CERES and MODTRAN VI, Cloud Radiative Effect

CERES is the acronym for the rather chewy title Cloud Earth Radiative Effect Satellites. Displayed above is the TOA net flux in blue. The net TOA cloud radiative effect is in red. The axes are inverted to maintain the energy lost/cools planet energy retained/warms planet format. The 2005 aberration is alive and well.

There is a tension here because the All Sky net data already includes cloud effects, yet the  two data series are working in opposite directions for the planetary energy balance. How can the all sky (including clouds) net flux, defined as incoming minus outgoing, be increasing; when the cloud only (net CRE) effect is pushing a roughly equal net amount of energy to space?

We can deconstruct the net CRE signal into its LW and SW components above. This makes no sense at all. Both the LW and SW components are trending towards less radiation to space, but the net effect in the first graphic is clearly towards more radiation to space.

When net CRE is compared with LW CRE above, it can be seen that they are opposing.

When net CRE is compared with SW CRE, there is far better accordance, but the trends are opposing. It might be noted that the SW trend is so miniscule as to be arguably no trend at all, but we stick with the data.

You may recall that solar adjustment improved the agreement between the planetary (not CRE) net TOA flux and the component LW and SW fluxes. So it is with the CRE net flux above. Basically, they are removing the solar aspect from the component fluxes. That’s ok, but it would be better to be more explicit.

Are clouds warming or cooling the planet?

If we look at the component fluxes, which are stripped of solar fluctuation:

LW all sky TOA flux, which includes clouds and simply measures LW photons leaving for space, has cooled the planet ~.5 W/m2 since the millennium, according to the trend.

LW CRE over the same period has warmed the planet by ~.4 W/m2.

We conclude that without the moderating effect of clouds, LW TOA flux would have cooled the planet by .9 W/m2. Bear in mind that clear sky LW TOA flux, which shows little trend, radiates to space 18 W/m2 more than all sky with clouds.

SW all sky TOA flux, which includes reflection from clouds and measures SW photons leaving to space, has warmed the planet ~1.3 W/m2 over the period, according to the trend.

SW CRE over the same period has warmed the planet by maybe .1 W/m2.

We conclude that absent solar fluctuation, reduced SW TOA flux is warming the planet, but little of the warming comes from clouds. Something else is increasingly absorbing SW radiation.

The net of LW and SW all sky TOA fluxes without solar fluctuation is to warm the planet ~1.3 W/m2. When the sun is included in the net TOA all sky flux, the warming is reduced to about .5 W/m2.

Now we are ready to see what the MODTRAN radiative transfer program predicts about clouds in the next post.



Posted in CERES, Climate | Leave a comment


We must begin by confessing that we have resorted to digitizing. The first CERES data were garnered by mousing over the visualization tool, writing in a notebook, and typing into Excel. We discovered that if we right clicked the tool and “inspected the element”, the data was there in a weird bracketed  format that could be copied. By doing this we saved lots of time copying and pasting a few series rather than scribing and typing, but this caused the visualization tool to crash. Subsequently, access to the data this way has been denied.

Since CERES “Net” data is defined as incoming minus outgoing, positive numbers indicate energy retention. The “Flux” data is all outgoing, and indicates energy loss. To correct for this we initially multiplied the flux data by -1. This pointed the data in the proper direction, but caused all sorts of confusion in comparison with Net Cloud Radiative Effect data, which uses the same definition of “net”, but is all negative numbers.

All the graphics in this series have been changed so the net flux axis is in positive numbers, but inverted to point the data correctly.


The above two graphics are the same with the net flux axis inverted so the trends make sense in terms of cooling or warming the planet. This format will be used henceforth.

The data above  is doing everything wrong according to greenhouse theory. The slopes are small, and probably not statistically significant, but they are definitely not pointed in the “right” direction. The “right” direction would be a decrease in Longwave and Net radiation to space, and an increase in Shortwave (reflected from increasing clouds) radiation to space.

How does the net flux increase when the SW flux decreases more than the LW increases?

This quandary is illustrated well above. SW and LW outgoing fluxes are added together and compared with inverted net flux. The SW flux decrease is twice the all sky longwave flux increase. Since incoming solar is fairly constant, and the SW outgoing decrease increases the net, SW+LW is a reasonable approximation of inverted Net.

There is a consistent aberration in 2005. Aberrations often hold important lessons, and the 2005 aberration may be the topic of a future post.

It turns out that solar variation becomes important at the subtle scales we are working with here. The solar variation adjustment to LW+SW outward flux above improves the correlation with inverted Net, but exacerbates the 2005 aberration. Another clue.

