Numerical Weather Prediction

DMI carries out research and development on short range numerical weather prediction (NWP) modelling for usage in operational weather forecasting, serving the needs in weather forecasting for the societies in Kingdom of Denmark, including Denmark, Faroe Islands and Greenland.  Research and application on NWP is also carried out in externally funded projects and for specialized products.

DMI issues a hourly updated 2-day short range probabilistic weather forecast for Denmark, Faroe Islands and an extended area in North and West Europe. The forecasts are based on gridded model products with 2.5 km resolution. In addition, DMI is in the process of developing a rapid update nowcasting system with high resolution. For Greenland, the forecasts are issued with 3h update, also with a grid resolution of 2.5 km. For the southern Greenland coastal regions around Tasiilaq and souther Greenland coast, with particularly challenging and frequent storm occurrence, DMI also issues a 6 hourly updated 2 day forecast with a spatial resolution of 750 m, one of the finest ones in the world.

Operational numerical weather prediction is a great challenge. The main NWP model system used in DMI is Harmonie, which is a non-hydrostatic, mesoscale forecast system jointly developed by the international NWP consortium, HIRLAM (HIgh Resolution Limited Area Model). HIRLAM is a consortium consists of 10 European weather services including DMI. HARMONIE is a model jointly developed by HIRLAM in collaboration with Meteo France and the ALADIN consortium, the latter with national weather services in 15 countries in Europe and North Africa.

For operational HARMONIE system, measurement data on meteorological variables from diverse observation platforms such as surface stations, aircraft measurement, radio sondes, ships, buoys, weather radar and satellites are used in a three-dimensional variational data assimilation. The results of which,  together with a corresponding surface analysis, form the initial state for the HARMONIE forecasts.

Areas covered by the operational HARMONIE-arome forecast setup at DMI. The domains in blue are those with 2.5 km grid resolution covering North Europe and Faroe Islands (NEA), Greenland and Iceland (IGB). The domains in red are those with 750 m grid resolution covering some coast areas in southeast and southern Greenland. NEA is run with probabilistic forecast with hourly update. IGB is run with 3-hourly update and the 750 m models with 6-hourly update. All runs are for 2 day forecast.

In order to assess model uncertainties, ensemble prediction systems covering both Denmark and Europe are used. 

A separate activity focuses on optimum use of weather radar observations from the DMI’s radar network.


BALTRAD (2008 – 2012) & BALTRAD+ (2011- 2014) (Link: english (external))

Project aims: The BALTRAD project, and its extension BALTRAD+, is concerned with the creation of a sustainable weather radar network for the Baltic Sea Region, operating in real-time, with high-quality data, and with demonstrated value to forecasters and decision-makers.

DMI's role: Development of a hydrometeor classifier (HMC) in which the radar returns are classified into a number of precipitation and false-echo.

Funding: The European Unions regional programme INTERREG IV B, the Baltic Sea Region.

Period: Oct. 2008 - May 2014.

Collaborators: SMHI, FMI, IMGW, LVGMC, DMI, LHMS, RHMC (Belarus), STUK, EEA, ÅU, Aarhus Water A/S (Denmark),, DWD.

Participants from DMI: Rashpal S. Gill, Martin B. Sørensen and Thomas Bøvith.

CARRA (Copernicus Arctic Regional ReAnalysis)

Project aims: A 24-year reanalysis with the same regional model setup will be implemented and run on two complementary domains to cover the European sector of the Arctic. This will be done with the HARMONIE-AROME model running at 2.5 km resolution. The project is planned for four years duration and starts with a design and setup phase, then the reanalysis enters production in Month 20. The reanalysis dataset will be continuously updated to provide a complete 24 years time series before the end of the project. The work also includes a proof-of-concept Pan-Arctic reanalysis to be run for one year duration.velopment of road weather forecasting system; 24/7 operations.

DMI's role: DMI leads WP1, the work package on system development. DMI participates in all most task within WP2, the work package on input data, and will lead the WP2 tasks 2.1, 2.2 and 2.3 on physiography, glacier input and albedo, and SST and sea ice, respectively. Additionally, DMI will be responsible for the reanalysis production (part of WP3) for one of two CARRA model domains. This domain covers Greenland, Iceland and a few of the adjacent islands in the high Arctic. will be responsible for the reanalysis production in the other domain that covers the Barents Sea, Svalbard and Northern Scandinavia. Finally, DMI will contribute to WP4 on user guidance and support, in which documentation and guidance on the available products from the service will be madeorecasting of weather dependent road conditions, in particular the formation of ice.

