Support the Arctic Sea Ice Forum and Blog

Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.

Topics - Michael Hauber

Pages: [1]
Arctic sea ice / Model projections vs observations.
« on: June 20, 2019, 03:13:07 AM »
A quick hunt for model predictions on Arctic extent finds Huang et al

Figure 3, to which I've added a circle to show that the models predict that we should currently be somewhere near 4m in September extent:

Last four years NSIDC September monthly:
2015             4.62
2016             4.53
2017             4.82
2018             4.71

Last four years JAXA minimum daily:
2015               4.26
2016               4.02
2017               4.47
2018               4.46

Obs copied from the June poll threads.

Antarctica / SH Polar Vortex
« on: November 23, 2018, 06:34:28 AM »
Also posted in weird weather thread in consequences.

Furthermore there is a remarkable string of strong low pressure systems belting across Australia and well into the Pacific.  Got some members of the weatherzone forum scratching our heads not sure what to make of it.  The pattern might have implications for ENSO, as I think its part of why a significant WWB is forecast to commence shortly.

I've been doing some hunting around to see if I can explain it as some sort of polar vortex variation.  Due to this forum I know more about the northern polar vortex than the southern which impacts on my weather.  Of course Antarctica hasn't had no real trend towards reducing sea ice, but has had a couple years of fairly low values.  Start of a similar change in SH to what has happened in NH?  Maybe some connection to changes in the Ozone hole....

Does anyone follow the SH polar vortex much and know if anything is going on that might explain part of this weird weather?

The current Arctic cyclone has been quite intense.  968 vs 963 for the GAC.  Looking at Canadian Analysis charts it covers only a fraction of the Arctic basin, compared to GAC covering the majority.  However the squeeze against the high pressure towards the Siberia Sea looks to have produced tighter wind gradients and presumably stronger maximum winds than the GAC, but over a relatively small area.

What struck me with this cyclone is the intensity of the associated warm air mass over Siberia.  After the central Arctic has generally been a little on the cooler side (compared to recent years, but still warm compared to longer term stats).  Temperature contrast plays a big role in cyclogenesis.  And while much has been said about Arctic amplification, and warm arctic cold continents, I think the situation may be reversed in summer.  During summer, particularly early summer the Arctic is still dominated by ice. This pins the surface air temperature close to 0 and the basin is pretty similar to what it was several decades ago.  The surrounding regions are getting warmer, thus we have warm continents, cold Arctic, and increasing frequency/intensity of Arctic cyclones.

And some seasons we see a transition from strong high pressure dominated weather early in the melting season to low pressure dominated weather.  2010 and 2011 really stood out as seasons that early on had severe melting weather, with some dramatic (at the time) early season stats, but then fell flat quite significantly as cyclone dominated weather took hold.  2013 was the year of the persistent arctic cyclone where the cyclones started early and just kept going.  While the surface temperature may be pinned close to 0, the atmosphere above has been getting warmer (eg 925hp temps).  But when we get a significant cyclone, mixing of the air column with the surface pinned at 0, plus clouds etc result in the relatively cool surface extending through a more significant portion of the atmosphere, thus increasing the warm continent cold arctic temperature contrast and making further cyclones more likely.

So negative feedback.  But perhaps only sometimes.  In 2012 we saw both early season cyclones that failed to establish a more persistent cyclone shield, but spread the ice and allowed intermittent high pressure weather to pump lots of heat into the mix of ice and open ocean, and then the very severe GAC.  The GAC was of course followed by massive loss of ice, although there is an argument that the ice had already been set up to melt by previous conditions and the GAC played a relatively minor role.

Science / Discussions on climate sensitivity at Ringberg
« on: May 13, 2015, 05:06:06 AM »
I thought the recent discussions on climate sensitivity at Ringberg were quite interesting.  I have summarised each one to the best of my understanding.  Note that these presentation are not peer reviewed science, but more the working hypothesis of researchers currently on the frontline, so a solid peer review may spot errors and issues with any of these claims.  I have tried to represent each presenters view as faithfully as I can and noted as a comment anything that is my opinion about these presentations:

