Mechanism for the spin-up of the MOC after Krakatoa and other late 19th century volcanoes

Before you read this, make sure you have read NoKrakatoaExperiment. All anomalies are annual mean (unless otherwise stated) ensemble mean with volcanoes minus ensemble mean without volcanoes.

The volcanoes increase the amount of aerosols in the stratosphere, which leads to increased short-wave scatter. Due to the increased availability of short-wave radiation, stratospheric ozone absorb more radiation and heat up the equatorial stratosphere as shown in fig 1a. This change in the meridional temperature gradient in the stratosphere means by geostrophic balance that there will be anomalous westerlies in the northern hemisphere, which leads to a positive AO/NAO. This mechanism is well known and can be found in the literature (Stenchikov et al (JGR, 2006, doi:10.1029/2005JD006286). The positive AO/NAO normally only persists for two winters following the eruption. Note that Alan found an NAO- in his long-term mean and so do I in the following two years annual means. Perhaps I need to look at DJF?

Fig 1. (a) Temperature anomalies on pressure levels and (b) mean sea level pressure years 2-3 after the eruption of Krakatoa. Significant anomalies are contoured.

The pressure anomalies lead to temperature anomalies over the North Atlantic, shown for years 5-6 in fig 2a. 950hPa winds are overlaid in the same plot and show how southerly winds bring warmth and northerly winds (especially in Barents Sea) bring cold. Associated with these temperature anomalies are sea-ice anomalies. There is increased ice-cover in Barents Sea and reduced ice cover in the Norwegian Sea (not shown). The decreased ice-cover in the Norwegian Sea exposes the ocean surface there to cold temperatures. This increases the surface water density and leads to increased convection as shown for year 6 March mixed layer depth in fig 2b. Convection is increased for several years, but the peak is near year 6.

Fig 2. (a) 1.5m temperature and winds for years 5-6 after eruption (note that the colour scale is not centred on zero) and (b) March mixed layer depth for year 6 after the eruption. Contours show significant anomalies.

The increased mixing leads to a cooling (see fig 3a) and freshening (not shown) of the top 1000m in the Nordic Seas by year 9-12 after the eruption. It also leads to an increased density of the flow over the Greenland-Iceland-Scotland sills, shown in fig 3b.

Fig 3. (a) Temperature in the top 1000m and (b) density at 2700m for years 9-12. Significant anomalies contoured.

By year 13-16 the subpolar gyre has become denser in the top 2000m (fig 4a). This spins up the subpolar gyre and increases the local MOC (sinking dense water increases the lower branch transport and the increased gyre the upper branch transport), see fig 4b. That the MOC is dominated by the western boundary can be seen in the meridional velocity field (not shown).
Fig4. (a) Top 2000m integrated density and (b) Atlantic MOC for years 13-16 after the eruption. Contours show significant anomalies.

So in summary, this is the proposed mechanism, from eruption to increased MOC 15 years later:

• Aerosols create a warm equatorial stratosphere within the first year.
• Through geostrophic balance an AO/NAO is induced.
• Surface temperature changes due to anomalous atmospheric circulation induce increased oceanic convection in the Norwegian Sea within 5 years.
• The density of the GIS-sill overflow increases within 10 years.
• The density of the subpolar gyre increases within 15 years.
• The last two points combine to increase the MOC locally from 40N-67N within 15 years.

There still remain many questions, like why does this not happen for Pinatubo? Has the atmospheric response to the aerosols changed? Are there any observations that could back up these mechanisms?

- LeonHermanson - 13 Feb 2009

• Figure 1: