Wednesday, November 22, 2006

The Sun causes warming? no, surely not!

Recently we showed that when Australian maximum temperatures increased, the actual temperature only increased when the sun was out. Likwise, from 1947 to 1970 when temperatures decreased, the decrease only occured when the sun was out. Hence, when maximum temperatures are up or down, we are not seeing an increase/decrease throughout the day of temperatures but only at the heat of the day (around 3pm) when the sun is at it's hottest.

Maybe the sun has something to do with the discrepancies in maximum temperatures perhaps? Well Scafetta and West's (2006) research seems that it could agree with us saying that:

The sun might have contributed approximately 50% of the observed global warming since 1900

50% Wow! That's like....half. Maybe there is something in this. Which would you bet causes more warming....CO2 levels or the sun? Hmmm....


millers said...

good blog

keep up the good work

Anonymous said...

There is certainly a bias in climate change research towards explaining it ie caused by man.

For what is caused by man can be controlled. Hence more research (funding) and government is the remedy.

Still, explain to me how the sun is hottest at 3 PM. Air temperature tends to peak about then. But air termperature is not sun radiation.

Jonathan Lowe said...

well maybe not the sun hottest at 3pm, but this is when the sun probably has the most influence over the temperature, and at around this time, we reach a maximum....

Steve said...

While I'm happy to see a thorough analysis of temperature variations, I'd like to make a point or two. Statisticians are not atmospheric physicists, or vice versa. I am neither statistician or atmospheric physicist. I am a solar-terrestrial physicist with a strong statistical background. Statistics, without full consideration of the context, are often very misleading. Any use of statistics is based on a set of assumptions, which may not be valid. Though the analysis may be statistically correct (or not), that doesn't mean it has scientific value.

For example, your use of the "local time", since many of your posts argue that this (rather than daily maxima or minima) is of great importance: It's important to realise that local civil time (as defined by time zone) is not the true local time of any particular weather station. Sydney, Melbourne and Mildura all have the same local civil time. However, looking at their longitudes tells us immediately that Sydney (151.2E), Melbourne (144.97E) and Mildura (142.2E) are at local times of GMT+10.08, GMT+9.66, and GMT+9.48 hrs respectively. This is a difference of 36 minutes between Sydney and Mildura, and 25 minutes between Sydney and Melbourne.
Further, local effects like low- and high-altitude cloud cover will vary through the day. Coastal sites are next to a large heat sink (with variable current temperatures and sea surface temperatures). Are fixed civil time air temperatures meaningful in this analysis?

If we then realise that the air temperature records constitute a persistant series (i.e. they are autocorrelated to some degree, more temporally than spatially), then a simple estimate of your p-value is unjustified. In a persistant series, your effective number of degrees of freedom is much reduced, thus any conclusions about the significance should be re-evaluated taking the level of autocorrelation into account.

As for the solar forcing of climate: I won't deny that there are clear links. But weather isn't climate, and there has been a substantial effort in recent years to examine the solar contribution to climate forcing. The majority of these studies have shown small increases in the total solar irradiance (the energy flux per square metre at Earth's mean orbital distance), though some have failed to find a significant trend in this data. Further, the appearance of the 11-year periodicity in solar activity (expected if the solar heating directly drives daily temperature) is contentious (at best) and little evidence has been found. There are complex lags in climate, and probably the most effective solar forcing of climate comes from cosmic rays. How? Earth is shielded from cosmic rays by the Sun's outer magnetic field. This is modulated on a 22-year cycle (polarity reversal every 11 years). The cosmic rays are hypothesised to seed low-level cloud formation, and to interact with the global atmospheric electric circuit (also affecting thunderstorms). As you can imagine, the feedbacks are complex. Still, current estimates appear to be converging on a solar forcing of about 30% of recent recorded changes in global mean air temperature.

But let's assume it's 50%, as the study you cited suggests. That means that half the change is not externally driven. Should we just shrug and say we should ignore it?

I'd be inclined to say that it's no surprise that the Sun impacts global temperature trends; That solar forcing is masking the true extent of the man-made problem, that's the real surprise.

I'll be bookmarking this blog.

Jonathan Lowe said...

Nice response Steve,
I agree with you entirely on the true time that measurements are taken, in regards to local and physical times. This is definetly something that I will be looking at more closely. HOwever, to do this the data needs to be a lot more accurate. There are several stations that record temperatures every 10 minutes or so, and only these therefore would be adequate for such an analysis. There are not many of these, and the analysis might somewhat be limited. But despite this it is something that I intend on doing. One can also look at similar placed stations where their physical time would be similar. This might solve this problem.

Irrespective of this however, keeping the time constant instead of just looking at maximums and minimums is a step closer to evaluating the true degree of temperature increase which has not, regretably been done before.

steve said...

Exploring temperature anomaly by time of day certainly is a nice approach. That it hasn't been done before is strange... although, as you pointed out, just because the main variation in temperature arises in the early afternoon (3 hour lag after Sun's highest position, as expected from a simple heating argument) doesn't mean the mean temperature isn't changing. Actually, if we're interested in the temperatures required for stable ecosystems, aren't we interested in the extremes? If it was -30 overnight and +50 through the day we'd have a mean temperature of 20 degrees, which sounds nice (but wouldn't make anyone very happy). Perhaps separating temperature anomaly by time of day will highlight the change.

To a science question (and I realise this may be a touch off topic). This graph shows that the daytime temperatures are increasing, but the overnight temperatures are not. It would seem reasonable (as a first estimate) that this may be due to more direct solar warming. Why should this be the case? In other words, what changed that more solar radiation is getting to the ground? Why has the local albedo decreased?

This is one for climate scientists to argue over, but albedo is strongly dependent on cloud cover (esp. the altitude of the clouds) atmospheric CO2, etc.

Anonymous said...

Okay, the sun is the source of the heat but it is the atmosphere that makes earth a place were u can live
Without the atmosphere the earth would be a cold place and it is the CO2 that makes the sunbeams stay so it is hot enough, but now it is to much CO2 so the temperature rise. That why is called global warming, the sun is still so hot it was for centuries ago.
Read about global warming to understand better
I am just 16 Swedish boy going in a ordinary school