IS-ENES workshop on Statistical downscaling of climate scenarios for the impact communities
- https://is.enes.org/archive-1/archive/events/is-enes-workshop-on-statistical-downscaling-of-climate-scenarios-for-the-impact-communities
- IS-ENES workshop on Statistical downscaling of climate scenarios for the impact communities
- 2012-10-16T09:00:00+02:00
- 2012-10-17T16:00:00+02:00
- What
- Project Event
- When
- Oct 16, 2012 09:00 AM to Oct 17, 2012 04:00 PM (Europe/Vienna / UTC200)
- Where
- UPMC, Jussieu, Paris
- Contact Name
- Sylvain Lassonde
- Add event to calendar
-
iCal
IS-ENES workshop
Statistical downscaling of climate scenarios for the impact communities.
A CMIP5 perspective.
16-17th October 2012, Paris
A workshop about statistical downscaling and bias correction have been organised in Paris ( Jussieu University campus) under the IS-ENES project (WP11-JRA5 workpackage) on October 16 and 17th, 2012. It was a very interactive workshop, with keynotes, contributions from each participant with short presentations, 2 round tables and a live discussion. The main objective of the IS-ENES JRA5 workpackage is to build a bridge between climate scientists and impact community users. This workshop discussed methodologies of statistical downscaling and bias correction that is being used to answer various impacts community needs.
The main highlights out of the presentations were that :
- Impactmodelsareusedwithinmanydifferentsectorsthatcouldbeaffectedbyclimatechange.Thesemodelscanbeascomplexthanclimatemodels,and they have their own uncertainties.
- Dynamicalandstatisticaldownscaling methods arecomplementary.They must be compared and canbecoupled together inordertoproducebetterdownscaleddata.
- Thequalityofdownscaledresultsdependsonthequality and time scale ofobservationsusedtotrainthestatisticalmodel.
- Itis important to check thephysicalconsistenciesoftheresultsbecausebiascorrectionandstatisticaldownscaling can introducespuriousrelationshipsbetweenclimatevariableswhentheyaretreatedindependantly.Somemethodsbetterpreservethetrends.
- Thehypothesisofstationarityisusuallyusedwithin bias correctionand statistical downscalingmethods:careshouldbetakentovalidateasmuchaspossible this hypothesis, especially concerning the predictor-predictandrelationship.
One of the working groups discussed the inventory of statistical downscaling and bias correction methodologies. These methods will be summarized using matrices of strengths and weaknesses for each method. The other working Group on uncertainties will summarize its discussions using a matrix of uncertainty.
The recommendations of the workshop are as follows:
- Increase the interaction of IS-ENES with other research projects on statistical downscaling and on impacts studies (COST VALUE, StaRMIP, etc.)
- Providing Probability Density Functions (PDF) of climate indicators can be as important as providing directly climate variables.
- One should be careful when downscaling to a very high resolution because uncertainties could increase at a very high level.
- It is important to characterize uncertainty at each step of the downscaling-correction method (observations, reanalysis, bias correction, downscaling, etc.)
- In order to better characterize uncertainty it is advised to analyse an ensemble of simulations and use a multi-model approaches.
- All methods are not equivalent for a given study. It is advisable to select the more suitable methods considering regional characteristics (for example a topology analysis).
Click here to read the Workshop minutes.
Other useful documentation:
Presentations:
E. Palazzi: Stochastic rainfall downscaling
Short contributions:
Working groups:
WG1: Methods and datasets requirements
WG2: Collection of ideas and uncertainties
Statistical Downscaling Classification Matrix
