Archive for August, 2016

Simple Time-Dependent Model of the Atmospheric Greenhouse Effect

Tuesday, August 30th, 2016

I thought it would be a useful followup to post a simple time-dependent energy balance model (spreadsheet attached) to demonstrate how infrared radiative flows affect the Earth’s surface temperature and atmospheric temperature. (I might have done this before…it sounds familiar).

The model is the simplest I could come up with to demonstrate how an atmosphere that absorbs and emits IR radiation ends up warming the surface, and itself as well, while maintaining an atmospheric temperature below that of the surface.

Here are the basic energy fluxes included in the model. The illustration is just schematic.

Fig. 1 Cartoon representation of the energy flows in the simple time-dependent energy balance model of the climate system

Fig. 1 Cartoon representation of the energy flows in the simple time-dependent energy balance model of the climate system

The energy input from the sun is fixed at an assumed 240 Watts per sq. meter. The radiative fluxes use the Stefan-Boltzmann equation (sigma T^^4), where T is either the surface or atmospheric temperature. Surface emissivity is assumed to be 1 (changing it to 0.95 or less make no difference to the conclusions, only the details).

You can adjust the IR absorptivity of the model in the spreadsheet, which is just a multiplier on the radiative flux coming from the atmosphere, and the radiative flux coming up from the surface and being absorbed by the atmosphere.

The model is initialized at absolute zero Kelvin, and heat capacities are prescribed so you can see the temperature changing over time as the model goes toward energy equilibrium. The heat capacity of the surface and atmosphere are assumed to be the same, equivalent to 1 meter of water for simplicity (the atmosphere is really more like 2 m of water effective heat capacity).

Using “0” for the atmospheric absorptivity leads to a surface temperature of 255 K, and zero atmospheric temperature (the model is radiative only, no convection, no conduction, so without any atmospheric absorption of radiation, the atmosphere cannot warm):


Then, to see how this “no-atmosphere” earth changes with an atmosphere that absorbs and emits IR, an IR absorptivity of 0.8 gives a surface temperature close to 290 K, and an atmospheric temperature of about 244 K.


If the model had dozens of atmospheric layers all interacting, it would produce much higher surface temperatures, and much lower temperatures in the upper atmosphere, producing a strongly super-adiabatic temperature profile (Manabe and Strickler, 1964). This is what causes atmospheric convection, which provides a net transport of heat from the surface to the middle and upper troposphere (not contained in this radiation-only model).

Again, this is an EXTREMELY simplified model of the effect of radiative flows on the global climate system. It is only meant to demonstrate the most basic components of the atmospheric “greenhouse effect”, which act to:

1) make the Earth’s surface warmer than it would otherwise be, and

2) keep the atmosphere cooler than the surface (since the atmosphere cools radiatively to deep space, but partially “blocks” the surface from cooling to space).

UPDATE: Based upon a few comments, it might be useful to point out:
1) the final equilibrium temperature does not depend upon the initial temperature assumed at the beginning of the model integration, it can be 0 K, 100 K, or 1,000 K.
2) the final equilibrium temperature does not depend upon the assumed heat capacities of the Earth’s surface and atmosphere…those just change how much time it takes for equilibrium to be reached.

Experiment Results Show a Cool Object Can Make a Warm Object Warmer Still

Sunday, August 28th, 2016

My previous post explaining a simple experiment to demonstrate that a cool object can make a warm object warmer still led me to give the experiment a try.

The purpose is to demonstrate that, energetically, the atmosphere’s greenhouse effect can make the surface of the Earth warmer than if the greenhouse effect didn’t exist even though the atmosphere is colder than the Earth’s surface. There is no violation of the 2nd Law of Thermodynamics, which states that the net flow of heat must be from higher to lower temperature, which does not preclude cooler object from emitting IR radiation in the direction of warmer objects.

If the atmosphere didn’t exist, the Earth’s surface would lose IR radiation directly to the cold depths of outer space, which is essentially at absolute zero temperature, and emits no energy back to the Earth; but instead the atmosphere, in effect, blocks some of that radiation, and emits some of its own IR radiation back towards the surface.

The net effect is that the surface and lower atmosphere cannot cool as rapidly to deep space, raising its average temperature.

The experiment shown below does not prove that greenhouse gases in the atmosphere perform such a function, only that it is not a violation of the 2nd Law of Thermodynamics for a cooler object emitting infrared radiation to keep a warm object warmer that it would otherwise be if the cooler object was not present.

The reason I am posting this is not to convince the rabid disbelievers, who are probably beyond hope. It is to reduce their influence on others. I’ve discussed all kinds of evidence here, the most convincing is just using a handheld infrared thermometer pointed upward at a “cold” clear sky, then measuring a warmer temperature when point it obliquely at the sky. That shows a cold object (the sky) can warm a surface (the thermopile in the handheld thermometer), even though the sky is colder.