You have probably noticed that this is an adventure, not a script. We will get to prior promises, but the next post must deal with cloud radiative effects.



Posted in CERES, Climate | Leave a comment


The tile of this series promises a comparison of CERES and MODTRAN. The groundwork of exploring the CERES data in the first three posts is done. At the conclusion of the first post, we concluded that CERES and MODTRAN could not both be right.

MODTRAN has no time dimension. It differs only on the basis of gas concentrations in the atmosphere, and it deals only with infrared (IR) light, the spectrum the earth radiates based on its temperature. MODTRAN can be related to CERES in time by adjusting atmospheric CO2 concentration to match the yearly values. We used the data from Mauna Loa in Hawaii. The comparison is further complicated because polar orbiting CERES data covers the entire planet, and MODTRAN data is broken down into tropics, mid latitude summer, mid latitude winter, subarctic summer, and subarctic winter.

CERES data defines the top of the atmosphere as 20km. This is easy to match in MODTRAN by setting the altitude. CERES TOA data is limited to clear sky and all sky. All sky, which is whatever the satellite saw, is what we would like, but MODTRAN has a smorgasbord of cloud options that would be difficult to relate.

Accordingly, we have matched CERES clear sky with MODTRAN clear sky.



We decided to simplify the problem by weighting the MODTRAN “tropics” at  30 north to 30 south latitude, which conveniently is 50% of the planet surface. The subtropics were given 30 to 50 degrees in both hemispheres, summer and winter were averaged, and weighted at 26%. We gave subarctic all the rest, averaged summer and winter, and weighted at 24%.

The result, at the top of the graphic above, predicts a modest but steady linear decline in IR radiation to space as CO2 concentration increases.

The CERES measured LW data at the bottom of the graphic is more jumpy, but definitely does not show a steady decline as MODTRAN predicts.

In the middle of the graphic, plotted on a second axis, CERES net clear sky data, which includes shortwave reflection, is inverted to show that increases in value result in net energy gain.

It is worth bearing in mind that the CERES time series basically corresponds with the “pause”, the lack of statistically significant warming in satellite lower troposphere datasets between the millennium and the 2016 el Niño. The lack of any reduction in CERES  clear sky radiation to space, as MODTRAN predicts, suggests that CO2 is not driving warming.

CO2 remains when the clouds are gone. Water vapor that has not condensed remains when the clouds are gone.

The CERES all sky data is what shows the reduction in TOA radiation to space. The difference between clear sky and all sky is clouds. It appears that clouds are warming the planet.

To continue this comparison of CERES and MODTRAN, we will need to figure out how to equilibrate the smorgasbord of cloud options in MODTRAN with “all sky” in CERES.

We will attempt this in the next post.



Posted in Climate | Leave a comment

Responses to the Questions Posed by Judge Alsup

Question 1. What caused the various ice ages (including the “little ice age” and prolonged cool periods) and what caused the ice to melt? When they melted, by how much did sea level rise?

Ice ages occur on our planet on two entirely different scales.

In the large scale, three “ice ages” of different durations, with warm intervals between lasting hundreds of millions of years, have taken place in the Phanerozoic, the “period of abundant life” of the last 700 million years. No known orbital parameters can explain these glacial episodes.

The small scale ice ages, to which you likely refer, are oscillations within the third, and most recent of the large scale ice ages above. We call this third large scale ice age the Pleistocene. It began about 2.5 million years ago, and we are still very much in it. Orbital variations, particularly obliquity, a ~41k year variation in the earth’s wobble, are clearly an influence–but not a controlling influence–on these oscillations. We simply do not know what caused the large scale ice ages, the glacial/interglacial oscillations within the Pleistocene, the little ice age, or the Roman and Medieval  warm periods. CO2 is not implicated in any of these changes.

The Oxygen isotope content of marine sediments cored in the Ocean Drilling Project show us the descent into the current ice age. We live at the warm peak at the extreme right side of the graphic. It is somewhat less warm than the three preceding “interglacial”  peaks. The preceding low points are times when there was a mile of ice on the Great Lakes.

To the left of the Mid Pleistocene Transition, the dominant periodicity is ~40k years. To the right of it, the period changes to ~100k years, and the amplitude of variation continues to increase.

Sea level has varied enormously through these glacial/interglacial oscillations.

It can be seen that sea level has risen some 130 meters since the most recent glacial maximum. It can also be seen that the most rapid increase ended about 8k years ago, and we are on the long tail. Current estimates of sea level rise vary between 1.5 and 3mm per year.

It can be seen that Greenland ice mass was lower in the 1930’s than it is today. According to NASA, Antarctic ice mass is currently increasing.