Funding: The Copernicus Climate Change Services (C3S) programme, which is a programme initiated by the EU commission. 

Period: Sept. 2017 - June 2021.

Collaborators: (det norske meteorologiske institutt), DMI, SMHI, Meteo France, IMO (Veðurstofa Íslands), FMI.

Participants from DMIKristian Pagh Nielsen, Xiaohua Yang, Jacob L. Høyer, Rasmus Tage Tonboe, Ruth Mottram, Bjarne Amstrup, Kai Sattler, Mats Dahlbom, Pia Nielsen-Englyst, Emy Alerskans, Peter L. Langen

Ice-nucleating bacteria

Project aims: This phd study investigates the sensitivity of cloud ice production in an NWP model to changes in bacterial ice nucleators and thereby develops a new parameterization of ice probability for usage in NWP models.  

DMI’s role: 
Supervision and model development.

Funding: Aarhus University

Period: 2011- 2014

CollaboratorsUniversity of Aarhus; TROPOS: Leibniz Institute for Tropospheric Research

Participants from DMI: Ulrik Smith Korsholm, Niels Woetmann Nielsen, Jens Havskov Sørensen and Maher Sahyoun

IEA SHC Task 46: Solar Resource Assessment and Forecasting: (Link:

Project aims: Advanced solar resource assessment and forecasting is essential for optimal use of all forms of solar energy. The project is the Danish contribution to IEA SHC Task 46 and includes historical and new measurement of solar irradiances and state-of-the the NWP modelling.

DMI’s role: 
Project management (Kristian Pagh Nielsen); NWP forecasting; responsible for the work with improving IEA standards for TMYs;  analysis of the directional radiance distribution measured at DTU; and theoretical calculations of spectral irradiances. DMI also performs pilot testing of equipment for measuring the diffuse and direct components of solar irradiance.

FundingDMI is supported with EUDP grants from the Danish Energy Agency.

Period2012-10 to 2016-12

CollaboratorsDTU Civil Engineering.

Participants from DMIKristian Pagh Nielsen, Peter ThejllKlavs Allerslev Jensen and Claus Petersen

E-GVAP (EIG EUMEETNET GNSS Water Vapour Programme; Link: E-GVAP )

Project aims: The purpose of E-GVAP is to provide estimates of atmospheric properties based on data from ground-ground GNSS receivers (GPS is one example) in near real-time for usage in operational numerical weather prediction and general weather forecasting. The main property estimated is the zenith tropospherica delay, which contains information about atmospheric humidity. The project is based on a tight collaboration between meteorology and geodesy.

DMI's role: DMI is the coordinator of the programme.

Funding: The programme is funded by the participating national metoffices
The programme is in addition funded implicitly by the geodetic institutions
delivering zenith total delay data at no cost.

Period: April 2005 - December 2018, with third phase running 2013-17, and latest extension (jan.2019 - dec.2023 - E-GVAP 4)

Collaborators: 17 European national metoffices
17 data analysis centers determining ZTD from raw GNSS data (mainly geodetic institutions, but also a few European metoffices and NOAA). A large number of regional GNSS network operators providing raw GNSS data for the analysis centers.

Participants from DMI: Henrik Vedel

EnviroChemistry on ECMWF: Enviro-HIRLAM / HARMONIE - Development and Test of an NWP Model System Accounting for Aerosol-Meteorology Interactions (Link: english (external))

Project aims: evaluation and testing of the online integrated Enviro-HIRLAM / HARMONIE modelling systems and sensitivity analyses of feedback mechanisms for chemical weather forecasting and numerical weather prediction. Two main application areas of the integrated modelling to be considered: (i) improved numerical weather prediction with short-term feedbacks of aerosols and chemistry on meteorological variables, and (ii) improved chemical weather forecasting with online integrated meteorological forecast and two-way feedbacks between aerosols/chemistry and meteorology.

