Wednesday, March 26, 2014

Might help more if the link supports your assertion

Back to Roger Pielke Jr., sadly. Readers of Stoat will see deep in the comments that RPJr did a follow-up to his original piece claiming a lack of a climate signal in disaster stats. Once again it wholly fails to deal with the primary issue, that estimating effect makes far more sense than trying to detect a signal in a very noisy environment. It also has a problem with being misleading:
One final note: Other readers raised questions about the role of technological change — such as evolving building practices — and its effects on disaster losses over time. This subject is well addressed in the literature, and has been deemed important in damage trends with respect to Australian cyclone damage and U.S. earthquakes, for instance, but not for floods, U.S. hurricanes or tornadoes.
He's got a link for "floods," a lengthy article that cites repeatedly to RPJr but not one that does much to support his assertion. Instead it has a significant section on how flood protection efforts over time have reduced damages (or sometimes make things worse when done wrong) and concludes:
Based on the evidence recently assessed in the SREX report (S12), one can assess at present that it is likely that there have been statistically significant increases in the number of heavy precipitation events (e.g. 95th percentile of 24-h precipitation totals of all days with precipitation) in more regions than there have been statistically significant decreases, but there are strong regional and sub-regional variations in the trends, both between and within regions. Based on cumulative evidence, there is additionally medium confidence that anthropogenic influence has contributed to the intensification of heavy precipitation at the global scale, though attribution at the regional scale is not feasible at present. Projected changes from both global and regional studies indicate that it is likely that the frequency of heavy precipitation, or the proportion of total rainfall from heavy falls, will increase in the 21st century over many areas of the globe, especially in the high-latitude and tropical regions and northern mid-latitudes in winter. Heavy precipitation is projected to increase in some (but not all) regions with projected decreases of total precipitation (medium confidence).  
Despite the diagnosed extreme-precipitation-based signal, and its possible link to changes in flood patterns, no gauge-based evidence had been found for a climate-driven, globally widespread change in the magnitude/frequency of floods during the last decades.

I find it somewhat difficult to reconcile the three areas I bolded, maybe there's a mixing of estimation and detection going on. That last statement is a thin reed for RPJr though and he's directly contradicted by the parts saying flood mitigation has significant effects on outcomes. He might note that he only referred to "technology" and that building levees isn't a technological change, but that's proof then that he's throwing sand to obscure the flaws in his attempt to normalize damages over time.

A lot of work to refute just one misleading claim. Thanks a lot, Nate Silver.


UPDATE:  thought I'd check the RPJr refs for US hurricanes and tornadoes not being mitigated by changes over time, but they're paywalled. Curious that changed practices can reduce Australian tropical cyclone damages but not US hurricanes. Either the Aussies are better at this than us or somebody's wrong.

25 comments:

  1. Not perhaps relevant to anything but do you recognise the name of the editor of that article on flooding?

    This is what AR5 WGI says about floods:

    'AR5 WGII assesses floods in regional detail accounting for the fact that trends in floods are strongly influenced by changes in river management'

    'In summary, there continues to be a lack of evidence and thus low confidence regarding the sign of trend in the magnitude and/or frequency of floods on a global scale.'

    Flood mitigation efforts have not just been about minimising damage from floods but actually preventing them happening in the first place. Central London is built on a flood plain but the Thames Barrier has basically put an end to flooding there. Elsewhere along the Thames I know there have been projects to dig extra river routes to enable lowering of water levels.

    Across the channel flood prevention in the Netherlands is getting more and more sophisticated.

    AR5 also states:

    'Although the most evident flood
    trends appear to be in northern high latitudes, where observed warming trends have been largest, in some regions no evidence of a trend in extreme flooding has been found, for example, over Russia based on daily river discharge (Shiklomanov et al., 2007).'

    As I understand it there has been major river engineering work in Russia - see (or not) the Aral Sea - albeit not necessarily with flood risk management as an aim.

    ReplyDelete
  2. The reason for the apparent differences is the key word "GLOBAL", ironically (or intentionally?) the same dodge RPjr used to weasel out of his misleading Senate testimony before being slapped down hard by John Holdren.

    So yes, significant differences in many locations. But no, not globally, because desert areas get more desertified and therefore less flooding.

    ReplyDelete
  3. Those three bold part can fit together. The first one just claims that increasing trends in severe precipitation have been observed.

    These trends do not have to be due to human influence. Thus the attribution mentioned in the second bold part is more difficult.

    Floods require rain rates well above the 95th percentile. That is a rain amount that happens every 20th day. That percentile is thus not yet very extreme, but because it is more frequent it is easier gather enough data for a statistically significant result. Rain rates that cause floods are much rarer and it is consequently harder to statistically significant results.

    Add to that the additional noise due to spatial variations in the quality of flood defences. And I can imagine that a flood is only counted as a flood if humans are affected. That would be an additional source of noise. And temporal changes in the flood defences make the trend unreliable.

    ReplyDelete
  4. Keep Documenting the Atrocities, Eli.

    Best,

    D

    ReplyDelete
  5. Regarding the third bolded part, I think it refers back to section 4.1, which states:

    'Hence, the SREX report (S12) corroborates Rosenzweig et al. (2007) and Bates et al. (2008) in stating that no gauge-based evidence has been identified for a clear climate-driven, globally widespread, observed change in the magnitude/frequency of river floods during the last decades. There is thus low confidence regarding the magnitude/frequency and even the sign of these changes.

    This low confidence directly reflects the limited evidence in many regions. Available instrumental records of floods at stream gauge stations are often sparse in space and short or interrupted in time. Moreover, changes in land use and river engineering further hamper the identification of climate-driven trends.'

    ReplyDelete
  6. To continue my tangential reference of the editor, the following paragraph shows evidence of his influence, ending: 'Acceptance of the long-term persistence hypothesis would lead to a dramatic increase in uncertainty in statistical estimation. This thwarts the trend detection.'

    ReplyDelete
  7. Thanks Paul for this "changes in land use and river engineering further hamper the identification of climate-driven trends"

    IOW changes in management affect the data, and therefore the data can't be used to say that changes in managment have no effect (what RPJr implies).

    I'm missing something you're saying about the editor - same person as the author?

    ReplyDelete
  8. Brian,

    Yes, the general message is clearly: "The data available aren't sufficient to say anything. Even if there were a reasonably large climate-related trend we may not be able to see it."

    And not: "No trend".

    But see the AR5 quote in my first comment. Apparently WGII will discuss attempts to account for those issues.

    The editor is one D. Koutsoyiannis. I thought he was an old favourite of the blogosphere - maybe more obscure than I believed? Has been quite pally with Pielke the Elder in the past.

    ReplyDelete
  9. RPjr is quite prolific

    Though usually failing to find things by choosing data that likely wouldn't show a thing, even if it were happening and choosing tests that likely wouldn't show a thing, even if it were happening even if the chosen data wasn't quite so woolly wouldn't demand quite so many articles.

    ReplyDelete
  10. If you look @ google scholar, not much in 2012, 2013 and 2014 to date.

    Interesting

    ReplyDelete
  11. Hi Brian,

    SkS has a post up on this very subject. Things are more nuanced than Jnr. wants people to think they are.

    The papers in question do not support Jnr's assertions, contrary to his cocksure claims or they are not relevant.

    That is but one of many issues with his 538 posts, however.

    Rumor has it that Pielke Jnr. has been engaging in some Moncktonian behavior behind the scenes against people who had the temerity disagree with his misleading claims.

    The folks at 538 must be pretty annoyed with Jnr. right now, wouldn't be surprising if UC and GOP are as well.

    ReplyDelete
  12. Paul S.

    "Central London is built on a flood plain but the Thames Barrier has basically put an end to flooding there."

    This may be relevant then.
    http://www.theweek.co.uk/uk-news/uk-storms/57290/infographic-how-often-has-thames-barrier-been-used

    AnonySpilopsylla

    ReplyDelete
  13. "Either the Aussies are better at this than us or somebody's wrong."

    With policy advisors like RPJr in the US, I'll go for the former.

    ReplyDelete
  14. Re the Thames Barrier (ref wikipedia):

    Construction cost at 2001 prices: £1.3bn (with > £100 million of additional flood defences).

    Closure Cost (per closure and frequency looks to be rising): £16000 (2008 prices).

    Estimated potential damage cost due to failed gate (see https://en.wikipedia.org/wiki/Thames_flood_barrier#cite_note-maib-sand-kite-14): £13bn (1999 prices).

    The North Sea recently had a surge that was higher (in places) than the 1953 flood (https://en.wikipedia.org/wiki/North_Sea_flood_of_1953#United_Kingdom) yet the damage was far less (don't know about cost-wise, but actual damage and loss of life-wise). The improved flood defences here and on the continent have made noticeable improvements, but will all need to be updated.

    The Jubilee River in the Thames (https://en.wikipedia.org/wiki/Jubilee_River) cost £110 million (early 2000s). It is blamed for increasing the damage in some areas, however as it diverts water from more expensive areas (Windsor, Maidenhead, Eton) to less expensive ones (eg Wraysbury) it probably reduces economic cost, even if it doesn't make much difference in actual damage. The diversion, as opposed to reduction, of damage may only be theoretical - I've not yet seen anything substantive (not that I've looked hard).

    ReplyDelete
  15. Adam, the North Sea surge of 5 Dec '13 was high but sea defenses kept it out everywhere. Damages by surge were effectively nil except for some objects outside the sea defenses (Hamburg might've had a couple 100k damages from that).

    This storm system would have drowned hundreds half a century ago (e.g. http://en.wikipedia.org/wiki/North_Sea_flood_of_1962 ).