2. What is the molecular difference by which CO2 absorbs infrared radiation but oxygen and nitrogen do not? 


When the earth’s surface absorbs shortwave radiation from the sun, it warms, and creates its own longer wave (IR) radiation according to the earth’s much lower temperature. CO2 is a very efficient absorber in the earth’s longwave emission spectrum.

CO2 is a linear molecule. Its modes of bending and stretching and the applicable wavelengths are shown proportionally above. CO2 is unusually dependent on its fundamental bend at wave number 667.4 or about 15 microns. It absorbs so strongly in this band that all the earth’s radiation is completely absorbed, even at preindustrial CO2 concentrations.

The MODTRAN model above shows the difference in radiation to space between 280 ppm (preindustrial) and a “doubled” 560 ppm concentration. We are currently about 410 ppm. The differences between the red and blue are very subtle.

CO2 definitely absorbs outgoing longwave radiation the earth emits. The amount of warming this causes is limited by the saturation of the fundamental bend. The differences of absorption between CO2, Nitrogen, and Oxygen are just brute facts of the material properties of the gasses.

3. What is the mechanism by which infrared radiation trapped by CO2 in the atmosphere is turned into heat and finds its way back to sea level?

“Heat” is the net kinetic energy of all the molecules of the atmosphere. When a molecule absorbs radiation, it bumps into its neighbors more, raising the temperature. Warmed air rises, so this “conduction” by kinetic interaction is not an efficient way to get absorbed energy from any significant altitude back to sea level.

Warmed molecules also radiate in all directions, including down. This radiation can reach the surface, but it is subject to reabsorption along the way.

Above is the MODTRAN program at one meter elevation looking up in blue and down in red at 400ppm with CO2 set as the only greenhouse gas. The red looking down shows no deviation from the Planck curve. This is the case up to about 100 meters, so up to this altitude the atmosphere radiates upward like a brick. Looking up in blue, the atmosphere is radiating downward at nearly surface temperature in the CO2 bands, but falls off quickly to very low temperatures and intensities outside the CO2 bands. This radiation warms the surface.

4. Does CO2 in the atmosphere reflect any sunlight back into space such that the reflected sunlight never penetrates the atmosphere in the first place?


CO2 is a very minor absorber of incoming short wave radiation from the sun. It will re emit radiation ~2000 nm it has absorbed, but it is colorless, and does not reflect at all.

5.  Apart from CO2, what happens to the collective heat from tailpipe exhausts, engine radiators, and all other heat from combustion of fossil fuels? How, if at all, does this collective heat contribute to warming of the atmosphere?

This energy, including the heat radiating from our bodies, is insignificant at planetary scale. It becomes important to the tune of several degrees C in big cities.

6. In grade school, many of us were taught that humans exhale CO2 but plants absorb CO2 and return oxygen to the air (keeping the carbon for fiber). Is this still valid? If so, why hasn’t plant life turned the higher levels of CO2 back into oxygen? Given the increase in human population on Earth (four billion), is human respiration a contributing factor to the buildup of CO2?

Humans exhale about 3 billion tons (Gt) of CO2 a year , or nearly 1 Gt Carbon. Human combustion is about 10 Gt Carbon. Soil respiration is about 60 Gt Carbon per year. The annual global Carbon cycle is hundreds of Gt per year. Atmospheric CO2 is increasing, but at only half the rate of human production. The planet is greening as a result of the increased atmospheric CO2.

7. What are the main sources of CO2 that account for the incremental buildup of CO2 in the atmosphere?

Human CO2, Soil respiration, land use reduction of chlorophyll, ocean outgassing; in that order. The exact proportions are unknown.

8. What are the main sources of heat that account for the incremental rise in temperature on Earth?

The answer to this question is unknown. Humans are certainly responsible for some increase in temperature, but the increase in temperature we have seen is not exceptional. Enormous variations in temperature took place long before humans, and these changes were not driven by CO2. Careful measurements in the ice cores extending back 800k years show that without exception, changes of temperature preceded and controlled changes in CO2.  

We come full circle to the first question. Until we understand why the planet will go from paradise to ice, and within these ice ages to oscillate between warmer periods like we live in and colder periods when the ice piles up; we cannot say with certainty what the main sources of heat are that account for the recent incremental rise in the temperature on earth.



Posted in Climate, Judge Alsup Questions | Leave a comment


In the first post in this series we began with an assertion that according to greenhouse theory, it is impossible to warm the planet when longwave radiation to space is increasing. In the second post we discovered that it is possible for the planet to warm while longwave radiation to space increases if shortwave absorption increases. The first post has been corrected.

