As I keep losing track of these, I want to put them all in one place:
UAH
http://vortex.nsstc.uah.edu/data/msu/t2lt/uahncdc.lt
Hadley
http://hadobs.metoffice.com/hadcrut3/diagnostics/global/nh+sh/monthly
GISS
http://data.giss.nasa.gov/gistemp/tabledata/GLB.Ts+dSST.txt
NCDC
ftp://ftp.ncdc.noaa.gov/pub/data/anomalies/monthly.land_ocean.90S.90N.df_1901-2000mean.dat
RSS
http://www.remss.com/data/msu/monthly_time_series/
STAR mid-troposphere data (this is not directly comparable with UAH, but can be compared with the RSS mid-troposphere data).
ftp://ftp.orbit.nesdis.noaa.gov/pub/smcd/emb/mscat/data/v2.0/monthly/
Wednesday, December 22, 2010
Monday, December 13, 2010
Oops ...
I have made an embarrassing discovery: I have been using the wrong rainfall data for Canberra. The rainfall data that I have been using was from here: http://www.bom.gov.au/climate/dwo/IDCJDW2801.latest.shtml
I had made the assumption that this data was the same as the date from here:
http://www.bom.gov.au/jsp/ncc/cdio/weatherData/av?p_nccObsCode=139&p_display_type=dataFile&p_startYear=&p_stn_num=070014
This is not the case, as I would have realised had I been paying attention:
"Source of data
Observations were drawn from Canberra Airport {station 070351}.
Weather observations for Canberra were previously taken at a nearby site {number 070014}. These can be seen on the "Canberra Airport" Daily Weather Observations; you may need to consult these to get all the relevant data."
My predictions were based on the data from station 070014, as the data from station 070351 only goes back to the middle of 2008. This can be seen here:
http://www.bom.gov.au/jsp/ncc/cdio/weatherData/av?p_nccObsCode=139&p_display_type=dataFile&p_startYear=&p_stn_num=070351
So, at the moment it is not clear that my prediction has been falsified. I think that it has, but I will not know for certain until some time early next year.
To be sure that this is the data that I used, please read this post:
http://evilreductionist.blogspot.com/2010/01/future-of-rainfall-in-canberra.html
You can check the 10-year averages for temperature and rainfall from the data obtained from site 070014 and see that the slope between them is practically identical to the one displayed in the above post. The very small difference is due to slight corrections in the 2009 data, the latter part of it being still provisional in January 2010.
To go back a step, you can bring up the data by entering the site number into this web page: http://www.bom.gov.au/climate/data/
So: we will soon see if my prediction has been falsified (which it likely has been).
Another consequence of this error on my part is that inflow for the rest of the year will be higher than it otherwise would have been, as the rate of inflow for megalitre has increased (due to there having been less rainfall at station 070014). Thus, my estimate for the inflow - and I must estimate it, because it cannot be measured as the dams are all above 100 per cent - will be greater. At present, that does not matter, as we have not had much rain since reaching 100 per cent. But rain is expected soon.
I had made the assumption that this data was the same as the date from here:
http://www.bom.gov.au/jsp/ncc/cdio/weatherData/av?p_nccObsCode=139&p_display_type=dataFile&p_startYear=&p_stn_num=070014
This is not the case, as I would have realised had I been paying attention:
"Source of data
Observations were drawn from Canberra Airport {station 070351}.
Weather observations for Canberra were previously taken at a nearby site {number 070014}. These can be seen on the "Canberra Airport" Daily Weather Observations; you may need to consult these to get all the relevant data."
My predictions were based on the data from station 070014, as the data from station 070351 only goes back to the middle of 2008. This can be seen here:
http://www.bom.gov.au/jsp/ncc/cdio/weatherData/av?p_nccObsCode=139&p_display_type=dataFile&p_startYear=&p_stn_num=070351
So, at the moment it is not clear that my prediction has been falsified. I think that it has, but I will not know for certain until some time early next year.
To be sure that this is the data that I used, please read this post:
http://evilreductionist.blogspot.com/2010/01/future-of-rainfall-in-canberra.html
You can check the 10-year averages for temperature and rainfall from the data obtained from site 070014 and see that the slope between them is practically identical to the one displayed in the above post. The very small difference is due to slight corrections in the 2009 data, the latter part of it being still provisional in January 2010.
To go back a step, you can bring up the data by entering the site number into this web page: http://www.bom.gov.au/climate/data/
So: we will soon see if my prediction has been falsified (which it likely has been).
