Dive In: Geoengineering the Southern Ocean

This is the fourth post in the Dive Into Science series.  Here I’ll be explaining results from recent scientific papers.  Dive Into Science gives you a glimpse of current research in an easy to read format that anyone can understand.  To read more, just use the Dive Into Science tag.

Today’s article is “How deep is deep enough? Ocean iron fertilization and carbon sequestration in the Southern Ocean” by Josie Robinson, et al, published in Geophysical Research Letters, 2014.


One of the main factors influencing climate change is the amount of greenhouse gases, specifically carbon dioxide, in the atmosphere.  I’m not going to go into how greenhouse gases increase global temperatures, or what the sources are.  Instead, let’s jump ahead to: what can we do about it?

One major option is that we can store a lot of it in the deep ocean.  This carbon sequestration keeps the carbon dioxide in the ocean for maybe hundreds of years, giving us more time to cope with climate change.  How do we increase the amount of carbon sequestration?  By geo-engineering the ocean.

When phytoplankton bloom, they take up sunlight and carbon dioxide.  Some of the phytoplankton are eaten, and the nutrients and carbon dioxide they take up are eventually returned to the surface ocean.  Some of the phytoplankton sink out of the surface ocean and are exported to deep waters, maybe even to the bottom, storing carbon dioxide and nutrients deep in the ocean for long periods of time.

To geo-engineer the ocean to increase carbon sequestration, we would need to stimulate phytoplankton to bloom more, over longer time periods, and ensure a large portion of that bloom sinks.  The Intergovernmental Panel on Climate Change (IPCC) has guidelines on carbon sequestration.  They suggest, that to be truly sequestered, the carbon needs to stay at least 1000 meters deep in the ocean for 100 years.

The prime location to geo-engineer is the Southern Ocean.  A large portion of it is iron-limited, which means we just need to add iron (called iron fertilization) to get more phytoplankton to bloom.  It also has certain locations where dense water sinks to the bottom of the ocean and stays there for hundreds of years.  Sounds promising, right?

This premise is what the authors of this paper tested.  Suppose we geo-engineer the Southern Ocean.  We add enough iron to the Southern Ocean at the right time, stimulating a massive bloom, which sinks to 1000 meters.  Everything goes right.  What happens to the carbon next?  Does it stay below the 1000 meter limit for 100 years?  Or do we need new guidelines for carbon sequestration?

Using an ocean model, the authors of the paper tested this theory.  They put in almost 25,000 tracer particles at 1000 meters all throughout the Southern Ocean.  Then they ran the model for 100 years to see where the particles ended up.  Particles that make it above the Mixed Layer Depth (MLD), or the depth to which surface forcing, such as wind, can mix the water, are considered to be exposed to the atmosphere and aren’t sequestered.

Figure from Robinson et al.  Panel a shows the starting location of all the particles that made it to the surface, and the color indicates how long it took them to upwell.   Panel b shows the percentage of particles in each block that stayed sequested for the entire 100 year simulation.

They find that after 100 years, 66% of the particles have been exposed to the atmosphere.  That means less than half of them that met the original standard actually stayed sequestered for 100 years.  On average, it took a particle 37.8 years to make it back to the surface.  However, if the particle got out of the Southern Ocean, it tended to stay in the deep ocean.  Of the ones that reached the surface, 97% were still in the Southern Ocean.

So what does this mean for carbon sequestration and iron fertilization?  Even if we manage to get all the geo-engineering aspects of iron fertilization right, it doesn’t mean that the carbon will stay down long enough for it to be useful.

From Robinson et al. Shows the percentage of particles that stayed sequestered at different time points over the model run, based on whether they started at 1000 or 2000 meters deep.

The authors also re-ran the same simulation, but this time set all the particles at 2000 meters, instead of 1000.  By requiring a deeper depth in order to be “sequestered”, they found that only 29% made it back to the surface ocean over 100 years, a large improvement.

Overall, this paper demonstrates the issues with geo-engineering the Southern Ocean.  It suggests that new guidelines are needed to define carbon sequestration – it must initially sink to a depth of 2000 meters instead of 1000 meters.  It also demonstrates that the Southern Ocean might not be the best place for geo-engineering, as the dynamics of the water cause a large portion of particles to rise to the surface, rather than keep sinking.

Dive In: Global sea ice cycles and trends

This is the third post in a new series called Dive Into Science.  Here I’ll be explaining results from recent scientific papers.  Dive Into Science gives you a glimpse of current research in an easy to read format that anyone can understand.  To read more, just use the Dive Into Science tag.

