Guest Post by Willis Eschenbach
Albedo is the percentage of incident light that is reflected by an object. For years, I’ve read claims that the loss of Arctic sea ice is a positive feedback. It is logical—warming leads to less ice, less ice reduces the surface albedo; reduced surface albedo means more sunlight is absorbed; more sunlight absorbed leads to increased warming. Positive feedback. What’s not to like?
For example, in 2019 the IPCC said:
Feedbacks from the loss of summer sea ice and spring snow cover on land have contributed to amplified warming in the Arctic (high confidence).
Wim Rost pointed me to an interesting 2007 NASA article about Arctic albedo which says:
Although sea ice and snow cover had noticeably declined in the Arctic from 2000 to 2004, there had been no detectable change in the albedo measured at the top of the atmosphere: the proportion of light the Arctic reflected hadn’t changed. In other words, the ice albedo feedback that most climate models predict will ultimately amplify global warming apparently hadn’t yet kicked in.
Kato quickly understood why: not only is the Arctic’s average cloud fraction on summer days large enough—on average 0.8, or 80 percent—to mask sea ice changes, but an increase in cloudiness between 2000 and 2004 further hid any impact that sea ice and snow losses might have had on the Arctic’s ability to reflect incoming light. According to the MODIS observations, cloud fraction had increased at a rate of 0.65 percent per year between 2000 and 2004. If the trend continues, it will amount to a relative increase of about 6.5 percent per decade. At least during this short time period, says Kato, increased cloudiness in the Arctic appears to have offset the expected decline in albedo from melting sea ice and snow.
Wim suggested that I take a look to see if this process, of the changes in cloud albedo counteracting the changes in surface albedo, had continued up to the present.
Fortunately, the CERES data allows us to calculate the trends in both the surface albedo and the top-of-atmosphere (TOA) albedo. First, here’s the trend in surface albedo in percent per year, on a 1° latitude by 1° longitude basis.
Figure 1. Atlantic and Pacific centered views of the trend in surface albedo, in percent per year. Seasonal variations removed.
As expected, due to the reduction in Arctic sea ice, the albedo in the Arctic has indeed decreased significantly over the 21-year period. It’s decreased at a rate of 0.28% per year, a total of almost 6% over the 21 year period. Note also that the poles are the only part of the surface with a significant trend.
Next, here’s the top-of-atmosphere (TOA) albedo trend.
Figure 2. Atlantic and Pacific centered views of the trend in TOA albedo, in percent per year. Seasonal variations removed.
Amazing. The increase in cloud albedo has almost totally counteracted the decrease in Arctic surface albedo. The change is only six-hundredths of a percent per year, basically lost in the noise. The effect of the clouds has brought the polar regions back into line with the rest of the planet.
This inspired me to look at the correlation of the surface albedo and the cloud albedo over the period. Positive correlation of two variables means generally that when one increases, so does the other. Negative correlation means that they move in opposite directions. Figure 3 shows that result.
Figure 3. Correlation, surface albedo and cloud albedo.
This is also most interesting. It shows that the cloud albedo not only counteracts the sea ice albedo changes. It also counteracts the changes in surface albedo from snow and land ice. Not only that, but in the area of the sea ice, the correlation is around -1, meaning that surface albedo and cloud albedo move in nearly total opposition..
Examining Figure 3, it is obvious that over the land the correlation is negative almost everywhere. However, over the ocean, the correlation is clearly related to the temperature. As the Figure 4 scatterplot below shows, wherever the ocean is below about 22°C, the clouds tend to oppose any change in surface albedo.
Figure 4. Scatterplot showing the correlation of cloud and surface albedo trends versus surface temperature. Data is the gridcell-by-gridcell 21-year average values. Yellow/black line is a LOWESS smooth of the data.
Again, in the sea ice area where 21-year average temperatures are around zero, the negative correlation is almost perfect.
Discussion
With those results in mind, let me return to the 2019 IPCC claim:
Feedbacks from the loss of summer sea ice and spring snow cover on land have contributed to amplified warming in the Arctic (high confidence).
Note that despite the IPCC claim of “high confidence”, the 2007 findings of Kato and the more recent CERES data shown above demonstrate that feedback from changes in sea ice and snow cover have NOT contributed in any significant way to amplified warming in the Arctic. Cloud changes offset these sea ice and snow changes almost entirely. In short, the IPCC claim is overstated.
This highlights the problem with the claim that we should all listen to the “97% consensus” … it’s meaningless. Science is the process of overthrowing the consensus.
My best to all on a lovely fall day,
w.
PS—As usual, I ask that you quote the exact words that you are discussing. For the reasons why, see here.
Planetary albedo is essentially atmospheric (cloud) albedo, with a small contribution from Antarctica. This is well known by anybody with an interest in albedo, and by climate models since at least CMIP3.
From Donohoe & Battisti (2011):
Comparing planetary albedo (b) with the atmospheric contribution to albedo (d) shows that they are almost the same.
