EGU 2018
4/9/18 - 4/13/18
Data Science
pangeo - work on using python in the HPC. Open source project.
psyplot - Interactive plotting
opendatacube - Satellite data analysis
Machine Learning
Extreme learning machines e.g. hpelm - High-Performance Extreme Learning Machines
Random forrest e.g. scikit-learn - a machine learning method
Sub-seasonal NAO
Ayarzagüena - ENSO on NAO in early winter via Gulf of Mexico by ppt anomalies in GoM and Central America.
Fromang - QGPV idealized mjo perturbation on NAO
Hong - ET forcing the 2015 MJO-ENSO
NAO
My talk - 2008/2009 winter had a majow SSW however it was not well forecast. See this paper (comment by Dr. Blanca Ayarzagüena-Porras) and this paper on the event. Could look into likelihood of one year not being forecast. Therefore, 08/09 may just be explained by noise (Comment by Tim Wollings). paper by Doblas-Reyes on multi-model NAO forecasting: 4 AGCM. Argue the use of model PC.
Bernat Jimenez-Esteve (Zurich) - Enso -> Auleition low -> waves in NAtl. Enso -> walker cell -> tropical atlantic -> waves. + NAO late winter El Nino. Transient eddies (2<T<8 bandpass filter); Q-stationary (T>10 days). Opposite response for El Nino and La Nina. paper
Tyrrell et al (Helsinki) - Oct 2016 weak polar vortex, warm anomaly. Advection, not sea-ice loss cause warming
Richard Greatbatch et al (Kiel, de) - Relaxation experiments to determine remote influence on NAO. Relax tropics in one experiment and another in the troposphere. Strat influence is lower that trop influence. MJO is a large influence in some years when it is suppressed. ECMWF seasonal forecasting is good at SSW.
Wolf et al (University of Reading) - Quasi-stationary waves on temperature extremes in Europe.
Rousi - NAO 'flavors' using clustering.
Mecking - Ocean vs atm on Europe summer temperatures
- Reduced November Barents-Kara sea ice is linked to a more negative winter NAO.
General Climate Science
PICO - retrograde earth experiences. Used paraview for the visualizations
Re-insurance
Chaucer re-insurance company.
Fathom Bristol/Uni Bristol - global flood hazard layer.
OASIS LMF Open source Loss Modeling Framework
Ocean Remote Sensing
Ardhuin - SKIM: Potential ESA mission for measuring surface currents and waves.
Bourassa - WaCM: Wave and current potential satellite mission.
Amores - limits of ocean eddy sensing
Ciani - surface currents in the Med
Decadal Forecasting
Martineau - ocean temps and weather extremes.
Li - Relationship between NAO and AMO.
Tsunami
Aoyanagi - Tsunami evacuation simualtion
Waves
Jean-Raymond Bidlot - sea-state dependency of air-sea fluxes in ECMWF Earth System Model. Charnock term is dynamic. Cd is influenced by Ch - heat. Janssen (1997) sea-state on heat flux. Tech Mem 239.
Sasmal - coastal waves in Sagami Bay, Japan. WW3 and SWAN.
Markina - NAtl EKE on Hs
Ardhuin - Hs spectrum looks like current spectrum. Collard paper. Climate change initiative starting off in IFREMER. Need currents for high resolution model. Testing global tides and CMEMS 1/12 hindcast.
Extratropical Cyclones
Stoll - Climatology of polar lows.
Priestley - ETC clustering.
Posters (see below)
Bertoncelj - Med sea-level/waves storms.
Kettle - North Sea storm surge
Waves
Stefanie Rynders - wave, current, time and sea-ice on offshore loads: update morison eq. and add sea-ice. Look at hazards in different regions. e.g. waves in North Sea, currents in shelf slope, tides in some coastal areas. Funding by SOS-SOS (Safer Operations at Sea - Supported by operational simulations) and here.
North Sea: AMM7, WW3 7km
Arctic: CICE, NEMO, SWARP
Git for science workshop
You can use git to version control your script locally
$ mkdir test
$ vi my_script.py # add 'test'
$ git init # Turned the current directory into a local git repository
$ git status # show status
$ git add my_script.py
# Similarly git rm my_script.py
$ git commit -m "initial commit" # -m is message
$ vi my_scrit.py # change to test2
$ git commit -a -m "added a line to the script" # Add the file and commit
$ git log # show changes
# Have a look at code state on previous commit
$ git checkout ...
# Back to maskter
$ git checkout master
# Revert latest changes
$ git revert HEAD # esc -> :z
You can use git to version control remotely
# sign up for github and create a new repository
# call it 'EGU_test' and description 'test repo for EGU course'
# Push an existing git repo here
$ cd test
$ git remote add origin https://github.com/USERNAME/EGU_test.git
$ git remote -v
# Push to the github repo
$ git push -u origin master
# Create a new branch
$ git checkout -b awesome_feature
# See branches
$ git branch
$ vi my_script.py # change to testb
$ git commit -a -m "working on a new feature"
# Switch back to master branch
$ git checkout master
# Push new branch to github repo
$ git push origin awesome_feature
# Click pull request button on github
# Update local master branch with github repo
$ git pull origin master
# Delete local branch
$ git branch -d awesome_feature
# Delete the remote branch
$ git push origin --delete awesome_feature
Skill Scores
Krzysztofowicz - Bayesian Approach to Statistical Post-Processing.
