Hao Zuo, Andrew Bennett, Andrew Dawson Research Department

836

Weakly coupled ocean–atmospheredata assimilation in the ECMWF

NWP system

Philip Browne, Patricia de Rosnay,Hao Zuo, Andrew Bennett,

Andrew Dawson

Research Department

To be submitted to Remote Sensing

December 18, 2018

Series: ECMWF Technical Memoranda

A full list of ECMWF Publications can be found on our web site under:http://www.ecmwf.int/en/research/publications

Contact: [email protected]

©Copyright 2018

European Centre for Medium-Range Weather ForecastsShinfield Park, Reading, RG2 9AX, England

Literary and scientific copyrights belong to ECMWF and are reserved in all countries. This publicationis not to be reprinted or translated in whole or in part without the written permission of the Director-General. Appropriate non-commercial use will normally be granted under the condition that referenceis made to ECMWF.

The information within this publication is given in good faith and considered to be true, but ECMWFaccepts no liability for error, omission and for loss or damage arising from its use.

http://www.ecmwf.int/en/research/publications

Weakly coupled DA in NWP at ECMWF

Abstract

Numerical weather prediction models are including an increasing number of components of the Earthsystem. In particular, every forecast now issued by the European Centre for Medium-range WeatherForecasts (ECMWF) runs with a 3D ocean model and a sea ice model below the atmosphere. Ini-tialisation of different components using different methods and on different timescales can lead toinconsistencies when they are combined in the full system. Historically, the methods for initialisingthe ocean and the atmosphere have typically been developed separately. This paper describes anapproach for combining the existing ocean and atmospheric analyses into what we categorise as aweakly coupled assimilation scheme. Here we show the performance improvements for the atmo-sphere by having a weakly coupled ocean–atmosphere assimilation system compared to an uncoupledsystem. Using numerical weather prediction diagnostics we show that forecast errors are decreasedcompared to forecasts initialised from an uncoupled analysis. Further, a detailed investigation intospatial coverage of sea ice concentration in the Baltic sea shows much more realistic structure givenby the weakly coupled analysis. By introducing the weakly coupled ocean–atmosphere analysis, theocean analysis becomes a critical part of the numerical weather prediction system and provides aplatform from which to build ever stronger forms of analysis coupling.

1 Introduction

As of June 2018, the European Centre for Medium-range Weather Forecasts (ECMWF) has a coupledforecasting system for all timescales in that every forecast from the high resolution 10 day forecasts, theensemble forecasts, the monthly forecasts to the seasonal prediction system runs an Earth system model.Specifically this means that there is a 3-dimensional ocean model and a sea ice model that runs coupledto the atmosphere, wave, and land surface components. Such a multi-component Earth system modelneeds initialisation of each of its components, and this manuscript is concerned with how the ocean andatmosphere are initialised together for the purposes of numerical weather prediction.

In order to advance numerical weather prediction, ECMWF is developing its modelling and its dataassimilation towards an Earth system approach [ECMWF, 2016]. When forecasts of medium-range tolonger-range are the focus, components of the Earth system that are typically slower than the atmospherebecome more important. This is both in terms of their presence in the model and an accurate specificationof their initial conditions [Bauer et al., 2015]. Such components include the ocean and sea ice, but alsothe land surface, waves, aerosols, and how they interact with each other and the atmosphere [Maclachlanet al., 2015]. Figure 1 represents the various components present in the system.

The land surface and waves are fully established components of the ECMWF systems. Aerosols aretreated separately within the Integrated Forecasting System (IFS) as part of the Copernicus AtmosphereMonitoring Service (CAMS). The ocean has been used in the IFS for seasonal applications since 1997[Stockdale et al., 1998], for monthly forecast since 2002 [Vitart, 2004] and in the Ensemble Forecasts(ENS) from the initial time step since 2013 [Balmaseda et al., 2013, Bauer and Richardson, 2014]. Inlate 2016 an interactive sea ice model was added to the ENS. Only now are these components starting tointeract with the atmospheric analyses.

In order to make the most of the Earth system approach in a forecast, the components should be somehowconsistent with one another. If the different components are not internally consistent they are sometimesreferred to as unbalanced. This lack of balance can lead to fast adjustments in the system in the initialstages of the forecast in a phenomenon known as initialisation shock. Initialisation shock can be reducedby initialising the various components together via coupled data assimilation [Mulholland et al., 2015].

Much of the literature on coupled assimilation has focused on initialisation of forecasts for seasonal to

Technical Memorandum No. 836 1


Figure 1: Components of ECMWF’s IFS Earth System. Along with the atmosphere, there are the ocean,wave, sea ice, land surface, and lake models.

decadal timescales. For example the Japan Agency for Marine-Earth Science and Technology (JAM-STEC) have a fully coupled 4D-Var system used for experimental seasonal and decadal predictions[Sugiura et al., 2008, Masuda et al., 2015, Mochizuki et al., 2016]. The National Oceanic and Atmo-spheric Administration Geophysical Fluid Dynamics Laboratory (NOAA/GFDL) have a coupled assim-ilation system based on the ensemble Kalman filter (specifically the ensemble adjustment Kalman filter)to initialise decadal predictions [Yang et al., 2013, Zhang et al., 2014]. The NOAA National Centersfor Environmental Prediction (NOAA/NCEP) have a coupled assimilation system [Saha et al., 2014] forsubseasonal and seasonal predictions, as well as reanalysis [Saha et al., 2010]. A prototype system wasbuilt in March 2016 in the Japan Meteorological Agency Meteorological Research Institute (JMA/MRI),designed to replace the ocean-only observation assimilation approach. The atmosphere component isupdated every 6-hours by 4D-Var with a TL159L100 uncoupled inner-loop model, while the ocean com-ponent runs on a 10-day cycle using 3D-Var with an incremental analysis update. Experimentation withthis system for coupled reanalysis and NWP is underway [Penny et al., 2017].

The U.S. Naval Research Laboratory (NRL) has a different focus to most centres, running a coupledmodel with most resources dedicated to the ocean component. Their global coupled model goes up to anocean resolution of 1/25 and is initialised by separate assimilation systems. In the near future they planto implement the interface solver of Frolov et al. [2016] to allow more coupling within the analysis.

Environment and Climate Change Canada (ECCC) have recently begun producing their global deter-ministic NWP forecasts using a coupled ocean–atmosphere model. This model is currently initialisedseparately in each of its components. The UK Met Office have a system to initialise global coupledNWP forecasts where the background for each component in a 6 hour assimilation window comes fromthe coupled model [Lea et al., 2015], although this is not yet operational.

Under the ERA-CLIM2 project ECMWF has piloted techniques for coupled ocean–atmosphere data as-similation that were applied in the context of reanalysis [Laloyaux et al., 2016, Schepers et al., 2018].These are the Coupled European Reanalysis of the 20th century (CERA-20C) and the CERA-SAT reanal-ysis using the modern day satellite observation system. The assimilation method developed for CERA

2 Technical Memorandum No. 836


involved “outer-loop” coupling within the 4D-Var algorithm of the atmosphere and the ocean. Thismethod has a high level of coupling in the analysis, which as Figure 2 shows, can mean that the wholeNWP system can be degraded by model biases in the ocean component of the coupled model.

Figure 2: Impact of first implementation of outer loop coupling (quasi strongly coupled data assimilation)on high resolution global NWP. The presence of model bias in the western boundary currents of the oceanshows itself as degradations (red) to the 24 hour forecast scores of vector winds at 1000hPa. Results areobtained from IFS cycle 45R1 at a resolution of 25km (TCo399) with a 0.25 ocean, based on globalouter loop coupling over the period 2017-06-01–2017-07-02.

As the first steps into coupled ocean–atmosphere data assimilation for NWP at ECMWF we have chosento adopt a weaker form of coupled assimilation than in the CERA system.

Following a World Meteorological Organisation (WMO) meeting on coupled assimilation, Penny et al.[2017] defined weakly and strongly coupled data assimilation (and variations thereof). Their definitionswere as follows:

• “Quasi Weakly Coupled DA (QWCDA) assimilation is applied independently to each of a subsetof components of the coupled model. The result may be used to initialize a coupled forecast.”

• “Weakly Coupled DA (WCDA): assimilation is applied to each of the components of the cou-pled model independently, while interaction between the components is provided by the coupledforecasting system.”

• “Quasi Strongly Coupled DA (QSCDA): observations are assimilated from a subset of componentsof the coupled system. The observations are permitted to influence other components during theanalysis phase, but the coupled system is not necessarily treated as a single integrated system at allstages of the process.”

• “Strongly Coupled DA (SCDA): assimilation is applied to the full Earth system state simultane-ously, treating the coupled system as one single integrated system. In most modern DA systemsthis would require a cross-domain error covariance matrix be defined.”

Hence QWCDA might be thought of as uncoupled assimilation to initialize a coupled model. Observa-tions in one component never influence the analysis of the other component. In WCDA, an observationof one component is not able to directly influence the analysis of the other component in the valid assim-ilation window. However, as a coupled forecast is used, the observational information gets propagated



to the background used for subsequent analysis cycles, hence there is a lag by which observations caninfluence different components. The CERA system falls under the QSCDA category, where observationsfrom each component can influence the analysis of the other within a single analysis window. SCDA issimply treating a coupled system as a multivariate assimilation problem and no special terminology ormathematical analysis is necessary.

In this paper we introduce a form of weakly coupled data assimilation which allows for the differenttimescales in the ocean and atmospheric analysis windows. The atmosphere and the ocean are coupledimplicitly at a frequency of 24 hours, determined by the frequency of the slowest component to update.

The remainder of this paper is organised as follows. In Section 2 we describe the various components ofthe Integrated Forecasting System and describe both uncoupled and weakly coupled ocean–atmosphereassimilation strategies. The experimental design is described in Section 3. Section 4 shows and discussesthe experimental results, and gives a detailed examination of local impacts to sea ice. Finally in Section5 we look to the future developments of the weakly coupled data assimilation system at ECMWF.

2 IFS

The ECMWF Integrated Forecasting System consists of multiple components. The main component forit all is the upper atmospheric analysis. The dynamical model which propagates the analysis from onecycle to the next contains a limited number of components. They are the atmospheric model [ECMWF,2017a,b], the land model [Balsamo et al., 2009], the lake model [Dutra et al., 2009], and the wave model[ECMWF, 2017c]. The atmosphere is represented on a 3D reduced Gaussian grid and its analysis isfound by using 4-Dimensional Variational data assimilation (4D-Var) in incremental form. A numberof outer loops are used and the minimisation is performed at increasingly high resolution. The numberof outer loops and resolution of the inner loops is dependent on the resolution of the nonlinear model.The land data assimilation component is weakly coupled to the atmosphere. They share the same modelto produce the first guess, and the state of the land surface and the state of the atmosphere are modifiedseparately [de Rosnay et al., 2014]. Subsequent forecasts are initialised using the latest analysis of theatmosphere and the land surface. This is archetypal weakly coupled assimilation as defined previously.Currently the land surface has various components. The snow analysis is performed using a 2D-OI, asis the soil temperature analysis. Soil moisture is analysed using a Simplified Extended Kalman Filter(SEKF). Similarly to the land, the wave analysis is weakly coupled to the atmosphere. However, thefirst guess used for the wave analysis is not the same first guess that is used for the atmosphere. Thefirst guess is the nonlinear trajectory of one of the outer loops of the atmospheric 4D-Var. Currently thefinal trajectory is used. This means that, in a given cycle, observations of the atmosphere will influencethe wave analysis in that given cycle. The opposite is not true – wave observations will not modify theatmospheric state during that cycle. These observations will only modify the atmospheric state at thesubsequent cycles due to the interactions in the forecasts that cycle the analysis.

The model that cycles the analysis does not contain a dynamical ocean model or sea ice model. For thepurposes of this paper we say it is uncoupled, referring to ocean–atmosphere interactions. The lowerboundary of the atmosphere needs to be supplied; the sea-surface temperature (SST) field and sea iceconcentration (CI) field is required.



Observations

Over 40 million observations are processed and used daily with the vast majority of these coming fromsatellites. These include polar orbiting and geostationary, infrared and microwave imagers, scatterome-ters, altimeters, and GPS radio occultations [English et al., 2013]. In addition to the satellite observationsthere are in situ observations used from aircraft, radiosondes and dropsondes, as well as observationsfrom ships, buoys, land based stations and radar [Haiden et al., 2018].

For the sea surface L4 gridded products are used to give global coverage of sea-surface temperaturesand sea ice concentrations. The L4 product used is the Operational Sea Surface Temperature and SeaIce Analysis (OSTIA) [Donlon et al., 2012], a 0.05 resolution dataset that is solely observation based.For SST OSTIA combines satellite data from the Group for High Resolution Sea Surface Temperature(GHRSST) and in-situ observations to produce a daily analysed field of foundation sea-surface temper-ature. Sea ice concentration fields in OSTIA are derived from the EUMETSAT Ocean Sea Ice SatelliteApplication Facility (OSI SAF) L3 OSI-401-b observations of sea ice concentration. Lake ice con-centration observations that are outside of the domain of OSI-401-b are taken from an NCEP sea iceconcentration product.

4D-Var, HRES and the EDA

The above observations are assimilated with the 4D-Var methodology (see e.g. Rabier et al. [2000])that uses, amongst other details, hybrid-B and a weak constraint term. A single high resolution (HRES)analysis and forecast is produced. The flow dependent component of the background error covariancematrix B comes from an Ensemble of Data Assimilations (EDA) [Bonavita et al., 2012] that solvessimilar 4D-Var problems but at a lower resolution and with stochastically perturbed observations. TheEDA currently runs with 25 members. The HRES analysis is performed twice daily over a 12 houranalysis window from 2100Z (0900Z) to 0900Z (2100Z). The 4D-Var is solved in incremental form with3 outer loops, such that each inner loop minimisation is performed on a lower resolution grid. From eachanalysis a 10 day coupled ocean–atmosphere forecast is produced. For more details on the configurationsee Haseler [2004].

OCEAN5

OCEAN5 is a reanalysis-analysis system with 2 streams - behind real-time and real-time. The 3 dimen-sional ocean Nucleus for European Modelling of the Ocean (NEMO) model and the Louvain-la-Neuve2 (LIM2) sea ice model are coupled and used as the model for within OCEAN5. The OCEAN5 analy-sis is initialized from a behind real time ocean and sea ice coupled reanalysis, known as ORAS5 [Zuoet al., 2018] (see purple boxes in Figure 3). The variables temperature, salinity, and horizontal currents(T,S,U,V ) are analysed using the 3D-Var First Guess at Appropriate Time (FGAT) assimilation tech-nique. The length of the assimilation window varies from 8 to 12 days and is split into two chunks (seeblue boxes in Figure 3), the first of which is 5 days long. In parallel, a separate minimisation is performedto analyse sea ice concentration using the same 3D-Var FGAT method.

Observations that are assimilated currently are in situ profiles of temperature and salinity, and satellitederived sea level anomaly and sea ice concentration observations. For SST a relaxation is performedtowards the OSTIA SST product.

The ocean and sea ice analysis system requires forcing fields in the form of a surface wind field, surface



temperature and humidity fields, and surface fluxes. These come from the HRES analysis (and forecastfor the final day of the OCEAN5 assimilation window). The surface fluxes consist of downward solarradiation, thermal radiation downwards, total precipitation, and snowfall. From the wave model the oceanrequires forcing fields of significant wave height, mean wave period, coefficient of drag with waves, tenmetre neutral windspeed, normalized energy flux into the ocean, normalized wave stress into the ocean,and Stokes drift. A full description is given in Zuo et al. [2018].

Uncoupled approach/work flow

From the above description, we can see that the HRES system can stand alone. It does not require anyinformation from the OCEAN5 analysis. OCEAN5 on the other hand, requires forcing fields from anatmospheric analysis to operate.

Under this system, observations in the atmosphere will modify the atmospheric state. This change inatmospheric state will lead to a change in the forcing fields by which the ocean analysis is driven. Thiswill lead to a change in the ocean analysis.

Observations of the ocean (unused by OSTIA, such as observations of currents) will not modify theatmospheric state as no information from the ocean model is propagated back to the atmosphere. Thissystem as a whole can be thought of as a “one way” coupled assimilation system. The flow of informationfrom the atmosphere to the ocean is depicted in the diagram in Figure 3 as orange arrows.

WCDA

We have seen that the atmospheric analysis requires the provision of an SST field and a sea ice field foruse as its lower boundary condition. Similarly the ocean analysis requires a set of atmospheric forcingfields to drive the ocean only analysis.

To form a weakly coupled ocean–atmosphere data assimilation system, fields from the OCEAN5 analysisare used as the lower boundary conditions for the atmospheric analysis over the ocean, rather than takingdirectly fields from the external OSTIA product. This will mean that observations of the ocean andthe sea ice which previously would only influence the ocean analyses will also modify the atmosphericanalysis via the lower boundary conditions. The effect is not within a given assimilation cycle, but it isdelayed.

Figure 3 shows the information flow between atmosphere and ocean. Consider an atmospheric observa-tion on day 11 (within the highlighted region). This will change the analysis fields of the atmosphere onday 11, which will lead to the forcing fields that drive the ocean to change. Hence this observation willhave an effect on the latest ocean analysis valid at the start of day 12 (which in this diagram has a 11 daywindow from day 1 to day 12).

Now instead consider an ocean or sea ice observation on day 11. This directly changes the ocean/seaice analysis for that day, but as there is no feedback to the atmosphere until the end of the window, theatmospheric analysis for day 11 is unchanged. The impact of that ocean or sea ice observation is onlydetected by the atmosphere on day 12, when the updated sea ice analysis is seen as the lower boundarycondition for the atmosphere (magenta arrow in Figure 3).

Recall the definition of Weakly Coupled DA (WCDA): assimilation is applied to each of the compo-nents of the coupled model independently, while interaction between the components is provided by the



Analysis day1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

Atmos, land, and waves Ocean BRT Ocean NRT chunk 1 Ocean NRT chunk 2

Figure 3: Weakly coupled assimilation system information flow. This is a simplified plot ignoring a 3hour offset of the systems. Orange arrows show the existing transfer of forcing from the atmosphere tothe ocean. Magenta arrows show the addition of taking the OCEAN5 fields and using them as the lowerboundary condition for the atmospheric analysis, thus forming the WCDA system. The highlightedregion is discussed in an example in the text.

coupled forecasting system. Clearly the assimilation is applied independently to the atmosphere and theocean/sea ice. Interaction between the components is not provided by a coupled model (i.e. a singleparallel task on the supercomputer) but by the coupled forecasting system. That is, over a 24 hour periodthe forecasting system passes forcings from the ocean to the atmosphere and lower boundary conditionspassed from the ocean/sea ice to the atmosphere. Hence we categorise this as a weakly coupled dataassimilation system for the ocean–atmosphere interaction.

Partial coupling

The analysis of SST and sea ice concentration from OCEAN5 may not always be better than the OSTIAproduct. In particular there are known deficiencies in the OCEAN5 analysis that can lead to degradationsin forecast performance. For example in the extratropics the position of western boundary currents suchas the Gulf Stream are known to be less accurate in the OCEAN5 analyses compared to OSTIA. Hencefor WCDA, as with the model, the flexibility to take ocean fields only over specific regions and notglobally has been developed.

The sub-optimality of the ocean analyses are due to a well known model bias in the ocean model which



has been recognised already in the coupled model used to produce the 10 day forecasts [ECMWF, 2014,Mogensen et al, 2019a,b, Keeley et al, 2019]. The solution to this problem has been to use, for the model,a “partial-coupling” approach, where the tendencies, rather than the absolute values, of SST are passedto the atmosphere. The partial coupling is required at latitudes (> 25) where the ocean model is unableto resolve eddies. Partial coupling can be described by the following equations.

SSTIFS(t) = SSTNEMO(t)+α(t)(SSTREF(0)−SSTNEMO(0)) (1a)

where α(t) is a function of lead time with α(0) = 1 decreasing to 0 by the end of the forecast. Thereference field at initial time, SSTREF(0) is given by

SSTREF(0) = βSSTOSTIA(0)+(1−β )SSTNEMO(0), (1b)

where β is spatially varying and is depicted in Figure 4.

0.0 0.2 0.4 0.6 0.8 1.0

Figure 4: β coefficient from (1b), indicating initial SST is taken from the ocean analysis in the tropics(20S to 20N) and OSTIA in the extratropics. The 5 transition regions can be seen as the bands consistingof intermediate colours.

Equation (1) evaluated at t = 0 gives an initial SST field as

SSTIFS(0) = βSSTOSTIA(0)+(1−β )SSTNEMO(0) (2)

which is the SST from NEMO in the tropics and the SST from OSTIA in the extratropics. For WCDAwe have therefore chosen to align the SST in the analysis with that used to initialise the forecast, i.e.following equation (2).

3 Experimental setup

A set of two experiments are conducted: the first a control which does not use weakly coupled assimi-lation, and the second an experiment with weakly coupled assimilation. Both experiments are based onIFS cycle 45R1. They run at a 9km global resolution (TCo1279), and both use the same uncoupled atmo-spheric EDA. The time period of the investigation is from 20170609 to 20180521. From each analysis a10 day coupled ocean–atmosphere forecast is produced.

Control - uncoupled assimilation

The initial conditions for the ocean component are taken from an ocean only analysis which takes itsforcing fields from the operational HRES system, i.e. the same resolution but atmosphere only and driven



by OSTIA boundary conditions. The atmospheric analysis uses OSTIA SST and CI fields globally as itslower boundary conditions.

Experiment - WCDA

An ocean analysis is run alongside the atmospheric analysis. The sea ice concentration field used for thelower boundary of the atmospheric analysis comes from the ocean analysis. Similarly the sea-surfacetemperature field comes from the ocean analysis, although this is restricted to the tropics only as de-scribed in equation (2) and Figure 4. That is, the SST comes from the ocean analysis between 20S and20N, OSTIA outside of 25N(S), and a linear interpolation of the two in the 5 degree band from 20N(S)to 25N(S).

The atmospheric analysis is otherwise identical to the control run. Similarly, except for the forcing fieldscoming from the atmospheric analysis rather than the operational HRES system, the ocean analysis isidentical to the ocean analysis used in the control experiment. A comparison of the WCDA experimentand the uncoupled control setup is shown in Figure 5.

Experimentatmosphere

analysis

Experimentocean

analysis

Controlatmosphere

analysis

Controlocean

analysis

10 DayForecast

10 DayForecast

SST/CIHRES

Uncoupled DA WCDA

Figure 5: Schematic of experiment design. On the right shows the experiment, with the weakly coupledDA passing information between atmosphere and ocean analyses. On the left shows the control atmo-spheric analysis getting its SST and CI from the external OSTIA product, and the forcing field for thecontrol ocean analysis coming from the operational HRES system.

4 Results and discussion

For brevity we choose to show normalized differences in RMSE as a measure of forecast errors [Geer,2016]. Normalised RMSE differences (dRMSE) for an experiment e compared to a control experimentc is defined as

dRMSE =||xxxe

f (T : T + t)− xxxea(T + t)||− ||xxxc

f (T : T + t)− xxxca(T + t)||

||xxxcf (T : T + t)− xxxc

a(T + t)||,

where xxx f (T : T + t) and xxxa(T + t) refer to forecasts of length t, and analyses, valid at time T + t, and thenorm || · || is the root mean square throughout number of samples through time.



Atmospheric performance

Figure 6 shows the impact of WCDA on atmospheric humidity and temperature. There are 3 distinctregions of impact - the tropics and the poles - coming from the separate influences of WCDA throughtropical SST and SIC respectively. The areas of hashed shading indicate that the differences in forecasterrors are statistically significant. It is clear that the impact from WCDA does not give long range impactsto the upper troposphere or to the spatial regions where WCDA is not active.

In the tropical region of Figure 6 we can see that the impact of the tropical SST is detected from thesurface up to around 850hPa, in both temperature and humidity. The hashed shading show that these im-provements, indicated by blue colours, are statistically significant. The maps in Figure 7 shows clearlythat the improvement from SST is restricted to the latitudinal band for which the WCDA SST is active.Within this band there are variations in how much benefit we get from the WCDA. For instance with tem-perature at 1000hPa the strongest positive impacts are seen in the Arabian Sea, and the eastern Atlanticand Eastern Pacific (associated with cold tongues).

In the Arabian Sea there is evidence of improvement to other variables, such as to low level winds andsignificant wave heights (not shown). This indicates that the SST WCDA has improved the position ofthe summer monsoon which is a known difficult feature to forecast well. The regions of positive impactin the Atlantic and the equatorial Pacific are regions that tend to have high cloud cover. This persistentcloud makes observing the SST from satellites very difficult, and so it may be that the use of the oceanmodel within the OCEAN5 analysis system is able to effectively fill the observational gap.

In the polar regions we see significant improvements in forecast errors due to the weakly coupled assim-ilation. Figures 8 and 9 show that the improvements due to sea ice encompasses the entire extent of thesea ice cover, and are not simply confined to the ice edge. The influence of the WCDA SIC extends upto roughly 700hPa (Figure 6). This is further vertically than the impact of SST seen in the tropics. Thiscan be explained by considering Figure 10 that shows the usage of microwave humidity soundings in thesouthern hemisphere. Such soundings are rejected over sea ice, and hence rejecting more contaminatedsoundings could lead to a more consistent atmospheric state in these regions. Further area averageddRMSEs of forecast error are shown in the appendix.



(a) dRMSE of humidity

(b) dRMSE of temperature

Figure 6: Latitude–pressure diagram of the normalised difference in RMSE between WCDA and controlfor humidity (top) and temperature (bottom) forecasts at 12h (left) and 24h (right) lead times, for theperiod 20170609 to 20180521. Hashed areas indicate statistically significant differences.



Figure 7: Spatial maps of the normalised difference in RMSE between WCDA and control for tem-perature at 1000hPa (top), skin temperature (middle), and sea-surface temperature (bottom) forecasts at12h (left) and 120h (right) lead times (48 hours lead time for temperature at 1000hPa), for the period20170609 to 20180521.



Figure 8: Spatial maps centred on the Antarctic of the normalised difference in RMSE between WCDAand control for 1000hPa temperature (top), 1000hPa humidity (middle), and sea ice concentration (bot-tom) forecasts at 12h (left) and 24h (right) lead times (120 hours lead time for sea ice concentration), forthe period 20170609 to 20180521.



Figure 9: Similar to Figure 8, but focused on the Arctic.



98.8 99.0 99.2 99.4 99.6 99.8 100.0number [%, normalised]

3

4

5

Cha

nnel

num

ber

0 1×106 2×106 3×106 4×106 5×106

number

a b

Instrument(s): FY3B MWHS Tb Area(s): S.HemisFrom 00Z 9−Jun−2017 to 12Z 21−May−2018

WCDA

Uncoupled analysis

Figure 10: Observation usage of MWHS-1 in the southern hemisphere for the WCDA experiment (black)and the uncoupled control (red).



Ocean performance

Surface Heat Flux: Damping(a) Global

2017

-06

2017

-07

2017

-08

2017

-09

2017

-10

2017

-11

2017

-12

2018

-01

2018

-02

2018

-03

2018

-044

6

8

10

12

14

16

W/

m2

(b) Tropics

2017

-06

2017

-07

2017

-08

2017

-09

2017

-10

2017

-11

2017

-12

2018

-01

2018

-02

2018

-03

2018

-04

4

6

8

10

W/

m2

(c) Northern Hemisphere Extratropics

2017

-06

2017

-07

2017

-08

2017

-09

2017

-10

2017

-11

2017

-12

2018

-01

2018

-02

2018

-03

2018

-040

10

20

30

40

W/

m2

(d) Southern Hemisphere Extratropics

2017

-06

2017

-07

2017

-08

2017

-09

2017

-10

2017

-11

2017

-12

2018

-01

2018

-02

2018

-03

2018

-04−20

−10

0

10

20

30

W/

m2

Figure 11: Monthly averaged ocean heat flux correction from the uncoupled analysis (red) and the weaklycoupled analysis (black). This is split by region showing global (top left), the tropics (top right), thenorthern hemisphere extratropics (bottom left) and the southern hemisphere extratropics (bottom right).

Figure 11 shows the area averaged heat flux correction in the ocean analyses. This flux correction repre-sents the strength of the relaxation towards the OSTIA SST product. One can see that the flux correctionin the extratropics is broadly similar in both experiments, with only the northern hemisphere extratropicsshowing a slightly smaller flux correction in July and August in the WCDA experiment. However in thetropics the flux correction is consistently substantially smaller in WCDA than uncoupled. This showsthat the SST coupling in the tropics is leading to an SST field that is more consistent than the uncoupledanalysis, requiring less corrections. We postulate that this could be due to the improved timeliness ofthe the SST that the atmospheric component uses (for the uncoupled analysis the OSTIA SST field isonly available on the day after its valid time), however this requires a more detailed investigation that isoutside the scope of this paper.

Figure 12 shows the area averaged surface temperature increments in the ocean from the uncoupled and



Assimilation Increment of Temperature at 0.50576m(a) Global

2017

-06

2017

-07

2017

-08

2017

-09

2017

-10

2017

-11

2017

-12

2018

-01

2018

-02

2018

-03

2018

-041

1.2

1.4

1.6

1.8

2

2.2·10−8

K/s

(b) Tropics

2017

-06

2017

-07

2017

-08

2017

-09

2017

-10

2017

-11

2017

-12

2018

-01

2018

-02

2018

-03

2018

-04

1.5

2

2.5

3

·10−8

K/s

(c) Northern Hemisphere Extratropics

2017

-06

2017

-07

2017

-08

2017

-09

2017

-10

2017

-11

2017

-12

2018

-01

2018

-02

2018

-03

2018

-04

0.5

1

1.5

2

·10−8

K/s

(d) Southern Hemisphere Extratropics

2017

-06

2017

-07

2017

-08

2017

-09

2017

-10

2017

-11

2017

-12

2018

-01

2018

-02

2018

-03

2018

-04−0.5

0

0.5

1

1.5

·10−8

K/s

Figure 12: As Figure 11 but for surface assimilation temperature increments.

weakly coupled assimilation experiments. Similarly to Figure 11, the larger benefits from WCDA ap-pear to be seen in the tropics with the coupling of SST, showing reduced increments compared to theuncoupled analysis. However in the northern hemisphere extratropics the weakly coupled assimilationincrements are consistently smaller than the increments in the uncoupled system. For the southern hemi-sphere extratropics the picture is mixed and it seems like the 2 systems are behaving very similarly.



Operational impacts - Baltic sea detailed sea ice investigation

Here we take a detailed look at the spatial distribution of sea ice in the Baltic sea, a particularly chal-lenging area for sea ice concentration analyses. Figure 13 shows different representations of the sea iceon 2018-02-17 from different sources. Figure 13a is a manually produced ice chart from FMI/SMHIshowing sea ice concentration at the north of the Gulf of Bothnia, as well as in the eastern end of theGulf of Finland. Figure 13b shows the available OSI SAF L3 sea ice concentration observations in thearea. Note that because of the geography of the area observations are only available in the centre of theGulfs - coastal contamination requires that those points near the coast be masked from the product.

Figures 13c and 13d show the sea ice from uncoupled assimilation and WCDA experiments. One can seethat the uncoupled analysis effectively smooths out the L3 observations and does not capture the highice concentrations along the northern coastlines. WCDA on the other hand does a much better job atcapturing the structures seen in the manual ice chart. In particular the use of the background informationcoming from the dynamical model gives a much more realistic spatial distribution of the ice field. Figure14 is similar but on 2018-03-05, the date of maximal sea ice extent in the Baltic sea for the 2017/2018season.

5 Conclusions and future plans

This paper has investigated the impact of weakly coupled ocean–atmosphere data assimilation on theECMWF forecasts. The WCDA approach allows components of the Earth system with different timescalesand assimilation methods to be linked together. As an alternative to using purely observation based L4products for the lower boundary of the atmosphere, with their associated latencies, WCDA allows dy-namical models of the ocean and sea ice to fill in the gaps in observations and propagate fields to theappropriate time. The results presented show the use of WCDA improves the coupled forecasts in theregions near to the interface of those variables being coupled.

ECMWF’s operational upgrade to cycle 45R1 in June 2018 saw the introduction of WCDA through seaice concentration. The forthcoming upgrade to cycle 46R1 is scheduled to also couple tropical SST, asper the experiments shown in this paper.

As well as surface temperature and ice concentration information, the ocean analysis system can providesurface current information. Surface currents can be important for the assimilation of scatterometer data.Scatterometers measure backscattering coefficient of the ocean surface. This coefficient is a function ofwind velocity relative to the ocean current. In the current usage of scatterometers at ECMWF we assumezero ocean currents, and so a WCDA system that has knowledge of the ocean currents should be able tobetter make use of scatterometer data.

There is plenty of scope to improve the partial coupling approach and its geospatial structure that de-termines the extent of SST coupling in the WCDA system. At the moment it is a simple function oflatitude. It may be beneficial for this to be basin dependent. Given the model biases in the westernboundary currents it may be beneficial to be different in the west and east of each basin.

These are the first steps in operational coupled ocean–atmosphere assimilation at ECMWF. A progressiveapproach towards implementation has been adopted, rather than introducing coupling in all variables in asingle system upgrade. Whilst the weakly coupled approach is being developed, in parallel the outer loopcoupling approach is being explored as a possible operational system which would give more immediateimpact across the various Earth system components.



Second thickest iceNäst grövsta isen

Toiseksi paksuin jää

Third thickest iceTredje grövsta isen

Kolmanneksi paksuin jää

Thickest iceGrövsta isenPaksuin jää

a b c

C =

CaCbCc =

SaSbSc =

FaFbFc =

Total ice concentration (in tenths)Totaliskoncentration (tiondelar)Jään kokonaispeittävyys (kymmenesosina)

PartialconcentrationDelkoncentrationOsittaispeittävyys

Stage of developmentIstjocklekJään paksuus

Form of ice / floesizeForm av is / flakstorlekJään muoto / lauttakoko

S

01234567891.

F

012345678

cm

-new ice< 1010 - 3010 - 1515 - 3030 - 20030 - 7030 - 5050 - 7070 - 120

diameter

-< 2 m2 - 20 m20 - 100 m100 - 500 m500 m - 2 km2 - 10 km> 10 kmfast ice

C

Ca Cb Cc

Sa Sb Sc

Fa Fb Fc

Vessels bound for Gulf of Bothnia ports in which traffic restrictions apply shall, when passing the latitude60°00'N, report their nationality, name, port of destination, ETA and speed to ICE INFO on VHF channel 78. This report can also be given directly by phone +46 31 699 100.

Vessels bound for ports in the Bay of Bothnia shall report to Bothnia VTS on VHF channel 67 20 NM before Nordvalen lighthouse.

The traffic separation schemes in the Quark are temporarily out of use from 25.01.2018.

The transit traffic west of Holmoarna is temporarily prohibited.arna is temporarily prohibited

Fartyg destinerade till hamnar med trafikrestriktion i Bottniska viken ska, vid passage av latituden 60°00'N,rapportera enligt instruktionerna för vintersjöfarten till ICE INFO på VHF-kanal 78 eller per telefon +46 31 699 100.

Fartyg destinerade till hamnar i Bottenviken ska 20 nautiska mil före Nordvalen rapportera till Bothnia VTS på VHF-kanal 67.

Trafiksepareringssystemen i Norra Kvarken är tillfälligt ur bruk från 2018-01-25.

Transittrafiken väster om Holmöarna är tillfälligt förbjuden.

Aluksen, joka on matkalla Pohjanlahden satamaan, jossa on voimassa liikennerajoitus, on ylittäessäänleveysasteen 60°00'N tehtävä talviliikenneohjeen mukainen ilmoitus ICE INFO:lle VHF-kanavalla 78 tai puhelimitse +46 31 699 100.

Aluksen, joka on matkalla Perämeren satamaan, on 20 mpk ennen Nordvalenin majakkaa ilmoittauduttava Bothnia VTS:lle VHF-kanavalla 67.

Reittijakojärjestelmät Merenkurkussa ovat tilapäisesti poissa käytöstä 25.1.2018 alkaen.

Kauttakulkuliikenne Holmöarnan länsipuolitse on tilapäisesti kielletty.

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

xx

x

x

x

x

x

x x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

xx

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x x

x

x

x

x

x

x

x

x

x

x

x

x

x

xx

x

x

x

xx

xx

xx

xx

x

x

35-5535-55

35-55

30-40

10-20

15-30

15-25

10-20

15-30

30-45

15-25

20-35

5-10

15-30

25-45

25-45

10-20

10-20

25-45

5-15

10-20

5-10

15-30

10-20

5-12

15-30 15-30

5-20

15-30

30-50

30-50

15-30

10-25

MUDJUG

SANKT-PETERBURG

ALE

KONTIO*

YMER

NOVOROSSISKIJ

OTSO

PROTECTOR

ODEN*

THETIS

SEMEN DEZHNEV

POLARISFREJ

ATLE

VOIMA*

KRUZENSTHERN

URHO

IZMAILOV

KAPITAN SOROKIN

YURI LISIANSKIJ

2°

2°

4°

1°

1°

1°

1°

1°

1°

1°

2°

2°

2°

2°

2°

2°

2°

3°

3°

3°

3°

3°

3°

3°

3°

4°

4°

5°

1°2°

4°

4°

4°

1,8°

2,2°

2,6°

2,8°

2,3°

2,4°

1,7°

1,9°

0,4°

0,6°

0,7°

0,2°

0,7°

0,5°

0,5°0,2°

0,2°

KundaLoksa

Wismar

Odense

Aalborg

Kiel

Sillamäe

Lidköping

STOCKHOLM

Gedser

Warnemünde

Paldiski

TALLINN

Schepelevski

Travemünde

SANKT-PETERBURG

OSLO

UMEÅ

Grums

Ystad

Larvik

Anholt

Kolding

Kantvik

Szczecin

Västerås

Uddevalla

Karlshamn Karlskrona

Ångermanälven

KOTKA PORVOO Borgå

LOVIISALovisa

TORNIO Torneå

KalixKarlsborg

NAANTALI Nådendal

HELSINKIHelsingforsMAARIANHAMINA

Mariehamn

Muuga

Vänersborg

Höganäs

PIETARSAARIJakobstad

Ajos

Åmål

Sikeå

GÄVLE

LULEÅ

Gdynia

Aarhus

Köping

Åhus

KØBENHAVN

Kalmar

Oskarshamn

SKELLEFTEÅ

HANKOHangö

SUNDSVALL

Västervik

Norrtälje

SÖDERHAMN

HUDIKSVALL

ÖRNSKÖLDSVIK

PITEÅ

Gdansk

Rozewie

Norrköping

Frederikshavn

INKOOIngå

Swinoujscie

Simrishamn

Oxelösund

Baltijsk

Karlstad

Trelleb.

Virpiniemi

HAMINAFredrikshamn

SALO

RIGA

KEMI

Ustka

Malmö

Rönne

Vysotsk

Liepaja

Darlowo

GÖTEBORG

Halmstad

KALAJOKI

HAAPSALU

Klaipeda

Hailuoto

VENTSPILS

Tolkmicko

Kolobrzeg

Strömstad

Mariestad

Merikarvia

Otterbäcken

Kristinehamn

VAASA Vasa

RAUMA Raumo

Kaliningrad

OULU Uleåborg

PORIBjörneborg

RAAHEBrahestad

KASKINEN Kaskö

KOKKOLA Karleby

UUSIKAUPUNKI Nystad

Trälhavet

Blackkallen

Fehmarn Belt

Romsö

Falsterborev

Lysekil

Gåsören

Strömmingsbådan

Ruhnu

Ulvöarna

Bol. Tyters

SuursaariHogland

Nötö

Sov.

Kumlinge

Naissaar

Kihnu

KotlinSeskar

Läsö

Iniö

Kökar

Mal. Tyters

Hel

Idö

Fårö

Mohni

SörveKolka

Tärnan

Brämön

Arkona

Glotovi

Bogskär

Harmaja

Landsort

Söderarm

Hammaren

HelsinkiPorkkala

Huvudskär

Östergarn

Kajakulma

St. F-ägg

Holmögadd

Skagsudde

Högbonden

Häradskär

Akmenrags

Yttergrund

Kalbådagr.

Finngrundet

Grundkallen

Storkläppen

Valassaaret

Gustaf Dalén

Ölands S gr.

Ölands N udde

Falkens grund

Norströmsgrund

Hallands Väderö

Agö

Mön

Ruden

Björn

Rauma

Stevns

Kallan

Isokari

Mersrags

Vilsandi

Utgrunden

Ulkokalla

Bjuröklubb

Hirstshals

Nahkiainen

Almagrundet

Rata Storgr.

Pater Noster

Tunö

Kiel

Saggö

Pakri

Hjälm

Rösnäs

Kullen

Ristna

Märket

Oulu 1

Hesselö

Tahkuna

G.Sandön

Osmussaar

Kopparst.

Darsser Ort

Svenska Björn

Lohm

Hoburg

Fladen

Rodser

Kemi 1

Malören

Drogden

VigrundVaindloJussarö

Norrskär

Tiiskeri

Rödkallen

Simpgrund

Utklippan

Irbenskij

Bengtskär

St. Karlsö

Sydostbrotten

Kokkola

Sälgrund

Storkallegrund

Örskär

Helsingkallan

Gran

Svartklubben

Farstugrunden

PÄRNU

Ust-Luga

Helsingborg

HÄRNÖSAND

Primorsk

NARVA

VYBORG

TURKUÅbo

Zap. Haapasaari Bol. Berezowyi

Mäntyluoto

Sandsänkan

Skagen

Blomman

Nordvalen

Trubaduren Knolls gr.

Säppi

Nygrån

Understen

30°

30°

29°

29°

28°

28°

27°

27°

26°

26°

25°

25°

24°

24°

23°

23°

22°

22°

21°

21°

20°

20°

19°

19°

18°

18°

17°

17°

16°

16°

15°

15°

14°

14°

13°

13°

12°

12°

11°

11°

10°

10°

66° 66°

65° 65°

64° 64°

63° 63°

62° 62°

61° 61°

60° 60°

59° 59°

58° 58°

57° 57°

56° 56°

55° 55°

54° 54°

© FMI & SMHI ISSN: 1796-0185

New ice (< 5 cm)Nyis (< 5 cm)Uusi jää (< 5 cm)

Nilas, grey ice (5-15 cm)Tunn jämn is (5-15 cm)Ohut tasainen jää (5-15 cm)

Open waterÖppet vattenAvovesi

Very open iceMycket spridd drivisHyvin harva ajojää

Open iceSpridd drivisHarva ajojää

Close iceTät drivisTiheä ajojää

Very close iceMycket tät drivisHyvin tiheä ajojää

Consolidated iceSammanfrusen drivisYhteenjäätynyt ajojää

Fast iceFastisKiintojää

Rotten fast iceRutten fastisHauras kiintojää

Ice freeIsfrittAvovesi

ATLE*

Ice thickness (cm)Istjocklek (cm)Jään paksuus (cm)

Jammed brash barrierStampisvallSohjovyö

Rafted iceHopskjuten isPäällekkäin ajautunut jää

Ridged or hummocked iceVallar eller upptornad isAhtautunut tai röykkiöitynyt jää

Strips and patchesSträngar av drivisAjojäänauhoja

Floebit, floebergIsbumlingAhtojää - tai röykkiölautta

FractureSprickaRepeämä

Fracture zoneOmråde med sprickorRepeämävyöhyke

Estimated ice edgeUppskattad iskant Arvioitu jään reuna

Icebreaker ( * coordinating)Isbrytare ( * coordinerande)Jäänmurtaja ( * koordinaattori)

Water temperature isotherm (oC)Vattentemperaturisoterm (oC)Veden lämpötilan tasa-arvokäyrä (oC)

2,1o Mean water temperatureYtvattnets medeltemperatur Meriveden pintalämpötilan keskiarvo(1971 - 2000)

20-40

-

7 - 10/10

10/10

< 1/10

1 - 3/10

4 - 6/10

7 - 8/10

9 - 9+/10

10/10

9 - 10/10

-

2o

ICE CHARTIskarta - Jääkartta

2018-02-17No. 80

ConcentrationKoncentrationPeittävyys

Ice typeIstyp Jäätyyppi

SymbolsSymbolerMerkinnät

Restrictions to Navigation, FINLANDTrafikrestriktioner, Finland – Liikennerajoitukset, Suomi

PortHamnSatama

Ice ClassIsklass

Jääluokka

Minimum tonnageMinsta tonnage

Minimikantavuus

First day of validityDatum för ikraftträdande

Voimaantulopäivä

Tornio, Kemi, Oulu IA 4000 2018-02-14

Raahe, Kalajoki, Kokkola, Pietarsaari

IA 2000 2018-02-12

VaasaIA, IBIC, II

20003000

2018-01-27

Kaskinen I, II 2000 2018-01-27

Kristiinankaupunki, Pori, Rauma, Uusikaupunki, Taalintehdas, Förby, Koverhar, Inkoo, Kantvik, Helsinki, Sköldvik

I, II 2000 2018-02-12

Loviisa, Kotka, HaminaIA, IBIC, II

20003000

2018-02-10

SköldvikIA, IBIC, II

20003000

2018-02-21

Loviisa, Kotka, Hamina IA, IB 2000 2018-02-21

Finnish Transport Agency (Liikennevirasto)

Restrictions to Navigation, SWEDENTrafikrestriktioner, Sverige – Liikennerajoitukset, Ruotsi

PortHamnSatama

Min Ice ClassMinsta IsklassMin Jääluokka


Minimikantavuus


Voimaantulopäivä

Karlsborg - Skelleftehamn IA 4000 2018-02-07

Holmsund - Örnsköldsvik IB 2000 2018-02-07

Ångermanälven IB 2000 2018-02-07

Härnösand - Skutskär II 2000 2018-02-05

Köping - Västerås IC 2000 2018-02-03

Östra MälarenIIIC

2000 1300

2018-02-06

Vänern, Trollhätte kanal, Göta älv

IIIC

2000 1300

2018-02-07

Swedish Maritime Administration (Sjöfartsverket)

(a) Manually produced Finnish-Swedish Ice Chartof the Baltic Sea. © FMI and SMHI [2018] Repro-duced with permission.

(b) OSI SAF 401-b product. Note missing data(grey) around coastlines due to coastal contamina-tion in satellite retrievals of sea ice concentration.

(c) Uncoupled analysis. Note the Gaussian natureof the ice field centred on the available OSI SAF L3observations.

(d) WCDA. Note the much more realistic structureand the good agreement with the manual ice chart.

Figure 13: A manual ice chart (top left) and sea ice concentration values in the Baltic sea on 2018-02-17from OSI SAF L3 observations (top right), uncoupled analysis (bottom left), and WCDA (bottom right).



Second thickest iceNäst grövsta isen

Toiseksi paksuin jää

Third thickest iceTredje grövsta isen

Kolmanneksi paksuin jää

Thickest iceGrövsta isenPaksuin jää

a b c

C =

CaCbCc =

SaSbSc =

FaFbFc =

Total ice concentration (in tenths)Totaliskoncentration (tiondelar)Jään kokonaispeittävyys (kymmenesosina)

PartialconcentrationDelkoncentrationOsittaispeittävyys

Stage of developmentIstjocklekJään paksuus

Form of ice / floesizeForm av is / flakstorlekJään muoto / lauttakoko

S

01234567891.

F

012345678

cm

-new ice< 1010 - 3010 - 1515 - 3030 - 20030 - 7030 - 5050 - 7070 - 120

diameter

-< 2 m2 - 20 m20 - 100 m100 - 500 m500 m - 2 km2 - 10 km> 10 kmfast ice

C

Ca Cb Cc

Sa Sb Sc

Fa Fb Fc

Vessels bound for Gulf of Bothnia ports in which traffic restrictions apply shall, when passing the latitude60°00'N, report their nationality, name, port of destination, ETA and speed to ICE INFO on VHF channel 78. This report can also be given directly by phone +46 31 699 100.

Vessels bound for ports in the Bay of Bothnia shall report to Bothnia VTS on VHF channel 67 20 NM before Nordvalen lighthouse.

The traffic separation schemes in the Quark are temporarily out of use from 25.01.2018.The transit traffic west of Holmoarna is temporarily prohibited.

Kalmarsund and Öregrundsgrepen: Tranist traffic for low powered vessels is not recommended.

Fartyg destinerade till hamnar med trafikrestriktion i Bottniska viken ska, vid passage av latituden 60°00'N,rapportera enligt instruktionerna för vintersjöfarten till ICE INFO på VHF-kanal 78 eller per telefon +46 31 699 100.

Fartyg destinerade till hamnar i Bottenviken ska 20 nautiska mil före Nordvalen rapportera till Bothnia VTS på VHF-kanal 67.

Trafiksepareringssystemen i Norra Kvarken är tillfälligt ur bruk från 2018-01-25.Transittrafiken väster om Holmöarna är tillfälligt förbjuden.

Kalmarsund och Öregrundsgrepen: Genomfartstrafik avrådes för maskinsvaga fartyg.

Aluksen, joka on matkalla Pohjanlahden satamaan, jossa on voimassa liikennerajoitus, on ylittäessäänleveysasteen 60°00'N tehtävä talviliikenneohjeen mukainen ilmoitus ICE INFO:lle VHF-kanavalla 78 tai puhelimitse +46 31 699 100.

Aluksen, joka on matkalla Perämeren satamaan, on 20 mpk ennen Nordvalenin majakkaa ilmoittauduttava Bothnia VTS:lle VHF-kanavalla 67.

Reittijakojärjestelmät Merenkurkussa ovat tilapäisesti poissa käytöstä 25.1.2018 alkaen.Kauttakulkuliikenne Holmöarnan länsipuolitse on tilapäisesti kielletty.

Kalmarsund ja Öregrundsgrepen: Läpikulkuliikennettä ei suositella heikkotehoisille aluksille.

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

xx

x

x

x

x

x

x x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

xx

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x x

x

x

x

x

x

x

x

x

x

x

x

x

x

xx

x

x

x

xx

xx

xx

xx

x

x

40-7040-70

25-40

20-40

15-35

15-35

20-35

15-30

20-35

10-25

15-35

20-35

10-20

20-40

20-40

40-50

10-20

15-30

5-15

5-1210-25

5-15

30-50

25-50

5-15

5-15

25-40

10-2010-20

10-20

15-25

15-25

40-60

40-6040-60

5-15

5-15

5-20

3-10

5-15

5-155-25

5-15

10-20

5-20

20-40

5-15

15-30

5-20

ZEUS*

SANKT-PETERBURGALE

KONTIO

YMER*

MURMANSK

OTSO*

PROTECTOR

ODEN*

THETIS*

POLARIS*

FREJ

ATLE

VOIMA*

I. KRUZENSTHERN

URHO

IZMAILOV

BOTNICAFYRBYGGAREN

FENNICA

SCANDICA

BALTICA

NORDICA

NOVOROSSIYSK

VARMA

ERMAKS. DEZHNEV

K. NIKOLAEV

K. PLAKHINTARMO

SISU

TRELLE

BONDEN

MUDYUG

YURIY LISYANSKIY

2°

2°

2°

2°2°

2°

2°3°

1° 1°

1°

1°

1°

1°

1°

1°1°

1°2°

1°

3°

3°

2,0°

2,1°

2,3°

2,5°

2,0°

1,9°

1,4°

1,4°

0,4°

0,3°

KundaLoksa

Wismar

Odense

Aalborg

Kiel

Sillamäe

Lidköping

STOCKHOLM

Gedser

Warnemünde

Paldiski

TALLINN

Schepelevski

Travemünde

SANKT-PETERBURG

OSLO

UMEÅ

Grums

PÄRNU

Ystad

Larvik

Anholt

Kolding

Kantvik

Szczecin

Västerås

Uddevalla

Karlshamn Karlskrona

Ångermanälven

KOTKA PORVOO Borgå

LOVIISALovisa

TORNIO Torneå

KalixKarlsborg

NAANTALI Nådendal

HELSINKIHelsingforsMAARIANHAMINA

Mariehamn

Muuga

Vänersborg

Höganäs

PIETARSAARIJakobstad

Ajos

Åmål

Sikeå

GÄVLE

LULEÅ

Gdynia

Aarhus

Köping

Åhus

KØBENHAVN

Oskarshamn

SKELLEFTEÅ

HANKOHangö

SUNDSVALL

Västervik

Norrtälje

SÖDERHAMN

HUDIKSVALL

ÖRNSKÖLDSVIK

PITEÅ

Gdansk

Rozewie

Norrköping

Frederikshavn

INKOOIngå

Swinoujscie

Simrishamn

Baltijsk

Karlstad

Trelleb.