The ice-albedo feedback was always a lame duck even to other climate scientists. There was never evidence that it played a significant role.
Sea-ice goes opposite to insolation and that works against such role, and cloud cover in the Arctic in summer is c. 80-90%, and c. 100% above open ocean areas. That’s the reason of the strong anticorrelation. The moment there is open ocean evaporation condenses and forms clouds. The atmosphere is quite cold even during the summer at the Arctic, so it holds very little humidity before condensing.
I’ve discussed that silly argument for years.
You are far too sensible to be a “climate scientist”..Good that you took up life science, where, outside of medical journals, the scientific method is still operating, although endangered by the corrosive effect of antiscientific “climate science”.
Clouds forming over open polar waters:
Frigid air blowing from Eastern Russia created dramatic cloud formations over the Sea of Okhotsk in late November, 2017. The Moderate Resolution Imaging Spectroradiometer (MODIS) aboard NASA’s Terra satellite acquired a true-color image of the stunning scene on November 25.
Snow covers the land of Eastern Russia in the west of this image, with a large bank of cloud overlying the land in the northwest. Long parallel rows of cumulus clouds blow off the snow-covered area and over the blue waters of the Sea of Okhotsk. These rows of cloud, known as “cloud streets” form as cold, dry air from the land blows across relatively warmer, much moister ocean water and create cylinders of spinning air. Where the air is rising, small clouds form. Where the air is descending, the skies are clear. The cloud streets align along the direction of the wind movement.
From:https://www.nasa.gov/image-feature/cloud-streets-in-the-sea-of-okhotsk
Increasingly, whenever anomalies are discovered in a scientific model of the climate and weather, clouds are found to be a large missing component, Clouds are very hard to study and model as they are so transient, so there is no rush to do anything about them. Nevertheless, they are a major player.
Nice work, Willis. Very interesting.
From the new IPCC report – AR6, page 1-51. No change from 2019.
”Feedbacks from the loss of summer sea ice and snow cover on land have contributed to amplified warming in the Arctic (high confidence) where surface air temperature likely increased at more than twice the global average over the past two decades.”
Willis, in you Figure 4 Scatter plot, it appears there might be a hint of a discontinuity in the scatter of data points around 1°C. What might the physics/meteorology of that be?
Noaaprogrammer: “a discontinuity in the scatter of data points around 1°C”.
WR: Interesting question. The difference between sea-ice and snow versus ‘no ice’ coverage makes the difference. When high-pressure areas over the poles lose some of their dominance (when ice and snow disappear), the role of water vapor is enhanced a lot and so is the role of low-pressure areas. Summer and winter situations over the Antarctic show the differences.
Figure 4 indeed gives an interesting shift between the two different states as shown by the blue spots.
noaa, the coldest parts around 1°C are where there is ice part of the year but not other parts of the year. As a result, it is discontinuous from the parts where there’s never any ice.
w.
That makes sense. A quibble about the LOWESS smoothing for that figure: Since LOWESS is a local averaging technique, it will not behave correctly near the discontinuity: It takes some of its values from the “other” side. Better to separate the data at zero and smooth each section, or use some sort of one-sided smoothing (backwards for <0, forwards for >0), or maybe even the “regime change” techniques from the time series literature (in this case applied to non-time series data). But as I said, this is a quibble, and does not detract significantly from very nice work.
22C – this is an incredibly important dicovery. It shows a planetary feedback system centered on 22C. This should show up as coincidences centered on 22C.
– the temp of houses before oil embargo.
– the temperature leaves rotate to maintain to optimize photosynthesis.
Clouds are the free variable that adjust the opacity of the atmosphere to achieve the goal of maximum surface warmth given the available post albedo solar energy. Of course, more clouds reduces the total solar energy, yet maximum warmth remains the goal. Maximum warmth occurs when the average emissivity is constant since changing the emissivity in either direction takes work that will not otherwise heat the surface.
The math shows discrete values of a steady state emissivity whose approaches have an infinite number of slopes, while all others have but one. These are the values of emissivity that conform to the equation
e^-2 +/- e^-1 = N
The infinite number of approaching slopes arises since values of e that conform to the above equation can be expressed as a infinite number of functions of e and powers of e without introducing any additional free variables.
For N = 1, e = 0.618023 which is the measured ratio between the average surface emissions per the SB Law and the average emissions of the planet.
I scroll right past any article that has “robust” in the title.
Then I see the author is Willis….
How deep do rabbit holes go ??
Probably slightly above your IQ that is very number is very low.
Yet, you took the bait 🙂
willis, question, did you consider the fact that at the poles in winter the sun don’t shine so that for albedo changes you only need to look at summer conditions. So how are the results when you only look at hemispheric summer?
Thanks, Hans. As you suggest, albedo only can be calculated when there is sunlight, so only those times are considered, and all other times are ignored.
w.
Willis, is the albedo in the maps measured as the amount of light striking a surface at right angles that gets reflected? If this is the case, then isn’t it true that at low angles, water reflects as much, or more, light than ice? As the light in the Arctic is almost always at low angles, intuitively I’d have thought that melting sea ice would increase albedo.
