Arguments
Australia’s hottest year was no freak event: humans caused it
Posted on 17 January 2014 by Guest Author
This is a re-post from The Conversation
The Bureau of Meteorology has confirmed that 2013 was the hottest year in Australia since records began in 1910.
Unusual heat was a persistent feature throughout the year. For the continent as a whole, we experienced our hottest day on record on January 7. Then January was the hottest month on record, and the 2012-13 summer was the hottest recorded for the nation.
The nation-wide temperature record set for the month of September exceeded the previous record by more than a degree. This was the largest temperature anomaly for any month yet recorded.
Averaged across all of Australia, the temperature for 2013 was 1.2C above the 1961-1990 average, and well above the previous record hot year of 2005 of 1.03C above average.
What caused these extreme temperatures? Climate scientists have a problem: because climate deals with averages and trends, we can’t attribute specific records to a particular cause.
But our research has made significant headway in identifying the causes of climate events, by calculating how much various factors increase the risk of extreme climate events occurring. And we have found sobering results.
We previously analysed the role human-caused climate change played in recent extremes across Australia.
For various record-breaking 2013 Australian temperatures, we investigated the contributing factors to temperature extremes using a suite of state-of-the-art global climate models. The models simulated well the natural variability of Australian temperatures.
Using this approach, we calculated the probability of hot Australian temperatures in model experiments. These incorporated human (changes in greenhouse gases, aerosols and ozone) and natural (solar radiation changes and volcanic) factors. We compared these probabilities to those calculated for a parallel set of experiments that include only natural factors. In this way, natural and human climate influences can be separated.
In our previous studies, we then applied an approach (known as Fraction of Attributable Risk) widely used in health and population studies to quantify the contribution of a risk factor to the occurrence of a disease. Health studies, for example, can quantify how much smoking increases the risk of lung cancer.
Using the climate models, the Fraction of Attributable Risk (FAR) shows how much the risk of extreme temperatures increases thanks to human influences.
In our earlier study of our record hot Australian summer of 2012-13, we found that it was very likely (with 90% confidence) that human influences increased the odds of extreme summers such as 2012-13 by at least five times.
In August 2013, Australia broke the record for the hottest 12-month period. The odds of this occurring increased again from the hottest summer. We found that human influence increased the odds of setting this new record by at least 100 times.
Recent extreme temperatures are exceeding previous records by increasingly large margins. The chance of reaching these extreme temperatures from natural climate variations alone is becoming increasingly unlikely. When we considered the 12-month record at the end of August, it was nearly impossible for this temperature extreme to occur from natural climate variations alone in these model experiments.
We have just completed a preliminary investigation of contributing factors for the record Australian temperature in the 2013 calendar year.
In the model experiments, it is impossible to reach such a temperature record due to natural climate variations alone. In climate model simulations with only natural factors, none of the nearly 13,000 model years analysed exceed the previous hottest year recorded back in 2005.
Australian annual temperature changes (relative to 1911-1940 average) for observations (dashed black) and model simulations with natural influences only (green) and with both human and natural influences (red). The grey plumes indicate the range of values simulated across nine global climate models used. Average Australian temperature anomalies are indicated for 2013 and the previous hottest year on record in 2005. David Karoly & Sophie Lewis
Probabilities of annual average temperatures for Australia from climate model simulations including natural influences only (green) and both natural and human climate influences (red) for model years 2006-2020. The vertical lines show the temperature anomalies observed in 2013 and in 2005 (the previous hottest year observed). David Karoly & Sophie Lewis
Our extensive catalogue of 2013 record-breaking events in Australia occurred in a global context of increasing temperatures that must be considered. Globally, 2013 will likely rank as the 6th hottest year recorded.
So to return to our question, what caused the 2013 record hot year across Australia? Simply put, our climate has changed due to human activities. Recent extremes, such as this hot year, are occurring well outside the bounds of natural climate variations alone.
bruiser,
reference to "1990 - 2013" is not my comment.
I'm not sure you read my post clearly. I didn't "go back 80 years." All the records were pre-1909, and half of them were from the late 1880s. The reason a handful of them are little more than 80 years long is because the record truncates in the 1980 or 90s. I took your reference to pre-1900s and early 1900s as the basis for the selction.
The majority of sites are rural. For many of them I checked the siting of the weather station.
In the records I saw 1888 or 89 were often warm, occasionally warmest, but for the most part the average decadal warmth of that period was less than in the last decade or two of the record selected, whether the period stretched to recent temps or the 1980s/90s.
The topic we were discussing was the influence on solar exposure on the 2013 record. Your argument seems to have shifted to a general proposition that Australia is no warmer than it was more than a century ago. Your argument is based on very spotty data with obvious problems (and you have cited precisely one weather station), vague assertions about UHI (which the Australian record attempts to correct for), and absolutely no number-crunching or even casual testing of your opinion as I did. As the global record is quite robust on warming from 1880, with all large-region (continental-scale) records showing significantly warmer temps now than then, your assertions are not at all convincing. (Not enough data for the Antarctic, but it shows warming from the 1950s, corroborated by analysis from 'skeptics')
I provided a list for you of Australian weather stations with long records. Amongst them is a usable subset that roughly covers the periods you have nominated. The station numbers are provided, which you can type into a box at the BOM page you linked me to, and immediately bring up the annual min/max temperture series, each of which BOM will chart for you with a click. Why don't you select all that fit the period, cherry-picking none, and see for yourself? It's no more than two hours' work. Confirming the locations for those stations takes ony a little bit of googling. There is a site that briefly describes the locations, and plots them on google maps for you. Frequently the description I read was in the vein of "the station is located 650 metres from an unnamed sealed road."
