She et al.
Operational Oceanography and Earth System Science
Frontiers In Earth Science | www.frontlersln.org
7
February 2020 | Volume 8 | Article 7
reanalysis production in RCO-SCOBI and NEMO-SCOBI (Axell
and Liu, 2016; Liu et al., 2017). Currently the Baltic Sea data
assimilation collaboration in CMEMS focuses on developing
physical and biogeochemical assimilation systems by using the
PDAF both for operational NRT forecast and also for reanalyses.
Assimilation schemes for SST, T/S, sea ice and nutrients are
relatively mature. New schemes are developed for assimilating sea
level and satellite ocean-color data (Tuomi et al., 2018).
Model Quality and Validation
Before 2009, the operational model validation and quality
assessment were mainly done at national level with different
methods and quality standards. Common cal/val methodology
had been developed and applied in the MyOcean projects
(2009-2015), including cal/val metrics definition and error
statistics calculation and presentation for new model version,
NRT validation and reanalysis quality assessment. The cal/val
has been part of the BAL MFC operational activities since
2015. Before release of each new BAL MFC model version—
both forecasting and reanalysis systems, a correspondent QUID
(Quality Information Document) report has to be released
to demonstrate the quality of the new version products
(e.g., Golbeck et al., 2017). The NRT validation for the
BAL MFC forecast products (ocean-ice-wave-biogeochemical
parameters) is shown online at the BOOS website. The
cal/val method and toolbox developed in the BAL MFC
is now further extended to a BOOS model quality and
validation cooperation.
Multi-Model Ensemble (MME) Forecasting
Based on NRT exchange of both model and observational data
via the BOOS ftp network, a MME forecast system has been
developed for sea level, SST, sea surface salinity (SSS), T/S and
currents (Golbeck et al., 2015). By weighting the individual
forecast related to its forecasting error, a weighted MME method
is used to generate the MME forecast. The results, shown online
at the BOOS website, demonstrate superior quality of the MME
forecast to the individual ones. The MME is a joint achievement
of ROOSs and MFCs in the Baltic and North Sea. Currently
the BOOS MME Working Group aims at extending the current
MME system for national forecasting use.
Major Challenges
Future direction of the operational modeling in the Baltic Sea
is seamless modeling in spatial, temporal, and state variable
dimensions (WMO, 2015; She and Murawski, 2018). In spatial
scales, the modeling capacity will be extended from basin scale to
local coastal-estuary scale and from mesoscale to sub-mesoscale.
In temporal scales, the synoptic and climate scales will be
resolved by the same operational modeling framework. For
state variables, future operational modeling capacity (including
forecast, reanalysis, and scenario-based projections) will be
extended to cover sedimentation, movements of pollutants
(either as particles or Eulerian tracers) and biological parameters.
In this dimension, operational ecological modeling will be
developed, different modeling sectors will be coupled, e.g.,
hydrological-ocean coupling, wave-ocean and wave-ice coupling,
and ocean-optical coupling.
There still exist many knowledge gaps toward the
establishment of the seamless operational service. Monitoring
and accurate modeling of water and biogeochemical mass
transport caused by coastal-estuary interaction, inter-sub-basin
exchange and meso- and submeso-scale eddies is still a challenge.
Capacities for precisely predicting currents, upwelling, extreme
sea level and waves in icing waters, skin temperature, algae
bloom, and oxygen depletion are yet to be improved.
BALTEX/BALTIC EARTH MARINE
RESEARCH
In the following, a few selected research highlights from
BALTEX/Baltic Earth are presented, documenting the progress
in physical oceanography of the Baltic Sea during 2003-2014
(Omstedt et al., 2004, 2014; see also BACC Author Team, 2008;
BACC II Author Team, 2015) and after. One of the main
motivations for the foundation of BALTEX in the 1990s was
the exchange of data between eastern and western Baltic Sea
countries. Due to the Iron Curtain after World War II, the
exchange of scientific information was limited. Hence, after the
rise of the Iron Curtain in the 1990s BALTEX aimed to enhance
international collaboration between the Baltic Sea countries and
to increase the exchange of especially observational data.
Process Understanding
With the help of project-oriented research data and process
modeling, our knowledge about oceanographic processes and
the interactions of the ocean with the other components of the
Earth system such as atmosphere, land surface and sediments
has considerably increased since the start of BALTEX (Omstedt
et al., 2004, 2014). For instance, the importance of surface waves
in air-sea interaction of heat, momentum, and matter is better
understood, and Stokes drift and Langmuir circulation have been
identified as likely playing an important role in surface water
mixing explaining the underestimation of mixed layer depth in
many Baltic Sea models (Reissmann et al., 2009).
Research on water exchange between the Baltic Sea and North
Sea and saltwater inflows into the Baltic has a long tradition.
Today we know that, on average, half of the total amount of
salt imported into the Baltic is supplied by barotropic inflows
of highly saline water (Mohrholz, 2018). In particular, the well-
observed, exceptionally strong Major Baltic Inflow (MBI) of
2014 (Mohrholz et al., 2015) enabled unprecedented, detailed
studies of the dynamics of saltwater inflow events and of their
implications for the ecosystem (e.g., Grawe et al., 2015; Schmale
et al., 2016; Holtermann et al., 2017; Bergen et al., 2018). In a
long-term average MBIs contribute 20 to 25% to the total salt
import into the Baltic (Mohrholz, 2018), beside this they are
the solely mechanism for deep water ventilation of the central
Baltic (Meier et al., 2006). Despite the decrease of nutrient supply
after the 1980s, recently observed oxygen consumption rates
are higher than ever observed (Meier et al., 2018b). According
to model results, oxygen consumption in the water column