Air pollution in Beijing, with orange alerts and high health risks made the news (again) recently.

One measure discussed to reduce pollution levels is to issue a partial ban for cars on the streets. This would certainly reduce some of the pressure, however, I am not quite sure whether this would significantly help improve the situation.

Looking at this 2005 energy flow Sankey diagram for China from World Resources Sim Center, one notices that the largest chunk of energy produced in China is from coal.

The energy from coal consumed in 2005 (purple boxes) was seven times higher than the energy consumed in the transportation sector (1.450 million tonnes of coal equivalent vs 230 million tonnes of coal equivalent, if I interpret this correctly). That is without counting losses that occur both at power stations as well as in vehicle engines.

Of course one might argue that traffic probably has risen enormously since 2005. I checked the latest available data for China (2011) at the International Energy Agency (IEA) website where you can access the national energy flow diagrams for more than 100 countries.

It confirms that both coal and petroleum consumption have risen, but coal is still predominant. Transportation makes up for “only” roughly 200 Mtoe (Millions of tonnes of oil equivalent) out of 1644 Mtoe final consumption, while the significant consumption is in industry and ‘other’ (probably private home heating).

So, even if burning of coal and gasoline has different levels of pollution (e.g. through efficient filtering technolgy), my guess is that the main reason for the smog in China are the coal-fired power plants and industrial furnaces. Reducing vehicle traffic will not lead to reduced coal-burning.

Anyone has data on GHG emissions from different sources in China? (preferably as a Sankey diagram….).

Color gradients seem to be the new like for Sankey diagrams. I already featured an example in yesterday’s post. The new e!Sankey 3.1 version has a color gradient feature now as well.

This diagram is taken from the samples included in the trial version. Traffic flows at a fictitious highway intersection is shown with the number of cars going from A to B, A to C, and so on.

It is not new at all (see this post) but has been pimped and now sports the color effect and some icons. Nice!

Nice to see that more and more applications now use Sankey diagrams rather than pie charts to visualize distributions. In this case it is a widget available in the dashboard of Boundary, a real-time application performance measurment (APM) suite.

Inbound traffic is on the left side in green, outbound traffic to the right in blue. Only the top five connected hosts by traffic are shown. Inbound and outbound traffic is about the same size, so they have the same height in the stacked part in the middle.

Similar to the cargo traffic Sankey diagram I did here, but in this case the traffic is not traffic of physical goods…

A reader of the blog pointed me to some Sankey diagrams available on the U.S. Department of Transport (DOT) Federal Highway Administration (FHWA) website. Sankey arrows are shown as a U.S. map overlay.


The first transport Sankey diagram shows the net tons of goods being transported on flatcars (either as trailer-on-flatcar, or container-on-flatcar) on the U.S. railway systems. The transport volume quantities are clustered into four groups shown with four different arrow widths. It is nice to see how in the eastern part you still have many railroad tracks, but with significantly less transport on them, while to the west coast you basically have 3 main lines, two of which carry more than 25 mio tons of freight per year.

The second one represents freight transport on railroad (bright red), inland waterways (blue) and national highways (dark red). Values also in million tons per year. I am not sure whether flows are to scale, or if transport quantities are also clustered into groups as in the first diagram. I can distinguish at least five different arrow magnitudes, even though only three sizes are given in the legend. A great transport Sankey diagram, as it shows that the East has most of the cargo, and has a much denser transport infrastructure. The reason why railroad transport on this second Sankey diagram differs so much from what is shown the first one, is probably due to the transport of bulk materials not shown in the intermodal transport quantity map.

Flows are not directional in both diagrams, so I assume that quantities for both directions have simply been added.

Can anybody confirm that the massive stream out of Wyoming by rail is coal?

Just discovered this neat Sankey diagram provided in the e!Sankey 2.0 trial version as a sample file. It shows the number of vehicles passing a (virtual?) intersection within a given time frame.

Not sure if traffic planners could make use of such a visualization, but it is yet another nice idea of how Sankey diagrams (originally used only to visualize energy efficiency) are being used for other purposes.