20 November 2014

The underlying truth in figures

I look at a lot of pictures, maps and statistics on water use. A good figure will help you understand, but a bad one will mislead you. What's the difference between good and bad? Sometimes it's intention, sometimes it's interpretation, but often it's an unfamiliarity with the underlying data.

Take this example from the Hamilton Project:

Three observations:
  1. Irrigation doesn't mean farming when you can grow crops with rain. You know that, but does this map help you understand where our food grows?
  2. Withdrawals have nothing to do with value. It's quite possible for heavy withdrawals for valuable -- or marginal -- use.
  3. How much can be taken out without threatening tomorrow's supplies? I don't see that.
Here's another, more complicated example from the USGS:

Fleck notes that (a) thermoelectric withdrawals are usually returned, (b) irrigation withdrawals do not reflect return flows, and (c) the recent dip reflects a drop in the economy. I'll add these notes:
  1. How can we know irrigation withdrawals when many surface diversions and groundwater withdrawals are unmetered?
  2. How do these uses occur locally, where excess demand has a local impact? National totals are almost meaningless, especially if one ignores seasonal variations in water supply.
  3. Residential water services require water, but they also require careful management and robust distribution systems. Are lower withdrawals a sign of better service?
Bottom Line: I'm glad to see more data on water use, but let's make sure we use these figures as the start, rather than the end of the conversation.

H/Ts to RM and MV

1 comment:

  1. While I don’t really disagree much with your discussion of the two charts below in “The underlying truth in figures”, I do have a bit of a “So what” response.

    It seems to me that neither of the charts seek to give definitive answers (or more importantly reasons) to anything. They are simply graphical representations of data. Any mischief, or misleading, would, in my view, only come from the interpretation/spin put in the text on the data of the charts. At that stage I think your cautions may well come into play, as it is the text and analysis of data where the sophistry usually comes in and the sneaky/unthinking logic jumps and twisting come into play.


Spam will be deleted. Comments on older posts must be approved.
If you're having problems posting, email your comment to me