity distributions reflect the combined effects of wind, baroclinicity, and topographic interaction. We also know that there is insufficient observational information for clearly separating the roles of atmospheric and thermohaline forcing in the Arctic Ocean. Through numerical modeling, however, the relative strengths of the circulations and major features of hydrographic fields arising from atmospheric driving and thermohaline driving can be assessed and compared

17. Goals for idealized experiments a) Understand model’s work better and improve models: Compare model results and understand sources of differences b) Separate different factors to understand their roles in the Arctic Ocean and sea ice dynamics and thermodynamics a) Understand model’s work better and improve models: Compare model results and understand sources of differences b) Separate different factors to understand their roles in the Arctic Ocean and sea ice dynamics and thermodynamics

18. “ WIND” experiment Major idea is to investigate the role of wind forcing in the processes of freshwater, circulation and hydrography Conditions: ocean is a closed domain without fluxes via ocean boundaries, no river runoff, precipitation and evaporation. There is no sea ice and only wind is a driving force Initial conditions: horizontally uniform water temperature and salinity fields with a vertical stratification corresponding to mean parameters corresponding to a) upper mixed layer, b) Pacific water layer, c) Atlantic water layer and d) deep waters. Forcing: Annual wind stresses calculated based on annual SLPs for a) 1989 and b) 2007 (AOMIP recommended algorithm) Run model for 20 years and analyze T, S, circulation, FWC and HC

19. Mechanical: Ekman pumping Anticyclonic wind Let’s consider a very simple experiment with horizontally uniform but vertically stratified ocean. If we apply anticyclonic winds with negative wind curl we will observe downwelling in the center of our basin and upwelling along continental slope and deformations of isohalines – i.e. accumulation of fresh water under the center of Arctic High and salinification of water in the “coastal” regions due to upwelling. In general the total freshwater content in the region will not change but salt will be redistributed. If wind forcing does not change in time, wind stresses (Ekman pumping) have to be balanced by horizontal pressure gradients associated with deformed salinity/density fields and friction and we will have a steady-state solution. Sea surface Density Ekman transport 400m Fresh water PUMP-I

20. Forcing Regimes Left – Anti-cyclonic; Right - Cyclonic H L H L L H L H

23. “ Thermo” experiment Major idea is to investigate a role of thermal forcing (air temperature drives variability of freshwater content and water circulation via sea ice and ocean trabnsofrmation Conditions: ocean is a closed domain without fluxes via ocean boundaries, no river runoff, precipitation and evaporation. There is no wind. Clouds are annual mean, and wind speed for calculation of heat fluxes is 5 m/s. Humidity is annual mean. Initial conditions: horizontally uniform water temperature and salinity fields with a vertical stratification corresponding to mean: a) upper mixed layer, Pacific waters, Atlantic water layer and deep waters. Forcing: Monthly air temperatures for a) 1989 and b) 2007 conditions Model were ran for 20 years with repeated monthly forcing

24. Thermodynamic: Cooling/warming Sea surface Density 400m Cooling, ice formation and salt release ICE. Mixing Freshening Fresh water PUMP-II Non-uniform seasonally and interannually changing Arctic cooling and warming accompanied by ice formation and melting results in the formation of horizontal water density and sea surface gradients and system of currents. Fresh water transformations due to this processes can be named “thermo-FW-pump”

25. Zubov’s model for calculation of ice thickness based on sum of freezing degree days: I*I+50*I=8R where I is the ice thickness (cm) and R is the number of freezing degree-days. Solid lines – ice thickness (meters) and dotted lines – annual mean air temperature Expected ice thickness after 20 years

26. Sea ice thickness (meters) in the COCO model after 20 years of thermo forcing It is warmer in 2007 then in 1989 and more Atlantic water goes into the Arctic in 1989 than in 2007

27. Sea ice thickness (meters) in the UW (left) and RAS model (right) after 20 years of 2007 thermo forcing

28. COCO model results (2007): 20 years UWmodel results (2007): 20 years Freshwater content (m) and surface circulation SAT, Ice thickness, FWC and SSH along A-B Salinity along A-B

