Michael Hauber

Reading through the articles carefully clears the confusion. China dramatically increased aerosol pollution up to about 2006 and has decreased since then. And the final article notes that the lunar eclipses since the late 90s have been much clearer than the lunar eclipse in 1992 which was at the height of the Pinatubo eruption. The atmosphere over the last decade is dirtier now than it used to be in general, has been getting clearer again over the last few years, and has not been as dirty as it was during the volcanic eruption in 1992.

There is still year to year variability. Although PIPs is an older model than PIOMAS, and its thickness data has not been validated, I don't think it is rubbish, but rather that the data from this model is often misrepresented. Note that a quick check of PIOMAS arctic anomalies shows similar anomalies now to early 2009, and late 2008. The presentation made at iceagenow shows only the amount of ice thicker than 2.5 metres, and excludes all the thinner ice at the edge. Ice thicker than 2.5 metres is not what I would classify as 'thick', as 1st year ice can often reach this thickness. It is quite possible that the 2.5 metre thickness value and the two dates chosen were deliberately chosen to make the increase look as significant as possible. Changing the thickness threshold or the two dates selected by a few months either side will make a big difference to the result. Current weather patterns are cold (compared to the last 5 years) in the central Arctic, and warm on the edge of the Arctic. It would be no suprise to me if the ice in the central region really is thicker than 2 years ago as shown by PIPS, at the same time as the ice towards the edge of the Arctic has extent noticeably lower than any equivelant date in the last 30 years we have had satellite measurements.

Yes but this year the warmest temperatures are over the Hudson and Baffin Bay areas where ice should be rapidly forming at this time of year. In 2009 a lot of the warmth was over the high Arctic where much more warmth would be required to melt any ice at all, and over the Arctic/North Atlantic border area. In this area the temperature gradient between the cold Arctic and the relatively warm waters of the Arctic end of the gulf stream is quite steep, and a lot of warming or cooling is required to move the ice edge very far. Similar amount of warmth this year as last, just ideally located to produce low ice extent.

I think an increase from 1 event to 10 events is too small a sample size to be meaningful. Consider the possibility that someone could have trawled through hundreds of different locations and time periods to find one that just by chance offers such a comparison. The number of days with more than 25 mm was about 45, and increased to just over 60, and being a larger data set, this does look a little more like it could be a real trend, and not nearly as spectacular as the >40mm case. Of course even for those numbers, if you looked through dozen of locations with purely random variations in rainfall it would not surprise me to see some increases of that size purely from chance. I do expect that the number of higher rainfall events should have increased in the last few decades, but I'd want to see more data before I'd consider this confirmed.

Both the reduction in weight of the ice sheet pressing the rock down, and the reduction in gravity pulling the ocean towards the ice sheet are basic facts of physics that don't need to be proven/disproven. The only question is how strong the effect is. The article doesn't state how great the effect is, and a general rule of thumb I use for the media is that if they leave out an important piece of information, it is usually because that piece of information is boring (eg the effect may be only a 5% reduction and hardly worth mentioning). Either that or they just don't know how strong the effect will be.

For the last year or so it seems as if the NW Atlantic has been unsually warm and the NE Atlantic while still above average, certainly a bit cooler than around 2007. At the same maps at ESRL PSD show that there has been higher pressure than normal SE of Greenland, which has weakened the westerly belt and seems to be allowing more warm Gulf Stream water to end up near Greenland and NE America, instead of in the far north Atlantic near Svalbaard. This is the time of year when growth in sea ice in Hudson Bay, and Baffin/Newfoundland areas dominates extent change, with growth on the Pacific side limited by the narrow Bering Strait (until near maximum where it spreads into NW Pacific), and growth in Atlantic limited mostly to the coast near Europe. This seems to be a fairly persistent pattern, and unless it breaks down soon (due to La Nina?) I would imagine Arctic Ice will continue to be very to record low. At the same time diversion of warmer Atlantic waters westward may help to increase melt in Greenland, and result in less warm water entering the Arctic from the Atlantic, and slow the reduction in summer sea ice.