   - feedbacks within most models become stronger over time, from about year 21 onward
   - different patterns of regional sea temperature change cause different feedbacks
   - for the 'slow pattern' equilibrium climate sensitivity is 5 degrees,
   - For the 1900-2012 observational pattern climate sensitivity is 2.3 degrees
Comment:  The slow pattern looks similar to El Nino.  The observed pattern looks more La Nina like.  The question is whether the observed pattern is due to natural variability occurring on a century timeframe to slow down warming , or whether this pattern represents a significant negative feedback.  If the pattern has been a response to Co2 it may continue and be a substantial negative feedback.  However the pattern may change over time.  In particular it could be a response to the rate of warming (eg related to the temperature difference between surface and deep ocean),  in that case the patterns would induce a negative feedback while warming is fast, but as we approach equilibrium and the rate slows the feedback may disappear, with the net result that sensitivity is still high but it takes a lot longer to reach.

Annan I
   - Compares model results with paleoclimate observations, particularly for the last glacial maximum. 
   - Earth System Sensitivity likely beteween 1.1 and 2 times Equilibrium Sensitivity, best estimate of 1.5.  Most likely Equilibrium Sensitivity is 3, Earth System Sensitivity 4.5
   - Modesl good at broad scale, poor at regional scale, no improvement from CMIP3 to CMIP5
   - Doubts that sensitivity outside CMIP range (say 1.5 to 5) could reproduce changes observed in LGM and other paleo epochs

Annan II
   - Discusses model independance and bayesian analysis. 
   - True independance would give a range of 2.9 to 3.5 for sensitivity (with a 95% confidence interval)
   - No one thinks the models are independant.
   - Although some have tried, no one has produced a useful definition of independance.

   - models show feedbacks that increase over time.  Energy balance models assume constant feedbacks. 
   - Therefore energy balance models may underestimate sensitivity should be considered a lower constraint

Armour Efficacy
   - Discusses energy balance calculations

   - Discusses 'fast adjustments'

   - discusses the usability of various sources of temperature data for estimating sensitivity

   - Finds observational evidence for an IRIS effect. 
   - But finds no negative feedback associated with this effect
   - Short wave effects from low clouds seem to oppose this effect
   - Sensitivity unlikely to be below 3 degrees

   - Discusses paleo, primarily PETM.  Sensitivity about 3K
   - higher sensitivity at warmer temperature due to tropical cloud feedbacks
Comment:  The model output looks to me like a stronger MJO, and an increased convective aggregation (IRIS effect)
   - no compelling reason to think that modern ECS is outside "canonical" 2.5-4.5 K

   - Discusses heat uptake of oceans

  - discusses energy balance modelling from paleo data
  - finds effective (transient?) sensitivity of 2.33 K

  - equilibrium sensitivity is 1.8-2.2 without considering cloud effects (comment: and excluding earth system issues such as changes in Co2 via vegetation, permafrost etc)
  - Once clouds are included it is 1.9 to 5.1

  - Based on paleo evidence equilibrium climate sensitivity is between 1.6 and 5.4

  - As the energy balance was unchanged during the hiatus there is no evidence for weaker +ve feedbacks
  - sea level rise during the "grand" hiatus 1945 - 1975 continued with no slowdown
  - Therefore the slow down was not externally forced, implying a greater role for PDO, smaller role for aerosol cooling. 
  - Comment: less aerosol cooling tends to imply a lower climate sensitivity as less current warming has been offset by this cooling effect

  - Describes cloud changes for higher sensitivity models

   - to some extent climate sensitivity can be 'tuned'
comment:  It is a favorite denier meme that the climate models can be tuned to get any result wanted.  Note that Golaz says 'to some extent'.  If they can be tuned to give any result wanted then why has no skeptical scientist tuned a climate model to produce a low sensitivity?

   - climate feedbacks increase over time.
   - therefore energy balance estimates may be low

   - Discusses diagnosis sensitivity from observations
   - good lower bound, not so good upper bound due to potential non-linearities (comment:  increasing over time is a non linearity)

   - Feedbacks change over time
   - Comment: The presentation here makes it look more like a fast feedback vs slow feedback issue

   - Energy accumulation during the hiatus has continued without slowdown, however energy balance model suggests that 'natural variability' would cause an increase in heat accumulation
   - Therefore current hiatus is internal variability not externally forced
   - Comment:  while this would be the case if natural variability involves moving heat from the ocean surface to the depths, it would not be the case if natural variability causes a change in the radiative balance.  Natural variability could affect the radiative balance through changes in clouds, or increasing warmth in snow covered areas resulting in an albedo shift.