Friederichs - modelling of spatial extremes.
Continuous Rank Probability Score (CRPS) here
van Straaten - stat post-proc of high-res ppt EPS.
Simon - prob forecasting of thunderstorms: generalized additive models; ECMWF will soon have a lightening diagnostic.
Thorarinsdottir - proper skill scoring: squared error; absolute error;
ignorance score (probabilistic) e.g. here and here ; CRPS.
statistical postprocessing of ensemble forecasts book
Peirce skill score; odds ratio skill score e.g.
High resolution modeling
PRIMAVERA - H2020 EU consortium on community wide high resolution modeling.
Haarsma13 TCs in EC-Earth 25 km AGCM. More TCs/ETCs in Western Europe in the future. Warm seclusion storms paper.
TC and ETC tracks will be available for PRIMAVERA.
TCs
Vidale - stochastic physics (SP) and resolution on TCs: No stochastic physics kills TCs. Stochastic physics is equivalent to increasing resolution. More TCs.
SP acts as vortex seeder? Not obvious relationship to vws even though they change.
S2S
Francisco J. Doblas-Reyes - S2S climate services: prodhomme15 land surface initialization on forecast. prodhomme16 mod res on seasonal forecasting; Equitable Threat Score. improved predictions for agriculture; S2S4E - S2S for energy; Lledo18 paper - wind anomaly on west coast of US.
Alice Grimm - SA monsoon and the influence of the MJO. RW from Central Pacific to SA. Lin08 eval skill to predict MJO. Bivarate correlation.
Yuejian Zhu - 3-4 week forecast GEFS. FV3GEFS. SubX May 1st 2014 - May 26th 2016. int every 7 days. Stochastic schemes. 2-Tiered SST (not coupled). SP improves in tropics. The schemes add skill in later weeks. FV3 dycore.
Seok-Woo Son - QBO on MJO. Vitart17 Son17. Moisture advection over the maritime continent important for MJO. Cloud long-wave radiation. BMSE amplitude and BMSE phase error. MJO better predicted during EQBO winters by about 5 days. What is seasonal cycle like of MJO?
Gilbret Brunet - wave processes across time-scales.
Laura Baker - over/under confidence of NAO in EUROSIP. Eade14 under-confidence - Signal is too weak. Unpredictable noise. Don't standardize there is a huge spread. GA3, GA6, MF Sys3, Sys4, JMA Sys2. Box based NAO e.g. Stephenson06. GA3 is best. Multi-model is slightly higher. RPC ratio of predictable components see Eade16. >1 is under-confident. More ensembles, more under-confident. All models have common drivers i.e. similar inter-annual variability. ECMWF low skill but not under-confident.
Christopher White - Applications of S2S: paper. s2sdata. Sub-seasonal drivers: SAM, blocking. Early warning, disaster risk. response to resilience. Q: are we getting ahead of ourselves with the lack of skill of science? Case studies can be useful but may over egg the skill. ask about shipping.
Carlo Buontempo - ECMWF Copernicus Climate Change Service: C3S seasonal. EU Seasonal hydrological forecast. Shipping with OSM. Carlo.Buontempo@ecmwf.inf
Ole Wulff - Subseasonal prediction of 2003 European summer heat wave. atm blocking -< SST anomalies; trop-extrop RW. Spring dry soil moisture. Split ensembles by choosing members that get the gph500 best. Also split based on soil moisture.
Mike DeFlorio - sub-seasonal skill of atmospheric rivers: ARcatelogue. AR anomaly as a function of MJO phase.
Michael Walz - Predictability of extreme wind speed over Europe. paper Stasitical entropy; predictive information; predictive power. >95th percentile. Integrated over time steps. Mostly correlated with NAO. Not a great study.
Chaim Garfinkel - Predictability of SSW based on MJO. Strat and MJO on NAtl paper.
Ben Green - Sub-seasonal errors in FIM-iHYCOM.
High Resolution Modeling
Stevens - Extreme Earth: Advancing global storm resolving models to usher in a new era of climate modeling and climate change science
Neumann - Storm-Resolving Simulations of the Climate System
Bauer - Energy-efficient Scalable Algorithms for Weather Prediction at Exascale
Roberts - TCs in PRIMAVERA
Chantry - Model precision
Mavilia - Resolution and stoch pyhys on Euro-Atl weather regimes
Satoh - NICAM model
Voigt - High-res of an ETC
Manganello - TC landfall in high-res
Vanniere - Hyd cyc in high-res
Budich - Models for next gen comp
Gettelman - variable res CCSM