Experimental Setup

The following setup (assisted by one of my daughters) includes a metal plate, painted flat black, and heated with a 250 W flood light. The heated plate has exposed to it a Styrofoam cooler containing ice. The hot plate is kept above the ice to minimize any air convection effects on the results.

Fig. 1. Experimental setup to demonstrate a cooler object can make a heated object even warmer still.

Fig. 1. Experimental setup to demonstrate a cooler object can make a heated object even warmer still.

Since a part of the heat budget of the heated plate is its loss of infrared energy to the cold ice, it should be possible to measure an increase in the temperature of the hot plate if the view of the ice is blocked with a second sheet (painted with very high IR emissivity paint, Krylon white #1502) at room temperature.

I tried two setups: the one on the left in Fig. 1 has the temperature probe scotch taped to the hot plate, but it did not stay firmly in contact with the plate, and also the steel plate I used exhibited large temperature gradients. On the right side of Fig 1, a better setup had the temperature probe firmly attached to the back with aluminum tape, and the metal sheet is aluminum flashing, painted flat black. The results presented below are for the arrangement on the right.

Flir thermal imager measurements of the setup late in the experiment are shown in Fig. 2, where we see the ice is generally below 32 deg. C F (it came from the deep freeze), and the heated plate (in the second phase of the experiment, results below) has a temperature around 110 deg F.

Fig. 2. Flir thermal imagery of ice and hot plate during the second portion of the experiment.

Fig. 2. Flir thermal imagery of ice and hot plate during the second portion of the experiment.

I recorded temperatures every 5 secs with the plate alternately exposed to a view of the ice for 5 minutes, then with the ice covered for 5 minutes. This cycling was repeated five times. The results are shown in Fig. 3. What we see is just what I would expect, that the temperature of the hot plate increases with time when its view of the ice is blocked by the room-temperature sheet.

Fig. 3. Time series of temperature of the hot plate as it was alternately exposed to ice, then a room-temperature high-emissivity sheet, every 5 minutes.

Fig. 3. Time series of temperature of the hot plate as it was alternately exposed to ice, then a room-temperature high-emissivity sheet, every 5 minutes.

Also shown in Fig. 3 are the results averaged over all five cycles, which smooths out some of the noise. The hot plate was so hot that just a small breeze of air from moving the room temperature sheet around the apparatus was found to cool the hot plate by a couple of degrees. The gradual warming trend was due to the ice slowly warming up over the ~1 hr period of the experiment.

Experiments like this often have sources of error when one tries to isolate a certain process. One concern would be whether the room-temperature sheet was slightly reflective to infrared radiation, which would cause the observed temperature effect by reflecting some of the hot-plate emitted IR radiation back on itself. I tested the sheet, which has very high infrared emissivity paint applied, by measuring its IR temperature with the Flir imager outside, both at right angles to the surface, and then at a ~45 deg. angle so the cold sky (today running in the mid-20s deg. F in the Flir measurements) could reflect off the sheet. There was no noticeable difference to a small fraction of a degree, so the paint appears to have an IR emissivity close to 1.0 and is indeed non-reflective in the infrared.

Another concern was the close proximity of the flood lamp to the ice and room temperature shield. If the shield was slightly more reflective to visible light than the ice, it could cause a little more heating of the hot plate. So I moved the lamp down and away from the setup so that there was little noticeable illumination of ice or shield (see the left side of Fig. 4). The hot plate cooled considerably since the lamp wasn’t nearly as close, but the effect of adding the shield was very clear with just one application, with temporary warming of the hot plate until the shield was again removed:

Fig. 4. Second setup where the flood light was place farther away from the plat it was heating, and the results of one cycle with the hot plate shielded from emitting IR toward the ice, then exposed to the ice.

Fig. 4. Second setup where the flood light was place farther away from the plate it was heating, and the results of one cycle with the hot plate shielded from emitting IR toward the ice, then exposed to the ice.

It should be noted that heating a surface with incandescent bulbs will be mainly through infrared radiation, since their emission spectrum is more in the IR than at visible wavelengths. Halogen bulbs, which I did not use, have a somewhat closer spectrum to the solar spectrum.


There is no violation of the 2nd Law of Thermodynamics in the experiment; a cool object can make a warm object even warmer still through infrared radiative effects. The phenomenon can only happen, though, if the cool object replaces something that is even colder, and thereby reduces the rate at which the warm object loses infrared energy to its surroundings. In this experiment, the room temperature plate takes the place of the ice which still emits at around 300 Watts per sq. meter; in the climate system, the atmosphere takes the place of deep space, which emits energy at close to 0 Watts per sq. meter.

The heated plate is placed above the ice so that there is essentially no intermingling of ice-chilled air (which will flow downward) with hot plate-heated air, which will flow upward. Ideally, the experiment would be carried out in a vacuum chamber, so that conduction effects by air would not be present.

There are changes that would make the experiment work better:

1) Dry ice in the place of water ice, to provide a colder target for the hot plate to lose IR energy to;

2) use an open chest freezer (running) covered with a single layer of plastic wrap, so the temperature of the ice won’t increase with time;

3) cover the hot plate holder with plastic wrap, which is about 90% transparent to IR, but will reduce the variations in heat losses due to air currents.