Virpiniemi

SALO

RIGA

KEMI

Ustka

NARVA

Malmö

Rönne

Vysotsk

Liepaja

Darlowo

GÖTEBORG

Halmstad

KALAJOKI

HAAPSALU

Klaipeda

Hailuoto

VENTSPILS

Tolkmicko

Kolobrzeg

Strömstad

Mariestad

Merikarvia

Otterbäcken

Kristinehamn

VAASA Vasa

Kaliningrad

OULU Uleåborg

PORIBjörneborg

RAAHEBrahestad

KASKINEN Kaskö

UUSIKAUPUNKI Nystad

Trälhavet

Blackkallen

Romsö

Falsterborev

Lysekil

Gåsören

Strömmingsbådan

Ruhnu

Ulvöarna

Bol. TytersNötö

Sov. Zap.

Kumlinge

Naissaar

Kihnu

KotlinSeskar

Läsö

Iniö

Kökar

Mal. Tyters

Bol. Berezowyi

Hel

Fårö

Mohni

SörveKolka

Tärnan

Brämön

Arkona

Bogskär

Harmaja

Landsort

Söderarm

Hammaren

Porkkala

Huvudskär

Östergarn

Kajakulma

Holmögadd

Högbonden

Häradskär

Akmenrags

Yttergrund

Kalbådagr.

Finngrundet

Grundkallen

Storkläppen

Gustaf Dalén

Ölands S gr.

Ölands N udde

Norströmsgrund

Hallands Väderö

Agö

Ruden

Björn

Rauma

Skagen

Stevns

Kallan

Isokari

Mersrags

Vilsandi

Nordvalen

Utgrunden

Ulkokalla

Trubaduren

Bjuröklubb

Hirstshals

Almagrundet

Rata Storgr.

Pater Noster

Tunö

Saggö

Pakri

Hjälm

Rösnäs

Kullen

Ristna

Märket

Oulu 1

Hesselö

Tahkuna

G.Sandön

Osmussaar

Kopparst.

Darsser Ort

Svenska Björn

Lohm

Hoburg

Fladen

Rodser

Kemi 1

Malören

Drogden

VigrundVaindlo

Norrskär

Tiiskeri

Rödkallen

Simpgrund

Utklippan

Irbenskij

Bengtskär

St. Karlsö

Kokkola

Sälgrund

Storkallegrund

Örskär

Helsingkallan

Gran

Nygrån

Svartklubben

Ust-Luga

Helsingborg

Kalmar

HÄRNÖSAND

Oxelösund

Primorsk

HAMINAFredrikshamn

VYBORG

TURKUÅbo

RAUMA Raumo

KOKKOLA Karleby

Fehmarn Belt

SuursaariHogland

Haapasaari

Idö

Glotovi

Helsinki

St. F-ägg

Skagsudde

Mäntyluoto

Sandsänkan

Valassaaret

Falkens grund

Mön

Blomman

Nahkiainen

Kiel

Knolls gr.

Jussarö

Sydostbrotten

Säppi

Understen

Farstugrunden

30°

30°

29°

29°

28°

28°

27°

27°

26°

26°

25°

25°

24°

24°

23°

23°

22°

22°

21°

21°

20°

20°

19°

19°

18°

18°

17°

17°

16°

16°

15°

15°

14°

14°

13°

13°

12°

12°

11°

11°

10°

10°

66° 66°

65° 65°

64° 64°

63° 63°

62° 62°

61° 61°

60° 60°

59° 59°

58° 58°

57° 57°

56° 56°

55° 55°

54° 54°

© FMI & SMHI ISSN: 1796-0185

New ice (< 5 cm)Nyis (< 5 cm)Uusi jää (< 5 cm)

Nilas, grey ice (5-15 cm)Tunn jämn is (5-15 cm)Ohut tasainen jää (5-15 cm)

Open waterÖppet vattenAvovesi

Very open iceMycket spridd drivisHyvin harva ajojää

Open iceSpridd drivisHarva ajojää

Close iceTät drivisTiheä ajojää

Very close iceMycket tät drivisHyvin tiheä ajojää

Consolidated iceSammanfrusen drivisYhteenjäätynyt ajojää

Fast iceFastisKiintojää

Rotten fast iceRutten fastisHauras kiintojää

Ice freeIsfrittAvovesi

ATLE*

Ice thickness (cm)Istjocklek (cm)Jään paksuus (cm)

Jammed brash barrierStampisvallSohjovyö

Rafted iceHopskjuten isPäällekkäin ajautunut jää

Ridged or hummocked iceVallar eller upptornad isAhtautunut tai röykkiöitynyt jää

Strips and patchesSträngar av drivisAjojäänauhoja

Floebit, floebergIsbumlingAhtojää - tai röykkiölautta

FractureSprickaRepeämä

Fracture zoneOmråde med sprickorRepeämävyöhyke

Estimated ice edgeUppskattad iskant Arvioitu jään reuna

Icebreaker ( * coordinating)Isbrytare ( * coordinerande)Jäänmurtaja ( * koordinaattori)

Water temperature isotherm (oC)Vattentemperaturisoterm (oC)Veden lämpötilan tasa-arvokäyrä (oC)

2,1o Mean water temperatureYtvattnets medeltemperatur Meriveden pintalämpötilan keskiarvo(1971 - 2000)

20-40

-

7 - 10/10

10/10

< 1/10

1 - 3/10

4 - 6/10

7 - 8/10

9 - 9+/10

10/10

9 - 10/10

-

2o

ICE CHARTIskarta - Jääkartta

2018-03-05No. 96

ConcentrationKoncentrationPeittävyys

Ice typeIstyp Jäätyyppi

SymbolsSymbolerMerkinnät

Restrictions to Navigation, FINLANDTrafikrestriktioner, Finland – Liikennerajoitukset, Suomi

PortHamnSatama

Ice ClassIsklass

Jääluokka


Minimikantavuus


VoimaantulopäiväTornio, Kemi, Oulu IA 4000 2018-02-14Raahe, Kalajoki IA 4000 2018-02-24Kokkola, Pietarsaari IA 4000 2018-03-03Vaasa IA, IB 2000 2018-02-26Kaskinen, Kristiinankaupunki, Pori, Rauma

IA, IBIC, II

20003000

2018-02-28

UusikaupunkiIA, IBIC, II

20003000

2018-02-26

Naantali, Turku I, II 2000 2018-02-26

Taalintehdas, FörbyIA, IBIC, II

20003000

2018-02-28

Hanko I, II 2000 2018-02-28

Koverhar, Inkoo, Kantvik, HelsinkiIA, IBIC, II

20003000

2018-02-28

Sköldvik IA, IB 2000 2018-03-04Loviisa, Kotka, Hamina IA 2000 2018-03-04

Finnish Transport Agency (Liikennevirasto)

Restrictions to Navigation, SWEDENTrafikrestriktioner, Sverige – Liikennerajoitukset, Ruotsi

PortHamnSatama

Min Ice ClassMinsta IsklassMin Jääluokka


Minimikantavuus


VoimaantulopäiväKarlsborg. Minimum load ordischarge 2000 tons

IA 4000 2018-03-05

Luleå - Skelleftehamn IA 4000 2018-02-07

Holmsund - Örnsköldsvik IA 3000 2018-03-03

Ångermanälven IA 2000 2018-03-05

Härnösand - Skutskär IA 2000 2018-03-05

Öregrund - Hargshamn IB 2000 2018-03-05

Grisslehamn - BergkvaraIIIC

20001300

2018-02-27

Mälaren IC 2000 2018-02-26

Vänern, Trollhätte kanal,Göta älv

IC 2000 2018-02-27

Vänern, Trollhätte kanal,Göta älv

IB 2000 2018-03-07

Mälaren IB 2000 2018-03-08Swedish Maritime Administration (Sjöfartsverket)

(a) Manually produced Finnish-Swedish Ice Chartof the Baltic Sea. © FMI and SMHI [2018] Repro-duced with permission.

(b) OSI SAF 401-b product. Note missing data(grey) around coastlines due to coastal contamina-tion in satellite retrievals of sea ice concentration.

(c) Uncoupled analysis. Note the Gaussian natureof the ice field centred on the available OSI SAF L3observations.

(d) WCDA. Note the much more realistic structureand the good agreement with the manual ice chart.

Figure 14: As Figure 13, but on 2018-03-05, the date of maximum sea ice extent in the region for theseason.



6 Acknowledgements

We are grateful to Patrick Eriksson of FMI for providing assistance with the FMI/SMHI sea ice charts.Our thanks go to Magdalena Balmaseda, Stephen English, Sarah Keeley, and Kristian Mogensen for theirhelp with implementations and insightful discussion of results.

References

M. A. Balmaseda, K. Mogensen, and A. T. Weaver. Evaluation of the ECMWF ocean reanalysis systemORAS4. Quarterly Journal of the Royal Meteorological Society, 139(674):1132–1161, 2013. ISSN00359009. doi: 10.1002/qj.2063.

G. Balsamo, A. Beljaars, K. Scipal, P. Viterbo, B. van den Hurk, M. Hirschi, and A. K. Betts. ARevised Hydrology for the ECMWF Model: Verification from Field Site to Terrestrial Water Storageand Impact in the Integrated Forecast System. Journal of Hydrometeorology, 10(3):623–643, 2009.ISSN 1525-755X. doi: 10.1175/2008JHM1068.1. URL http://journals.ametsoc.org/doi/abs/10.1175/2008JHM1068.1.

P. Bauer and D. Richardson. New model cycle 40r1. ECMWF Newsletter No. 138 - Winter 2013/2014,(138):3, 2014. URL www.ecmwf.int/publications/newsletter.

P. Bauer, A. Thorpe, and G. Brunet. The quiet revolution of numerical weather prediction. Nature, 525(7567):47–55, 2015. ISSN 0028-0836. doi: 10.1038/nature14956. URL http://www.nature.com/doifinder/10.1038/nature14956.

