Sunday, 30 November 2008

Carbon Capture & Storage in the Netherlands.

Earlier this week, the Dutch ministry VROM (responsible, among other things, for the environmental policy), in a combined effort with the ministry of economy, released a press release that they'll spend no less than 60 million € for two CO² underground storage pilot projects in the Netherlands.

The aim as stated in the press realease is clear :
The aim was to develop two separate projects each enabling at least 2 million tons of CO2 to be stored over a maximum ten-year period. The state will pay for each tonne of CO2 stored.

The first selected project is designed by Shell Holland. The company has created a website (Dutch only) which explains their plans in Barendrecht. The FAQ could be interesting for who wants to learn some more on the subject. This pdf (Shell brochure, Dutch only) also gives a nice overview of what sequestrion is all about.

The second pilot comes from a consortium called GTI Zuidoost, which according to my information, will look for the storage possibilities in abandonned mining facilities in the province of (Dutch) Limburg.


Even though personnally i'm not convinced CCS is the answer to comabt AGW, as a geologists i'm very interested in what the results of the projects will be. Yet, to really decrease CO2-levels in the atmosphere, i'm convinced there's only one possible route to follow : decrease emissions.

Friday, 14 November 2008

Nature Indicators.

In an area like Flanders with a population density of 445 inhabitants/km² it is clear nature is under high stress because of all sort of environmental issues.

Nature report
In a previous post i mentioned the Flemish environmental policy is monitored by the MIRA-reports which make an inventary & evaluation of the policy.

For Nature, similarly, the Flemish Research Institute for Nature and Forest (INBO) publishes a two-yearly Natuurrapport (in Dutch) which gives an overview of biodiversity in Flanders, searches the causes of the changes detected and evaluates the actions taken. The report contains a chapter about climate change (pdf in Dutch only) as of course it is one of the factors changing ecosystems and thereby biodiversity.

Nature indicators
One of the ways to evaluate the nature policy, is by using natuurindicatoren (English version on the same URL): of course it is impossible to measure every little parameter or species alive. Therefore, for each domain a set of parameters have been chosen that are considered to be indicative for the entire (sub)-domain. e.g. for the field of acifidication/overfertilisation => indicators are the concentrations of Phosphor, Nitrogen & the oxygen demand (both BOD & COD).

The graphic below shows the number of observations that have been made during the past 20 years of a number of Dragonfly-species that in Belgium are being considered to be 'Southern species', meaning they are/were considered to be non-indigeneous in Belgium. (Lestes barbarus seems to have a stable population since the middle of the 90ies)


In the domain of climate change, all indicators are :

  • the spring arrival data of migratory birds (graph)
  • the spring-index for dragonflies (graph)
  • the appearance of Southern species of dragonflies (see image above and here)
  • the peak moment of pollen-production in birch-trees (graph)
  • a monitoring program for following-up the budding of several species of trees (still in it's testing phase)

Nature Indicator global warming dragonflies Belgium
When having a closer look at the flying-season of some indigeneous species of butterflies & dragonflies, there's no simple conclusion. Some species haven't changed their season, while for some others there are really significant changes :

In spring, the spring-index (first appearance) for Inachis Io (day peacock eye / dagpauwoog) is 42 days earlier than it was 20 years ago.

In autumn, the Plebeius agestis (Brown argus / bruin blauwtje) prolongued it's season no less than 62 days. Above this, this species in spring appears one month earlier, meaning it's yearly cycle nowadays is three months longer !

Regression analysis has shown there's a clear correlation between the flying season and temperature for 20 of the 23 monitored species of butterflies.

The INBO report Biodiversity indicators 2007 presents 22 Flemish biodiversity indicators according to the focal areas of the Convention on Biological Diversity and the EU biodiversity headline indicators, in order to track progress towards the achievement of the 2010 target.

Nature Indicator global warming Belgium

Wednesday, 12 November 2008

A layman’s guide on cherry-picking

Lately, a lot of attention has gone to the supposed fact that global warming stopped and that this proves IPCC’s projections to be wrong. To evaluate this claim, we need to split the presented subject into two separate questions :
  • did temperatures really stop rising ?
  • does the halt in warming (assuming there is one) disprove climate science ?
In this post i want to have a closer look at the first question : The claim AGW has stopped is one appearing in many slightly different ways. Looking at each of those individual claims would be rather time-consuming. Rather I seek to provide the reader some more background on climate science so he/she will be able to make valid judgments on such 'global warming is over' claims himself, or at least understand the more mathematical approaches on the web which are linked in the bottom.


Finding a decreasing temperature trend in data sets : easier than learning your ABC

I want to demonstrate in this post how easy it is, with a little bit of cherry-picking, to FABRICATE a trend line (we’ll presume all over this text it’s a linear trend) suited for making a exactly the claim you wanted to make before even looking at the data. Sure this claim will be very dubious, but 99% of your readers will not have the background to be able to detect the errors you deliberately inserted.

I want to look at the addressed issue from a certain viewpoint : Picture yourself to be a climate skeptic wanting to prove there’s no warming anymore. What would be the logical thing to do :

You’ll need to take the dataset showing earth’s surface temperatures and search if there’s a chance fabricating a negative trend line in it somewhere. In a set with so much noise and natural variability in it as the one of earth’s temperature, not surprisingly, choosing data suited for drawing the trend line you like best is fairly simple.

Simply draw a trend line between two carefully selected points on the graph so that you’ll have a downward trend. Whether this trend line actually is meaningful is another question which will be addressed later on. But first things first and let’s start by fabricating the trend line.



The art of shutting up on the things you don’t want the reader to know
It might sound a little silly, but when drawing a trend line, the important thing to remember is what the trend line you draw is all about. A trend line on temperature is a result of ALL processes that have an influence on earth’s temperature.

Climate shows a relatively big natural variability and when judging whether or not there’s an anthropogenic climate change, one will need to have a closer look at the importance of all those factors influencing climate.

There are several datasets that try to give an estimate for earth’s temperature. As you can see in the graph below, there are slight differences between the different sets, but those differences are quite small and averaged out of the years, and not relevant for the story to follow. (Actually, the data-sets all describe slightly different things, which explains)

When applying temperature trend lines on a graph like the one above, I suppose almost everybody will automatically agree you cannot simply compare two consecutive years and draw conclusions, because the differences between those two years will rather be the result of variability and noise than being indisputably part of a long-term climate trend.

Because two years clearly is a period way too short to compare, you’ll need to find a longer period to be able to draw meaningful conclusions : the image shown clearly shows there’s a lot of natural variability on the temperature signal, and as mentioned in my previous post, on top of that there’s a number of natural cycles also having an impact on the result. Using a reference period which is too short will result in comparing signal-noise instead of looking a real trend.

When reading articles on climate data, any reference period being remarkably short (like only a decade or so) should make you suspicious even though, when there’s simply lack of sufficient data, of course it is possible people are obliged to use shorter periods.

On a longer time scale, there are many cycles having an impact on earth’s temperature, the Milankovic cycles probably being the best known of them. Throughout history, temperature records clearly relate to those cycles. Yet, the importance of the Milankovic cycles on a scale of a century or so is rather ignorable because in such a short time span the changes or too small to be detected.

Yet besides the Milankovic-cycles, there’s of course other things having an impact an earth’s climate, like solar variability, aerosols, volcanism, etc… All these other factors and cycles do have a measurable effect on the data sets and therefore cannot simply be ignored as if they weren’t there.

It’s the sun stupid !

One of the factors having an impact on our climate is variability in solar radiation.

