Friday, February 23, 2007

Proof that warming in Western Australia is all about the sun

I promised you some good analysis on Southern Pilbara (central-west Western Australia) yesterday when I analysed the temperatures and I keep to my promise so here it is:

Yesterday we came up with the very unusual result that despite the fact that there was no significant increases in temperature from 9pm to 6am in southern Pilbara - and not even any recogniseable patterns – we found that the area recorded a significant increase in minimum temperature. This is very surprising. We also found that in the area temperatures were significantly higher from 9am thru to 3pm when the sun is at it’s hottest.

At first look it’s quite clear that the reason southern Pilbara is heating up is solely due to the sun. The sun is just getting damned hotter. So why the increase in minimum temperatures, especially when there is no difference in temperatures throughout the night?

I decided to have a look at the differences in temperature anomalies for neighboring times over the years. In other words, I looked at the temperature anomalies for Midnight minus 9pm. As previously noted, 9pm doesn’t have a lot of data for Southern Pilbara so it is hard to find a pattern, and the statistics suggest no difference (t = 0.24, p = 0.81). In other words, Midnight has not been heating up at a quicker rate than 9pm over the years in Southern Pilbara.

The same goes for 3am minus Midnight, only this time we have a good amount of data. There is no pattern. Temperatuers at 3am have not been heating up quicker or slower than Midnight over the years (t = 1.4, p = 0.15).

But wow, look at this. When looking at 6am minus 3am, we find that temperatures have increased at a significantly higher rate at 6am than 3am (t = 5.3, p < 0.01). Keeping in mind that we did not find a significant increase in temperature at 6am, however we have found that the temperature increase at 6am is significantly greater than 3am. This has occurred with especial magnitude in the last 10 years as shown in the graph.

And a similar pattern occurs when looking at the 9am – 6am anomalies. A significant increase occurs (t = 6.04, p < 0.01). With temperatures increasing on average 0.5 degrees more in the past 10 years. If you look at the 6am and 9am graphs, this makes sence, as in the last 10 years 6am was about average temperature whilst 9am temperatures were about 0.5 degrees above the norm.

Surprisingly this is where the large boom stops. Analysis of Noon minus 9am temperatures suggest no increase or decrease in temperature (t = -0.6, p =0.55). A cyclic pattern can be seen, but this could be purely due to random variation.

And now for the even more surprising results. 3pm minus Noon saw a significant decrease in anomaly temperatures (t = -8.1, p < 0.01). The decrease is almost perfectly linear and is clearly obvious. So despite significant increase in temperature for south Pilbara for Noon and 3pm, 3pm isn’t heating up as much as it has been 3 hours before.

The obvious linear line in the graph of 6pm to 3pm is startling. So perfect is the line, that the amount of variability is very small. The decreasing trend is very significant (t = -10.8, p < 0.01).

There was no significant difference in temperatures when looking at the differences between the 9pm and 6pm anomalies, however a smaller database for temperatures at 9pm could have been a factor (t = 1.3, p = 0.18).

So what does all this mean? How is this all relevant. Well let’s summarise what we have just found. The rate of increase or decrease in temperatures with respect to the time 3 years prior makes little difference at night, but when the sun is a factor the difference is significant. Temperature anomalies at 6am and 9am have increased significantly with respect to the time 3 hours previous over the years, whilst temperatures at 3pm and 6pm have decreased significantly with respect to the time 3 hours previous.

But to understand fully what this means, we have to discuss very briefly how daily temperature works. Basically we reach a maximum temperature, and from that point on the temperature generally decreases. When the sun as set, there is no general reason why temperatures will be on the rise, so they decrease until the sun makes an appearance again. In southern Pilbara the sun rises in general half way between 6am and 9am depending on the season.

We suggested before that is was relatively obvious that the sun was a major factor in determining the average maximum temperature in Southern Pilbara, as we saw no increases in temperature over night. We have also proven that the sun has made significant increases in maximum temperature rather than minimum during years when Australia is heating up. So would a stronger sun therefore also have an influence on minimum temperatures? Our analysis suggests so.

Whilst no increases in temperature were found for Southern Pilbara from between 9pm and 6am when the sun was set, at 6am, the temperature increase as compared to 3am was highly significant. The sun, whilst not risen, quite possibly was warming up neighbouring areas more than normal, and adding heat to the air above at a greater rate than normal. Thus paving the way for when it makes its grand entrance at sunrise.

Hence, whilst not making a significant temperature difference at 6am before sunrise, it made a significant increase in temperature compared to 3 hours prior at a time where normally the minimum temperature would have been reached. This extra layer of heat paved the way for massive increases in temperature at 9am after the sun had risen. Following, due to the massive increases in temperature 9am, Noon and 3pm also had increasing temperatures due to the intensity of the sun, but not quite as dramatic as that when the sun rose. Hence we have a situation where we have a significant increase in temperature at 3pm but a highly significant decrease in temperature at 3pm with relation to Noon. I guess there’s only so much difference that the sun can make.