Another consequence of this error on my part is that inflow for the rest of the year will be higher than it otherwise would have been, as the rate of inflow for megalitre has increased (due to there having been less rainfall at station 070014). Thus, my estimate for the inflow - and I must estimate it, because it cannot be measured as the dams are all above 100 per cent - will be greater. At present, that does not matter, as we have not had much rain since reaching 100 per cent. But rain is expected soon.
Tuesday, December 7, 2010
Prediction Shortgevity
Following on from my discussion about longevity comes something about the opposite: my short-lived prediction regarding Canberra rainfall. We have received a staggering 899 mm thus far this year, blowing past my 850 mm limit. Thus, my statistical analysis that pointed to technical desert conditions for Canberra by 2050 has been proven false.
So the key now for me is to keep watching to see where the evidence points. What is obvious now is that my conclusion was not warranted from the data, and I needed more data - data which I now have. But that is how science works: you build a model from observations and use that model to make predictions about the future, understanding that falsifying those predictions falsifies the model.
I will continue to track the rainfall and inflow (although with Canberra dams now at 100 per cent, tracking excess inflow is a little difficult - I will have to make some assumptions about extra inflow, and I am looking at what those assumptions might be at the moment.) We have received 142,000 megalitres of inflow thus far this year. What I might do is slightly increase the overall average to account for lost water. It should be pointed out that this amount of inflow is still significantly lower than the average.
So the key now for me is to keep watching to see where the evidence points. What is obvious now is that my conclusion was not warranted from the data, and I needed more data - data which I now have. But that is how science works: you build a model from observations and use that model to make predictions about the future, understanding that falsifying those predictions falsifies the model.
I will continue to track the rainfall and inflow (although with Canberra dams now at 100 per cent, tracking excess inflow is a little difficult - I will have to make some assumptions about extra inflow, and I am looking at what those assumptions might be at the moment.) We have received 142,000 megalitres of inflow thus far this year. What I might do is slightly increase the overall average to account for lost water. It should be pointed out that this amount of inflow is still significantly lower than the average.
Wednesday, December 1, 2010
Life expectancy for my age cohort
I have just been doing some calculations on my age cohort based on the observed improvements in life expectancy for Australians over the last 20 years. If those improvements are replicated every 20 years over the next century, the median life expectancy of all those aged 40 becomes 127.5, with those people living most of the last 40 years of their lives with a health approximating those in their early 70s today. They would have reached statistical immortality (see a previous post on this) at around age 90. Some of that cohort should live much longer than that, and there is a slim possibility that some of them could be alive hundreds of years from now.
That gives me reason to hope that I might be alive - doddering, perhaps, and dreaming of the past but alive - in the year 2100, an interesting milestone to me because (a) I am human and love nice round numbers and (b) it is a year about which there is much speculation in science fiction novels and roleplaying games (see (a)).
While I am a pessimist regarding climate change, I am also an optimist: while I believe that humans are going to cause a lot of suffering for ourselves and other species over the next century, I think that we as a species will pull through it and have an amazing history to write on our planet, on other locations in the solar system and among the stars. I would like to see more of that history. :)
That gives me reason to hope that I might be alive - doddering, perhaps, and dreaming of the past but alive - in the year 2100, an interesting milestone to me because (a) I am human and love nice round numbers and (b) it is a year about which there is much speculation in science fiction novels and roleplaying games (see (a)).
While I am a pessimist regarding climate change, I am also an optimist: while I believe that humans are going to cause a lot of suffering for ourselves and other species over the next century, I think that we as a species will pull through it and have an amazing history to write on our planet, on other locations in the solar system and among the stars. I would like to see more of that history. :)
Tuesday, November 30, 2010
Topping 700 mm - rainfall and inflow update
I have been holding off for a little while waiting for rainfall to top 700 mm for the year. And it has done so with a bang: we have now had 744.6 mm of rainfall this year, which is a very good year indeed, and we still have a month to go, which means that 800 mm looks to be well within reach. And it is possible that we will top 850 mm, blowing my statistical predictions out of the water, so to speak. But we will see.
Regarding inflow, the full inflow from the last two days rain has not yet been measured, but we are currently sitting at around 122500 megalitree for the year. This is still below what we would have expected from such an amount for rain, but it is still almost triple what we had last year.
Regarding inflow, the full inflow from the last two days rain has not yet been measured, but we are currently sitting at around 122500 megalitree for the year. This is still below what we would have expected from such an amount for rain, but it is still almost triple what we had last year.
Wednesday, November 24, 2010
Statistical immortality
Okay: first up, statistical immortality does not mean that you will live forever. But it does open up the possibility for people to live very long lives indeed.
What is statistical immortality? Statistical immortality, as I define it, is where the pace of increase in life expectancy reaches parity - in other words, life expectancy increases by a year every year.