   - Southern ocean obs suggests a 100 year cycle, modelling can produce a 400 year cycle.

   - Many observational estimates of climate sensitivity are flawed (AMO, bad priors too high aerosol forcing).  (comment: that is all the higher ones)
   - Paleo estimates are generally uncertain
   - models warm 3 times faster than obs for 1988 to 2012
   - therefore the lower estimates of climate sensitivity from obs are the only good estimates and ECs < 2.2
 Comment:  while the criticisms he makes of other estimates may be (at least in part) genuine he ignores weaknesses in the estimates that he prefers.  It is not very useful or even particularly clever in science to point out that particular methods have weaknesses or uncertainties (particularly in a field like climate science). 
 It is more interesting to point out how these weaknesses may be corrected and what happens when you do.  As an example models overestimate recent warming.  What happens if you try to correct for this by selecting model runs that match the recent slow warming?  You get almost exactly the same warming rate.

   - not all forcings have the same effect.
   - adjusting for this effect on observational studies results in an increase in ECS as calculated from energy balance methods from 1.9 to 3.1

   - General discussion on issues around climate sensitivity

   - Slower warming in SH since 1979 which is not explained by models
   - Possible mechanism is that ozone reduction has caused an increase in Southern Ocean winds.  This has increasing aerosols, leading to an increase in cloud that has caused regional cooling

   - discussion on cloud feedback issues (both +ve and -ve).  Seems incomplete.

   - some (not yet entirely convincing) reasons why Equilibrium climate sensitivity is between 2.0 and 3.5
   - cloud feedbacks are positive, but not strongly positive
   - aerosol forcing is not likely to be high
   - no strong negative feedbacks to push ECS under 2.0

Sutton Hawkins
   - Constraining CMIP models based on observations results in a 10-20% reduction in 21 century projections
   - using a different reference period for climate projections changes the result

  - Discussion on detailed modelling of convection to investigate low cloud feedbacks.  No results.

   - Discussion on a variety of mechanisms for cloud feedback

  - cloud feedbacks are not likely to be strongly negative.

Consequences / NH Snow cover loss
« on: February 02, 2015, 10:30:45 PM »
Following on from some arguments about Antarctica vs Arctic in the other thread I went on a hunt for comparison between CMIP5 modelled and observed NH snow extent and found This presentation by Environment Canada.

This shows a fairly dramatic decline in June extent, which to me looks much more significant than the drop in Arctic sea ice at minimum, both in size, and in departure from model.  However April extent is pretty closely in line with the models.

Science / Comparing modelled and observed warming rates
« on: November 24, 2014, 04:19:24 AM »
GISS warming rate from 1980 to now:  0.155 degrees/decade.
CMIP Multi-model mean warming rate for RCP8.5 from 1980 to end of 2014:  0.233 degrees/decade.

Over the last 35 years models have overestimated the warming rate by 50%.  The multi-model mean anomaly vs 1950-1980 average for 2014 is 0.89.  Monthly values for GISS for 2014 range from 0.43 to 0.78.  Despite all the 'Its breaking records even in a neutral year' rhetoric, temperatures are still cooler than the multi-model mean.

A common non-scientific response is to assume that if the warming rate is slower than modeled over 35 years it must be slower than modeled for any foreseeable future. 

Another non-scientific response is to pretend the fact does not exist, while at the same time pointing to a few recent discoveries of why climate sensitivity might be 'higher' than something else and assume this 'higher' means the multi-model mean.

The correct scientific response is to start looking for reasons why there is a difference.  If co2 causes less warming then presumably someone will find a factor that the models have overlooked, plug it into the models and say 'look the models now match reality'.  But that hasn't happened (so far).  What has happened is that various confounding factors have been found to have contributed some cooling over recent years, which should be considered when trying to make the model vs reality comparison.  Looking at the variety of different factors I sometimes wonder whether combined they could push the warming rate up by enough to get significantly above the multi-model mean.  But no one has done any such calculation that I am aware of.  And if they did, would that mean we simply conclude that Co2 causes more warming than modeled?  Or do we then search for further confounding factors that may have caused some additional warming over the last 35 years?