DMI’s role: 
improve, apply and evaluate online coupled Enviro-HIRLAM integrated model; sensitivity analyses of feedback mechanisms

FundingECMWF High Performance Computing Project

Period1 Apr 2012 – 31 Dec 2014

Collaboratorscoordinator - Prof. Alexander Baklanov - Danish Meteorological Institute, University of Copenhagen (Denmark), and PhD and MSc students involved into Enviro-HIRLAM/ HARMONIE research and development tasks.

Participants from DMIAlexander Baklanov, Ulrik Korsholm, Xiaohua Yang, Ashraf Zakey, Kristian Pagh Nielsen and Alexander Mahura

ESCAPE: Energy-efficient SCalable Algorithms for Weather Prediction at Exascale

Project aims: The overall objective of the ESCAPE project is to achieve world-class, extreme-scale computing capabilities for European operational numerical weather prediction (NWP). The project is lead by ECMWF.

DMI's role: Code optimization, multigrid option for radiation and dissemination activities.

Funding: EU-Horizon2020

Period: 1. Okt. 2015 - 30. Sep. 2018.

Collaborators: ECMWF

Participants from DMI: Bent Hansen Sass, Kristian Pagh Nielsen, Jacob Weismann Poulsen, Per Berg and Alexander Mahura


Project aims: Forecasting air traffic management (ATM) related risks for EUROCONTROL

DMI’s role: Development of meteorological models of relevance to ATM and operational execution of an ensemble system for Europe coupled with downstream ATM risk model. Communication of ATM risks to EUROCONTROL.


Period: 2013-2015

Collaborators: Metron Aviation

Participants from DMIHenrik Feddersen, Ulrik Smith Korsholm and Bent Hansen Sass


Project aims: HIRLAM (HIgh Resolution Limited Area Model) is a research cooperation of 10 European meteorological institutes, with the aim to develop and maintain a numerical short-range weather forecasting system for operational use by the participating meteorological institutes. HIRLAM-collaboration was initiated in 1985 at DMI and has gone through numerous phases in the past two decades. Since 1 January 2011, the programme entered a phase, HIRLAM-B, with the focus to develop and maintain the non-hydrostatic convection permitting model HARMONIE suitable for kilometer-scale weather focus with particular focus on extreme weather and on probabilistic forecast- HARMONIE is being developed in close cooperation with the ALADIN consortium (lead by Météo-France), as well as European Center for Medium Weather Forecast (ECMWF). 

DMI's role: DMI staffs has been active players in the HIRLAM cooperation throughout all phases of the programme. Each year DMI contributes with an equivalent of 40 person-month on activities coordinated by the HIRLAM programme, mainly on probabilistic forecast, data assimilation, physical parameterisation, system and application and programme management. Since March 2014, DMI has been acting as a RCR (Regular Cycle of Reference HARMONIE) center with a HARMONIE-model for Denmark and neighbouring area.

Funding: DMI as member service contribute to HIRLAM-B with agreed share of work amount as specified in the HIRLAM-B workplan. The work as project leader is funded by the HIRLAM-B programme, which in turn is funded by the annual membership fee by all participating member services.

Period: Jan 2011-Dec 2015.

Collaborators: member services and associate partners of the HIRLAM-consortia; member services of ALADIN consortia; ECMWF, Copenhagen University.

Participants from DMI: Xiaohua Yang, Mats Dahlbom, Bjarne Stig Andersen, Ulrik Korsholm, Kai Sattler, Kristian Nielsen, Niels Woetman Nielsen, Shiyu Zhuang and Alexander Mahura

HydroCast – Hydrological Forecasting and Data Assimilation: (Link: english (external))

Project aims: HydroCast is a joint research project undertaken by a project team of fifteen research scientists, three PhD students and three end-users across ten institutions. In the HydroCast project new forecasting and data assimilation tools are developed and tested that combine different on-line data sources with meteorological and hydrological modelling for provision of probabilistic hydrological forecasts, covering short-range, medium-range and seasonal forecasting.  The tools are tested on three test cases but they will be generic and applicable in a wide range of cases. 

DMI’s role: 
DMI is involved  in two work packages  and  a test study.  In the first work package a quantitative precipitation forecast  system that combines weather radar and high-resolution short-range NWP modelling is  developed. Assimilation of weather radar is based on a nudging scheme that is implemented in the NWP model. The project will develop more effective quality control algorithms for state-of-the-art dual polarisation radars. In particular, major improvement is expected in identifying and removing false echoes in the weather radar data.  In the second work package DMI contributes to probabilistic precipitation forecasts using DMI’s atmospheric ensemble prediction system for precipitation.  DMI’s forecasts are used in a test study for an area around the new Silkeborg highway.