M. Bonavita, L. Isaksen, and E. Holm. On the use of EDA background error variances in the ECMWF4D-Var. Quarterly Journal of the Royal Meteorological Society, 138(667):1540–1559, jul 2012. ISSN00359009. doi: 10.1002/qj.1899. URL http://doi.wiley.com/10.1002/qj.1899.

P. de Rosnay, G. Balsamo, C. Albergel, J. Munoz-Sabater, and L. Isaksen. Initialisation of Land SurfaceVariables for Numerical Weather Prediction. Surveys in Geophysics, 35(3):607–621, 2014. ISSN01693298. doi: 10.1007/s10712-012-9207-x.

C. J. Donlon, M. Martin, J. Stark, J. Roberts-Jones, E. Fiedler, and W. Wimmer. The Operational SeaSurface Temperature and Sea Ice Analysis (OSTIA) system. Remote Sensing of Environment, 116:140–158, jan 2012. ISSN 00344257. doi: 10.1016/j.rse.2010.10.017. URL http://linkinghub.elsevier.com/retrieve/pii/S0034425711002197.

E. Dutra, V. M. Stepanenko, G. Balsamo, P. Viterbo, P. M. A. Miranda, D. Mironov, and C. Schar. Impactof lakes on the ECMWF surface scheme. ECMWF Technical Memorandum, (November 2009):1–15,2009.

ECMWF. PART V: ENSEMBLE PREDICTION SYSTEM. Number 5 in IFS Documentation. ECMWF,2014.

ECMWF. The Strength of a Common Goal: A Roadmap To 2025. Technical re-port, ECMWF, 2016. URL https://www.ecmwf.int/sites/default/files/ECMWF_Roadmap_to_2025.pdf.

ECMWF. PART III: DYNAMICS AND NUMERICAL PROCEDURES. Number 3 in IFS Documentation.ECMWF, 2017a.


http://journals.ametsoc.org/doi/abs/10.1175/2008JHM1068.1

http://journals.ametsoc.org/doi/abs/10.1175/2008JHM1068.1

www.ecmwf.int/publications/newsletter

http://www.nature.com/doifinder/10.1038/nature14956

http://www.nature.com/doifinder/10.1038/nature14956

http://doi.wiley.com/10.1002/qj.1899

http://linkinghub.elsevier.com/retrieve/pii/S0034425711002197

http://linkinghub.elsevier.com/retrieve/pii/S0034425711002197

https://www.ecmwf.int/sites/default/files/ECMWF_Roadmap_to_2025.pdf



ECMWF. PART IV: PHYSICAL PROCESSES. Number 4 in IFS Documentation. ECMWF, 2017b.

ECMWF. PART VII: ECMWF WAVE MODEL. Number 7 in IFS Documentation. ECMWF, 2017c.

S. English, A. McNally, N. Borman, K. Salonen, M. Matricardi, A. Horanyi, M. Rennie, M. Janiskova,S. Di Michele, A. Geer, E. Di Tomaso, C. Cardinali, P. de Rosnay, J. Sabater, M. Bonavita, C. Albergel,R. Engelen, and J.-N. Thepaut. Impact of Satellite Data. Technical Memoradum ECMWF, (711):46,2013.

FMI and SMHI. Baltic sea ice chart, 2018. URL http://cdn.fmi.fi/marine-observations/products/ice-charts/20180217-full-color-ice-chart.pdf.

S. Frolov, C. H. Bishop, T. Holt, J. Cummings, and D. Kuhl. Facilitating Strongly Coupled OceanAt-mosphere Data Assimilation with an Interface Solver. Monthly Weather Review, 144(1):3–20, 2016.ISSN 0027-0644. doi: 10.1175/MWR-D-15-0041.1. URL http://journals.ametsoc.org/doi/abs/10.1175/MWR-D-15-0041.1.

A. J. Geer. Significance of changes in medium-range forecast scores. Tellus A, 68(0):1–21, 2016. ISSN1600-0870. doi: 10.3402/tellusa.v68.30229. URL http://www.tellusa.net/index.php/tellusa/article/view/30229.

T. Haiden, M. Dahoui, B. Ingleby, P. D. Rosnay, C. Prates, E. Kuscu, T. Hewson, L. Isaksen, D. Richard-son, H. Zuo, and L. Jones. Use of in situ surface observations at ECMWF. Technical MemoradumECMWF, 834(November), 2018.

J. Haseler. The early-delivery suite. Technical memorandum 454, ECMWF, 2004.

S. Keeley et al. The ECMWF Coupled Atmosphere-Wave-Ocean-Ice Model as Implemented in CY45R1:Part 3: Ice model performance. ECMWF Technical Memorandum, 2019.

P. Laloyaux, M. Balmaseda, D. Dee, K. Mogensen, and P. Janssen. A coupled data assimilation systemfor climate reanalysis. Quarterly Journal of the Royal Meteorological Society, 142(694):65–78, 2016.ISSN 1477870X. doi: 10.1002/qj.2629.

D. J. Lea, I. Mirouze, M. J. Martin, R. R. King, A. Hines, D. Walters, and M. Thurlow. As-sessing a New Coupled Data Assimilation System Based on the Met Office Coupled Atmosphere-LandOceanSea Ice Model. Monthly Weather Review, 143(11):4678–4694, 2015. ISSN 0027-0644.doi: 10.1175/MWR-D-15-0174.1. URL http://journals.ametsoc.org/doi/10.1175/MWR-D-15-0174.1.

C. Maclachlan, A. Arribas, K. A. Peterson, A. Maidens, D. Fereday, A. A. Scaife, M. Gordon, M. Vel-linga, A. Williams, R. E. Comer, J. Camp, P. Xavier, and G. Madec. Global Seasonal forecast systemversion 5 (GloSea5): A high-resolution seasonal forecast system. Quarterly Journal of the RoyalMeteorological Society, 141(689):1072–1084, 2015. ISSN 1477870X. doi: 10.1002/qj.2396.

S. Masuda, J. Philip Matthews, Y. Ishikawa, T. Mochizuki, Y. Tanaka, and T. Awaji. A new Approach toEl Nino Prediction beyond the Spring Season. Scientific Reports, 5:1–9, 2015. ISSN 20452322. doi:10.1038/srep16782.

T. Mochizuki, S. Masuda, Y. Ishikawa, and T. Awaji. Multiyear climate prediction with initializationbased on 4D-Var data assimilation. Geophysical Research Letters, 43(8):3903–3910, 2016. ISSN19448007. doi: 10.1002/2016GL067895.


http://cdn.fmi.fi/marine-observations/products/ice-charts/20180217-full-color-ice-chart.pdf



http://journals.ametsoc.org/doi/abs/10.1175/MWR-D-15-0041.1

http://journals.ametsoc.org/doi/abs/10.1175/MWR-D-15-0041.1

http://www.tellusa.net/index.php/tellusa/article/view/30229

http://www.tellusa.net/index.php/tellusa/article/view/30229

http://journals.ametsoc.org/doi/10.1175/MWR-D-15-0174.1

http://journals.ametsoc.org/doi/10.1175/MWR-D-15-0174.1


K. Mogensen et al. The ECMWF Coupled Atmosphere-Wave-Ocean-Ice Model as Implemented inCY45R1: Part 1: technical implementation. ECMWF Technical Memorandum, 2019a.

K. Mogensen et al. The ECMWF Coupled Atmosphere-Wave-Ocean-Ice Model as Implemented inCY45R1: Part 2: Ocean model performance. ECMWF Technical Memorandum, 2019b.

D. P. Mulholland, P. Laloyaux, K. Haines, and M. Alonso Balmaseda. Origin and Impact of InitializationShocks in Coupled Atmosphere-Ocean Forecasts. Monthly Weather Review, 143:4631–4644, 2015.ISSN 0027-0644. doi: 10.1175/MWR-D-15-0076.1. URL https://www.ecmwf.int/sites/default/files/ECMWF_Roadmap_to_2025.pdf.

S. G. Penny, S. Akella, O. Alves, C. Bishop, M. Buehner, M. Chevallier, F. Counillon, C. Draper,S. Frolov, Y. Fujii, A. Karspeck, A. Kumar, P. Laloyaux, J.-F. Mahfouf, M. Martin, M. Pena, P. DeRosnay, A. Subramanian, R. Tardif, Y. Wang, and X. Wu. Coupled Data Assimilation for IntegratedEarth System Analysis and Prediction: Goals, Challenges and Recommendations. Technical report,World Meteorological Organisation, 2017.

F. Rabier, H. Jarvinen, E. Klinker, J.-F. Mahfouf, and A. Simmons. The ECMWF operational imple-mentation of four-dimensional variational assimilation. Part I: Experimental results and diagnosticswiht operational configuration. Q. J. R. Meteorol. Soc., 126:1143–1170, 2000. ISSN 1477-870X. doi:10.1002/qj.49712656417.

S. Saha, S. Moorthi, H. L. Pan, X. Wu, J. Wang, S. Nadiga, P. Tripp, R. Kistler, J. Woollen, D. Behringer,H. Liu, D. Stokes, R. Grumbine, G. Gayno, J. Wang, Y. T. Hou, H. Y. Chuang, H. M. H. Juang, J. Sela,M. Iredell, R. Treadon, D. Kleist, P. Van Delst, D. Keyser, J. Derber, M. Ek, J. Meng, H. Wei, R. Yang,S. Lord, H. Van Den Dool, A. Kumar, W. Wang, C. Long, M. Chelliah, Y. Xue, B. Huang, J. K.Schemm, W. Ebisuzaki, R. Lin, P. Xie, M. Chen, S. Zhou, W. Higgins, C. Z. Zou, Q. Liu, Y. Chen,Y. Han, L. Cucurull, R. W. Reynolds, G. Rutledge, and M. Goldberg. The NCEP climate forecastsystem reanalysis. Bulletin of the American Meteorological Society, 91(8):1015–1057, 2010. ISSN00030007. doi: 10.1175/2010BAMS3001.1.

S. Saha, S. Moorthi, X. Wu, J. Wang, S. Nadiga, P. Tripp, D. Behringer, Y. T. Hou, H. Y. Chuang,M. Iredell, M. Ek, J. Meng, R. Yang, M. P. Mendez, H. Van Den Dool, Q. Zhang, W. Wang, M. Chen,and E. Becker. The NCEP climate forecast system version 2. Journal of Climate, 27(6):2185–2208,2014. ISSN 08948755. doi: 10.1175/JCLI-D-12-00823.1.

D. Schepers, E. D. Boisseson, R. Eresmaa, C. Lupu, and P. D. Rosnay. CERA-SAT: A coupled satellite-era reanalysis. ECMWF Newsletter, (155):32–37, 2018. doi: 10.21957/sp619ds74g.