If you take the trouble to investigate for yourself, I will repeat what I did the other day, this time noting down the weather station numbers and providing them to you so that you can cross-check that I selected appropriately.
To reemphasise a point, our excursion has little to do with the analysis given in the main article, which would be unaffected by pre-1910 data. If you're not sure why, I recommend reading it carefully.
Something nopbody has mentioned is that solar radiation reaching the surface is only one factor in how much energy is available to heat the overlying air (which is where air temperature comes from).
Nobody has mentioned the role of soil moisture, which plays a critical role in the division between thermal transfer to the atmosphere and evaporation (latent heat) transfer to the atmosphere. Dry soil means less evaporation means more thermal transfer means hotter air.
An less evaporation combined with warmer air often means less cloud, which lets more solar radiation through to the surface. HIgher solar radiation could be the result of the hot, dry conditions, not the cause of it.
And "more water vapour means more cloud" is not necessarily true, either. To form cloud, you need to take moist air and cool it to the condensation point. If the air is both more moist and warmer, it may be further from the condensation point (dew point temperature) rather than closer.
On a global scale I have been assuming that the increased atmospheric water vapour would lead to more clouds in general (and more precipitation). Is that incorrect? I did point out that effects would be different regionally, and was thinking of evaporative potential.
How does this work in the tropics? I have spent a lot of time in very hot places like Malaysia, and the air is very warm day and night with lots of cloudiness.
There are many confounding factors. If high clouds tend to augment the greenhouse effect, and low clouds tend to feedback negatively by reflecting solar energy, then a reduction of one more than the other has different consequences for surface temperatures. A change in solar exposure largely due to fewer high clouds would have an effect inverse to what we've been assuming. I don't know to what degree solar exposure is determined by high/low cloud changes.
I guess it must be near impossible to isolate and tally attributing components to global, and especially regional temperature on annual and decadal periods. The long-term probability estimate seems to be a more robust way of estimating the impact of natural and anthroopogenic factors. But bruiser raised an interesting point on solar exposure.
barry - On a global scale it appears that relative humidity is quite constant over warming, closely following the Clausius–Clapeyron relation with constant relative humidity. For example, Dai 2006 reviewed observations from some 15,000 weather stations and ships finding that "During 1976–2004, global changes in surface RH are small (within 0.6% for absolute values)...". Clouds are tied to that relative humidity; globally changes in RH are small.
There are, however, significant regional variations - the wet getting wetter, the dry getting drier, as increased absolute humidity and climate energy amplify the hydrological cycle. And the vertical structure of the atmosphere, affecting the balance of high (warming) and low (cooling) clouds is changing regionally as well. And regional changes such as the ones bruiser has raised could quite possibly be tied to those.
barry:
KR has mentioned relative humidity (RH). While it has its place in humidity measurement, it is not an absolute quantity (the hint comes from the use of "relative" in the title), and misses out some key components of water vapour in the atmosphere.
The question is "relative to what?", and the answer is "the saturation humidity at the current temperature". The Clausius-Clapeyron relation tells us the saturation quantity as a function of temperature - it's roughly exponential, with higher values at higher temperatures. A relative humidity of 65% means that the absolute humidity is 65% of the saturation value at the current temperature.
Humidity can be measured in several related combinations of units that aren't "relative":
- vapour pressure (partial pressure of the water vapour gas)
- specific humidity (ratio, weight of water vapour to weight of dry air)
- absolute humidity (ratio, weight of water vapour to weight of moist air)
- dew point (temperature at which the current absolute humidity equals the saturation value)
Now, to get back to the cloud issue:
- warmer air can hold more water vapour before it reaches saturation
- air that isn't saturated needs to be cooled to the saturation point before clouds can form
in comparing two masses of air, the one that is "warmer and more moist" may reach saturation at a higher, lower, or the same temperature, depending on how much warmer and how much wetter (and this also depends on where along the exponential Clausius-Clapeyron curve you are).
In weather/climate, the three common ways of cooling air to form clouds all involve adiabatic cooling: as air rises, the pressure drops, and cooling occurs. The three ways of getting air to rise are:
- free convection (heating from the ground, heated air rises through overlying cooler air, due to density differences)
- push the air up over a mountain (orographic precipitation)
- push warm, moist air up over cold dense air (along fronts between air masses. Happens in cyclonic storms.)
So, if the air becomes more moist (in absolute terms), we also need to know if it is warmer or cooler. Let's take one example where the air is further from the saturation point. You can get a combination of:
- clouds won't form (doesn't cool enough as it rises - or in other words, doesn't rise enough)
- clouds form at a greater height
- clouds form a a similar hieght, but in a different form
Same options (in reverse) if the air is closer to saturation.
In summary, yes it is complex, and this is why cloud feedbacks are difficult to estimate wrt climate change. All evidence so far is that cloud feedback effects are small, however (globally-averaged).
I knew that clouds were a complicated and uncertain factor and this adds some helpful detail. Thanks for the comments.