Actually, a claim frequently made by skeptics is that present day changes in surface temperatures are not a result of man-made GHG’s, but changes in temperature are simply following changes in solar radiation. Science disagrees on this claim because those changes simply are too small to (fully) explain present day warming, yet changes in solar radiation DO have a (minor) impact on earth’s temperature. So this factor cannot explain why earth’s temperatures have been rising lately and the skeptics claim is disproven. Above that, the last couple of years have seen a decrease in solar activity while temperatures were soaring

But back to the trend line we wanted to draw :

When we put a decreasing trend like the one from solar radiation into a graph we’ll get an negative trend line like the blue line in the image below.

Clearly, as said the trend to be expected from the change in solar variability (blue line) is negative for the period shown in the graphic from point A to B.

Suppose the temperature data as they were recorded (the red arrow) does follow that negative trend, but not as much as expected. Then this implies there must be second factor having an influence on the temperature (p.ex. AGW), yet not enough to compensate the negative trend set by the change is solar variability. If the second factor would've been bigger, the actual trendline even could've become positive, totally hiding the cooling trend from the sun...

The thing to remember is : Even though in the selected period the temperature-trend is negative, what it really shows is that besides the negative solar-effect there actually IS (the green area) a second factor with a positive temperature effect ! You can't present your result without mentioning this.

Of course if you do shut your mouth about all the underlying science and simply present the image above, but with nothing on it but the red trend line, you can happily present your conclusion like p.ex. global warming stopped (silently suggesting man has no effect on earth’s climate).

Sure, the facts you presented are incomplete, but very useful anyway in non-scientific environments. The power of knowing what to not say can never be underestimated !

Unfortunately, the method presented to fabricate a decreasing trend as described here isn’t sufficient because solar variance is too small. We’ll need to look a bit further. All we need to find is something with an effect that’s big enough to help us to reach our goal.


Oscillations Schmoscillations

In my previous post, I mentioned there’s a natural cycle called the El Nino Southern Oscillation (ENSO) which does have an measurable impact (and a rather big one) on earth’s surface temperature.

Typically, an oscillation looks like shown in the image below. Clearly, the ENSO-oscillation will not be as perfect as the one presented here, but the image is suited for continuing our search towards a negative temperature trend.

When looking at the image, what’s the long term trend on the oscillation do you think ?

If your answer is : it’s a status quo (the green line), this simply means you are an alarmist who isn’t able to understand real science.

The correct answer of course is the image shows a downward trend. And I’ll prove it !

Not by altering data or making errors in calculations or anything. I’ll use exactly the same data set as the picture above. Yet, what I want to do is to cleverly pick the right spots to support my claim.

Remember, we want to prove there’s a decreasing trend in temperature, so the best thing to possibly do is start your trend line on the highest possible point in the graph below (like, p.ex. point A)

The point where your trend line stops has to be as low as possible. I’ve chosen three similar points on the bottom of the graph (B, C, D) and drawn the trend lines. Later on we’ll decide which one suits us best.

Notice there’s something very awkward with those trend lines : even though they all three connect points on the same level, the slope of the trend line differs. I’ve isolated the trend lines so it’ll be easier to see :

The longer the time period between the points that have to be connected, the more horizontal it becomes. Or in other words : the more it will approach the real trend (green line) until at some point it will become indistinguishable from the real trend. (Science actually has a couple of methods to transform the data and make it easier to distinguish trends and noise)

This difference in slope of the trend lines of course is the reason in climate science you need to take periods that are long enough (an absolute necessity is to overcome the ENSO oscillation) otherwise you’ll only be comparing noise instead of the real long-term trend.

Because in this post the aim was to draw a trend line that decreases as much as possible, what we need to do is ignore the paragraph above and take the trend line which suits us best. So we’ll go for the trend line on a very short time span : trend line n° 1 is the one we want to use !


Now, with all the things we've learned, we are almost ready to take the real data of surface temperatures and find the points we want to connect to get our desired trend.

A last thing we need to do is to select the years we want to connect : preferably starting from a year with a temperature above average (an El Nino year would be perfect for this) and a colder year (like a La Nina year).

