Monday, 4 March 2013

No fact-checking when it's most needed

It's always interesting seeing the birth of a myth. Myth works in quite a specific way, within the hallowed halls of the anti-wind lobby. They begin simply - a piece of research or information is misrepresented, mutated or mauled. This is published, repeated, re-blogged and emailed until the myth becomes reality. A great example is the absurd claim that wind farms contribute to the effects of global warming. 

Occasionally, you can catch these myths in their embryonic form - just prior to moment they reproduce furiously across the internet. Let's have a look at a media release from David Ridgway, current Liberal Party Leader of the Opposition in the South Australian Legislative Council.

Electranet is a Transmission Network Service Provider (TNSP) in South Australia. Last Thursday, they testified as part of a South Australian parliamentary inquiry into wind farms. The transcript isn't available - the only information I've been able to find on it has been a media release on David Ridgway's website, and a few breathless tweets:

Gout's twitter profile lists journalism as his profession - he is currently chief of staff for David Ridgway. As the transcript for the hearing is currently unavailable, it's impossible to determine exactly what was said during the hearing. Nevertheless, we can still analyse the accuracy of the statements on Ridgway's media release. 
"No Wind Power When it's Most Needed
A parliamentary inquiry into wind farms in SA has heard that only eight per cent of the State’s installed wind generating capacity is available on hot summer days when demand is greatest."
Availability can mean a few things, and without the transcript, we don't know the context in which the statement was made.

With regards to the National Electricity Market (NEM), availability refers to the capacity for a generating unit to generate. So, to suggest that 92% of all wind turbines in South Australia are offline and unable to generate on every single 'hot summer day' is absurd - it's too silly to debunk. 

What's more likely is that the individual they quote was referring to 'firm' availability - a quantity of power that can be guaranteed to be available - as AEMO state in their '2012 South Australian Electricity Report':

"AEMO has increased its determination of the proportion of installed wind generation capacity that can be considered to be firmly available to meet maximum demand from 5.0 per cent to 8.3 per cent for summer...The increase is due to more wind farms operating in the state, providing greater geographical diversity, and also improvements to AEMO’s wind data analysis methodology."

This hypothetical assumption, used as part of the Electricity Market Dispatch Engine, is not representative of the actual supply of wind energy on hot days - it's a conservative mathematical operator, not a historical measurement of supply. 

This is not what Ridgway states in his press release. The title of the press release is blatant in its assertion: "No Wind Power When it's Most Needed".

There are fourteen wind farms in South Australia - their total installed capacity is 1,205 megawatts. Wind energy is variable - sometimes, wind farms operate at full capacity, at other times, they operate at a lower level. The same logic applies to hot, summer days: on some days, they are generating a large proportion of their installed capacity, and other days, they are generating less.

Let's have a look at the summer that just passed - Australia's hottest on record. Below is daily maximum and minimum temperature, measured in Adelaide, South Australia, sourced from the Bureau of Meteorology website. I've superimposed a red line for 30 degrees - a maximum temperature above that could well be classified as a 'hot day'. 

Now, let's have a look at the average daily capacity factor of wind farms, over the course of the month - the data below are sourced from the AEMO 'Market Management Systems' database. 

The data are as you'd expect - wind varies from day to day. No one claims wind energy correlates precisely with demand levels. But, is wind energy at lower levels on hot days, as put forward by Ridgway? Let's combine these two data sets - temperature and capacity factor - on a scatter plot: 

Again, no correlation between temperature and wind energy supply. Again, this is exactly what we expect - what we can say for sure is that there isn't less wind when temperature are high. Those who claim a strong negative correlation (Ridgway), and those who claim a strong positive correlation (no one) are wrong - it's shotgun scatter.  Squint hard enough, and you can see any pattern you want. 
“We've got more than 1000 megawatts of installed wind power sitting uselessly, their blades not turning, when demand is at its greatest”
What Ridgway seems to be implying is that wind energy is specifically muted on hot days. Let's look at wind energy supply on the 10 hottest days of summer, in Adelaide:

Note one important concept - An average South Australian 4-person household consumes 18.2 KWh per day, in summer. When 18.2 KWh of energy are exported to the grid, that unit is assumed to be equivalent to powering one household. The power output of generators changes over time, as does the consumption of power by households and industry. The number of homes powered is a static benchmark from which to gauge how much energy is produced from a particular source, over a time period.

Wind farms do contribute a significant quantity of energy to the grid. The statement that only 8% of wind energy capacity is 'available' on hot days is indefensible. The generation and temperature data make that pretty obvious. 

Despite the demonstrably inaccurate nature of both assertions, the myth has already flourished online. The Hansard for the inquiry will take two to three weeks to be publish online - when it is, we'll be able to compare the media releases to the remarks made in the hearing. I suspect their statements were slightly more nuanced than is made out by Ridgway and his advisers.  

Data Sources
Click here for the full Excel file used to generate the data (includes the raw data) - 32 MB


  1. Ketan, your analysis is sound, but you are looking at daily average wind contribution on a hot day.

    Electranet, SAPN and AEMO in their planning/forecasting roles only care about the peak few half-hourly periods per year (as this is what drives the need for investment in generators and transmission infrastructure).

    Possibly look at the contribution of wind when SA's demand is over 3,000MW, and the it's a hot day (it's rare demand that high isn't on a hot day in SA).

    Lastly, I think you'll find that ElectraNet, SAPN and AEMO refer to the 8% of wind available as "firm capacity", i.e. capacity that can be relied upon for planning purposes.

    Happy to talk with you further, but don't want to put my name on a website (as I may or may not work for one of these organisations).

    1. Hi anon,

      Thanks for the comment. You raise some good points, and I have a follow up post coming soon, so please watch out for that.

      In the interim, I might address your points:

      - Totally agree that Electranet are interested in 'firm' (rather than average) reliability - this is a massively important consideration. There still needs to be careful elaboration and accuracy when using the term 'availability', ie, we need to ensure that the word 'firm' is right there along with it. If it isn't, there's the potential for misunderstanding and misinterpretation.

      - AEMO's calculations actually involve a high-demand percentile (rather than temps, like I've done here) - which means their data are an interesting picture of how wind energy performs, versus the fluctuations in demand in the market - see my new post (when it gets posted).

      Thanks again for the thoughtful comments.