FundingDanish council for strategic research

PeriodJanuary 2012 – December 2015

Collaborators:Danish Hydrological Institute, Geological Survey of Denmark and Greenland, DMI, Aalborg University, ECMWF, Delft University, Danish Road Directorate, Knowledge Centre for Agriculture, Danish Nature Agency

Participants from DMIBent Hansen Sass, Rashpal Gill, Thomas BøvithClaus Petersen and Henrik Feddersen

Nordic CRUCIAL “Critical steps in understanding land surface atmosphere interactions: from improved knowledge to socioeconomic solutions”

Project aims: are continuing (after pilots-themes of the CRAICC-PEEX project) and strengthening research activities in understanding land surface atmosphere interactions and involving key Nordic and Russian research organizations of atmosphere-cryosphere research, research infrastructure facets, and linking further to socio-economical analysis; focusing on research and infrastructure development, science education activity - capability building - mobility.

DMI's role: contribute to research and infrastructure development, and science education activity - capability building – mobility, realization of the PEEX scientific plan, research proposals, participation in join CRUCIAL events, training/ consulting of young researches with Enviro-HIRLAM/ HARMONIE research and development and in particular: for urbanization processes impact on changes in urban weather and climate on urban-subregional-regional scales with contribution to assessment studies; for effects of various feedback mechanisms on aerosol and cloud formation and radiative forcing on urban-regional scales for better predicting extreme weather events with contribution to early warning systems; for GIS estimating of pollution from continuous emissions and industrial accidents for better assessment and decision making.

Funding: NordForsk – Top-Level Research Initiative (TRI); total budget – 2.964 Mnok (DMI - 327 Knok)

Period: 1 Aug 2016 – 31 Dec 2017

Collaborators: 16 partners from 4 Nordic countries and Russia: 8 partners - University Helsinki (Coordinator – Prof. Markku Kulmala; Finland), Norwegian Institute for Air Research, Stockholm University (Sweden), University of Århus (Denmark), Lund University (Sweden) Finnish Meteorological Institute, Nansen Environmental and Remote Sensing Center (Norway), Danish Meteorological Institute, and 8 Russian partners - Moscow State University, Russian State Hydrometeorological University, Tver State University, Tumen State University, AEROCOSMOS, High School of Economics, A.N. Severtsov Institute of Ecology and Evolution, Institute of Atmospheric Optics SB RAS.

Participants from DMIAlexander MahuraRossella FerrettiKristian P. NielsenNiels W. Nielsen 


Project aims: Improved forecasting of heavy precipitation and other hazardous meteorological phenomena including slippery roads and fog. Improved daily weather forecasting with usage in e.g. air traffic management, energy industry and specialized products.  

DMI’s role: 
Development of new methods for assimilating radar derived precipitation, cloud properties and other observational data. Operational execution of rapid update cycles using hydrostatic and non-hydrostatic NWP models in high spatial resolution.

Funding: DMI, Danish Council of Strategic Research, Danish Ministry of Environment (MUDP)

Period: Since 2012 - ongoing activity.

Collaborators: DMI, Aalborg University, DTU-RISØ, Copenhagen University, DHI, GEUS, Knowledge Centre for Agriculture, Danisk Nature Agency, Danisk Road Directorate, Delft University of Technology, European Centre for Medium-Range Weather Forecasts (ECMWF), HOFOR, BIOFOS, and Krüger A/S.

Participants from DMI: Ulrik Smith Korsholm, Henrik Vedel, Claus Petersen, Rashpal Gill, Thomas Bøvith, Martin Sørensen, Mats Dahlbom and Bent Hansen Sass.

OMOVAST (Operativ model til varsling og styring)

Project aims: To enable better handling of waste and runoff water in cases of severe precipitation in the Copenhagen area. Including improved prediction of rain, more precise determination of the amount of water in sewage pipes and on street level, resulting in improved decision making when controlling the pipesystem, use of pumps and reservoirs, and the waste water plants.

DMI's role: DMI improves forecasting of precipitation, by including cloud and radar data in NWP, and by increasing the updatae frequency and time resolution of the NWP forecasts.

