Satellite Inputs to Numerical Weather Prediction (NWP)
Applications
Numerical Weather Prediction (NWP) is a technique that combines current observations of Earth's atmosphere and oceans with mathematical models to forecast future weather. Satellite observations from polar and geostationary orbits play a crucial role in these models, providing continuous and global coverage of Earth's weather.
NWP creates a digital model of Earth’s atmosphere, dividing it into a three-dimensional grid. The model is initialised with a set of conditions that describe the current state of the atmosphere, using observations from satellite and in-situ sensors. Within each grid cell, the laws of physics are applied to calculate properties such as temperature, pressure, wind speed, and humidity, which are then projected forward in time to simulate how the atmosphere will evolve. These models are physics-based systems that numerically solve the fundamental equations governing atmospheric motion and thermodynamics, in contrast to emerging data driven or machine learning models. To represent small-scale processes that cannot be directly resolved, models use parametrisation – simplified equations that approximate their influence on the larger-scale flow. 1) 2) 3)
NWP models have a variety of meteorological applications, including regional and global weather forecasts, air quality analysis, and cyclone tracking. Data used in the models is obtained from satellite, airborne, and ground sensors such as weather balloons, aircraft, buoys, ships, and ground stations. 1) 3) 4) 5) 6)
Satellite Inputs to NWP Models
NWP models receive inputs from a wide variety of satellite sensors, each specialised in collecting data on different atmospheric variables. These include passive sensors such as optical and infrared imagers, and infrared and microwave sounders, as well as active sensors such as radar altimeters and scatterometers, lidar, and GNSS radio occultation instruments. These sensors provide NWP input measurements such as vertical profiles of atmospheric temperature and humidity, cloud profiles, precipitation, wind fields, and surface parameters like temperature. 1) 3) 7) 8)
Imagers, such as those onboard the Himawari and Joint Polar Satellite System (JPSS) series of satellites, observe the Earth’s surface, clouds, and atmosphere, providing data on cloud temperature and height, surface temperature, and aerosol and volcanic ash properties. 1) 3) 7) 8)
Infrared (IR) sounders, most commonly present onboard polar orbiting satellites, provide vertical profiles of temperature and humidity measurements by detecting IR radiation emitted by atmospheric gases. The Infrared Sounder (IRS) onboard the Meteosat Third Generation - Sounder (MTG-S) satellites is important for NWP models as it tracks the three-dimensional structure of atmospheric water vapor and temperature, enabling more accurate short-range weather forecasting. 1) 3) 7) 8) 9) 10)
Microwave sounders, for instance the Advanced Technology Microwave Sounder (ATMS) onboard JPSS, measure microwave emissions from the Earth’s atmosphere and surface. They provide measurements of temperature and humidity, precipitation rate, and snow and ice properties, as well as wind speed. 1) 3) 7) 8) 10)
Geostationary and Polar-Orbiting Satellite Inputs
To produce accurate forecasts, NWP models require inputs from both polar-orbiting and geostationary satellites, which together offer comprehensive spatial and temporal coverage. 1) 3) 7) 8)
Geostationary satellites, such as the Geostationary Operational Environmental Satellite (GOES) series, Himawari, and Meteosat Third Generation (MTG) provide continuous data with very high temporal resolution over an entire hemisphere of the planet. This is important for monitoring rapid changes, nowcasting, and tracking the movement of storms. Despite this, their high altitude and typically fixed inclination results in coarser spatial resolution and limited coverage of polar regions. 1) 3) 7) 8)
Polar-orbiting satellites, including the Joint Polar Satellite System (JPSS) and the MetOp series, orbit at lower altitudes and complete multiple orbits per day. Their low altitude allows for higher spatial resolution and the use of microwave and infrared sounders to capture accurate vertical profiles of temperature and humidity, which are needed for global model initialisation. Polar-orbiting satellites tend to have lower temporal resolution, in comparison to geostationary observations. 1) 3) 7) 8)
Example Products
The type of input data used in NWP models directly influences the scale, resolution, and maximum forecast duration of the NWP model. Different models are optimised to utilise different data types to achieve specific forecasting goals.
