Introduction
In these exercises we will look at a case study of the interaction between a Hurricane and Atlantic cut-off low in the context of ensemble forecasting. During the course of the exercises, we will explore the scientific rationale for using forecast ensembles, why they are necessary and how ensemble forecasts can be visualized. We will also use principle component analysis (PCA) and clustering to analyse the behaviour of the ensemble. In this interesting case study, the HyMEX field campaign was also underway. We will see how the forecast products were used during this campaign.
Pantillon, F., Chaboureau, J.-P. and Richard, E. (2015), 'Vortex-vortex interaction between Hurrican Nadine and an Atlantic cutoff dropping the predictability over the Mediterranean, http://onlinelibrary.wiley.com/doi/10.1002/qj.2635/abstract
Caveat on use of ensembles for case studies
In practise many cases are aggregated in order to evaluate the forecast behaviour of the ensemble. However, it is always useful to complement such assessments with case studies of individual events, like the one in this exercise, to get a more complete picture of IFS performance and identify weaker aspects that need further exploration.
Recap
ECMWF operational forecasts consist of:
- HRES : T1279 (16km grid) highest resolution 10 day deterministic forecast
- ENS : T639 (34km grid) resolution ensemble forecast (50 members) is run for days 1-10 of the forecast, T319 (70km) is run for days 11-15.
Saving images and printing
To save images during these exercises for discussion later, you can either use:
"Export" button in Metview's display window under the 'File' menu to save to PNG image format. This will also allow animations to be saved into postscript.
or use the following command to take a 'snapshot' of the screen:
ksnapshot
Virtual machine
If using the metview/openifs virtual machine with these exercises the recommended minimum memory is 4Gb. These exercises use a relatively large domain with high resolution data. Some of the plotting options can therefore require significant amounts of memory. If the virtual machine freezes when running metview, please try increasing the memory assigned to the VM.
Outline of the exercises
TO BE DONE (flowchart as last year?
- Study track & development of Hurricane Nadine and the N.Atlantic cut-off low
- Study HRES & ENS forecasts of the time
- Study HRES & ENS forecasts using the 2016 ECMWF ensemble forecast
- .... PCA analysis ..
- ...etc...
Starting up metview
To begin:
metview
Please enter the folder 'openifs_2016' to begin working.
Exercise 1. The ECMWF analysis
Learning objectives
- Study development and behaviour of Hurricane Nadine
- ...TO BE DONE
Hurricane Nadine and the cut-off low
For these tasks, use the metview icons in the row labelled 'Analysis'
an_1x1.mv : this plots horizontal maps of parameters from the ECMWF analyses overlaid on one plot.
an_2x2.mv : this plots horizontal maps of parameters from the ECMWF analyses four plots to a page (two by two).
an_xs.mv : this plots vertical cross-sections of parameters from the ECMWF analyses.
Task 1: Mean-sea-level pressure and track
Right-click the mouse button on the 'an_1x1.mv' icon and select the 'Visualise' menu item (see figure right)
After a pause, this will generate a map showing mean-sea-level pressure (MSLP).
Now drag the mv_track.mv icon onto the map. This will add the track of Nadine. Although the full track of the tropical storm is shown from the 10-09-2012 to 04-10-2016, the ECMWF analyses (for the purpose of this study) only show 15-09-2012 to 25-09-2012.
In the plot window, use the play button in the animation controls
to animate the map and follow the development and track of Hurricane Nadine.You can use the 'Speed' menu to change the animation speed (each frame is every 6 hours).
Close unused plot windows!
Please close any unused plot windows if using a virtual machine. This case study uses high resolution data over a relatively large domain. Multiple plot windows can therefore require significant amounts of computer memory which can be a problem for virtual machines with restricted memory.
Task 2: MSLP and 500hPa geopotential height
Right-click the mouse button on the 'an_1x1.mv' icon and select the 'Edit' menu item.
An edit window appears that shows the Metview macro code used to generate the plot. During these exercises you can change the top lines of these macros to alter the choice of parameters and plot types.
#Available parameters: # mslp,t2,wind10,speed10,sst # t,z,pt,eqpt [850,700,500,200] # wind,speed,r[925,850,700,500,200] # w700, vo850, pv320K
The surface fields (single level) are: MSLP (mean-sea-level-pressure), 2-metre temperature (t2), 10-metre wind arrows (wind10), wind-speed at 10m (sqrt(u^2+v^2): speed10), sea-surface temperature (sst).
The upper level fields are: temperature (t), geopotential (z), potential temperature (pt), equivalent potential temperature (eqpt), wind arrows (wind), wind-speed (speed), relative humidity (r).
These fields have a list of available pressure levels in square brackets.
To plot upper level fields, specify the pressure level after the name. e.g. z500 would plot geopotential at 500hPa.
Some extra fields are also provided: vertical velocity at 700hPa (w700), relative vorticity at 850hPa (vo850) and potential vorticity at 320K.
Wind fields are normally plotted as coloured arrows. To plot them as wind barbs add the suffix '.flag'. e.g. "wind10.flag" will plot 10m wind as barbs.
With the edit window open, find the line that defines 'plot1':
#Define plot list (min 1- max 4) plot1=["mslp"] # use square brackets when overlaying multiple fields per plot
Change this line to:
plot1=["z500.s","mslp"]
The suffix '.s' means plot the 500hPa geopotential as a shaded plot instead of using contours (this style is not available for all fields)
As above, click the play button and then animate the map that appears.
Questions: TO DO
Task 3: Changing geographical area
Right-click the mouse button on the 'an_1x1.mv' icon and select the 'Edit' menu item.
In the edit window that appears
#Map type: 0=Atl-an, 1: Atl-fc, 2: France mapType=0
With mapType=0
, the map covers a large area centred on the Atlantic suitable for plotting the analyses and track of the storm.
With mapType=1
, the map also covers the Atlantic but a smaller area than for the analyses. This is because the forecast data in the following exercises does not cover as large a geographical area as the analyses.
With mapType=2
, the map covers a much smaller region centred over France.
Change, mapType=0
to mapType=1
then click the play button at the top of the window. Repeat using mapType=2
to see the smaller region over France.
These different regions will be used in the following exercises.
Animate the storm on this smaller geographical map.
Task 4: Wind fields, sea-surface temperature (SST) and satellite images
The 'an_2x2.mv' icon allows for plotting up to 4 separate figures on a single frame. This task uses this icon to plot multiple fields.
Right-click on the 'an_2x2.mv' icon and select the 'Edit' menu item.
#Define plot list (min 1- max 4) plot1=["mslp"] plot2=["wind10"] plot3=["speed500","z500"] plot4=["sst"]
Click the play button
at the top of the window to run this macro with the existing plots as shown above.Note that each plot can be a single field or overlays of different fields as in the an_1x1.mv macro.
Wind parameters can be shown either as arrows or as wind flags ('barbs') by adding '.flag' to the end of variable name e.g. "wind10.flag".
Plotting may be slow depending on the computer used. This reads alot of data files.
Task 5: Satellite images
Open the folder 'satellite' by doubling clicking (scroll the window if it is not visible).
This folder contains satellite images (water vapour, infra-red, false colour) for 00Z on 20-09-2012 and animations of the infra-red and water vapour images.
Double click the images to display them.
Use the an_1x1.mv and/or the an_2x2.mv macros to compare the ECMWF analyses with the satellite images.
Questions
Q. What do you notice about the SST?
TODO
Task 6: Cross-sections
The last task in this exercise is to look at cross-sections through Hurricane Nadine and the cut-off low.
Right click on the icon 'an_xs.mv', select 'Edit' and push the play
button.This generates a plot with a map of MSLP, a red line and underneath a cross-section plot along that red-line.