    ReplyDelete
  16. "...the North Sea surge of 5 Dec '13 was high but sea defenses kept it out everywhere"

    CRK, I can assure you that the defences did not keep the sea out of everywhere, unless you are referring solely to the NL coastline and explicitly not the east coast of England?
    Significant flooding occurred along the Humber and along the Lincolnshire coast (e.g. South Ferriby and Boston), with several hundred homes flooded and with many people still out of their homes 4 months down the line (as is the way of these things)

    Hasis

    ReplyDelete
  17. The Dec 5th storm's tidal surge damage is discussed here:

    https://en.wikipedia.org/wiki/North_Sea_flood_of_2013#Storm_surge


    More generally, this is a useful primer on UK spending on flood protection:

    http://www.parliament.uk/briefing-papers/SN05755.pdf

    (pdf)

    ReplyDelete
  18. Adam, I stand somewhat nationalistically corrected :)
    Yes, I didn't think of the English e.g. Tyne coasts - that afternoon-into-night news on that didn't really make it across to the continent.
    I am a little bit shocked it happened to actually thousands of homes in that region. The region lost hundreds killed in the '53 surge disaster, haven't they built against the sea like sensible continental Europe did??

    The wiki link mentions "The Eastern Scheldt storm surge barrier closed all its 62 locks on Thursday night and several areas around Rotterdam experienced some flooding.[77] Minor flooding was also reported in Dordrecht and Vlaardingen.[78]"

    These floodings (quays and car parks mainly) were outside the sea defense line and did not count for me.

    ReplyDelete
  19. Some places are better protected than others. Note that places like Boston flooded when their defences were over-topped. Those defences are (generally, I wouldn't want to commit too strongly) better than in 1953.

    ReplyDelete
  20. How should I read this? Nate Silver responds to the negative reaction of Paul Krugman to his new venture:
    To be sure, the difference in Mr. Krugman’s views could reflect a decline in quality for FiveThirtyEight. The web site has brought on almost two dozen new employees and contributors. And it has expanded its coverage beyond politics into sports, economics and other areas.

    Sounds like he is saying that his new hires are not as good as he is. Interesting, Curry would write.

    ReplyDelete
  21. Now I checked that Boston video.
    Interesting how the surge there went above the '53 event. Dutch records were not broken, but 26 Feb 1990 ('Vivian') did a good test here. I believe we can carry a surge a metre over 1953 with no harm done.

    Looking at the British east coast, well, 'I dunno' is what they say, what with sea level rise... Otoh both Xaver and '53 featured a special coincidence: surge and spring tide. But I would view neither as synoptically perfect events for that coast while '53 was certainly that for the Belgian and Dutch (for us only the southwestern area) while Xaver was especially well formed for the Elbe mouth.

    No AGW trend in this type of surge. They ever so slowly become more dangerous from sea level rise. Otherwise it is waiting for a bit of a freak for a true test of the different countries' defences.

    ReplyDelete
  22. @cRR Kampen

    Yes, there's a lot of factors involved that determine where gets the worst impacts. The problem is, while these events are synoptic/tidal coincidences, they will occur every now and then, and each time the sea is a bit deeper, as you say.

    So yes, there probably is no/little AGW trend in those sorts of events, there are two ways they can be affected by the changes.

    Firstly in the slr - some of those flood defences coped by less than the projected rise over the next twenty years (Ipswich, IIRC was one example). This is not just from sea level rise, but also changes in tidal effects from that rise. Maribo has an article on a paper that discusses this for NW Europe. Some will get mitigating effects, but some will be worse off from high tides - and thus this sort of event. IIRC the expected slr in the Bristol Estuary is about another >=20cm by 2030. The SE is sinking as well, due to isostatic rebound, so places like the Thames and Norfolk/Suffolk could get more (I've not looked into the figures).

    Secondly, more intense rainfall can add loading to tidal rivers, which means you get increased flow in both directions.

    I don't really have any research to hand on this, but if anyone has anything on the latter effects (esp. some UK rivers) I'd be grateful for a link (I will probably get round to looking for it myself, but day job and all that...).

    ReplyDelete
  23. the North Sea surge of 5 Dec '13 was from a different angle of arrival in coastal geology like in radio waves the measurement of the angle of arrival is a method for determining the direction of propagation and the intensity of a wave incident.....

    with radio-waves no problemo

    with oceanic or mediterranean waves is a more complex problem

    but the 1953 incident is a very complex and poorly documented incident

    ReplyDelete
  24. EFS_Junior27/3/14 3:49 PM

    Brian,

    Paywalled references?

    http://www.sciencedirect.com/science/article/pii/S1462901108000142
    http://antoine.leblois.free.fr/Taf/Chap_1/Revue_recente_INTRO/MultiriskCropInsurance/Statistics_risks_methods/Australian_normalized_catastrophicEvents.pdf

    http://ascelibrary.org/doi/abs/10.1061/(ASCE)1527-6988(2008)9:1(29)
    http://www.nhc.noaa.gov/pdf/NormalizedHurricane2008.pdf

    Took like 60 seconds to find.

    ReplyDelete
  25. EFS - thanks for the assistance. Anyway, looks like Skeptical Science has it covered.

    Still would like to know how mitigation protects against storm landfall in Australia and not in the US.

    ReplyDelete

Dear Anonymous,

UPDATE: The spambots got clever so the verification is back. Apologies

Some of the regulars here are having trouble telling the anonymice apart. Please add some distinguishing name to your comment such as Mickey, Minnie, Mighty, or Fred.

You can stretch the comment box for more space

The management.