KiK lunch 13 december 2010 Wat hebben we aan nòg gedetailleerdere klimaatscenario’s? R. Boers
KiK lunch 13 december 2010 We will leave all nuances behind…….. Purpose / hope: start thinking for yourself about vision It is NOT an attack on certain research programs at KNMI!!! Introduction, then a number of propositions open for discussion
KiK lunch 13 december 2010 ENSEMBLES run of Upper Tropospheric Water Vapor Question: If you start now: How long would you need to measure WV before the trend would be statistically significant? Answer: 50 years, and the estimated trend would not be closer to the truth than 20%!
KiK lunch 13 december 2010 But, also to ask another question:………….. Info can be used to inform decision making on investing in an expensive instrument to measure a relevant climate variable What is the skill of this ENSEMBLE run? A skillful prediction is one that improves upon a prediction based on a naieve baseline [such as climatology, or persistence] [see Pielke paper]
KiK lunch 13 december 2010 Answer: For the present, none whatsoever! And if in 50 years the computed trend does indeed confirm to the observations, we may never know whether this is caused by a true predictive skill or just randon chance
KiK lunch 13 december 2010 Four ENSEMBLES runs [r1,r2,r3, MIROC] Total price (several M Euros) 36 runs based on [r1,r2,r3] Total price 1 Eurocent 36 runs based on [r1,r2,r3,MIROC], Total price 1 Eurocent True climate ? What does the questionmark error bar tell us about skill? Can you prove that the skill of these new predictions generated for 1 Eurocent is less than those generated for several M euros??
KiK lunch 13 december 2010 Our ignorance in understanding future climate is irreducable The skill of climate prediction is zero Dessai et al [2004] [the principle of irreducable ignorance]
KiK lunch 13 december 2010 Cascade of uncertainty
KiK lunch 13 december 2010 focus on ‘low regret strategies: i.e. strategies that work well over a range of conditions faced now and potentially faced in the future [with or without climate change] In the absence of good predictions, what will decision makers do? They will make decisions, but…………..
KiK lunch 13 december 2010
In this diagram climate scenarios are completely absent
KiK lunch 13 december 2010 Sea defense in future [does not depend much on climate scenarios]: Prepare for high water: 1)Increase dyke top levels by [say] 1.3 m 2)Yawn and go to sleep Low Regret Example for the Netherlands Or The Principle of Robust Decision Making ! The Veerman solution !
KiK lunch 13 december 2010 Water management in future [does not depend much on climate scenarios]: Prepare for dry and wet summers: 1)Improve water storage capacity [risk>> dry] 2)Improve overflow capacity [risk>> wet] 3)Improve water efficiency [risk>> dry] 4)Improve regional understanding of climate controls and land surface feedbacks Low Regret Example for the Netherlands
KiK lunch 13 december 2010 Occurrence of heat waves in the future, using the ENSEMBLES predictions
KiK lunch 13 december 2010 # heat waves, where a heat wave is defined based on reference local climatology Time period Reference Climate # of occurrences of heat waves in consecutive 30 year time periods
KiK lunch 13 december 2010 Prepare population for heat waves [does not depend much on climate scenarios]: Prepare for extremely hot days: 1)Make useful predictions of maximum temperature and dewpoints three days in advance 2)Improve advance communication to prepare population for heat wave onset 3)Yawn and go to sleep (!) Low Regret Example for the Netherlands Acceptance that the Dutch population will quickly adapt to a new [and perhaps] warmer climate
KiK lunch 13 december 2010 A climate model at higher spatial resolution [think HARMONIE] will only change the precision of a climate forecast, not its accuracy ECEarth RACMO HARMONIE
KiK lunch 13 december 2010 Proposition 1: As choices for adaptation will rely more and more on Robust Decision Making, the value of climate scenarios in policy advice will quickly diminish: KNMI will need to adjust to a vision for the future that does not rely on climate scenarios as relevant output Proposition 1: Keuzes voor aanpassing van beleid zullen meer en meer gebruik maken van robuste beslissingsstrategieen: De waarde van de klimaatscenarios als beleidsinstrument zal dus afnemen. KNMI zal zich dus ook moeten aanpassen, en wel naar een toekomst waarbij klimaatscenarios minder nadrukkelijk als relevante output voor beleid wordt gepresenteerd [Principle of Robust Decision making]
KiK lunch 13 december 2010 Proposition 2: We do not need any more high resolution climate scenarios, because they do not bring us closer to the truth Proposition 2: We hebben geen hoge resolutie klimaatscenarios nodig omdat die ons niet dichter bij de waarheid brengen [Principle of Irreducable Accuracy]
KiK lunch 13 december 2010 Proposition 3: The KNMI-NEXT scenarios will not provide added value in comparison to the presently published scenarios Proposition 3: De KNMI-NEXT scenarios zijn een herhaling van zetten, ze hebben geen ‘added value’ t.o.v. eerder gepubliceerd materiaal. [Principle of Irreducable Ignorance]
KiK lunch 13 december 2010 KNMI time series of fractional cloudiness [1950 – 2009] SPOT THE DISCONTINUITY
KiK lunch 13 december 2010 Schiphol time series of a)visibility [Top] b) Relative humidity [Bottom] Is visibility change caused by chances in aerosol, or changes in land use?
KiK lunch 13 december 2010 De staat van het klimaat; Een koele blik op een verhit debat [Marcel Crok] Gaat voornamelijk over het gebrekkige bewijs voor de stelling dat klimaatverandering wel plaatsvindt, c.q. heeft plaatsgevonden
KiK lunch 13 december 2010 Proposition 4: Most if not all debate over climate change is about proving that it is real and not imagined. So why don’t we spend much more time on that issue? Proposition 4: Het meest intense debat over klimaat gaat over of het wel of niet plaatsvindt / heeft plaatsgevonden. Daar moeten we dus veel meer tijd aan spenderen.
KiK lunch 13 december 2010 Proposition 6: Op het KNMI besteden we -te veel tijd aan het uitwerken van klimaatscenario’s en -te weinig tijd aan de keten van meten naar weten
KiK lunch 13 december 2010 Proposition 5: KNMI climate research is MODELCENTRIC Als je het niet gelooft; check de acronyms die gebruikt worden! HIRLAM, ECMWF model, ECEarth, RACMO, TM3, TM5, Lotos Euros, ECHAM, HARMONIE, ARPEGGE, ALADIN, AROME, IFS, MOZART, prep- IFS, OASIS, MOCAGE Proposition 5: KNMI klimaatonderzoek is te MODELCENTRIC