The default plot shows potential vorticity (PV) and potential temperature roughly through the centre of the Hurricane and the cut-off low.
Changing forecast time
Cross-section data is only available every 24hrs, not every 3 hrs for the horizontal maps.
This means the steps in the macro is only valid for the times: [2012-09-20 00:00], [2012-09-21 00:00], [2012-09-22 00:00], [2012-09-23 00:00], [2012-09-24 00:00], [2012-09-25 00:00]
Changing fields
Note that a smaller set of fields is available for cross-sections: temperature (t), potential temperature (pt), relative humidity (r), potential vorticity (pv), vertical velocity (w), wind-speed (speed; sqrt(u*u+v*v))
Changing cross-section location
#Cross section line [ South, West, North, East ] line = [30,-29,45,-15]
The red line indicating the cross-section location can be changed in this macro by defining the end points of the line as shown above. Remember that if the forecast time is changed, the storm centres will move and the cross-section line will need to be repositioned to follow specific features. This is not computed automatically, but must be changed by altering the coordinates above.
This completes the first exercise. You have now learnt how to use the key macros, alter fields for plotting and animate fields. The next exercises use similar macros.
Exercise 2: The operational HRES forecast
Recap
The ECMWF operational deterministic forecast is called HRES. At the time of this case study, the model ran with a spectral resolution of T1279, equivalent to 16km grid spacing.
Only a single forecast is run at this resolution as the computational resources required are demanding. The ensemble forecasts are run at a lower resolution.
Before looking at the ensemble forecasts, first understand the performance of the operational HRES forecast of the time.
Available forecast
Data is provided for a single forecast starting from 20th Sept 2012, as used in the paper by Pantillon et al.
HRES data is provided at the same resolution as the operational model, in order to give the best representation of the Hurricane and cut-off low interations.
Questions
- How does the HRES forecast compare to analysis and satellite images?
- Was it a good or bad forecast? Why?
Available plot types
For this exercise, you will use the metview icons in the row labelled 'HRES forecast' as shown above.
hres_rmse.mv : this plots the root-mean-square-error growth curves for the operational HRES forecast for the different lead times.
hres_1x1.mv &
hres_2x2.mv
: these work in a similar way to the same icons used in the previous task where parameters from a single lead time can be plotted.
hres_to_an_diff.mv : this plots a single parameter as a difference between the operational HRES forecast and the ECMWF analysis. Use this to understand the forecast errors from the different lead times.
Parameters & map appearance. These macros have the same choice of parameters to plot and same choice of mapType
, either the Atlantic sector or over Europe.
Getting started
Task 1: Forecast error
In this task, we'll look at the difference between the forecast and the analysis by using "root-mean-square error" curves as a way of summarising the performance of the forecast. Root-mean square error curves are a standard measure to determine forecast error compared to the analysis and several of the exercises will use them.
Using the hres_rmse.mv icon, right-click, select 'Edit' and plot the RMSE curves for MSLP (mean-sea-level pressure). Repeat for the 10m wind-gust parameter wgust10.
Repeat for both geographical regions: mapType=0 and mapType=1.
Q. What do the RMSE curves show?
Task 2: Compare forecast to analysis
a) Use the hres_to_an_runs.mv icon (right-click -> Edit) and plot the MSLP and wind fields. This shows a comparison of 3 of the 4 forecasts to the analysis (the macro can be edited to change the choice of forecasts).
b) Use the hres_to_an_diff.mv icon and plot the difference map between a forecast date and the analysis.
We suggest looking at only one forecast lead-time (run) but when working in teams, different members of the team could choose a different forecast.
If you want to change the default date, edit the following line:
#Model run run=2013-10-24
Task 3: Team working
As a team, discuss the plots & parameters to address the questions above given what you see in the error growth curves and maps from task 2.
Look at the difference between forecast and analysis to understand the error in the forecast, particularly the starting formation and final error.