I personally find the second statement a lot scarier than the first. I like to think of the methane situation as a bunch of people huddled around a fire struggling for warmth (carbon consumption fuels our society). Someone then points out that the fire is built against a massive gas tank, which could explode (methane could be released abruptly and cause a disaster). A scientist works out some sums and says that with the thickness of the metal, and lack of oxygen supply to the gas tank the fire is quite safe (current calculations on methane suggest a 0.5 degree warming in next century). The people need the fire for their well-being and its not easy to just put it out. In such a situation I wouldn't feel completely safe with the fire close to the gas tank no matter how many assurances I had from scientists that it will not blow up. I think it would also be silly to run around in a panic expecting the tank to explode any second. Prudent would be to look for an alternate source of heat as soon as reasonably possible.

I'd be pretty sure that snow would be an insulator on the ice and help keep the heat in from the water below. Do you have any reason to believe that there has been more snow than normal this year? Cold temperatures could mean less snow as colder air holds less water vapour content. My impression on Arctic conditions of late is that a +ve dipole has been trapping cold air in the high Arctic, which I would assume would be helping the ice there to thicken. As the cold has been trapped in the high Arctic, the edges of the Arctic have been warmer, and so the ice extent hasn't increased as much. The pattern now seems to be changing to more of a dipole pattern with a strong blast of wind from the Pacific, which will spread the cold into the Atlantic and increase the ice extent, but also move more older/thicker ice out into the Atlantic. As the ice edge on the Pacific side is now fairly narrow as it approaches the Bering Strait the slower growth or possibly slight retreat of the ice edge in that area won't matter much. So I expect the next few days of conditions will be good for ice growth, but bad for the health of the central ice pack, kind of opposite to what we've seen in the last few weeks.

The water below has always been fluid with a large amount of heat relative to what is required to melt the thin ice above. This supplies a continuous heat source onto the ice from below. Cold temperatures above the ice let this heat escape, and the colder it is the faster this heat wants to escape. And at the same time thicker ice acts as a better insulator and slows down the rate of escape. So as surface temperature drop, the thickness of the ice increases until the thicker ice becomes a better insulator and balances out the loss of heat into the cold air. Decreasing temperatures at the surface from -10 to -20 will result in the ice getting thicker. On the flip side if the ice is at equilibrium at -20, and the temperature increases to -10, then the rate of heat escaping from the top decreases, and the heat supplied from the water from below can then melt some ice, until the insulation factor is reduced enough to balance out the warmer temperatures above, and the ice will melt from below for a temperature increase, even at temperatures well below zero. As for predictions of open water north of 80 deg, the NSIDC already tried that one in 2008. They noted that the ice near the north pole was first year ice, and prior to 2008 most first year ice melted out each summer, and predicted that the North Pole may (not will) be ice free that year. At the time a reasonable thing to speculate on. But the last few years have all seen very low amounts of multi-year ice, and significant amounts of first year ice surviving the summer melt to partly make up for that loss. I see no reason why 2011 will be any different, unless we get a repeat of 2007 type conditions, which I think would be enough to produce large amounts of open water north of 80 deg.

Does PDO affect Arctic Sea ice? During the 80s, sea ice high, no obvious drop, average PDO index +0.80 During the 90s, sea ice dropping slowly, average PDO index +0.33 During the 00s, sea ice dropping faster, average PDO index -0.10 Perhaps a cooling PDO since 1980 is the reason that sea ice is being lost faster than most models predict should be caused by Co2 alone. A possible mechanism would be changes in circulation which boost the trans-polar drift. edit: the other interesting point is to try and pick when the freeze restarts. Looking at MODIS it does look like the ice near Greenland is consolidating, and is taking on the appearance of a solid sheet with cracks from stress, instead of a bunch of floes jammed together. To me this hints that the refreeze and consolidation of the pack is starting, and minimum is imminent. However fiords on the north of Greenland and Ellesmere still seem to be most ice free and I would have guessed these would freeze up earliest as being the smallest bodies of water they would lose heat quickest....

Standard Co2 theory of feedback: Co2 increases the amount of radiation trapped in the atmosphere. This leads to changes in the climate system, that act to further modify the amount of radiation trapped/received/not reflected by earth, with a net effect that earth has a higher tendency to absorb radiation than can be explained by Co2 alone. The two most important feedback factors are an increase in water vapour (not cloud), which is a greenhouse gas, and a decrease in ice cover. Cloud cover is a third important factor and the IPCC do not know if clouds make things better or worse, and this accounts for most of the uncertainty in estimates of Co2 warming. I have read Spencer's paper and do not fully understand it, but from what I can tell, his approach is to measure the changes in radiation, and see how much temperature changes as a result of these changes in radiation. However the measurement of changes in radiation already takes into account any feedbacks that may be occuring, and so this appproach cannot measure Co2 feedback, as Co2 feedback acts to alter the radiative forcing that Spencer uses as in input into his calculations. So what is Spencer actually measuring? He himself states that he is only measuring what he calls 'feedback' on a short term basis - up to a few weeks. And on such time frames, what happens when the radiative balance of the earth changes? Some of this goes into increasing the temperature of the earth, and some is absorbed by the ocean's large heatcapacity. I believe Spencer is measuring the ocean's capability to absorb short term changes in radiation as heat content.