4) if the ice can cover a greater portion of the hemisphere that the plate is losing IR energy too, the effect will be magnified. An open chest freezer would accomplish this, too.

NOTE TO COMMENTERS: I intend to delete any comments which include personal insults.

NOTE TO READERS OF COMMENTS: Some commenters here throw around technical terms and make grand assertions and detailed arguments which I consider fallacious. I do not have time to counter them all every time they arise, although I have addressed virtually all of them in other posts over the years.

Simple Experimental Demonstration that Cool Objects Can Make Warm Objects Warmer Still

Friday, August 26th, 2016

I continue to receive emails (not to mention the hundreds of sometimes nasty blog comments) objecting to what I just expressed in the title of this article. So, I thought it would be useful to propose a simple experiment that demonstrates the concept.

This is a considerably simpler task than my recently proposed experiment to measure the warming effect of adding CO2 to the atmosphere (which I now believe was not possible, at least as I originally proposed it). This experiment would be easy enough for high school students to perform, and maybe even junior high students. It probably does require a good multi-probe temperature monitoring and data logging device. I use the Extech SD200 three-probe temperature sensor. Alternatively, a $50 handheld IR thermometer might be used in a pinch, if you are careful.


One of the supposed arguments against atmospheric greenhouse gases keeping the Earth’s surface warmer than if those gases were not present is the claim that, since the atmosphere is colder than the surface, it would violate the 2nd Law of Thermodynamics for a cold object (the atmosphere) to increase the temperature of a warm object (the Earths surface).

The Wikipedia entry for the 2nd Law of Thermodynamics includes the following statement from Rudolph Clausius, who formulated one of the necessary consequences of the 2nd Law (emphasis added):

“Heat can never pass from a colder to a warmer body without some other change, connected therewith, occurring at the same time.”

The statement by Clausius uses the concept of ‘passage of heat’. As is usual in thermodynamic discussions, this means ‘net transfer of energy as heat’, and does not refer to contributory transfers one way and the other.

The italicized words are important, and have been ignored by my critics: while it is true that the net flow of heat must be from higher temperature to lower temperature, this does not mean that the lower temperature object cannot (for example) emit radiant energy in the direction of the warmer object, and thus increase the temperature of the warmer object above what it would otherwise be.

But, this statement I just made will lead to endless arguments and objections (watch the comments, below), with hand waving qualitative statements about absorbing and emitting molecules and photons and entropy and perpetual motion and such.

So, let’s envision a simple experiment that will mimic what happens in the climate system, using visible light to heat a warm surface which then cools through infrared radiation toward a cold surface.

Energy Balance of the Global Climate System

The sun’s energy that is absorbed by the Earth’s surface raises its temperature. The surface then loses energy through both (1) non-radiative loss to the atmosphere (conduction, evaporation, and convection), as well as (2) infrared radiative loss to the atmosphere and to outer space.

What is interesting is that the clear atmosphere is mostly transparent to the sunlight passing through it and warming the system, but it is not transparent to the infrared radiation trying to escape back out to cool the system. This is the basis of the atmospheric “greenhouse effect”.

Now, in order for the climate system to maintain a roughly constant average temperature, there must be energy balance: the rate at which the earth-atmosphere system gains energy from the sun must match the rate at which the system loses infrared energy to the cold depths of outer space, which has a radiating temperature of almost absolute zero. If you can reduce the rate at which it cools to outer space, the climate system will increase its temperature until it emits enough infrared energy to restore radiative balance. This is the basis for global warming theory: increasing carbon dioxide in the atmosphere reduces the rate of IR energy loss to deep space, resulting in some warming. (The warming is actually in the lower atmosphere, while the upper atmosphere cools).

The Experiment

We can mimic these radiative processes by continuously heating one surface with halogen light bulbs (which more closely approximates the solar spectrum of light than incandescent bulbs). The hot surface will then radiatively cool toward a second surface which is chilled (e.g. with dry ice inside a cooler). The heated surface will also lose heat through conduction to the surrounding air, too, but we will reduce that effect with Styrofoam insulation….what we are looking for is a radiative effect.

The following cartoon shows the basic setup.


The heated surface is painted black to absorb as much visible light as possible and raise its temperature. The chilled surface is painted with Krylon white #1502 which has an infrared emissivity close to 1.0 (allowing it to efficiently absorb IR energy from the heated surface) while the white color also reflects visible light and so avoids heating from the lamps.

At some point, energy balance will be reached when temperatures stabilize. (Of course, eventually all of the dry ice will be used up…so there is limited time to do the experiment…maybe an hour or two). I suggest putting the heated surface on top so any heated air goes upward and away from the experimental setup. Similarly, the chilled surface will have chilled air spilling down the sides.

