Say Hello to La Niña Conditions

Well I for one have never been comfortable accepting numbers emanating from a louvered owl / bat box sitting some 60 inches above the ground, however high the ground sits, as some credible measure of earth’s “global Temperature”, or even its global Temperature anomaly.
Despite multiple repetitions of the thermodynamics admonition that anything that is at a Temperature higher than zero kelvin, emits thermal radiation depending on its Temperature (and generally the fourth power of that Temperature), climate scientists and others keep insisting that normal gases do not. Normal in the sense of mono-atomic gases (like the noble gases) or diatomic like H2, O2, N2.
Well nobody claims they emit black body radiation; nothing does that.
But real bodies at some uniform stable Temperature > 0.000 K can approximate the theoretical BB radiation, along with an emissivity less than 1.000, or more commonly a spectral emissivity < 1.000 that also varies with frequency / wavelength.
But certainly, solids and liquids are more common thermal radiation sources.
Well keep in mind that the definition of a black body, is a body that absorbs 100.000 % of any and all electromagnetic radiation from zero frequency up to zero wavelength.
NO real body absorbs 100% for even one single wavelength, let alone all possible wavelengths.
When was the last time that you observed any gas in the process of absorbing 100% of any radiation. Obviously, gases with their low atomic / molecular density are not strong absorbers, so at Temperature equilibrium with the radiation field, they are also not strong emitters; just very low emissivity emitters.
So bottom line, I prefer to think of the solid or liquid surfaces of the planet, as being the actual radiant emission sources, and it is their Temperature which should be measured; not the air five feet off the ground.
At the hottest SSTs the standard BB spectrum should peak at around 9-10 microns when plotted against wavelength.
Note that wavelength scaled radiation spectra are normally plots of Spectral Radiant Emittance measured in Wm^-2 per micron (m ?) of wavelength increment.
When plotted against frequency or wave number (f or nu) the graph is a plot of wm^-2 per wave number increment (cm^-1).
Remember that there are more microns available at longer wavelength scales, and more wave numbers available at higher frequency scales, so a wavelength spectrum that peaks at say 10 microns wavelength would give a wavenumber plot that peaks closer to 15 microns; which happens to be near the CO2 bending mode band.
Chemists seem to like wave numbers, and physicists seem to like wavelengths. (well this one does).
BUT I also have an aversion to the ground based Temperature anomaly data gathering systems, as opposed to the newer (since 1979 ?) satellite based systems, which apparently measure Temperatures at various atmospheric altitudes. I don't know if the satellites can directly measure ground and SST numbers.
My preference for the satellite numbers, is based on a deep suspicion that the surface station measurements are highly flawed from the point of view of sampled data system theory. The Nyquist gremlin.
The Nyquist sampling theory dates from 1928 (Bell Labs). 1928 and Bell Labs is also the origin of Harley's Law that relates channel capacity and bandwidth of communication systems. The two are intimately related.
Communications Engineers are familiar with the concept of "periodic" sampling of a band limited continuous function (signal), presumably of a bandwidth (B). Well it's a theoretical concept anyway, because no such thing as a "band limited" signal exists.
We know from Fourier theory, that a single frequency signal (sine wave) which has zero bandwidth, also exists for all time. You can't turn it on or off, and have it remain band limited.
Real signals start at some point in time, and end at some later time, perhaps (T1) seconds later, so they cannot be single frequency sine waves.
Nevertheless, standard sampling theory says that a real band limited (to the extent it can be) signal can be sampled at a minimum frequency of (2B), and from those samples, the original continuous signal can be completely recovered. NOTE: the samples are considered to be instantaneous time values, although in practice a finite sampling pulse width is normally used. So a sampled signal can be considered to be an amplitude modulated periodic stream of unit area impulse functions.
So when we consider the ground based sampling in both time and space, of the Temperature anomalies at some selection of station points; it is immediately apparent, that this process is neither strictly periodic in either time or space.
So how can these sparse observations (non satellite) be remotely valid samples of the real global temperature function ??
Well they can't; and they aren't.
Well the practitioners of these aberrant procedures respond that they do not wish to reconstruct and recover the original continuous band limited data signal.
They only wish to generate the response of various and sundry statistical mathematical algorithms to these inadequate sampled observations.
So they want averages or trends, and who knows what else; but they assume they can recover those from the improperly sampled observations.
The average is easy to deal with, because the average value of the original continuous function, is simply the zero frequency component of the continuous band limited signal; the DC component.
Well standard sampled data theory shows us, that if we only sample at a frequency of (B) or less, instead of (2B) or more, then the reconstructed output will contain spurious aliasing noise components at DC (and negative frequencies), so such a sampling regimen, will not even allow us to determine the average value, without error.
Many years ago (1960s) it was found that the requirement for strictly periodic sampling at a frequency (2B) or higher was overly restrictive. Within limits, the samples could be randomly spaced in time.
How do I know this ?? Well because back then, Tektronix manufactured a Random Sampling Oscilloscope, based on that idea. Why ??
Well a conventional sampling oscilloscope, needs to be triggered from the signal or the source of the signal, and trigger circuits can be very tricky, because they tend to have to out perform the vertical amplifier circuits, and too often, trigger circuits failed miserably to do that.
Also, with a triggered system, you then need to delay the signal before applying it to the vertical deflection plates, because you need to get the sweep started sweeping across the screen; so you need a delay line to delay the vertical signal, and that also can limit the bandwidth.
Apparently with random sampling, you let the sampler free run at some mean frequency with some controlled spread, and eventually it will catch the required information in a valid fashion, and somehow the delay line need disappears too.
So evidently strict periodic samples at a spacing of 1/2B, are not necessary.
So surface Temperature monitoring stations, may not be close enough to each other; but how far apart can they be ?
Well Dr. Hansen says they can be 1200 km apart and still work.
A more general statement of the Nyquist sampling theorem reveals the following.
For ANY group of INDEPENDENT samples.
Any 2B samples per second will permit correct recovery of a signal band limited to a frequency (B).
This can be stated another way. To recover an original signal that was T1 seconds long, you need a total of 2BT1 independent samples (or pieces of information).
I presume that the same more general condition is also required to be able to extract any of the statistical mathematical algorithmic bric-a-brac, that one might seek to create out of such observations.
So I don't know if RSS and UAH can actually measure true surface Temperatures (NO I am not bamboozled by arguments that their satellite thermometers are not calibrated; that's a red herring). I just don't know if they do report surface data; but I think their satellite scanning capability is far more important than whatever their restrictions are.
G