A lot of the ice, particularly north of Greenland shows a fairly bright white in MODIS, so would have a frozen surface, with no surface melt. Older ice and snow has a melting point of 0, and the surface will stop melting when air temperatures reach 0. However new ice in salty water needs -2 to -3 to form, and the ocean temperatures will lage the air temperatures. From watching last year's freeze on MODIS, generally sea ice does not start forming until nearby land has a quite solid cover of snow, and the fringes of Ellesmere and Greenland have only sketchy snow cover so far. The ocean needs to be a couple degrees below 0 before sea ice can form. Another indicator I'm watching is the Alert temps which I think tends to be somewhat indicative of the coldest part of the Arctic, and yesterdays range was about -1 to -4 probably quite close once a bit of lag for ocean cooling is alowed for. The final indicator I'm watching is the large area of lower concentration ice between the North Pole and Svalbaard. I'm pretty sure this is the coldest part of the Arctic with exposed ocean, and so far MODIS visual shows clear water and no sign of this water freezing over. The colder parts of the Arctic would be more between Greenland/Ellesmere Island and the North Pole. This has been a fairly solid mass for a few weeks now, which I think is due to compaction, and I've seen no sign of this solid area expanding in the last week or two.

From this document: 'To what degree the warmer climate of the Medieval Warm Period may have simplified westward progress through the Northwest Passage during the earlier phases of Greenland's Viking history remains unknown' The strongest evidence provided in this document that the Northwest passage was open is the finding of Norse Artifacts that indicate a Viking presence off the south east coast of Ellesmere Island which is in at the east entrance to the NW passage, which in recent years has opened up a lot earlier than the rest of the NW passage.

My attempt at a summary of this Arctic melt season so far: Seen was set by a record -ve dipole in winter. This caused a very strong clockwise circulation in Arctic which pushed multi-year ice into the Beaufort and Chutki seas. It also piled up first year ice against islands north of Siberia, and from thickness 'scientific guesses' such as PIPs this first year ice represented a larger and thicker area of ice than the multi-year ice north of Greenland. The same pattern imported warmer air into the Arctic and restricted overall ice growth, with low to record low extent for dates though much of winter. Late in winter the pattern broke down, and a burst of ice growth, especially in the NW Pacific saw extents quickly rise to values quite high compared to the last few years. The melt season started with a pack that was quite wide, and had a lot of thin ice, particularly outside the core Arctic Ocean area. Within the Artic Ocean the ice distribution was unusually even, instead of a more normal pattern of thin ice on the Siberia side and thicker ice on the Greenland side. Spring then saw perfect melt conditions, with strong high pressure north of Canada pumping sun into the Arctic, and sucking warm southerlies from the Pacific/Russia. With the thin ice in the fringes this melted very rapidly, and within a couple months we went from quite high to a long way below the record. In late June there was a dramatic change in pattern with a persistent low pressure dominating the Arctic. This pattern coincided fairly closely with the Russian heatwave, and I believe was part of the same large scale pattern, and contributed to the Russian heatwave by trapping cold air in the Arctic. At this time the melt was entering the Arctic basin, and in years such as 2007 the strongest melt is in the Siberian side, and a more even distribution of ice resulted in thicker first year, and also some multi-year ice in this region to help slow the melt. At the same time this low pressure resulted in a stalling of the Beaufort Gyre. Without this circulation driving ice movement, the movement became much more chaotic, and ice tended to disperse in different directions, resulting in very unusual polynyas to open up very close to the north pole. The cooler/cloudy conditions, combined with ice dispersal resulted in an extremely low rate of loss of ice extent, and extent In early August conditions improved, first with high pressure returning to the Arctic, and then a dipole pattern building to kick start ice movement towards the Fram, and import warm late summer air from further south. Melt has been moderate (compared to recent years), but not spectacularly record breaking as it was for similar patterns early in the melt season, despite what appears to be an exceptionally disperse pack, as judged by visual satellite images such as MODIS. One possible factor is that although the ice is disperse, it may still be thicker than we might otherwise expect, and that a remnant of multi-year ice spread along the Beaufort/Chutki sea region is helping protect thinner first year ice behind it. Another is that the cold period in July resulted in less build up of ocean heat, which is the dominant cause of melt later in the season. Now its time to see how much of the tongue of ice towards Siberia melts out in coming weeks. It was thick to start with, but is starting to break down. I don't think there is time for it to disspear completely, but it could contribute to steady late season melt. This area is also further from the pole than the rest of the pack, so will see a later freezing point.

