Cold Ocean-Warm Land (COWL) Index/Time Series (Updated annually)

Last updated on 30 Apr 2022.

Note (10 Mar 2019): Starting this year, I separate the time series into Nov/Dec (ND) and Jan/Feb (JF) means instead of DJF means I did previously. The motivation are some papers of mine like this https://journals.ametsoc.org/doi/full/10.1175/BAMS-D-17-0020.1. So I should follow my own recommedation. Additionally, I found that COWL time series for ND and JF are not highly correlated (Corr Coef<0.4).

Note (30 Apr 2022): Started using JFM instead of JF.

1. Temperature-based

The ascii file cowl_ncepair925_nd_1948_2021.txt and cowl_ncepair925_jfm_1948_2022.txt contain COWL index values calculated using air temperature from the NCEP-NCAR Reanalysis. The COWL index here is defined as the area average air temperature on the 925 hPa level over land minus that over sea in the area poleward of 40N for ND or JF. In my GRADS script, there is a line like:

'cowl=aave(lnd*air,lon=-180,lon=180,lat=40,lat=90)-aave(sea*air,lon=-180,lon=180,lat=40,lat=90)',

where lnd has undefined values for sea, and sea has undefined values for land. The values given here are not normalised or 'anomalised', and have a unit of K or C.

2. Thickness-based

The ascii file cowl_ncepzthick_nd_1948_2021.txt and cowl_ncepzthick_jfm_1948_2022.txt contains COWL Index based on the 1000 to 500hPa thickness (unit = m), but otherwise similar to the calculation in 1. above.

Synopsis for DJF 2017

The Arctic has been quite warm, up to 8K averaged over the whole DJF! The warmest regions in the Arctic appear to be in the Fram Strait and Barents Sea in the Atlantic sector, and the Chukchi/Beaufort Seas in the Pacific sector; these are regions with seasonal sea ice variations. Under a negative COWL anomaly, temperature in Europe and Siberia is expected to be cold. However, temperature in Europe over the whole DJF wasn't impacted much. This is maybe because the cyclonic anomaly over northern Canada was further east than in the canonical COWL pattern, creating an effect similar to positive NAO for Europe (warming of). This latter effect happens to be consistent with the concurrent La Nina teleconnection. Further, a region in northwest Russia is even anomalously warm. I think this is really a thermodynamic effect due to the warm Barents Sea because the easterly anomaly towards the Barents Sea cannot explain the warm anomaly in this region. Lastly, there is some weak cooling in Siberia. It would be interesting to investigate what caused this particular warm Arctic winter. Is it attributable to global warming? Surface preconditioning (ice, ocean, snow) from the previous autumn? Or simply due to an above normal number of stratospheric sudden warmings that occurred?

Synopsis for winter 2020/21

This was really a winter of two halves. In southern Norway we had typical mild and wet weather in the late autumn and early winter. But then in January to February we enjoyed many weeks of dry, cold and sunny weather. All the lakes and ponds were frozen, and there was some skating and ice hokeying on them. This is not rare, but I had only experienced it another time in the past 10 years. COWL shows these two halves of the winter clearly, with near-normal value in Nov/Dec and then dips to negative anomalous COWL (ie warmer ocean and colder land condition) in Jan/Feb. Was this due to sea ice loss? Sudden stratospheric warming? Warmer Nordic seas? Random atmospheric blocking over supolar areas?

COWL References

1. Wallace JM, Y Zhang, L Bajuk, 1996, Interpretation of interdecadal trends in Northern Hemisphere surface air temperature, J. Climate, 9, 249-259.

2. Broccoli AJ, NC Lau, MJ Nath, 1998, The cold ocean-warm land pattern: Model simulation and relevance to climate change detection, J. Climate, 11, 2743-2763.

3. Molteni F, MP King, F Kucharski, D Straus, 2010, Planetary-scale variability in the northern winter and the impact of thermal land-sea contrast, Clim Dyn, 37, 151-170.