It is imperative that the “warming” have its strongest effect in the arctic regions, because they are difficult for people to visit and see any changes. Now we know, there is no there, there.
This might be of interest, as it shows that a different approach leads to similar conclusions:
“The radiative interaction between clouds, surface temperature, and surface albedo, that is, the cloud–radiation feedback, is such that there is no significant trend in the net radiative flux during winter, spring, summer, or autumn, even though there are trends in cloud and surface properties. It appears that during the sunlit part of a year, the decreasing trends in sea ice extent and surface albedo that result from surface warming modulate the increasing cloud cooling effect, resulting in little or no change in the surface radiation budget.”
Arctic Surface, Cloud, and Radiation Properties Based on the AVHRR Polar
Pathfinder Dataset. Part II: Recent Trends, Wang and Key, 2004.
Data from 1982-1999.
https://doi.org/10.1175/JCLI3439.1
As you are aware Willis, Pistone et al. published ‘Observational determination of albedo decrease caused by vanishing Arctic sea ice’ in 2014. Any comments on this more recent work? https://www.pnas.org/content/111/9/3322
I read a study months back which showed that since the mid 1990’s there has been a decrease in cloud cover over Arctic land masses, but an increase in cloud cover over the Arctic Ocean. That would explain the mostly cooler than average summer temperatures since 2001 north of 80°.
http://ocean.dmi.dk/arctic/meant80n.uk.php
I’ve often looked at graphs of temperature over the Arctic (this might refer to areas above 80 degrees North latitude) as a function of Julian date. These graphs show a rapid rise from January through about mid-May (when the temperature reaches 0 C), but then a plateau at about +2 to +3 C (275 – 276 K) through June, July, and August, until the temperature starts to decline in September. There is a lot of year-to-year variation during the cold times of year (January to April, then October – December), but summer temperatures are remarkably stable from year to year.
This may seem surprising, since May through August is the time of maximum insolation in the Arctic, but the open water that forms near the coasts (northern Russia, Alaska, the Canadian archipelago, and Scandinavia) enables evaporation of water that is frozen in place the rest of the year. This increased moisture leads to the formation of clouds and/or precipitation, which reflects away most of the sunlight over the Arctic during summer.
Open water along the Arctic coasts can decrease albedo temporarily, but this allows for evaporation and cloud formation, which tends to increase summer albedo.
“…feedback from changes in sea ice and snow cover have NOT contributed in any significant way to amplified warming in the Arctic. Cloud changes offset these sea ice and snow changes almost entirely.”
While the data show that cloud changes offset the effect on albedo of sea ice and snow changes, the data does not show what the effect on surface temperatures is of these changes.
Albedo is not the only factor in surface temperatures – and clouds have other effects on surface temperatures that are not related to albedo. In the polar regions such as the arctic, additional clouds tend to create additional insulation, slowing the rate at which heat dissipates from the surface into space. Given that the polar regions have a strongly negative heat balance, with heat being advected into those regions from regions closer to the equator, the insulation effects of additional clouds can contribute to raising surface temperatures in the polar regions.
Additional clouds could well underlie the relative warming seen in arctic regions. It would be interesting to examine the times of the year when additional clouds are present in the arctic regions and their relationship to the presence of open water or ice beneath.
(PS I note that in the maps, only the sea areas off Antartica seem to have any cloud albedo effect – it does not apply to the main landmass, presumably because the polar landmass is still ice covered at all times. This also appears to apply to Greenland in the northern hemisphere.)
Excellent analysis. It’s so obvious the climate systems don’t work the way their models work, parametrizing cloud albedo as static when it absolutely is not. I wonder how much of the error in models is driven by this theory error.
“Science is the process of overthrowing the consensus.”
Just beautiful, thank you.
Thanks Willis E
You have always a good point and makes sense.
An open sea surface in a cold climate makes the vapour leave and form a cover to block out the sun.
This Year ice over is on its way to cover the surface tree weeks ahead of last Years, that was record low this time of the year.
Tipping point is not proven in real life and You can show us why-thanks.
Do we know that the albedo of the arctic ocean reduces as the ice melts? Ice is white and has high albedo. Water does not scatter in the same way. It is dark in all directions except where direct light reflects when it is very bright. Has the maths been done to show that the integrated scattering from ice is indeed greater than the directional reflection of water? In addition, the amount of sunlight falling on the poles is tiny in the summer, zero in the winter. Surely, the overall effect must be very small.
Thanks, Pat. First, the CERES data on albedo includes both directional and non-directional reflection.
Next, on average the TOA insolation is 340 W/m2. The average at the very poles is 171 W/m2, about half of the overall average, but hardly “tiny”.
Now, the amount of solar energy actually absorbed at the surface is indeed small … but that’s because of clouds and albedo. However, even there the effect is not “tiny”. On average about 46% of TOA solar makes it to the surface. In the Arctic, it’s about 25%.
My best to you,
w.