Consequences / Ecological disruption and human welfare
« on: November 03, 2014, 11:08:36 AM »
My personal view is that the average and most likely case for climate change is not something to be strongly worried about - from a human welfare point of view (unless we get unlucky with a climate change induced war or something).  Very different if you hold strong natural conservation ethics as I think dramatic ecological disruption is highly likely. 

Excuse me, but how does dramatic ecological disruption not impact human welfare?

I believe that we will see dramatic ecological disruption, being a mass extinction event and serious juggling of all of earth's ecosystems.  I don't think the direct impacts on human welfare are particularly serious, in particular compared to other impacts such as food security, and in high end scenarios the possibility of the atmosphere becoming too hot to support human life without air conditioning in some parts of the world on some occasions.

This is not a particularly well research opinion.  My belief in the lack of seriousness of human impacts is based largely on the fact I've never seen anyone put together a convincing case.  But I've never really seen anyone try either, so who knows?

Food security and increase in pathogens are areas I am concerned about, and it is kind of a grey area whether this is ecological disruption, or whether it is the direct affects of climate change on human welfare.  When I say ecological disruption I'm think of a flow on effect  such as where species A goes extinct, and B which normally eats A goes extinct.  Then C which is usually eaten by B undergoes a boom, and C happens to be a major pest which causes a disease in humans or eats lots of our crops.  Heat increasing the natural range of a pathogen, or drought reducing the amount of wheat we can harvest is not what I consider ecological disruption.

I definitely do not say that such an effect is impossible.  I consider it a low probability high consequence scenario that I would gladly pay a significance insurance premium in carbon taxes to avoid if it was my choice.

Consequences / Climate Shift?
« on: August 26, 2014, 12:41:44 PM »
It appears that the North Pacific has warmed up a lot:

That is a fairly big jump compared to anything else in the record, and suggests that maybe something a bit unusual is happening.  At the same time the North Atlantic is cooling a little, and now appears to be cooler than the North Pacific.

This is potentially interesting in light of a recent paper which suggests that the pause may relate to a warm North Atlantic pushing the Pacific into a La Nina like state  (link).  If this change in temperature balance between the North Pacific and North Atlantic persists, perhaps we'll see an end to the pause.  Of course perhaps a different explanation for the pause will turn out more accurate.  Or the current spike will reverse in a few months....

At the same time I note that the warm water volume for the west Pacific is at an all time August record by a significant margin.  (link).  Although the west WWV has been bigger in the past this has always been around the new year, and there is a strong seasonal component to WWV, as with most ENSO related statistics.  Furthermore the previous August records for WWV west have been set in 1989 and 2011 - just after strong La Ninas, which are associated with a build up of the western warm pool.  Whereas the current high anomaly is just after a massive Kelvin wave which has discharged a significant amount of warm water from the western warm pool.  In previous El Nino years the western warm pool had significantly shrunk by August.

Finally the PDO after several years of significant negative has gone substantially positive.  It peaked at 1.8 in May, which is higher than any time during the previous cool PDO phase (50s to mid 70s).  This spike in PDO has occurred despite and ahead of a mediocre spike in ENSO - nino 3.4 has barely made it past 0.5.  PDO often follows ENSO in the short term, but tends to by in sync with ENSO, or a little behind, and rarely leads as it has for this event.  While far too early to be confident of a sustained shift to a positive PDO, it is still interesting.

Arctic sea ice / Who's right - The models or the recent trend?
« on: July 30, 2014, 10:07:53 AM »
Its a central belief of some people that the models underestimate ice melt in the Arctic.  The recent trend towards stronger than predicted ice loss in the Arctic means a disaster and that the scientists have made a bad mistake with their models.

Its a central belief of other people that the models overestimate warming of the globe.  The recent trend towards slower than predicted warming means that everything is ok and that the scientists have made a bad mistake with their models.

Ever heard of natural variation?

Arctic sea ice / May melt ponds do not matter
« on: June 18, 2014, 11:41:35 AM »
I refer to recent paper which seems to imply that the September minimum is largely determined by the melt pond fraction in May.  However my analysis is that the strength of the relationship is an effect of overfitting the model, and that in actual fact melt pond fraction in late June is a better indicator of final minimum, and that it is not a significantly better predictor than a simple extrapolation of the long term trend.