Global Models
Global models, such as the European Centre for Medium-Range Weather Forecasts' (ECMWF) Integrated Forecasting System (IFS) and National Centers for Environmental Prediction's (NCEP) Global Forecast System (GFS), are NWP systems that simulate the atmosphere over the entire Earth. They rely heavily on polar-orbiting satellites and in-situ data sources. Microwave and infrared sounding satellite instruments are used in global models to infer the vertical structure of atmospheric temperature and humidity, which are essential for model initialisation. 7) 11) 12) 13)
Regional Models
Regional models, such as NCEP’s North American Mesoscale (NAM) Forecast System and Rapid Refresh/Rapid Update Cycle (RAP/RUC) primarily utilise data generated by geostationary satellites to provide forecasts over individual countries or regions. These satellites provide continuous data with frequent updates, which is required for monitoring of rapid changes. The satellites track cloud and water vapor characteristics to simulate wind fields for nowcasting and very short-range forecasts. 7) 12)
Short-Range Models
Short-range forecasts cover the close future, most commonly from a few hours up to 3 days. This includes ‘nowcasting,’ which describes the current weather and predicts changes on an even smaller timescale, typically 0 to 6 hours. Nowcasting relies on real-time observational data, mainly from ground-based Doppler radars and geostationary satellites, to track the current movement and development of storms. 14) 15)
Medium-Range Models
Medium-range forecasts typically cover a period from 3 days up to 10 or 15 days ahead. The ECMWF's medium-range ensemble (ENS) forecasts, for instance, predict weather up to 15 days ahead, while their high-resolution single forecast (HRES) lasts up to 10 days. 16)
Extended-Range and Reanalysis Models
Extended-range forecasts provide weather predictions typically up to 46 days in advance. Seasonal models simulate even further into the future, projecting atmospheric and oceanic conditions over several months to a year. Unlike shorter range forecasts, seasonal predictions depend less on the exact state of the atmosphere at the start and more on slowly evolving factors such as sea-surface temperatures, soil moisture, and snow cover, which influence larger-scale climate patterns. 17) 18)
Climate reanalysis is a process that combines past observations with models to generate consistent time series of multiple climate variables. Reanalysis generates a dataset of the past climate, allowing researchers to understand long-term trends, study past weather events, and based on that improve climate monitoring. For instance, the ECMWF's ERA5 reanalysis provides a detailed history of the global climate from 1940 to the present. 19)
NWP model | Source | Coverage | Maximum forecast length | Resolution |
European Centre for Medium-Range Weather Forecasts (ECMWF) | Global | 15 days | N/A | |
Deutscher Wetterdienst (DWD) | Global and regional | Short-term | N/A | |
UK Met Office | Global and regional | N/A | N/A | |
Japan Meteorological Agency (JMA) | Global | Daily | N/A | |
National Centers for Environmental Prediction (NCEP) | Global | 9 months | 56 km | |
National Centers for Environmental Prediction (NCEP) | Global | N/A | N/A | |
National Centers for Environmental Prediction (NCEP) | Global | 16 days | N/A | |
National Centers for Environmental Prediction (NCEP) | Global | 16 days | 28-70 km | |
National Centers for Environmental Prediction (NCEP) | Regional | Hourly | 12 km | |
National Oceanic and Atmospheric Administration (NOAA) | Global | Seasonal | N/A | |
National Centers for Environmental Prediction (NCEP) | Regional | Hourly | 13 km | |
Navy Operational Global Atmospheric Prediction System (NOGAPS) | United States Navy’s Fleet Numerical Meteorology and Oceanography Center | Global | N/A | N/A |
National Centers for Environmental Prediction (NCEP) | Regional | N/A | 90 km |
Artificial Intelligence and Machine Learning Approaches
Artificial intelligence (AI) and machine learning (ML) are increasingly being applied to weather prediction alongside traditional physics-based methods. While physics-based numerical models solve the fundamental equations of atmospheric motion and thermodynamics, AI models learn directly from large archives of satellite and observational data to emulate or accelerate these processes. Operational centres such as ECMWF and NOAA are developing hybrid forecasting systems that integrate machine learning for tasks such as data assimilation, bias correction, and sub-grid process representation. Several research and technology organisations, including private sector developers, have demonstrated global AI-based forecasting models capable of generating accurate predictions in seconds. These developments suggest that future weather prediction will combine the interpretability and physical consistency of traditional NWP with the speed and adaptability of data-driven methods.