At present, life expectancy is increasing by around one year every three years for those in rich Western nations like Australia. Most of this increase *not* in reduction in infant mortality - we have almost reached the limit of improvement there. Rather, most of it is coming at the other end of life: we are not dying when we used to.
To illustrate exactly what I mean, I will go through an example. Imagine a person who is 60. They have a life expectancy of a further 22.9 years. What this means is that some people of their age will die prior to 82.9 (and some will die before reaching 61!) and some will live longer, but that the average age of death for the group as a whole will be 82.9.
Looking at the table from the ABS, around 53 per cent would still be alive at 82.9. However, let us assume that by the time this cohort was 65, their life expectancy had increased five years to 87.9. Only around 3 per cent of them would be dead at this point. If we take them through five year steps, this pattern repeats, with a small per cent of them dying and the life expectancy of the rest extending further and further into the future.
Even with this small death rate, however, eventually the whole cohort would be dead. But this would take a significant amount of time. From an original cohort of 100,000 at age 60, there would still be around 50,000 alive after 23 steps - 115 years. So we are looking at a median age (the age by which half of them will be dead) of death for this cohort of 197.9.
Now, 197.9 is not immortality. So why would I call it statistical immortality? For two reasons: firstly, if you had an infinitely sized population (mathematicians like infinity) some of that cohort would be expected to survive forever (in fact, an infinite number of them :)); and secondly, this is so far beyond the usual life of a human being that it moves significant numbers of the population (50 per cent of this particular cohort) into a world that we can barely begin to imagine - one in which all sorts of other pathways would almost certainly open up for them.
Returning to climate change, the rapid increases in life expectancy that we are experiencing in the West at present makes it almost certain that, if you are reading this, you will be alive to see some of the worst effects. And then life expectancy may start to drop again ...
What is statistical immortality? Statistical immortality, as I define it, is where the pace of increase in life expectancy reaches parity - in other words, life expectancy increases by a year every year.
At present, life expectancy is increasing by around one year every three years for those in rich Western nations like Australia. Most of this increase *not* in reduction in infant mortality - we have almost reached the limit of improvement there. Rather, most of it is coming at the other end of life: we are not dying when we used to.
To illustrate exactly what I mean, I will go through an example. Imagine a person who is 60. They have a life expectancy of a further 22.9 years. What this means is that some people of their age will die prior to 82.9 (and some will die before reaching 61!) and some will live longer, but that the average age of death for the group as a whole will be 82.9.
Looking at the table from the ABS, around 53 per cent would still be alive at 82.9. However, let us assume that by the time this cohort was 65, their life expectancy had increased five years to 87.9. Only around 3 per cent of them would be dead at this point. If we take them through five year steps, this pattern repeats, with a small per cent of them dying and the life expectancy of the rest extending further and further into the future.
Even with this small death rate, however, eventually the whole cohort would be dead. But this would take a significant amount of time. From an original cohort of 100,000 at age 60, there would still be around 50,000 alive after 23 steps - 115 years. So we are looking at a median age (the age by which half of them will be dead) of death for this cohort of 197.9.
Now, 197.9 is not immortality. So why would I call it statistical immortality? For two reasons: firstly, if you had an infinitely sized population (mathematicians like infinity) some of that cohort would be expected to survive forever (in fact, an infinite number of them :)); and secondly, this is so far beyond the usual life of a human being that it moves significant numbers of the population (50 per cent of this particular cohort) into a world that we can barely begin to imagine - one in which all sorts of other pathways would almost certainly open up for them.
Returning to climate change, the rapid increases in life expectancy that we are experiencing in the West at present makes it almost certain that, if you are reading this, you will be alive to see some of the worst effects. And then life expectancy may start to drop again ...
Tuesday, November 23, 2010
Life expectancy
One of the interesting things - to me - about the climate change debate is how many people seem to think that the effects of climate change will not be experienced by them but rather by their children or grandchildren.
These statements are even made, perhaps rhetorically, perhaps not, by people who are authorities on the science. Storms of my grandchildren is a book written by James Hansen, the head of the Goddard Institute and the man who runs one of the five major global temperature data sets, GISSTemp. While I am sure that James Hansen is aware of what climate change is doing to the world now, the emphasis is on what will occur many decades into the future. (To be fair, Hansen is 69, so his grandchildren are likely around 10 or so).
However, even someone who is 69 and who lives in the wealthy west has a reasonable chance of living for another 20 years.
And that leads me to the point of this post: examining life expectancy.