T. N. Stockdale, D. L. Anderson, J. O. Alves, and M. A. Balmaseda. Global seasonal rainfall forecastsusing a coupled ocean-atmosphere model. Nature, 392(6674):370–373, 1998. ISSN 00280836. doi:10.1038/32861.

N. Sugiura, T. Awaji, S. Masuda, T. Mochizuki, T. Toyoda, T. Miyama, H. Igarashi, and Y. Ishikawa.Development of a four-dimensional variational coupled data assimilation system for enhanced anal-ysis and prediction of seasonal to interannual climate variations. Journal of Geophysical Research:Oceans, 113(10):1–21, 2008. ISSN 21699291. doi: 10.1029/2008JC004741.

F. Vitart. Monthly Forecasting at ECMWF. Mon. Wea. Rev., 132(1986):2761–2779, 2004. ISSN 0027-0644. doi: 10.1175/MWR2826.1.





X. Yang, A. Rosati, S. Zhang, T. L. Delworth, R. G. Gudgel, R. Zhang, G. Vecchi, W. Anderson, Y. soonChang, T. Delsole, K. Dixon, R. Msadek, W. F. Stern, A. Wittenberg, and F. Zeng. A predictableAMO-like pattern in the GFDL fully coupled ensemble initialization and decadal forecasting system.Journal of Climate, 26(2):650–661, 2013. ISSN 08948755. doi: 10.1175/JCLI-D-12-00231.1.

S. Zhang, Y. S. Chang, X. Yang, and A. Rosati. Balanced and coherent climate estimation by combiningdata with a biased coupled model. Journal of Climate, 27(3):1302–1314, 2014. ISSN 08948755. doi:10.1175/JCLI-D-13-00260.1.

H. Zuo, M. A. Balmaseda, K. Mogensen, and S. Tietsche. OCEAN5: the ECMWF Ocean ReanalysisSystem and its Real-Time analysis component. Technical Report 823, ECMWF, 2018.



A Appendix: area averaged dRMSE of forecast errors

CI: SH −90° to −20°, sfc

0 1 2 3 4 5 6 7 8 9 10Forecast day

−0.8

−0.7

−0.6

−0.5

−0.4

−0.3

−0.2

Nor

mal

ised

diff

eren

ce

CI: Tropics −20° to 20°, sfc

0 1 2 3 4 5 6 7 8 9 10Forecast day

1.0

1.2

1.4

1.6

1.8

2.0

9−Jun−2017 to 21−May−2018 from 674 to 693 samples. Verified against own−analysis. Confidence range 95% with AR(2) inflation and Sidak correction for 4 independent tests.

WCDA − Uncoupled analysis

CI: NH 20° to 90°, sfc

0 1 2 3 4 5 6 7 8 9 10Forecast day

−0.50

−0.45

−0.40

−0.35

−0.30

−0.25

−0.20CI: SH −90° to −20°, sfc

0 1 2 3 4 5 6 7 8 9 10Forecast day

−0.8

−0.7

−0.6

−0.5

−0.4

−0.3

−0.2

Nor

mal

ised

diff

eren

ce


0 1 2 3 4 5 6 7 8 9 10Forecast day

1.0

1.2

1.4

1.6

1.8

2.0




0 1 2 3 4 5 6 7 8 9 10Forecast day

−0.50

−0.45

−0.40

−0.35

−0.30

−0.25

−0.20CI: SH −90° to −20°, sfc

0 1 2 3 4 5 6 7 8 9 10Forecast day

−0.8

−0.7

−0.6

−0.5

−0.4

−0.3

−0.2

Nor

mal

ised

diff

eren

ce


0 1 2 3 4 5 6 7 8 9 10Forecast day

1.0

1.2

1.4

1.6

1.8

2.0




0 1 2 3 4 5 6 7 8 9 10Forecast day

−0.50

−0.45

−0.40

−0.35

−0.30

−0.25

−0.20

CI: SH −90° to −20°, sfc

0 1 2 3 4 5 6 7 8 9 10Forecast day

−0.8

−0.7

−0.6

−0.5

−0.4

−0.3

−0.2

Nor

mal

ised

diff

eren

ce


0 1 2 3 4 5 6 7 8 9 10Forecast day

1.0

1.2

1.4

1.6

1.8

2.0




0 1 2 3 4 5 6 7 8 9 10Forecast day

−0.50

−0.45

−0.40

−0.35

−0.30

−0.25

−0.20

Figure 15: Area averaged diagram of the normalised difference in RMSE between WCDA and control forsea ice concentration in the southern hemisphere (left), tropics (centre) and northern hemisphere (right)for forecasts at lead times of up to 10 days.

SST: SH −90° to −20°, sfc

0 1 2 3 4 5 6 7 8 9 10Forecast day

−0.035

−0.030

−0.025

−0.020

−0.015

−0.010

−0.005

Nor

mal

ised

diff

eren

ce

SST: Tropics −20° to 20°, sfc

0 1 2 3 4 5 6 7 8 9 10Forecast day

−0.080

−0.075

−0.070

−0.065

−0.060

−0.055

−0.050



SST: NH 20° to 90°, sfc

0 1 2 3 4 5 6 7 8 9 10Forecast day

−0.020

−0.018

−0.016

−0.014

−0.012

−0.010

−0.008

Figure 16: As Figure 15 but for sea-surface temperature.



Z2T: SH −90° to −20°, sfc

0 1 2 3 4 5 6 7 8 9 10Forecast day

−0.15

−0.10

−0.05

0.00

0.05

Nor

mal

ised

diff

eren

ce

Z2T: Tropics −20° to 20°, sfc

0 1 2 3 4 5 6 7 8 9 10Forecast day

−0.030

−0.025

−0.020

−0.015

−0.010

−0.005

0.000



Z2T: NH 20° to 90°, sfc

0 1 2 3 4 5 6 7 8 9 10Forecast day

−0.030−0.025

−0.020

−0.015

−0.010

−0.005

0.0000.005

Figure 17: As figure 15 but for atmospheric 2 metre temperature.



R: SH −90° to −20°, 100hPa

0 1 2 3 4 5 6 7 8 9 10−0.015

−0.010

−0.005

0.000

0.005

0.010

0.015

Nor

mal

ised

diff

eren

ce

R: Tropics −20° to 20°, 100hPa

0 1 2 3 4 5 6 7 8 9 10−0.004

−0.002

0.000

0.002

0.004

0.006R: NH 20° to 90°, 100hPa

0 1 2 3 4 5 6 7 8 9 10−0.015

−0.010

−0.005

0.000

0.005

0.010

0.015

R: SH −90° to −20°, 500hPa

0 1 2 3 4 5 6 7 8 9 10−0.006

−0.004

−0.002

0.000

0.002

0.004

0.006

0.008

Nor

mal

ised

diff

eren

ce


0 1 2 3 4 5 6 7 8 9 10−0.006

−0.004

−0.002

0.000

0.002

0.004

0.006R: NH 20° to 90°, 500hPa

0 1 2 3 4 5 6 7 8 9 10−0.006

−0.004

−0.002

0.000

0.002

0.004

0.006

R: SH −90° to −20°, 850hPa

0 1 2 3 4 5 6 7 8 9 10−0.010

−0.005

0.000

0.005

0.010

Nor

mal

ised

diff

eren

ce


0 1 2 3 4 5 6 7 8 9 10−0.008

−0.006

−0.004

−0.002

0.000

0.002

0.004R: NH 20° to 90°, 850hPa

0 1 2 3 4 5 6 7 8 9 10−0.008

−0.006

−0.004

−0.002

0.000

0.002

0.004

0.006

R: SH −90° to −20°, 1000hPa

0 1 2 3 4 5 6 7 8 9 10Forecast day

−0.04

−0.03

−0.02

−0.01

0.00

0.01

Nor

mal

ised

diff

eren

ce


0 1 2 3 4 5 6 7 8 9 10Forecast day

−0.06

−0.05

−0.04

−0.03

−0.02

−0.01

0.00



R: NH 20° to 90°, 1000hPa

0 1 2 3 4 5 6 7 8 9 10Forecast day

−0.025

−0.020

−0.015

−0.010

−0.005

0.000

0.005

Figure 18: As Figure 15 but for atmospheric relative humidity and stratified by level, 100hPa (top row),500hPa (second row), 850hPa (third row) and 1000hPa (bottom row).



T: SH −90° to −20°, 100hPa

0 1 2 3 4 5 6 7 8 9 10−0.010

−0.005

0.000

0.005

0.010

0.015

0.020

0.025

Nor

mal

ised

diff

eren

ce

T: Tropics −20° to 20°, 100hPa

0 1 2 3 4 5 6 7 8 9 10−0.006

−0.004

−0.002

0.000

0.002

0.004

0.006

0.008T: NH 20° to 90°, 100hPa

0 1 2 3 4 5 6 7 8 9 10−0.015

−0.010

−0.005

0.000

0.005

0.010

T: SH −90° to −20°, 500hPa

0 1 2 3 4 5 6 7 8 9 10−0.010

−0.005

0.000

0.005

0.010

0.015

0.020

Nor

mal

ised

diff

eren

ce


0 1 2 3 4 5 6 7 8 9 10−0.010

−0.005

0.000

0.005

0.010T: NH 20° to 90°, 500hPa

0 1 2 3 4 5 6 7 8 9 10−0.015

−0.010

−0.005

0.000

0.005

0.010

0.015

T: SH −90° to −20°, 850hPa

0 1 2 3 4 5 6 7 8 9 10−0.010

−0.005

0.000

0.005

0.010

0.015

0.020

Nor

mal

ised

diff

eren

ce


0 1 2 3 4 5 6 7 8 9 10−0.015

−0.010

−0.005

0.000

0.005

0.010

0.015T: NH 20° to 90°, 850hPa

0 1 2 3 4 5 6 7 8 9 10−0.015

−0.010

−0.005

0.000

0.005

0.010

T: SH −90° to −20°, 1000hPa

0 1 2 3 4 5 6 7 8 9 10Forecast day

−0.08

−0.06

−0.04

−0.02

0.00

0.02

Nor

mal

ised

diff

eren

ce


0 1 2 3 4 5 6 7 8 9 10Forecast day

−0.04

−0.03

−0.02

−0.01

0.00

0.01



T: NH 20° to 90°, 1000hPa

0 1 2 3 4 5 6 7 8 9 10Forecast day

−0.03

−0.02

−0.01

0.00

0.01

Figure 19: As Figure 18 but for atmospheric temperature.


Hao Zuo, Andrew Bennett, Andrew Dawson Research Department

Documents

Transcript of Hao Zuo, Andrew Bennett, Andrew Dawson Research Department