Related Missions
Geostationary Satellites
Geostationary Operational Environmental Satellite (GOES)
The Geostationary Operational Environmental Satellite (GOES) is a programme of the US National Weather Service (NWS) which provides data for weather forecasting and storm tracking. GOES is a joint mission between the National Oceanic and Atmospheric Administration (NOAA) and NASA. GOES satellites stay in a fixed position relative to the Earth in geostationary orbit, ensuring continuous observation of weather events over the western hemisphere. 21) 22) 23)
Himawari
Himawari is a series of satellites developed and operated by the Japan Meteorological Agency (JMA). The satellites provide imagery and data for weather forecasting, tropical cyclone tracking, and research. JMA conducted the operational satellite switchover from Himawari-8 to Himawari-9 on December 13, 2022. Himawari is the successor to the Geostationary Meteorological Satellite (GMS) series. 24) 25)
Meteosat Third Generation (MTG)
Meteosat Third Generation (MTG) is a new constellation of geostationary satellites from EUMETSAT and ESA. The mission includes two types of satellites, MTG-I (Imager) and MTG-S (Sounder) satellites. The MTG-I satellites are equipped with a Flexible Combined Imager (FCI), which provides improved detection capabilities compared to its predecessor, and a Lightning Imager (LI), which captures real-time information on lightning for more precise forecasts of storms. The MTG-S satellites carry an Infrared Sounder (IRS) and an Ultraviolet Visible Near-infrared (UVN) Sounder. IRS is important for NWP models, since it tracks the three-dimensional structure of atmospheric water vapor and temperature, allowing for significantly more accurate nowcasting and short-range weather forecasting. 25)
Polar Orbiting Satellites
Joint Polar Satellite System (JPSS)
The Joint Polar Satellite System (JPSS) is a series of environmental satellites developed by NOAA and NASA. JPSS satellites orbit Earth approximately 14 times a day, providing full global coverage twice daily. Each JPSS satellite carries a variety of instruments, including the Advanced Technology Microwave Sounder (ATMS), the Cross-track Infrared Sounder (CrIS), the Visible Infrared Imaging Radiometer Suite (VIIRS), and Clouds and the Earth’s Radiant Energy System (CERES). JPSS provides high-resolution imagery and data to NWP models, improving the accuracy of weather forecasts and storm detection. 26) 27)
MetOp-Second Generation Program (MetOp-SG)
The MetOp-Second Generation Programme (MetOp-SG) is a successor to the first-generation MetOp polar-orbiting satellites, developed by ESA and EUMETSAT. The programme consists of two complementary satellite series, MetOp-SG-A and MetOp-SG-B. The A-series satellites carry instruments for atmospheric sounding and imaging, including the Infrared Atmospheric Sounding Interferometer - New Generation (IASI-NG), Meteorological Imager (METImage), Microwave Sounder (MWS), and Sentinel-5. The B-series satellites focus on ocean and ice observations with instruments such as the Scatterometer (SCA) and Microwave Imager (MWI). MetOp-SG aims to provide new and more advanced environmental measurements and improve NWP models. 28)
Polar Operational Environmental Satellites (POES)
The Polar Operational Environmental Satellites (POES) series, developed by NOAA and the European Organisation for the Exploitation of Meteorological Satellites (EUMETSAT), provided global coverage for weather and environmental monitoring. POES satellites orbited the Earth approximately 14 times per day at an altitude of 830 km, covering the entire globe twice daily. The constellation provided data used for weather analysis, forecasting, climate research, and global sea surface temperature measurements. The final satellite of the constellation, NOAA-15, was decommissioned on August 19, 2025. 29) 30)
FengYun