First, the correlation between melt pond fraction and September minimum is quite impressive at over 0.8.  Hindcasts made with this method have an average error of 0.33 million km2.

However I've always been suspicious as there is usually almost no melt pond visible from satellite in the central Arctic area until early to mid June.  Figure 1 of the paper reveals that the typical melt pond fraction at the end of may is around 2% at the end of May.  Figure 2 shows that the area of the Arctic for which melt ponds are measured excludes the Hudson Bay and Sea of Okhotsk, but includes much of the Bering Strait, the Baffin bay, Greenland Sea, Barents Sea and Kara see.  At the end of May surely nearly all the melt ponds must be in such fringe areas if the total fraction is only 2%.  How can this determine the fate of the central ice in September? 

A key issue that is not immediately obvious, is the fact that the melt pond fraction is not a straightforward calculation of total melt pond area divided by total ice area.  A geographic weighting is applied with a different rating for each of 1000s of grid squares.  This raised the possibility that the strong correlation is not due to causation, but is due to overfitting the model due to having too many variables available to tweak.  In particular imagine any grid square that has melt ponds in one particular year, but not in any other year.  The weighting for this square can then be tweaked to change the model prediction for that year without affecting the prediction for other years.  Obviously there is a limit to this otherwise the correlation would be perfect instead of good.  But it may cause a stronger correlation than a purely physical causal relationship would otherwise suggest.  I suspect this may explain why there is a better correlation early on in May - there are more opportunities to find grid squares that affect only 1 or a small number of years whereas by July most grid squares would affect many or most years.

One good way to prevent such overfitting is to train the model on data for some of the years, and attempt to use the model to predict the result in other years.  This is done in the paper by making a 'forecast' for each year by using only data for years before that year.  This is effectively what would have been forecast if the method in the paper was used to actually forecast that year when the results for that year (and future years) are not yet available.

The results are much less impressive and suggest that some type of overfitting effect is at play.  The standard error for the end of May melt pond fraction prediction increases from 0.33 to 0.5.  The error for the prediction using melt pond fraction up to June 25 goes from being worse than end of May to being better (0.36 to 0.41).  A prediction error of 0.5 to my eyes looks to be no better than the average error for making a prediction based purely on extrapolating the long term trend, although I'm not motivated enough to try and download the data and perform a calculation to confirm this.

Arctic sea ice / On this day in history
« on: June 16, 2014, 11:38:18 AM »
Comparing Jaxa views  On 15th June:

2014: Melt ponding visible on ESS fast ice, and in a corner of the Beaufort.
2013: Significantly less open water in Chukchi, Beaufort and Laptev seas.  Some melt ponding from Beaufort through to ESS.
2012: Lots of melt ponding over Pacific side of Arctic basin, less open water Laptev, more Beaufort, and similar in Chukchi.  Much less ice in Kara, although as a fringe region this may have less impact.
2011: Melt ponding noticeable on ESS fast ice only.  Similar amount of open water in Laptev, Beaufort and Chukchi.  Much less ice in Kara
2010: Melt ponding on ESS fast ice, and some in ESS.  Less open water in Laptev and Chukchi, more in the Beaufort.
2009:  Melt ponding visible on ESS fast ice, and corner of Beaufort.  Less open water in Laptev, similar in Chukchi and Beaufort.
2008: Significant melt ponding in Beaufort.  Almost no open water Laptev, similar in Chukchi, major open water in Beaufort.
2007: Suprisingly little melt ponding visible in Jaxa, but massive melt ponding around this day in IUP and CT.  Less open water in Laptev, more in Chukchi, similar in Beaufort.  Dramatic trail of open water behind Wrangel Island as the ice is pushed strongly away from Pacific side towards Greenland.
Early 1980s (from CT):  Usually no open water in Laptev or Beaufort, lower ice years comparable to higher modern years in Chukchi, but generally more ice.  Dramatically more ice in Kara almost always full and Barents sometimes full.  Bafin, Hudson and East of Greenland, roughly similar to modern conditions on casual inspection, but probably more ice if I looked close enough.

Pages: [1]