Now, if we simply insert a piece of room-temperature cardboard in between the heated surface and the chilled surface, we should see an increase in the temperature of the heated surface despite the fact that we just used a cooler (room temperature) object to make a warmer object even warmer still, in apparent violation of the 2nd Law of Thermodynamics. The cardboard can probably just be laid on top of the cooler. Or, a sheet of Styrofoam might work just as well, if not better. The temperature of the air between the heated and chilled surfaces could be monitored with the third probe from the SD200 to answer any objections that the intervening cardboard is somehow reducing the mixing of air between the hot and cold surfaces (which shouldn’t happen anyway, if the heated surface is above and the chilled surface is below).

The intervening cardboard (or Styrofoam) sheet mimics what the atmospheric greenhouse effect does, at least in terms of energy flows (but it’s a solid surface, rather than a gas, so maybe it’s more analogous to the greenhouse effect of thick cirrus clouds, which completely block the transfer of infrared light).

I don’t know just how much the observed temperature increase in the heated surface will be when the cardboard sheet blocks its view of the chilled surface. Maybe 1 deg. F, maybe 10 deg. F. But it should be observable. The effect will be greater the bigger the temperature difference that can be maintained between the heated and chilled surfaces, and the closer you can get them together so the heated surface “sees” mostly the chilled surface, instead ot the room-temperature surroundings with which it is also exchanging infrared radiation.

Now, this experiment does not prove that gases can do what the cardboard has done; that is a separate issue that is much more complicated to demonstrate with an experimental setup. It only answers the 2nd Law violation claims some have made against a cool object (here, the cardboard sheet) causing a heated object to be warmer than if the cool object was not present, which is what the Earth’s greenhouse effect does.

NOTE TO COMMENTERS: I intend to delete any comments which include personal insults.

NOTE TO READERS OF COMMENTS: Some commenters here throw around technical terms and make grand assertions and detailed arguments which I consider fallacious. I do not have time to counter them all every time they arise, although I have addressed virtually all of them in other posts over the years.

Observational Evidence of the “Greenhouse Effect” at Desert Rock, Nevada

Tuesday, August 23rd, 2016

Weather balloon measurements have been made twice daily at Desert Rock, Nevada for many years. In 1998, a surface radiation (SURFRAD) measurement facility was also installed there, which allows new kinds of analysis of how the radiation budget is affected by atmospheric profiles of temperature and humidity.

One portion of the SURFRAD instrumentation network at Desert Rock, NV.

Fig. 1. One portion of the SURFRAD instrumentation network at Desert Rock, NV.

The location is arid, minimizing the influence of clouds and precipitation, making it an ideal site for analysis of downwelling infrared (IR) sky radiation and how it influences surface temperature.

A example of the main radiation components measured every 3 minutes at Desert Rock is shown in the following graph, for July 1, 1998. I have also annotated the approximate times that the radiosonde ascents are made:

3-minute radiation budget data at Desert Rock during July 1, 1998.

Fig. 2. 3-minute radiation budget data at Desert Rock during July 1, 1998.

There are many more measurements than this in the Desert Rock data archive, such as temperature, wind, relative humidity, barometric pressure, ultraviolet radiation, etc.

Of particular interest is the question: How does the downwelling IR intensity depend on the vertical profiles of temperature and humidity? Obviously, IR intensity depends upon temperature….but there has to be an atmospheric emitter of IR, and that is primarily water vapor.

We can examine the issue using nighttime data, so that we don’t have to deal with the huge fluxes of solar energy during the daytime. If I average the SURFRAD fluxes of downwelling IR between 00 UTC and 12Z every day, and compare them the to average of the 00 and 12 UTC radiosonde profiles in ~100 mb thick atmospheric layers, I can do correlations to see how the nighttime “sky radiation” variability is related to atmospheric temperature and humidity variability.

The results are very interesting:

Correlation coefficients between daily variations of nighttime average downwelling IR and atmospheric temperature and humidity in different layers, during July 1998 at Desert Rock, NV.

Fig. 3. Correlation coefficients between daily variations of nighttime average downwelling IR and atmospheric temperature and humidity in different layers (vertical axis is pressure altitude, mb), during July 1998 at Desert Rock, NV. Several days that experienced more than scattered clouds and some precipitation were omitted.

The dominant influence of humidity variations on downwelling IR is clearly seen; the greater the humidity, the lower in the atmosphere the downwelling IR radiation measured at the surface originates, and thus the warmer the emitting temperatures and greater the IR intensity.

In contrast, the correlations of downwelling IR variations with temperature variations themselves are rather poor, probably because the temperature variations are so small. Clearly humidity variations dominate the downwelling IR signal, moving the effective radiating altitude up and down as humidity decreases and increases, respectively.

Now, the downwelling IR flux (the dashed line in Fig. 2, above) is what a few of our friends claim does not exist. They claim that there is no “greenhouse effect”, and that the sky (which is almost always colder than the surface) cannot emit IR in the direction of the surface because that would violate the 2nd Law of Thermodynamics.