Today’s article is “Global sea ice coverage from satellite data: Annual cycle and 35-yr trends” by Claire Parkinson, published in Journal of Climate, 2014.

Sea ice is one of the hot topics in climate science right now, with a particular emphasis on Arctic sea ice.  In recent years, Arctic sea ice has been decreasing, both in how much area the ice covers, and in how thick the ice is.  This has serious implications on everything from climate change to global trade to animal welfare.

However, this is only one half of the issue.  On the opposite end of the earth, sea ice in Antarctica is increasing.  And, given that the average amount of sea ice at each pole is similar in magnitude, many people think (wrongly) that the two poles cancel each other out.  And so, instead of focusing on one pole, the author of this paper looked at both poles combined to calculate the global trend in sea ice and set the record straight.  (Spoiler: It’s still decreasing overall.)

Now, let’s back up for a minute and talk about exactly what I mean by “sea ice”.  The term itself is rather obvious – ice that is on the ocean.  This does not count ice that is attached to ice on land – no glaciers, ice shelves, or ice sheets here.  When sea ice is measured, a satellite measures the amount of ice in a particular area, essentially one square in a grid that covers the earth.  This measurement is returned as a percentage – so the percentage of water that is covered by ice in each particular box.  To get the total amount of ice, we say that each box that is more than 15% ice is ice covered.  Then we add up the areas of all the ice covered boxes.  15% may seem like a random threshold, but it is widely used in the scientific community as a standard.

In general, sea ice in the Arctic is very different from the Antarctic.  There are two main characteristics we can use to compare sea ice at each pole (indeed, any sort of environmental data).  When we measure sea ice, we can separate the data into two parts – a seasonal cycle, and a trend.  The seasonal cycle is just what it sounds like.  As it gets warm in summer, the ice decreases, and then increases again during winter.  Almost everything in the environment has a seasonal cycle.  We can calculate the seasonal cycle, and subtract it from the data.  What we are left with is the trend, or how the amount of sea ice changes from year to year.  In terms of magnitude, the seasonal cycle is typically much larger than the trend, so it is hard to see a trend without removing the seasonal cycle.

Sea ice seasonal cycles. The seasonal cycle is the same for every year data was taken.  Note that while both poles seems to have a similar amount of sea ice, the extremes in the Antarctic are larger than those in the Arctic (compare the maxes and then the mins).  Thus, the global seasonal cycle hits a minimum the same time as the Antarctic.  Figure 2 from Parkinson, 2014.

In the Antarctic, the seasonal cycle is large in magnitude.  In the (austral) summer, almost all of the ice melts away, but comes back again the following winter.  In terms of trend, different regions of the Antarctic have different trends.  If you average them out, the overall trend is slightly positive = increasing amounts of sea ice over time (years).

In the Arctic, the seasonal cycle is smaller in magnitude.  It is opposite sign of the Antarctic, because the seasons are switched going from southern to northern hemisphere.  Ice tends to grow slightly in winter, but sticks around during summer.  A good portion of the ice in the Arctic is multi-year ice – it has been around for several years.  The trend here is decidedly negative = decreasing amounts of sea ice over time.

Sea ice trend.  This is the part that is left over after removing the global seasonal cycle from the data.  The data is still noisy – not all points fall exactly on the trend line – but it is easy to see that the global trend is clearly negative.  (A statistical analysis shows that it is significant as well.)  Taken from Figure 3 of Parkinson, 2014.

Now, if we add the sea ice information from both poles together, we get an idea of what sea ice is doing globally.  Looking first at the seasonal cycle, the global cycle follows the same basic pattern as the Antarctic.  This makes sense.  The Antarctic cycle has more extremes than the Arctic one, and this dominates when you combine the data.  But, when you look at the global trend, it follows the negative pattern of the Arctic trend.  Since the Arctic is losing sea ice much faster than the Antarctic is gaining sea ice, it follows that globally sea ice is decreasing.

Thus, with a relatively simple (but necessary) analysis, the author showed that the negative sea ice trend in the Arctic is not canceled out by the positive sea ice trend in the Antarctic.  Overall, from a climate perspective, global sea ice is decreasing – a solid indicator of climate change.

You may have noticed the title of the paper specifically says a 35-yr record.  That’s all the longer we have the satellite data to make these sorts of calculations.  It is entirely possible that the 35 year trends found here are not trends at all, but small sections of a much longer cycle.  This is why climate scientists use other sources of data and climate models to help confirm findings from the satellite record.

Dive In: Oceans melting ice shelves

This is the second post in a new series called Dive Into Science.  Here I’ll be explaining results from recent scientific papers.  Dive Into Science gives you a glimpse of current research in an easy to read format that anyone can understand.  To read more, just use the Dive Into Science tag.