If shortwave absorption is increasing, the questions become, “How?” and, “Where?”.

NOAA ESRL has a long time series of atmospheric precipitable water (TPW) extending back to 1948. This time series shows a very small increase to the present and can be characterized as moving from higher values after WWII to a nadir in the late 70’s/mid 80’s, and increasing again to the present; much like surface temperature.

In the graphic above this TPW series is compared with RSS satellite lower troposphere (TLT) data. The correlation is not impressive, but there is reasonable agreement.

RSS has its own shorter TPW time series that excludes polar regions, and it is compared with ESRL TPW above. The correlation is poor, as will result from the trends being inverse from 1989 to 1997; but once again there is reasonable agreement. The burning question is what the satellites would have seen between 1948 and 1988. A reasonable supposition is that they would have generally tracked ESRL, with a poor correlation.

When the RSS TPW is compared with RSS TLT above, the correlation is good, but the trends appear divergent.

The point of all this is that if our increased shortwave absorption from CERES takes place in the atmosphere, it is probably going to be from increased atmospheric water.


We colored this Wikipedia image blue in the water shortwave absorption bands a while back. Oxygen, particularly in transition to ozone is a significant absorber in the SW solar spectrum. We will see when we get to MODTRAN that there is significant ozone formation near the surface as well as the stratosphere. CO2 absorption in the SW spectrum is limited to small areas near 2000 nm, shown green.

In the bands where the surface irradiance is decreased by atmospheric absorption, but some light still makes it through; adding more gas will increase the warming of the atmosphere and further decrease the warming of the surface. In the water bands where surface irradiance is already zero, adding more water will have no effect on the surface. These bands are about 19% of the total absorption by water.

The amount of light can’t change much. Like the sun, the earth radiates at an intensity and wavelength governed by temperature. In this sense one can think of the earth as the sun for outgoing longwave, and the atmosphere is caught in a crossfire between the two.

The crossfire becomes particularly wicked for reflected shortwave from the sun. In this case, the graphic above must be turned upside down. The atmosphere gets a second chance to absorb the photons that made it through the first time, and the longwave from the earth and the shortwave from the sun are both heading for space. The longwave that makes it out reduces the temperature of the planet, the shortwave that makes it out does not.

Even after all the atmospheric absorption and reflection from clouds, over half of the incoming solar SW makes it to the surface. Something like 7% is reflected from the surface, leaving just less than half absorbed by the surface. The surface is warmed, and radiates longwave out. Seventy percent of the surface is water, but as the graphic above shows, the shortwave absorption by water is at its lowest at the peak intensity of solar SW radiation.

The peak of solar intensity is in the visible wavelengths, and any diver knows that visible light penetrates to considerable depth, depending on turbidity. The ocean is still absorbing SW, albeit less efficiently than it might.

We really don’t have the tools yet to determine whether the atmosphere or the surface is where the CERES increase in SW absorption is predominantly taking place. Sea surface temperature, atmospheric temperature, and atmospheric precipitable water are all rising; together, as the laws of physics decree they must.

According to NASA (and Mr. Trenberth), the surface is radiating more energy up as longwave than the TOA receives as shortwave. We can address this, and longwave radiation to space, with MODTRAN in the next post.


Posted in Climate | 2 Comments


The earth does not burn at 5778 K like the sun, and therefore produces no shortwave radiation of its own. The earth radiates at about 288 K, a temperature that produces longwave radiation. When molecules absorb radiation according to their individual properties, higher energy shortwave radiation tends to produce “electronic” transitions where electrons jump to higher energy levels. Lower energy longwave radiation tends to produce rotational and vibrational transitions, where the atoms within the molecules move in relation to each other. Shortwave radiation is more likely to scatter than longwave.

In the last post we showed that longwave radiation from the top of the atmosphere has increased over the period of the CERES data. The earth is believed to receive a steady 340 W/M2 in shortwave radiation from the sun. About 30% of this is reflected back to space from light colored surfaces. The possibility remains that if this reflectivity or albedo were to decrease, increasing the shortwave energy influx, the increase in longwave radiation might not cool the planet.

Fortunately, CERES gives us a net TOA flux that includes shortwave in, shortwave out, possibly some longwave in from cosmic microwave radiation, and longwave out.


The net flux is defined as incoming minus outgoing, so positive values mean the planet is absorbing energy and negative values mean we are losing energy to space. This is opposite the LW flux, in which higher values mean greater energy loss to space. We see that net energy absorption, like LW loss to space, is higher in the northern hemisphere summer months, and that the planet loses energy in the southern hemisphere summer. The Min/Max values show a total range of some 27 W/M2. Photosynthesis consumes 5-10 W/M2 annually, depending on assumptions of conversion efficiency, and the far larger extent of photosynthesis in the northern hemisphere higher latitudes can explain part of this swing as increased SW absorption.