But, of course, it is the net IR (the sum of upwelling from the warmer surface plus the downwelling from the cooler sky) which must flow from higher to lower temperature, which it does.

So, what, in their minds, is actually being “measured” by these instruments for downwelling IR? Whatever it is, Fig. 3 clearly shows it’s closely related to the humidity of the atmosphere (correlations up to 0.88 for mid-tropospheric humidity), but not very well related to temperature variations in the atmosphere. Barring some sort of conspiracy between all of the atmospheric radiation experts in the world (as well as most of us “skeptics”) it is difficult to imagine how such a “fictitious” measurement, so sensitive to atmospheric humidity, could be constructed by mistake.

But what influence do these variations have on nighttime cooling of surface temperatures? For that is how the “greenhouse effect” is usually expressed: the increase in surface temperatures caused by greenhouse gases compared to if those gases did not exist. It is not possible to answer that question in an absolute sense with measurements because we do not have a full-depth atmosphere with no greenhouse gases we can experiment on. Instead, we can only examine how surface temperature changes by relatively small amounts when the amount of greenhouse gas changes by relatively small amounts.

This can be seen in the next plot, where I have compared the change in surface temperature from 00 UTC (late afternoon) to 12 UTC (early the following morning), to the average downwelling IR during the night:

Nighttime cooling of surface temperatures at Desert Rock during July, 1998 becomes less as downwelling IR intensity increases.

Fig. 4. Nighttime cooling of surface temperatures at Desert Rock during July, 1998 becomes less as downwelling IR intensity increases.

While the relationship is noisy because there are many factors governing nighttime surface cooling (wind speed, storage of solar energy in the soil during the previous day(s), etc.), we still see that the surface temperature drop during the night becomes less as the downwelling IR increases. This follows our daily weather experience that nighttime temperatures cool off more when humidity is lower, all other weather variables being roughly equal.

While the above analysis is preliminary, and there are many more relationships that could be examined (with many more years of data), the results clearly show that increasing greenhouse gas concentration in the atmosphere (in this case, water vapor) increases downwelling IR radiation from the sky, and increases surface temperature. And, while I have used nighttime data to isolate the effect from the complications introduced with daytime solar heating, it should be remembered that infrared effects on surface temperature are occurring 24 hours a day.

Downwelling IR from the sky continuously maintains surface temperatures well above what they would be without greenhouse gases (while at the same time cooling the upper atmosphere well below what it would be without those gases). Surface temperature is a function of energy gain (from the sun) and energy loss (which is reduced by greenhouse gases).

It’s not’s just physics.

NOTE TO COMMENTERS: I intend to delete any comments which include personal insults.

NOTE TO READERS OF COMMENTS: Some commenters here throw around technical terms and make grand assertions and detailed arguments which I consider fallacious. I do not have time to counter them all every time they arise, although I have addressed virtually all of them in other posts over the years.

Major Hurricane Gaston? A Potential U.S. Threat

Friday, August 19th, 2016

With the climatological peak in hurricane activity only 3 weeks away, the Atlantic has been fairly quiet so far, despite seasonal forecasts of a more active than normal season.

But recent forecast model runs have been consistently predicting that a low pressure wave in the tropical eastern Atlantic will become Tropical Storm Gaston in the next 5 days or so. Then, it looks like it could intensify into Major Hurricane Gaston, with 110 kt sustained winds by Sunday evening, August 28, which would make Gaston a strong Category 3 hurricane (graphic courtesy of

GFS model forecast of seal level pressure and wind speeds for Sunday evening, 28 August 2016. Graphic courtesy of

GFS model forecast of seal level pressure and wind speeds for Sunday evening, 28 August 2016. Graphic courtesy of

Of course this is very prelimnary, being almost 10 days out, and the system is currently not even a tropical depression yet. The predicted path of (potential) Gaston is especially uncertain. Interests along the Atlantic and Gulf coasts should monitor this system in the coming days.

Causes of Lake Superior Record Water Level Rise 2013-2014

Friday, August 19th, 2016

As a followup to my cursory analysis suggesting increased precipitation was the probable cause of the record rise in Lake Superior water levels during 2013-2014, the GLERL folks pointed me to a relatively recent paper they published (Hydrologic Drivers of Record-Setting Water Level Rise on Earth’s Largest Lake System) which provides a detailed analysis of all of the hydrologic inputs and outputs to the levels of the separate Great Lakes (over-lake precipitation and evaporation, land precipitation runoff into the lake, river and channel inflows and outflows).

The following plot from their paper provides their statistically optimized estimates of the various hydrologic components that cause levels to change on Lake Superior. I suspect the most accurate measurements are the lake levels and outflow through the St. Marys River. Precipitation would be less well measured, and evaporation would be even more uncertain. Use the top portion to see the water level rise over the January 2013 thru December 2014 period, and use the bottom plot to understand the components that went into the rise, where arrows pointing up increase lake levels, and those pointing down decrease lake levels, compared to the long-term averages for those months (click on image for large version).