Well…
There are plots of global warming altitude vs temperature which show the effect is strongest in the surface layer, attenuates for 8 km, is back up to the surface level for 8 km then goes rapidly negative.
I interpret this as the atmosphere above 100 meters has only 1/2 the expected influence and half or more of the effect is “throttled”.
The TLT as I understand it is really the 3 km temperature.
The ground stations have strong UHI effects and station aging effects that allow algorithmic gaming to produce artificial warming.

PA July 18, 2016 at 11:05 pm
Well…
There are plots of global warming altitude vs temperature which show the effect is strongest in the surface layer, attenuates for 8 km, is back up to the surface level for 8 km then goes rapidly negative.
I interpret this as the atmosphere above 100 meters has only 1/2 the expected influence and half or more of the effect is “throttled”.
The TLT as I understand it is really the 3 km temperature.
It rather depends on which TLT you mean.
Spencer doesn’t use the old LT version 5.6 any more but that had a weighting curve unto ~12km peaking at ~2km, whereas the new version 6 goes up to ~14km and peaks about 4km.
RSS TLT had a similar weighting function to the version 5.6 whereas their TTT is similar to version 6. The TLT product proved to be difficult to adjust for drifting measurement times and ‘should only be used with caution’.

“When the gods want to punish you they answer your prayers.”

I calculate from ***UAH TLT Global***
1998 was 5.79 deg.*months (from the baseline) anomaly.
So far 2016 is 3.70 deg.months, through June.
That means than the remaining 6 months must be +0.35 on average, per month, to tie 1998.
The current UAH value is 0.34, and dropping like a rock. So getting back up to 0.35 and holding there is not plausible.
There is one way that 2016 can be the “warmest year evah”. If the Alien Invasion Fleet arrives next month and uses their hyperblasters to melt our cities and boil the oceans, that would make it warmer. Other than that, not likely.

OTOH, 2010-2011 was a moderate La Niña, and we had one of the heaviest snowfalls in the Lake Tahoe area on record that winter, roughly twice the average snowfall at Squaw Valley, CA, much of it coming in March:
http://squawalpine.com/skiing-riding/weather-conditions-webcams/squaw-valley-snowfall-tracker/?tab=ui-id-6