Funding: MUDP

Period: Jan. 2013 - Dec. 2014

Collaborators: BIOFOS, HOFOR, Kruger (coordinator), DMI

Participants from DMI: Henrik Feddersen, Claus Petersen, Bent Hansen Sass and Henrik Vedel

SLIM (Source Localization by Inverse Methods)

Project aims: In early October 2017, the International Atomic Energy Agency (IAEA) was informed by Member States that low concentrations of Ru-106 were measured in high-volume air samples in Europe from routine monitoring networks. However, no information was given that an accidental release of Ru-106 had taken place. Such events signify that there is a need for prompt and accurate responses from national radiation protection authorities in such cases. This requires that methodologies, suited for operational use, are developed for localization of the source of contamination based on available monitoring data, and furthermore, that the source term is characterized as well as possible in terms of source strength and time dependence of the release.

For operational use, nuclear decision-support systems (DSSs) should be extended with modules handling such monitoring data automatically and conveying them to the national meteorological centre accompanied by a request to run an atmospheric dispersion model in inverse mode. The aim would be to determine a geographical area in which to find the potential release point as well as the release period. The DSS user should subsequently have the ability to request forward calculations from the potential release sites. These forward runs would involve fitting the dispersion model results to the monitoring data, and the resulting source characterization data should be returned to the DSS. Obviously, the latter facility can be applied also in cases where the release location is in fact known, and hence, the objective is to estimate the source term and the timing of the release.

In the previous NKS-B project MUD, a methodology was developed for quantitative estimation of the uncertainty of atmospheric dispersion modelling stemming from the inherent uncertainties of meteorological model predictions. In SLIM, the inherent meteorological uncertainties will be taken into account by applying the MUD methodology to the inverse modelling approach both with respect to localizing the source, and to deriving the source characteristics, the source term. Previously, due to lack of computational power, such methods could not be applied in operational real-time decision support. However, with modern supercomputing facilities available e.g. at national meteorological centres the proposed methodology is feasible for real-time use, thereby adding value to decision support.

DMI's role: Coordination, development of methodology using the DERMA atmospheric dispersion model, application to selected cases.

Funding: Nordic Nuclear Safety Research (NKS)

Period:  Jan 2019 - Dec.2020.

Collaborators: DEMA (Danish Emergency Management Agency), SSM (Swedish Radiation Safety Authority), STUK (Finnish Radiation and Nuclear Safety Authority), Met Norway, SMHI (Swedish Meteorological and Hydrological Institute), DTU Nutech, PDC-ARGOS, FMI

Participants from DMIJens Havskov Sørensen  ; Henrik Feddersen,  Kasper Tølløse

SolarPACES Guidelines for creation of meteorological data sets for CSP performance simulations

Project aims: Proper knowledge of the availability and variability of solar irradiance - in particularly solar direct normal irradiance (DNI) - is the single most important thing for estimating the performance of planned concentrating solar power (CSP) systems. Classically, such performance simulations have been based on some form of typical meteorological year (TMY) data set that includes 12 typical months of data taken from a longer data series of quality controlled meteorological measurements. This approach, however, is lacking in the sense that the typical meteorological months in the data series do not accurately represent the variability of the solar irradiance. By utilizing multi-year data sets, probabilistic modeling can be performed that accounts for the full variability in a longer data series. Best practices for making and utilizing such data sets are needed. Making such guidelines will be the focus of this project.

DMI's role: Project management and investigation of the future variability in the solar resources with a particular focus on the risk of major volcanic eruptions.

Funding: The IEA SolarPACES program.

Period: 2015-04 to 2016-09

Collaborators: Suntrace Gmbh (Germany), Solar Radiation Monitoring Laboratory at the University of Oregon (USA), CIEMAT (Spain), ARMINES (France), CSIRO (Australia).

Participants from DMI: Kristian Pagh Nielsen

Visibility in the Baffin Bay

Project aims: Investigation of the potential for improvements in visibility forecasting for the Baffin Bay area.

DMI’s role: GAP analysis; operations of new measurement station on Eddefugleøerne; suggesting improvements for visibility forecasting in the Arctic.

FundingMaersk Oil.

Period: 2013-2014

Participants from DMIUlrik Smith Korsholm and Bent Hansen Sass