Today’s article is “Ocean variability contributing to basal melt rate near the ice front of Ross Ice Shelf, Antarctica” by Isabella B. Arzeno et al, published in Journal of Geophysical Research: Oceans, 2014.

If you think about sea level rise and future climate predictions, and all that, one of the biggest unknowns scientists are trying to understand is the rate at which the ice sheets are melting.  See, it’s not exactly a straight-forward process to figure this out.  You would think that if you knew the temperature of the air, you could calculate how fast the ice melts, and how much melted water goes into the ocean.  But, you’d be wrong.

When we talk about melting ice sheets, we really mean the change in the ice sheet mass balance.  This is exactly what it sounds like.  To figure out how much ice is lost or gained from the ice sheet, we balance out all the sources and sinks.  In Antarctica, the major source of ice to the ice sheet comes in the form of snow over the continent.  This snow eventually compacts to form ice, and adds mass to the ice sheet.

Now, there are several ways an ice sheet can lose mass.  I’ve already mentioned melting ice from the top of the ice sheet – this does not have a large effect (true for Antarctica, not quite true for Greenland).  It is rarely warm enough to melt significant amounts of ice, and the water must also evaporate, or it will just refreeze at a later time.  More important are the outlets of the ice sheet, or where the ice meets the ocean.  Here, ice loss can occur through pieces of ice falling off as icebergs, or through the ocean melting the ice it touches.

Cartoon illustrating ice sheet mass balance.  Here, we’re concerned with the Antarctica side.  Image by Hugo Ahlenius, UNEP/GRID-Arendal

This brings us to the topic of ice shelves.  Ice shelves are the part of the ice sheet that floats in the ocean.  Trick question: How much would sea level change if all the ice shelves melted?  Answer: No change, because they are floating.  BUT, it’s even more of a trick question!  Even though the ice melted from the ice shelf doesn’t contribute to sea level rise, it causes more ice to flow in from the ice sheet and increases the ice loss term of our mass balance.  And, if the balance shows that ice is being lost, well, that does contribute to sea level rise.

So these ice shelves then, act as plugs to the much larger ice sheet.  We need to understand how they melt, by how much, and how that affects the ice sheet.  In a best case scenario, the ice shelves melt a little bit, lose some icebergs off the edges, and are replaced by ice from the ice sheet to form a stable configuration.  In a worst case scenario, the ice shelves melt enough that they no longer act as stable plugs, and the ice sheet becomes unstable and eventually drains out through the open hole.  Even for the worst case, the process would take several hundred years for all the ice in Antarctica to drain out.  But, what is happening now with the ice shelves may determine the eventual fate of the ice sheet.  And, wouldn’t it be nice to know what that is?

Figure taken from Arzeno, et al. 2014.  Data from the mooring under the iceshelf.  A) is current speed, b) is temperature difference from freezing, c) is calculated melt rate from temp, salinity, and current, d) is melt rate from tide velocities only, and e) is the tide velocities.  Notice how a higher temperature leads to a higher melt rate (b vs. c), and how peaks in the tide-only melt rate correspond to high tidal velocities (d vs. e).

This paper is a significant contribution to our understanding of one ice shelf in particular, in the Ross Sea.  (All the ice shelves behave differently, based on size, shape, what water they are in contact with, etc…)  It is the first time scientists have been able to simultaneously collect data on temperature, salinity, and currents under the Ross Ice Shelf.  These three variables are what you need to directly predict ice shelf melt.  The temperature tells you how much heat is available for use in the ocean water, the salinity helps determine what the temperature of ice melt actually is, and the current speeds give you an idea of how long it takes to replace this water.  As the ice melts, the underlying water grows colder and fresher, and the ice melt slows down.  In order to keep melting rates up, the water needs to be mixed around or replaced.

The authors use results from their mooring underneath the ice shelf, and from ocean models to better understand how the melt rate of the ice changes.  They find that tides play a large role, both in the data, and in the ocean model.  As the measurements were taken near the edge of the ice shelf, the movement of the tides back and forth under the ice shelf helps bring the cold melt water out, and replace it with warmer ocean water.  The authors also notice longer time signals in the currents that affect the melt rate, but don’t have enough information to identify them.

Overall, this paper presents rare data on ice shelf melt rates for the Ross Sea, and also investigates the processes that contribute to the melt rate.  This information, along with research from other groups, can be used to better understand, and eventually predict, the fate of the ice shelves and how that affects the Antarctic Ice Sheet.