29. COCO model results (2007): 10 years COCO model results (2007): 20 years Freshwater content (m) and surface circulation SAT, Ice thickness, FWC and SSH along A-B Salinity along A-B

30. RAS model results (1989): 20 years COCO model results (1989): 20 years Flux of Atlantic water inflow west of Spitsbergen is well developed. There is a lot of heat of AW layer in the RAS model but most of heat in COCO model is disappeared and ice is much thicker in the COCO model Note AW boundary flow west of Spitsbergen

31. Evolution of AW flow and its changes at section between Alaska and the White Sea during 10 years. After 20 years circulation became stronger but a lot of heat has disappeared due to continuation of ice growth. COCO model results (1989): 10 years COCO model results (1989): 20 years

32. UW model results (2007): 20 years (thermo) COCO model results (2007): 20 years (thermo) Flux of Atlantic water inflow west of Spitsbergen is well developed. There is a lot of heat of AW layer in the RAS model but most of heat in COCO model is disappeared and ice is much thicker in the COCO model

33. Left panels: redistribution of the oceanic fresh water due to ice formation under influence of only thermal forcing in 1989 and 2007. Note accumulation of freshwater along continental slopes and formation of a small BG FW reservoir. Right panels: Mean currents (cm/s) in the upper 200-m water layer forced by only thermal forcing in 1989 and 2007

34. Beaufort Gyre Observing System McLane Mooring profiler Upward-looking sonar Bottom Pressure recorder Acoustic Doppler Current Profiler Ice- Tethered Profiler

35. 2003 – 2010 expedition cruise tracks 2009 2008 2010

36. BG freshwater interannual variability The major cause of the large FWC in the region is the process of Ekman Pumping generated by the climatologic anticyclonic atmospheric circulation over the Canada Basin with a center in the Beaufort Gyre. In total, during 2003-2009 the Beaufort Gyre has accumulated approximately 5,300 km 3 of freshwater (from 17,000 km 3 in 2003 to 22,300 km 3 in 2009), which is 5,800 km 3 larger than in climatology of the 1970s. In 2010, based on preliminary processed data, the BG reservoir has reduced a bit and we think that the major cause of freshwater content release was the reduction of Ekman pumping due to domination of a cyclonic circulation regime over the Arctic in 2009. FW 2010: 21.8

37. Mean currents in the upper 200m layer: COCO model, 1989 thermo- experiment

38. Mean currents in the upper 200m layer: COCO model, 2007 thermo- experiment

39. Concluding remarks Recent findings show an unprecedented increase of liquid freshwater (FW) accumulated in the Arctic Ocean’s Beaufort Gyre (BG) in 2003- 2009. The release of this FW to the North Atlantic can significantly influence global climate via reduction of the ocean meridional overturning circulation. In this sense the BG as a major FW reservoir is “a ticking time bomb” for climate. Surface layer waters in the BG region in the 2000s are much fresher than in the 1970s. In total, during 2003-2009 the Beaufort Gyre has accumulated approximately 5,000 km 3 of freshwater (from 17,300 km 3 in 2003 to 22,300 km 3 in 2009), which is 5,800 km 3 larger than in climatology of the 1970s.

40. Future studies: There are several hypotheses that have to be tested based on observations and employing numerical models. 1.Beaufort Gyre reversibility: we have good observations for 2003- 2009 period when the BG region has accumulated freshwater. It is important to continue observations to observe where and how freshwater can be released and where it goes under reversed forcing. 2.Ocean mixing under conditions of sea ice melting: sea ice melt increases water stratification which has to reduce vertical mixing but sea ice growth rates should increase vertical convection. 3.Pacific and Atlantic water layer dynamics and possibility of their heat release under changing climate conditions. To solve these problems Arctic studies have to be supported by observations and new approaches for observations including new methods of observing and new instrumentation specifically designed for arctic conditions.