The bottom line is that the record rise in lake levels was mostly the result of above-normal precipitation (the blue [lake precip] and green [runoff from land precipitation] bars extending above the zero line). But also important was reduced evaporation (red bars) from the very cold winter of 2013-2014, which led to extensive ice cover and unusually cold lake water during the following summer.

Finally, note the grey bars, which indicate increased outflow through the St. Marys River at Sault Ste. Marie, MI, starting in mid-2013, which acted to limit the lake water rise.

As an aside, there is an interesting analogy between water storage in lakes, and heat storage in the ocean. There are inputs and outputs affecting each, and when there is a huge change (imbalance between inputs and outputs) it takes time for things to either depart from normal or go back to normal. Since the lake is small, that can happen in only a few years. In the case of heat storage in the ocean, it can take decades if not centuries for changes to be felt.

Lake Superior Water Levels: The Role of Precipitation

Wednesday, August 17th, 2016

The last couple years have seen exceptional erosion along portions of the south shore of Lake Superior, especially where the ground is very sandy. The following photo was recently taken west of Whitefish Point in the eastern Upper Peninsula of Michigan, of a cabin built in the 1950s:

Erosion on the south shore of Lake Superior (Ben Musielak, Ben Musielak Photography, Paradise, MI).

Erosion on the south shore of Lake Superior (Ben Musielak, Ben Musielak Photography, Paradise, MI).

While water levels have been on a slow, irregular decline for decades, there was a sudden rebound during 2013-2014 to near-record high levels:


What caused this rapid rise? In fact, what controls the water level of Lake Superior on a year-to-year basis?

The hydrology depends upon many factors, both natural and human. Precipitation over the lake and the drainage basin feeding the lake, evaporation, and outflow through the rapids on the St. Marys River at Sault Ste Marie are the primary natural processes.

But locks built in the Sault in the 1800s also altered the natural lake levels, as well as contruction of a canal which feeds a hydroelectric power plant. Subsequent dredging of the shipping channels has also altered the flow. (Here’s an amazing high-resolution 1905 photo during a celebration of the locks, it still looks the same today…except for fences to keep idiots from falling in, and an observation deck).

Gates at the head of the power canal are raised and lowered to provide some human control over lake levels, and are raised to allow more water to flow out when lake levels are high. Many factors are weighed in deciding to adjust the flow out of Lake Superior, including the water level in the rest of the Great Lakes downstream, as well as Canadian concerns. The governing body for these decisions is the International Joint Commission (IJC).

For example, there has been much debate over water levels on Lake Erie, which have also been running high. If you let more water out of (much larger) Lake Superior, then the coastal interests along Lake Erie (which I suspect are much larger in number and politically more powerful) are going to be very concerned. The total population living along the lower lakes is about 15-20 times that living along Lake Superior, so you can see why interests along Superior aren’t the only ones deciding how much extra water will be allowed to flow out of the lake.

How Much Control Do We Have Over Lake Superior Levels?

Given that the Sault locks are there to stay, just how much control could we have over the water level in Lake Superior, if we wanted to?

The answer, it turns out, is not very much.

The river discharge out of Lake Superior through the St. Marys River has been running around 100,000 cu. ft./sec in the last couple of years. If you assume that we could change that rate at will by, say, 10%, and divide that into the number of square ft. covered by Lake Superior (~884,000,000,000), you will find that such a change in river discharge would only change the lake water level by 4 inches in one year. The above graph shows that much larger changes occur on the lake than this. Obviously, natural influences on water level are much larger than human influences.

I used to live on the lower St. Marys River, and can attest that when the lake level is high, so is the level of the lower river. I specifically remember the summer of 1973 (see above graph) when water on some days went up on our lawn and came close to our front door. Then, years after I moved away, low lake levels led to the shoreline going out about 100 ft. further from the house than normal.

The reason I mention this is that, when lake levels are high, there is an increase in river discharge out of the lake. That is the direction of causation: high lake levels => increased outflow. It is not: decreased outflow => high lake levels.

In other words, human manipulation of water flow out of Lake Superior is not the cause of high water levels…although we have some small amount of control to mitigate changes in lake levels, after they have occurred.

The Primary Control Knob: Precipitation

While I’m not a lake hydrology expert, I suspect that the balance between precipitation and evaporation is the governing factor in Lake Superior water level.

If we look at the average yearly precipitation departures from normal from the National Centers for Environmental Information website for the Upper Peninsula and northeast Minnesota averaged together, we find that the Lake Superior water level rises and falls depending upon excesses and deficits in precipitation:


In fact, during the most recent rapid rise in water levels, a 2-year excess in precipitation of 11 inches led to a 22 inch rise in lake levels! This is pretty spectacular.

Why would the lake rise more than the precipitation? Because the drainage basin for Lake Superior is 1.55 times larger than the lake itself, and so some of the excess water that falls on the surrounding land flows into the lake, potentially more than doubling the lake level rise due to a precipitation increase:

How well this quantitatively explains things, I don’t know. I’m ignoring changes in evaporation, which would have to be estimated through some modeling assumptions.