Luckily, 2007 & 2008 both have been partially La Nina years and therefore are very useful to be used as an end point in our trend line. Just our luck ! The last thing is to find a year warm enough to make sure our trend goes down. Like an El Nino year. And this, as we've learned, in a time span that is as short as possible without becoming so short even laymen will ask questions about the reliability of the trend you're presenting. Whereas one year might be a bit exaggerated; surely something in between 5 a 10 years is more than long enough !

Now it's time to look at the actual data : immediately we see the strong El Nino effect in 1988, which is exactly ten years ago. It's a bit an outlier in the data sets. (see image 1), but we’ll use it anyway because it will give us the trend we wanted in the first place ...


There’s one more thing left to do : look at the actual data and see if your carefully picked data really have a downward trend. You still need to verify this because of what i explained under the solar variance paragraph : it's still possible there's a factor causing an upward trend which overrules the trend you've picked.

There are several data sets : if one doesn't show the desired result, try your luck by picking another one (p.ex. HadCRU instead of GISS). No layman will ever find out about the existance of the other sets contradicting your conclusion, so it's absolutaly safe to ignore them.

If the trend line, despite your carefully selected data still isn’t going downward, look for another starting point. If necessary, shorten your time span from 10 years to 9 years. Or even less. Don't be shy.



What happens if you make the time span of your trend line too short is clearly shown on this graphic :

This aren’t meaningful trend lines any more but Mikado. Shortening the time period to seven years naturally makes it even worse :

(El Chicon & Pinatubo are volcanoes which in the given years had major eruptions, causing a temporary global cooling effect which could be detected in the data set)

The last two images are taken from this article from the excellent website RealClimate and is well worth reading. With the explanation above, I hope even people with no scientific training are able to understand what’s written in the RC article.


If you want to read even more, I recommended these links :

Robert Grumbine on cherry picking

Maybe a bit more difficult, but still worth reading : statistician Tamino on ‘global warming stopped in 1988’ and the use of short term trends in this three posts :

Garbage is forever

Wiggles

Global Temperature from GISS, NCDC, HadCRU

Saturday, 8 November 2008

Jos Verhulst on Hurricanes and Climate Change

In this post (in Dutch) on Belgium's biggest forum about politics Jos Verhulst, who is co-author on the conservative blog The brussels journal mentions that the the past two years have been below average on hurricane-activity.Jos Verhulst klimaatverandering

Verhulst writes :
Het voorbije jaar 2007 was, in schril contrast met allerlei deskundige voorspellingen, behoorlijk slap op orkaangebied. Het ACE-totaal kwam voor het noordelijk halfrond uit op 73% van het gemiddelde voor de laatste 26 jaar. En 2008 is tot op heden evenzeer in mineur verlopen.

in English, his word would mean something like (my translation) :
In sharp contrast with what experts predict, the last two years have been minor ones on the field of hurricanes. The ACE-total for the Northern Hemisphere ended on 73% of the average level of hurricane activity of the past 26 years. Above, 2008 also has shown a small hurricane-activity.

In science, hurricane activity in isn't measured by simply adding the number of hurricanes because that number doesn't say anything about the strenght of the hurricanes. Instead, scientists use an index developed by the American National Oceanic & Atmospheric Association (NOAA) which is called the ACE-index (Accumulated Cyclone Energy). This index takes into account the number, strength and duration of all tropical storms in the season.

Verhulst shows a graphic from Florida State University to support his findings :

Is Verhulst's claim correct that climate science's predicted increased hurricane activity is falsified by this graphic which shows the last two years have shown lower activity?

Well, err, no of course not !


Hurricane activity is closely related to the sea-surface temperature (SST), and this SST is something that shows huge interannual variability, most notably because of the El Nino and La Nina events, or as scientists call it : The Southern Oscillation (often referred to as the El Nino Southern Oscillation ENSO)

The KMMI (Royal Dutch Meteorological Institute) gives a nice explanation in Dutch : click

The reason the Nino/Nina events are important enough to be taken into account is the fact they alter the SST in the areas where hurricanes are formed. And the change is this big, that it clearly has an impact on hurricane activity : when you look at the years there were Nino/Nina years, and compare this to the image below, there's no doubt at all : these events have a dramatic impact on SST temperatures in areas where Hurricanes are formed.