Let us examine the life expectancy by age tables published by the ABS here:
http://www.abs.gov.au/ausstats/abs@.nsf/Products/381E296AFC292B6CCA2577D60010A095?opendocument
What they show us is that an Australian male (and James Hansen is American, but the difference will not be all that great) aged 69 has a life expectancy of a further 15.7 years.
The tables towards the bottom of the page show something even more interesting. They show that as you get older the age at which you are expected to die increases quite signficantly.
For example, someone who was 40 in 1989 was expected to die at age 75.9. Those members of that demographic who reached the age of 60 in 2009 were expected to die at age 82.9, an increase of seven years in a 20-year period.
If you think about, this is at least partly to be expected. If you survive from age 40 to age 60, the most obvious conclusion is that you have not died. Thus, you have successfully avoided the dangers that have taken the lives of others in your demographic. Those who died were taken into account in working out the expected age of death of 75.9. They no longer exist, and so the expected age of death for the survivors must be higher than 75.9.
However, an increase of seven years seems quite large. Think about it this way: once we reach a point where our average age of death increases by one year for every year that goes by, we will have reached statistical immortality (in a way - there will still be deaths, but they will be compensated for, in the statistical sense, by faster and faster increases in life expectancy). Seven in 20 is a reasonable step towards that mark. And the figures for those aged 60 in 1989 who survived to 80 in 2009 are even more interesting: the expected age at death increased by 10 years in those 20 years.
What is going on here? Well, apart from death winnowing out people from the second set of statistics (ie, not everyone is making it to 80), medical technology is improving quite rapidly. This is expanding life expectancy, and it is particularly doing so for those aged 40 or above.
Having done some calculations based on these tables, it is my conclusion that someone who is aged approximately 40 today has a 25 per cent chance of living to 120. These calculations assume the continuation of the steady increase in life expectancies, with no spectacular breakthroughs.
It is also my calculation that 'statistical immortality' will be acheived in 100 years, with those aged around 20 today having about a 30 per cent chance of reaching that point.
And I will write a more detailed post about what I mean by 'statistical immortality' in the near future.
These statements are even made, perhaps rhetorically, perhaps not, by people who are authorities on the science. Storms of my grandchildren is a book written by James Hansen, the head of the Goddard Institute and the man who runs one of the five major global temperature data sets, GISSTemp. While I am sure that James Hansen is aware of what climate change is doing to the world now, the emphasis is on what will occur many decades into the future. (To be fair, Hansen is 69, so his grandchildren are likely around 10 or so).
However, even someone who is 69 and who lives in the wealthy west has a reasonable chance of living for another 20 years.
And that leads me to the point of this post: examining life expectancy.
Let us examine the life expectancy by age tables published by the ABS here:
http://www.abs.gov.au/ausstats/abs@.nsf/Products/381E296AFC292B6CCA2577D60010A095?opendocument
What they show us is that an Australian male (and James Hansen is American, but the difference will not be all that great) aged 69 has a life expectancy of a further 15.7 years.
The tables towards the bottom of the page show something even more interesting. They show that as you get older the age at which you are expected to die increases quite signficantly.
For example, someone who was 40 in 1989 was expected to die at age 75.9. Those members of that demographic who reached the age of 60 in 2009 were expected to die at age 82.9, an increase of seven years in a 20-year period.
If you think about, this is at least partly to be expected. If you survive from age 40 to age 60, the most obvious conclusion is that you have not died. Thus, you have successfully avoided the dangers that have taken the lives of others in your demographic. Those who died were taken into account in working out the expected age of death of 75.9. They no longer exist, and so the expected age of death for the survivors must be higher than 75.9.
However, an increase of seven years seems quite large. Think about it this way: once we reach a point where our average age of death increases by one year for every year that goes by, we will have reached statistical immortality (in a way - there will still be deaths, but they will be compensated for, in the statistical sense, by faster and faster increases in life expectancy). Seven in 20 is a reasonable step towards that mark. And the figures for those aged 60 in 1989 who survived to 80 in 2009 are even more interesting: the expected age at death increased by 10 years in those 20 years.
What is going on here? Well, apart from death winnowing out people from the second set of statistics (ie, not everyone is making it to 80), medical technology is improving quite rapidly. This is expanding life expectancy, and it is particularly doing so for those aged 40 or above.
Having done some calculations based on these tables, it is my conclusion that someone who is aged approximately 40 today has a 25 per cent chance of living to 120. These calculations assume the continuation of the steady increase in life expectancies, with no spectacular breakthroughs.
It is also my calculation that 'statistical immortality' will be acheived in 100 years, with those aged around 20 today having about a 30 per cent chance of reaching that point.
And I will write a more detailed post about what I mean by 'statistical immortality' in the near future.
Subscribe to:
Posts (Atom)