Do you know of any evidence that the halocline in the Arctic is actually eroding? If the Souterhn Ocean works with a 'seasonal' Halocline layer, can the Arctic work the same way? The southern ocean is much more exposed to giant swells from the stormy weather over a wide open ocean. I think we lost enough of the old Perenial by 2007 that losses since then don't really matter much. I don't think the loss of multi-year ice is going to cause anything much more than what we have already seen since 2007. If the reduction in multi-year ice was helping to increase the overall rate of ice loss then this period is over, and the rate of loss may now slow. However looking at PIOMAS and extrapolating the trend, we could reach an almost zero ice volume any year now; if June's loss's had been repeated in July and a little extra through August, we would have been basically at zero. So it will be interesting to see whether cryosat confirms this spectacular loss of volume. I'm half expecting that the satellite will show that we've only lost half as much volume as PIOMAS shows, which will be an opportunity for propogandists to say 'look the models are wrong yet again, and we aren't going to lose all the ice in 10 years'. And avoid any mention of the fact that we are going to have to wait 20 years instead to lose all the ice. One thing I have noticed this season is that some of the thickest ice at the end of winter was near Siberia, where the big -ve AO of last winter piled the first year ice up against the islands in that region. As far as I can make out that ice has drifted a little towards the pole, is thicker than ice at the pole, or the multi-year remnants near Greenland, and is set to become 2nd year ice next year. I think the thicker first year sea ice in this region has been far more relevant to holding up this years melt than any of the multi-year ice.

This was the image I had in mind, which is a trend from 1980 to 2009. The fact that your images show 2000-2009 similar or a little warmer than 1980-1989 suggests to me that the cooling trend is not all that consistent and that 'no definite trend' may be the fairest assessment for the southern ocean closest to the ice, for the period 1980 to 2009.

As I mentioned previously a cooling trend is visible in the southern most regions of the ocean in GISS data from 1980 to 2010, as can be seen from their map generating service at http://data.giss.nasa.gov/gistemp/maps/ However with a 250km smoothing applied, it can be seen that the measurements are all (or mostly?) outside the sea ice area. The paper you linked to is based on NCEP reanalysis data and shows a warming over the sea ice area, and mixed warming and cooling outside. This paper also shows a trend up to 2004, and if GISS is truncated at 2004 , then the cooling signal for the southern ocean (outside the ice area) is more ambiguous and looks reasonably similar to what the NCEP reanalysis in your paper shows. So why would the ocean outside the sea ice be cooling, and the area of sea ice be warming? Is the ocean cooling and air warming? GISS over Antarctica land area over this period has mixed cooling and warming. Perhaps the reanalysis data over the sea ice is wrong? Or maybe the GISS data showing 30 years of cooling in the southern ocean is wrong. Or perhaps a change in ocean/wind/precipitation processes is causing an increase in sea ice, which releases extra heat and this is a source of warming limited to the sea ice area. Or some other process that can cause warming limited to near the Atlantic continent?

A quick check of GISS shows that in general the Southern Ocean is cooling over the last 30 years, which corresponds to the period over which we are measuring an increasing sea ice trend. However a GISS trend for 100 years does show some of the strongest ocean warming for the entire globe in the Southern Ocean. I have also read somewhere that reports from whaling ships in the early 20th century suggest that the edge of the Antarctic Sea Ice used to be a lot further north. However I have also read that recent thinning of Antarctic Ice shelves is due to warm water from below, so I'm not sure how these facts fit together. Perhaps the ozone hole has tightened the polar vortex and cooled the air down a lot, and this is cooling the surface, but the subsurface water is warmer?