Team members can limit to a certain date and choose particular parameters for team discussion.
Remember to save (or print) plots of interest for later group discussion.
- Plot and animate MSL + 500hPa maps showing track of Nadine
> 1 : Nadine MSLP and T2m (or better SST) tracking 15-20 september
> 2 : Satellite views on the 20th (provided by Etienne, if possible to put on the VM)
> 3 : Studying of the horizontal maps (analysis + forecasts)
> 4 : Studying and building of the vertical x-sections (analysis + forecasts)
Notes from Frederic: email 7/4/16
day 1
day 2
I suggest to focus on the period before the 20-25 September and to study
Nadine and the cutoff, not yet at the impact o the Mediteranean area (we
leave that for later)
Agreed. The first exercise will be to examine the track & changes in the storm using the analyses.
*Here are some inputs concerning Day2 and Nadine's study on day 2:
*1)* T1279 Analysis 0920 + t+96 deterministic forecast 0924 (t+96h) -->Agreed.
focusing on the interaction between Nadine and the cutoff. Maybe an
extra plot of the forecasted rainfall at t+96 over France ?
*2)* Ens T639 forecasts : I saw that T639 is the 2012 operational
ensemble resolution, so we will see the same bifurcation in the
scenarios as explained in Pantillon : the visualization of the spread,
the plumes, the spaghettis, ... will help here. I am sure you have great
ideas on this topic. Maybe we can propose some horizontal maps of each
(or some) members ?
I think the exercises we used last year will fit well here. I will start drafting the exercises on the wiki and ask you to help and comment.
I presume the exercises should be in English? Or should we do a side-by-side English/French version?
*3)* PCA and clustering : if you manage to put it in Metview this will
be great lo look at the 2 distinct patterns. I asked Florian Pantillon
his NCL sources to do the trick. I'll use it to build an extra NCL
exercice with PCA, clustering and compositing, if we have time. The file
format needed will be netcdf.
We think our PCA code can be used to reproduce fig 5. We thought we could also reproduce fig 6 but instead of dots plot the ensemble number. Then the students can build the clusters (Fig. 7) but grouping the ensemble members together?
We (here) need to try this and see how far we can reproduce the rest of the figures in the paper.
*4)* Ensemble runs : initial (EDA+SV) and model (SPPT+SKEB) : same as
last year
SST experiment might be too much, except if we shorten the ensemble study...
See above. My preference after talking with people here is to use the comparison between 2012 operational ensemble and 2016 operational ensemble. The lower res (T319) ensembles; control (EDA+SV), (SPPT_SKEB) ensembles for this case are running now and we can include the data (as long as filesize does not become an issue). But honestly, I do not think there will be time. I will leave it to you to decide!
We counted 7hrs total for the practicals (not including the SCM). Part of that time the students will need to prepare some plots for the discussion on Friday.
I am concerned about the time available. Perhaps my talk on weds 9.30-10 could be shortened to 15mins.
*Day 3
SCM experiments
For the SCM we thought that it might be interesting to use the SCM for a point near Toulouse that experienced very heavy rainfall during HYMEX. Then we get the students to adjust the entrainment rates (similar to the convection exercises here) to see what impact it has on the precipitation?
Exercise 2.
As I am writing I am beginning to wonder if we should not make 2 groups : one for task 4 and one for task 5. Tasks 1-3 would be for all students. This would allow to keep the CDO task. What do you think ?
As I am writing I am beginning to wonder if we should not make 2 groups : one for task 4 and one for task 5. Tasks 1-3 would be for all students. This would allow to keep the CDO task. What do you think ?
On 04/05/16 15:25, FERRY Frédéric wrote:
*T1279 Analysis* : 20121020 0000UTC to 20121025 0000UTC --> Only the 20
september analysis will be looked at but I assume you need to get the
other analysis to compute the RMSE in day 2 ?