If you look at the graphic presented by Jos Verhulst there's a very straight correlation with the Southern Oscillation :

  • The low parts of the graphic are the years there's a La Nina
  • The top parts are the years there's a El Nino event.
This means there's a very close relation between SST and hurricane activity !


This implies:
  1. science's claim that increasing temperatures could result in a higher ACE-index isn't all that weird
  2. Verhulst's claim on the ACE-index of 2007/2008 is incomplete : he forgets to take into account in which phase ENSO is at the moment, for this matters.
2007 & 2008 were years there's a ... La Nina.

Or in other words: those two years have had lower SST's because of the ENSO, and therefore showed lower hurricane activity.

Therefore, Verhulst's claim that the last two years have known hurricane below the average off the last 26 years is ... exactly what one would expect.

One again, we can see that in climate science, you cannot draw conclusions out of very short time periods. Verhulst's base period simply is way too small to conclude anything.

Shamefully, Verhulst made this post exactly one day after i explained to him his baseperiod on another topic (sea level rise) was ... too short. And this by far hasn't been the only error Verhulst has made on politics.be on the field of climate change.

My conclusion is very clear :
Despites his high level of arrogance towards the scientific community, Jos Verhulst demonstrates time after time after time he fails to understand climate science. Verhulst instead is just another libertarian who gets blinded by his bias. Hereby helped by the known inreliable websites he frequently cites like e.g. co2-science

Verhulst concludes his post with :
Men kan nu een eenvoudige geestelijke oefening uitvoeren. Beeld u in, dat tijdens de laatste twee jaren de orkaanactiviteit sterk bovengemiddeld ware geweest. Denkt u, dat de media over deze bovengemiddelde activiteit dan dezelfde radiostilte in acht zouden nemen, die we nu vaststellen in verband met de sterk verlaagde activiteit? Denkt u, dat men dan zou zeggen dat twee jaar een te korte periode is om conclusies te trekken?

De vraag stellen is ze beantwoorden. Let maar op het propagandistisch wolvengehuil dat zal opstijgen, zodra er weer een jaar komt met meer orkanen dan gemiddeld.

or, in English (once again, my translation) :
Time to make a little mental exercise : picture that the last two years, hurricane activity would have been above average : do u think the media would have been as silent about the topic as they are now activity is below average ?
Asking the question means you have to answer it : be sure that the wolves' cries of propaganda will rise again as soon as there'll be a year activity is above average.


As Sarah Palin would say : it's all about the liberal media elite.

Monday, 3 November 2008

MIRA-klimaat (Dutch only)

De opvolging van het milieubeleid in Vlaanderen gebeurt via de jaarlijkse MIRA-rapporten die drie doeleinden hebben:
  • MIRA beschrijft, analyseert en evalueert de bestaande toestand van het milieu
  • MIRA evalueert het tot dan toe gevoerde milieubeleid
  • MIRA beschrijft de verwachte ontwikkelingen van het milieu volgens een aantal relevant geachte scenario’s, zowel bij ongewijzigd als bij gewijzigd beleid

De MIRA-rapporten worden opgezet rond een aantal milieuthema's, waarvan klimaatverandering er één van is. In het kader daarvan verscheen eerder dit jaar een wetenschappelijk achtergronddocument dat een bundeling vormt van de basis waarop ons klimaatbeleid stoelt en daarom een interessante lectuur vormt om te kijken wat Vlaanderen op dit terrein allemaal uitspookt.

Het rapport is bijzonder leesbaar geschreven en is daarom niet alleen beleidsmatig interessant, maar is tevens een goed vertrekpunt voor wie meer over het onderwerp klimaatverandering zelf wil weten