When we compare LW flux with net flux above, we see that net flux is always positive, in spite of increasing LW flux’s efforts to cool the planet. There appears to be a signal in both LW and net fluxes to the moderate el Niños of 2003 and 2010; and a surprisingly strong (and delayed) response to the weak 2006 el Niño. The very strong 2016 el Nino shows a strong LW response, but no net response. Unfortunately, the record does not extend back to include the comparable “very strong” Ninos of 1983 and 1998.



When SW flux is compared to net flux it becomes apparent that a very considerable decrease in SW loss to space (=increased absorption) has taken place in the CERES record. The increase in LW flux to space is on the order of 1 W/M2, and the decrease of SW flux to space is on the order of 1.5 W/M2 over the period of record.



It must be noted again that January and February data for the year 2000 and October to December data for 2017 have been infilled according to the “nearest living relative” approach; 2001 and 2016 data respectively. The 2017 data will eventually be corrected, the 2000 data probably not.

What we see above is that the SW flux added to the LW flux shows a decline generally in agreement with the increase in net absorption, as one would expect. There was some sort of aberration in 2005.

In the next post we will explore what can be discerned about where this increased SW absorption is going, before leaving SW radiation to focus on LW radiation and the disagreement between CERES and MODTRAN.


Posted in CERES, Climate | Leave a comment


You don’t radiatively warm the third rock from the sun without reducing long wave radiation from the top of the atmosphere to space. Period. [Unless shortwave absorption increases]

For far too long, we have ignored CERES data here at the Trunkmonkey Research Institute. Partly, this is because the data is only available in netCDF format developed by the “University Corporation for Atmospheric Research”. This format is good for painting raster images with color shaded values on 1×1 degree grid cell maps, but it is difficult to extract numeric values for use in other sorts of analysis.

At https://www.unidata.ucar.edu/software/netcdf/software.html, UCAR has published a long list of free and proprietary software for reading netCDF. We tried netCDF4Excel as there was some internet chatter that it worked well and we wanted the data in EXCEL anyway. We downloaded the latest version, only to be greeted with a popup that “this app cannot be used on this computer”. We also tried a cloud based app perporting to read netCDF to EXCEL, but EXCEL refused t open the output, claiming it was corrupt.

We use ARCMAP all the time and it is straightforward to paint color shaded rasters, but the ESRI toolbox utility supposed to create a table from netCDF  does not work with CERES data.

You can go to the CERES data and look at visualizations, and when you click on them this apparently very helpful screen pops up as above. Unfortunately, the “Save as ASCII” and “Save as PNG” buttons do not work. What you see is a screenshot.

Normally we would just digitize this, but the one thing that does work is when you mouse over the above screens, the data pops up. This particular graphic is so important that we deemed it worth the trouble, and while entertaining vile thoughts about the NASA nerds, duly recorded all the monthly values in a notebook.

The CERES data above comes from the TERRA (before 2002) and then the AQUA Moderate Resolution Imaging Spectroradiometers  (MODIS) instruments. The satellites are just above 70 Km altitude in sun synchronous polar orbits that cross the equator at the same time every day. The TOA is considered to be 20 Km altitude.

The first striking feature of the above data is the large swings in TOA longwave radiation to space between the northern and southern hemisphere summers. As the min/max values in the screenshot show, the range is some 12.5 W/M2, with the northern hemisphere always radiating more to space than the southern hemisphere in their respective summers. Land must radiate to space more effectively than ocean.

The second striking feature of the above data is that it does not seem to be decreasing as MODTRAN predicts from the some 2 W/M2 additional absorption from human CO2.

It is actually increasing. Above is the raw data, but this overstates the case a bit because the data ends in September, 2017. We wish the data continued to finish 2017, and as we approach the end of the first quarter 2018, it does not seem unduly burdensome for NASA to provide the data. Anyway, there are various mathematical tricks one might use to correct this, all with drawbacks. What we have done below is arbitrarily complete 2017 with 2016 monthly data.

The result looks like this. There has been a substantial increase in the long wave radiation to space in the CERES data. This is how you cool the planet, not warm it. This data glaringly contradicts MODTRAN to the extent that one or the other is wrong. Exploring this discrepancy will be difficult because the formats are different, but it is very important. The discrepancy between CERES and MODTRAN will be the subject of this series.

Posted in Climate | 3 Comments