*Extended analysis* : 15-20 September just for MSLP and T2m (or better
the SST) --> Nadine tracking before the 20th
15-20th ok. We'll produce all the variables for this date range as the students can animate the fields if they have them. If data size is an issue we'll restrict 15-19th to just T2m & MSLP.
The SST is a good point as this will be interesting to compare between the analysis and the forecasts. I'll add it to the list of fields.
I assumed the students would look at the analyses on all days. The first task would be to look at the track of Nadine & the cut-off low and animate different fields across the date range. This is very easy to do with the metview macros. We can also get them to plot difference maps between the forecasts and the analyses (and yes we need all the days to compute RMSE) if you think that's instructive.
As the data is time-consuming to extract (it will take ~ 7 days), I would rather get more than we think we might use in the exercises to allow the students to explore their own ideas during the workshop.
> *T1279 Deterministic forecast* : 20 000UTC analysis and 20121025 0000
> UTC deterministic forecast (t+120).
Ok. We'll use 6hrly data throughout.
*Extended deterministic forecast* : 20-28 September just for MSLP :
Etienne told me that the ECMWF model of the 20 000UTC proposed a very
extreme situation on the 28th, with a storm over Gibraltar. This would
be a way to illustrate the limits of a deterministic approach.
I'd rather not add another 3 days of data (26-28) to show a single forecast. If this is just a short example to the students during the workshop, can we plot these and include just the figures? (ie. not an interactive exercise)
*Horizontal maps* (analysis + forecast) :
All ok apart from:
3 : Equivalent potential temperature at 850 hPa + winds at 850 hPa +
vertical velocity at 600hPa + MSLP in background --> focussing on the
Can we use 700hPa VV instead of 600, to be consistent with data on other levels? We will need VV on multiple levels in order to plot the x-sections (see below), though these will only be available 00Z on each day. The horiz. maps will have VV available 6hrly but on selected levels only (we're proposing 200, 500, 700, 850).
*Vertical x-sections* in the cutoff and in the low :
**PV + winds (preferably normal winds) + if possible potential
temperature --> to look at the cold core or warm core structure of the
systems on the vertical and the signature in PV and winds.
PV + relative humidity + vertical velocity --> a more classical
x-section that we use to see if a PV anomaly is accompanied with
vertical motion or not.
For these x-sections we can choose 3 or 4 times that appear to be
interesting. Interactivity would be good, to make the students look a
little bit in the code.
We will not have enough space to store all pressure levels for the x-sections 6hrly, so we propose only each day at 00Z. Once we have the data extracted from the archive, we might need to review this.
In Metview it is possible to graphically draw a line on a horizontal map (e.g. MSLP) to identify the x-section location. The end-pts can then be used to plot the x-section itself. The students can experiment finding the centre & plotting x-sections from the different days.
Q: How do you want to plot the ensemble data? Plot individual members, groups of members or x-section of the cluster of members
*Satellite* : we have the satellite images of the situation (IR, WV,
cloud classification, IR-Visible composite). We can send them to you to
put on the VM.
Great!
Proposed tasks for Day 1 :
5 : Beyond D+5 deterministic scenario : MSLP only
See above. I think this will be too much data for us to include beyond day 5.
Concerning the ensemble runs, 6 hourly data is OK. If you have space on
the VM it would be interesting to go up to D+10 (or D+15). This would
allow to try and look at the extreme member over Gibraltar on the 28
September.
Task 1 : forecast error
Task 2 : compare forecast to analysis
we think task 2 fits better on day 1 when the students visualize the analyses & HRES forecast. If time, they look at HRES forecast error though that could be day 2. Day 2 then becomes mostly about the ensemble.
Task 3 : visualize ensembles (plumes, ensemble spread, spaghetti, stamp,
CDO)
--> These 3 tasks from last year are very interesting. To gain time
maybe that we should put a group on each item for task 3 or suppress
task 2 ? The CDO adds a "statistical" taint to the workshop do you think
we can adapt it to our case ?