So what can we conclude about Southern Pilbara? That increases in the suns intensity has caused the area to heat up during the day, to stay at a constant temperature during the night, and most importantly, that increased intensity in the sun has not only caused increases in maximum temperatures but also increases in minimum temperatures.

Who would have thought that the sun has something to do with heat?


Anonymous said...

Very good analysis Jonathan but be cautious about saying "the sun is getting hotter" when the real cause might be variations in cloud cover.

Sadly the ISCCP data has ended and I can't find a replacement anywhere. It was very useful data because it split the cloud into three layers and then into cloud types. What you describe could be due to a reduction in low level cloud (dense and cooling) or maybe an increase in thin high level cloud which traps heat.

From about 1999 the 12m running average of low level cloud cover from 30 to 90 degrees latitude in both directions decreased while mid-level increased. This could explain a lot...

Jonathan Lowe said...

nice comment, anonymous, do you have any extra links with this information and do you know of any reasons for this change in cloud cover apart from random variation?

Tom Davidson said...

Pedantic quibble: Your header states "I intend to provide statistical analysis into the area to prove if the recent well advertised rise in temperature is at all statistically significant."

Empirical science does not normally start out intending to "prove" anything, but rather to establish what the reality is. Statistical analysis can indeed determine whether anecdotal reports are significant, and is probably one of our best tools for this purpose.

The difficulty at the time appears to be that many people (scientists included) are innumerate (illiterate with respect to quantitative analysis of data), and completely incapable of coping with a situation in which a single effect can have multiple contributing factors that act concurrently.

Then there is also the problem of extrapolation of mathematical models beyond the range of applicability of the assumptions upon which the model is designed.

Roland Mahoney said...

Further to anonymous' cloud cover comment, a change could indeed explain a lot. But, the effects of cloud cover are unlikely to be limited to the daylight hours. The increase in temperature that Jonathan noted during the day should have been observed at night, if cloud cover were a significant explanation.

Jonathan Lowe said...

Good point Tom with regards to the header, a pedantic point as you note but something I should change, you might also want to read my tongue in cheek article about model extrapolation amongst many things with regards to temperature differences at 1966. Link:

And Ronald good point indeed, can "ananymous" reply to this point please?

phil_b said...

I recall minimum temperatures are generally reached just after dawn, before the sun's heat is greater than radiative cooling.

What you seem to have shown is an increase in the sun's energy early in the day when the sunlight must travel through more of the atmosphere relative to later in the day.

Decreased cloud cover could explain this, as would a decrease in atmospheric particles (dust).

I'm familiar with the Pilbara and human impacts are minimal (a few mines and the odd sheep station), which makes me think the phenomena is not local.

Have you looked trying to relate this to 'global dimming' caused by increased particles in the atmosphere and its recent reversal?

Nice blog, keep up the good work.

Count Iblis said...

This is 100% nonsense. The increase in temperature due to climate change is too small to be detected by your analyses. You would have to average over many hundreds of weather stations to see an effect like 0.6 degrees C per century.

If you do see a significant temperature change at some location then that can only be a large temperature change due to changing local circumstances which has nothing to do with climate change, the Sun or whatever.

Phil_B said...

Count Iblis, you clearly haven't done much in the way of statistics.

Count Iblis said...

"Count Iblis, you clearly haven't done much in the way of statistics"

I admit that I don't have a lot of "hands on" experience with statistics, but I do know what I'm talking about (I'm an expert in theoretical physics, I know a lot about statistical physics).

But statistics of temperature records is completely irrelevant here. We already know from observations of the Sun that there is no significant increase in the solar output that could even account for the rather small observed global temperature increase, let alone the huge increase Jonathan is talking about.

Such an increase would have to affect the whole Earth, not just Australia and it would necessarily have to have enormous effects on the climate.

The problem is not that I don't have experience with statistics, but rather that Jonathan doesn't know much about basic physics.

Anonymous said...

I think what Jonathon is trying to say that if CO2 were the issue then we would get relatively constant warming all day and everywhere. Doesn't seem to be happening in Australia with some times not increasing, some increasing, some decreaasing.

Your comment about too small a sample to discover climate change doesn't really make sence. If most local places were chaning temperatures due to local conditions only, then if we analyse them all together we conclude that its CO2 levels because we have a bigger sample size? Why couldn't it still be local conditions?

Count Iblis said...

Apart from reducing statistical fluctuations, if you average over the whole Earth then the result becomes insensitive to redistribution of heat. Suppose that local conditions change here because warmer of change in wind direction which brings cooler or hotter air here. But that heat comes from another place...

The signal from climate change is visible only if you average over many hundreds of weather stations spread out over the whole Earth. It is not visible in the data from each individual weather station.

Anonymous said...

so count, you are suggesting that local conditions play a greater part in climate change than CO2 levels? (based on your last posts). Interesting.