The bottom line is that Mother Nature is largely in control of water levels on Lake Superior. Humans can help mitigate it somewhat by adjusting flow through the St. Marys River, and I believe that is already done, but I suspect that coastal interests along the lake simply have to live with the changes, which we have very little control over.

I have sent a few questions to the experts at the Great Lakes Environmental Research Lab (GLERL), and will update this post if I find out any additional information.

Suggested Backyard Experiment to Measure the Greenhouse Effect of More Carbon Dioxide

Sunday, August 14th, 2016

In my continuing battle to keep people from being led astray by bad science, I sometimes try to think of new ways to demonstrate the existence of the Earth’s so-called greenhouse effect (GHE).

I won’t bore you with all of the many evidences I’ve already offered over the years, but instead get right to the experimental setup, which involves putting pure carbon dioxide in one tube, air in a second, identical tube, and measure how much the temperature at the bottom of the tubes cools down during a clear night. I believe that, despite the small path length of CO2, it might be possible to measure a reduction of cooling the the CO2-filled tube versus and identical tube containing air.

I would prefer to have a setup like this where the additional, incremental effect of more CO2 on the atmosphere “partially blocking the view of cold outer space” on an actual near-surface air temperature is measured. More of a real-world demonstration. It would be easier to measure how CO2 (or, say, water vapor) blocks the IR emission from a hot target in a laboratory, but that has already been measured thousands of times, to very high precision at many wavelengths and for many gases.

The widespread claim on the web that a temperature warming effect of CO2 can be measured with CO2 in a jar is totally bogus, and Anthony Watts demonstrated how Bill Nye erred in trying to demonstrate such a thing. You have to (1) have long path lengths, and (2) the CO2 must be “partially blocking” the view that a warm object has of the radiatively “cold” sky. It can’t be done inside.

The trouble with actually measuring a temperature effect from the GHE is that the broadband IR effect (over all wavelengths, not just near absorption lines) of greenhouse gases in the atmosphere really only shows up over long path lengths, so that there is appreciable amounts of greenhouse gas. These long path lengths, combined with the vertical temperature structure of the atmosphere, are necessary for the effect to become appreciable. (Right on CO2 absorption lines, the effect can be large over very short path lengths, but these very narrow bands affect only a tiny fraction of the total IR energy being transferred).

Further complicating any simple experimental setup is that the atmosphere is already pretty opaque to IR radiative transfer, making the practical measurement of a temperature change from adding a little more absorbing gas pretty difficult.

If we were to condense all 400 ppm of CO2 in the atmosphere down to the surface, it would fill a column about 4 meters deep. So, if we could fill a 4 ft tall concrete form tube with pure CO2, I calculate that would be about 120 ppm equivalent increase, or 30% increase, in CO2 compared to the total atmosphere.


The question is, would this be enough additional CO2 to change the radiative budget in the tube enough to measurably reduce the nighttime cooling at the bottom of the tube? It might be.

Of course, one could make the tube taller, and so increase the CO2 path length, but the problem is that the bottom of the tube can only see a small portion of the sky (through the top of the tube) to radiatively cool. Thats why the tubes should be lined on the inside with highly IR-reflective aluminum foil, which has IR reflectivity of 0.95-0.97. This will help the high-emissivity bottom of the tube see more of the cold sky, although multiple reflections off the walls might be required.

How might one fill the tube with CO2? Other than purchasing a cylinder and getting it filled from a supplier, one easy way might be to put a block of dry ice (solid CO2) in the bottom and just wait for it to sublimate. The cold gas will also tend to settle in the bottom, and push air out the top, maybe through a pinhole in the highly transparent (>0.9 transparency) plastic wrap cover. I calculate one 1 ft x 4 ft tube has about 0.1 cubic meter volume, which would require about 0.5 lb of dry ice to fill with CO2.

One issue is that CO2 has somewhat less thermal conductivity than air. But if you look up the numbers, you will find that the thermal conductivies involved are so tiny that they can be ignored in the energy budget of the air in the tubes. Air and CO2 are extremely poor conductors of heat.

I probably won’t attempt this till fall or winter, when air mass humidity goes down. From monitoring the Goodwin, MS SURFRAD site, I’ve noticed that the difference between upwelling IR and downwelling IR increases from only about 40 W/m2 in the summer to over 80 W/m2 on cold dry days in the winter:


One would need as big a cooling potential as possible so see the effect of adding CO2 in the tube, and that difference between upwelling IR and downwelling IR is what is important for the experiment to work. When that difference approaches zero, say with dense low cloud cover, there will be no radiatively-induced temperature change anyway.

Good insulation is absolutely necessary so that the effect of the surrounding air temperature has minimal impact on the results and the differences in tube interior temperatures will be dominated by IR radiation transfer. Insulation results in a temperature drop below ambient temperature. I have blogged previously about producing 4 deg. F temperature drops below ambient in a Styrofoam box covered with plastic wrap, and have since achieved up to 8 deg. F drop in a standard Styrofoam cooler. The effect shows up strongly within an hour or so of sunset.