Maybe we can put the students into the role of the forecaster during HyMEX; given the CDF plots & clusters, what would the students have decided to do and why?
We did this last year asking the question what would the participants do if the Queen had a birthday garden party and the forecast showed a risk of severe weather It was a good discussion.
Perhaps this could be led by Etienne based on Hymex experience?
Task 4 : PCA and clustering
Thanks for this. It will also be interesting to see what the latest operational ensemble does with this case (we do not yet know!). Comparing the two will be interesting.
To sum up : 15-25 for the analysis, 6 hourly. The 15-19 analysis will be for Nadine study mainly. The 20th is the beginning of the forecast. Forecast range is 5 days (20-25). So we need also the 20-25 analyses for the comparison analysis VS forecast. If you can archive some figures of MSLP for the deterministic forecast just to check the storm over Gibraltar. Concerning the VV field 700 is OK to avoid extracting another level. For the x sections 00Z is sufficient, 200 500 700 850 should be enough to see something I hope. I did not know the x-sections was easy to do in Metview, this will save some time ! I’ll ask Etienne to get the satellite images for the 20th. Sorry for the tephigrams ! Day 1 synopsis : I agree that we have to do the comparison deterministic VS analysis on the first day to concentrate on the ensemble on the second day. Task 1 : Animating analysis from 15 to 20, mainly tracking Nadine Task 2 : studying of the deterministic forecasts from 20-25 september, and the interaction between Nadine and cutoff Task 3 : compare the deterministic forecast to the analysis. I like your idea of plotting the difference maps, this could be a better way to localize the area where the model was wrong. We can leave the forecast error and the RMSE if time I am not sure I understand you question about the ensemble data plotting ? I was thinking that on the second day we could visualize the classical ensemble products : plumes, spread, spaghettis, stamps + the figures in Pantillon concerning the cluster composites and the cluster tracks. What do you mean about the x-section of the cluster ? Maybe you have other suggestions.
So day 2 « menu » would be : Looking at the ensemble products and the cluster products and making a decision for Hymex field campaign —> They will have Etienne's forecaster feedback the day after. I’ll ask Etienne his ideas for the workshop tasks on this topic. Looking at the impact of ocean coupling on the ensemble prediction. Tell me if you manage to redo the clustering and the composites in Metview, I hope it will work. If you manage to redo figures 5 6 7 8 and 10 I think I’ll have to tell Jean-Pierre to focus more during his presentation on the vortex-vortex interaction and the CRM sensitivity experiments he made. This will leave the cluster analysis for the students to discover.
Hello Glenn, Here are a few comments concerning your previous emails : 1- I can provide the satellite images from 18th 00 UTC to 24th 21 UTC by a 3h or 6h step (except for the 22nd, which is not available...). 2- Véronique Ducrocq could play the role of an HyMeX operation director being the client of the students' forecast. This forecasting exercise could be done by the 8 students following the forecasting option (with me as their "teacher"), whereas the 18 others (informatic or statistic options) could keep doing more sensitivity tests while manipulating the code of the model (with Frédéric and you). 3- It would be very interesting to briefly tackle with the ECMWF Data Targeting System which was one of the observation strategies used during HyMeX SOP1. I precisely asked Véronique Ducrocq to speak about DTS during her presentation on Day 1. 4- ARPEGE and IFS deterministic charts are available at the French Met School between 18th and 24th sept (except the 20th runs unfortunately !). As far as I was the HyMeX forecaster myself before the 24th sept. event, I would be very interesting in the MSLP fields from the 20th 00UTC run between 25th and 28th sept. , in order to be able to illustrate (in my own Day 3 presentation) the propagation of this impressive "Gibraltar storm" I mentionned into my daily meeting report. A 6h step would be perfect, even if it is only a paper-scanned version...
Appendix
Datasets available
The following datasets are available on the Virtual Machine for this workshop:
Acknowledgements
Exer