The experiment would probably need to be repeated with the tubes swapped since they cannot be constructed exactly the same.

Suggestions are welcome. I’m only going to attempt things that are easy, cheap, and not time consuming.

Finally, no matter how accurately the temperature effect is that is measured (if any) it can’t be used to quantitatively estimate the effect of adding CO2 to the atmosphere. There are too many differences between the experiment and what happens when the extra CO2 is spread out through the full depth of the atmosphere, AND the atmosphere has a chance to respond through changes in clouds, evaporation, precipitation, etc. The simple experiment is meant to simply demonstrate that the GHE exists…that is, that atmospheric greenhouse gases cause a warming tendency of surface temperatures.

(Edited for clarity).

Tropics Cool by 5 deg. F in One Week! (In the Stratosphere)

Thursday, August 11th, 2016

Capt. Kirk checks out the weather at 110,000 ft altitude.

Capt. Kirk checks out the weather at 110,000 ft altitude.

Most meteorologists consider the stratosphere to be a pretty boring place: no warm fronts, cold fronts, low pressure systems, and (almost) no clouds.

But there are a couple of things that happen there which are pretty dramatic. Sudden stratospheric warming in the polar regions is one. Another, but lesser-known, type of event is sudden temperature changes in the tropical stratosphere.

For example, in the last week the tropical middle stratosphere, as measured by channel 14 of the AMSU instrument flying on NOAA-19, has cooled by over 3 deg. C (about 5 deg. F):

Daily departures from average in the temperature of the tropical (25N-25S) mid-stratosphere, through 10 August 2016.

Daily departures from average in the temperature of the tropical (25N-25S) mid-stratosphere, through 10 August 2016.

The area covered is over 40% of the surface area of the Earth, so that’s a huge region.

The reason why such large temperature changes can occur in the stratosphere has to do with its vertical temperature structure, combined with dynamically-forced changes in the vertical circulation of air in the stratosphere.

The unique vertical temperature structure in the stratosphere is due to ozone being created and heated by ultraviolet solar radiation. The ozone layer then shields the air below it from UV radiation, so the layer is maintained at a “warmer” temperature as it continues to absorb UV energy.

Since the temperature increases with height in the stratosphere, any forced ascent of the air will cause a large drop in temperature at any given pressure altitude (since dry air ascent at any altitude will result in a temperature drop of 9.8 deg. C per km). Similarly, forced descent will cause a large temperature rise.

Now, the global stratosphere experiences a slow vertical circulation called the Brewer-Dobson circulation, with slowly rising air in the tropical stratosphere and slow sinking of stratospheric air outside the tropics. This circulation explains why, even though most stratospheric ozone is created in the tropics where sunlight (per square meter) is most intense, the greatest concentrations of ozone are found outside the tropics, as it is exported out of the tropical stratosphere.


Weather activity generated in the troposphere is what is believed to force this circulation (e.g., see here). As Rossby waves and gravity waves in the lower atmosphere propagate upward and equatorward, changes in their activity can cause changes in the strength of the Brewer-Dobson circulation, leading to the large temperature changes seen in the time series plot, above.

So, stronger rising of the stratospheric air is what led to cooling in the last week, but that will be almost exactly matched by warming outside the tropics where air is sinking faster than normal. (Note that rising air in one location and altitude must be almost exactly matched by sinking air elsewhere at the same altitude, otherwise there would be huge pressure differences leading to tremendous winds which will act to remove the pressure difference…this is what happens in hurricanes to some extent, where atmospheric mass is removed from the center of the hurricane faster than it can be replaced by in-flowing winds.)

I suspect these stratospheric temperature variations changes have no measurable effect on weather in the tropical troposphere, even though they might change the tropical radiative energy budget by a fraction of a Watt per sq. meter.

Will We Reach 4,000 Days Since a Major Hurricane Strike?

Monday, August 8th, 2016


In less than two months (October 6, 2016) it will be 4,000 days since the last time a major hurricane made landfall in the U.S., which was Wilma on October 24, 2005.

Wilma was a record-setter, being the strongest Atlantic hurricane on record, with peak estimated sustained winds of 183 mph and lowest surface pressure of 882 mb. That surface pressure corresponds to a 13% removal of atmospheric mass in the core of the hurricane compared to normal sea level pressure.

But after the record-setting 2005 Atlantic hurricane season, with a whopping 27 named tropical storms, the bottom pretty much dropped out of hurricane activity since then.

After an unusual January hurricane this year (which I don’t meteorologically count as part of the 2016 season), we’ve had one system (Earl) that briefly achieved hurricane status before making landfall in Belize several days ago:


Will we reach October 6 without a major (Cat3 or higher) hurricane strike? No one knows. The Atlantic is quiet right now, and there has been no significant trend in global tropical cyclone activity since satellite monitoring started in the early 1970s.