FengYun (FY) are China's meteorological satellites, operated by the China Meteorological Administration (CMA). The programme includes satellites orbiting a Sun-Synchronous Orbit (SSO) at an altitude of approximately 900 km. The satellites are designed to increase the accuracy of global meteorological and climate monitoring by providing data captured by instruments such as the Medium Resolution Spectral Imager (MERSI) and Microwave Temperature/Humidity Sounders. The Fengyun programme began with the launch of FY-1A in 1988. 31)
GPM
The Global Precipitation Measurement (GPM) mission is a joint satellite mission between NASA and the Japan Aerospace Exploration Agency (JAXA), designed to provide next-generation global observations of rain and snow. The GPM Core Observatory satellite, which orbits at a non-sun-synchronous orbit, carries the Dual-frequency Precipitation Radar (DPR) and a multi-channel GPM Microwave Imager (GMI). The GPM Core Observatory is used as a reference to calibrate and compare precipitation measurements from a constellation of partner satellites, providing global precipitation maps every 2-3 hours. 32)
Aeolus
Aeolus is a satellite mission designed and developed by the ESA. Aeolus was the first satellite to carry a Doppler Wind Lidar (DWL) in space. It also carried the Atmospheric Laser Doppler Instrument (ALADIN) which used a laser to measure wind profiles. These measurements provided observations of wind speeds along the instrument's line of sight, which are important for NWP models. Aeolus significantly improved the quality and accuracy of short and medium-range weather predictions by providing highly-accurate wind data. 33)
References
1) NOAA, “Numerical Weather Prediction (Weather Models)”, URL: https://www.weather.gov/media/ajk/brochures/NumericalWeatherPrediction.pdf
2) USA Today, “Supercomputer quietly puts U.S. weather resources back on top”, URL: https://eu.usatoday.com/story/weather/2016/02/22/supercomputer-reston-noaa-cray-ibm/80290546/
3) NCEI, “Numerical Weather Prediction”, URL: https://www.ncei.noaa.gov/products/weather-climate-models/numerical-weather-prediction
4) Kristine Harper, Louis W. Uccellini, Eugenia Kalnay, Kenneth Carey, and Lauren Morone, “50th Anniversary of Operational Numerical Weather Prediction”, URL: https://journals.ametsoc.org/view/journals/bams/88/5/bams-88-5-639.xml
5) A. J. Simmons and A. Hollingsworth, “Some aspects of the improvement in skill of numerical weather prediction”, URL: https://rmets.onlinelibrary.wiley.com/doi/10.1256/003590002321042135
6) ECMWF, “ECMWF’s 4D-Var data assimilation system – the genesis and ten years in operations”, URL: https://www.ecmwf.int/sites/default/files/elibrary/2008/17509-ecmwfs-4d-var-data-assimilation-system-genesis-and-ten-years-operations.pdf
7) Met Office, “Numerical weather prediction models”, URL: https://www.metoffice.gov.uk/research/approach/modelling-systems/unified-model/weather-forecasting
8) NCEI, “Numerical Weather Prediction”, URL: https://www.ncei.noaa.gov/products/weather-climate-models/numerical-weather-prediction
9) WMO Oscar, “ATMS”, URL: https://space.oscar.wmo.int/instruments/view/atms
10) WMO Oscar, “IRS”, URL: https://space.oscar.wmo.int/instruments/view/irs
11) ECMWF, “Integrated Forecasting System”, URL: https://www.ecmwf.int/en/forecasts/documentation-and-support/changes-ecmwf-model
12) Met Office, “Seasonal and climate models”, URL: https://www.metoffice.gov.uk/research/approach/modelling-systems/unified-model/climate-models
13) NCEI, “Global Forecast System (GFS)”, URL: https://www.ncei.noaa.gov/products/weather-climate-models/global-forecast
14) WMO Oscar, “Nowcasting / Very Short-Range Forecasting”, URL: https://space.oscar.wmo.int/applicationareas/view/2_3_nowcasting_very_short_range_forecasting
15) Met Office, “Nowcasting”, URL: https://www.metoffice.gov.uk/binaries/content/assets/metofficegovuk/pdf/data/nowcasting-datasheet_2019.pdf
16 )ECMWF, “Medium-range forecasts”, URL: https://www.ecmwf.int/en/forecasts/documentation-and-support/medium-range-forecasts
17) ECMWF, “Sub-seasonal-range forecasts”, URL: https://www.ecmwf.int/en/forecasts/about-our-forecasts/sub-seasonal-range-forecasts
18) ECMWF, “Seasonal forecasts”, URL: https://www.ecmwf.int/en/forecasts/documentation-and-support/seasonal
19) Copernicus, “Climate reanalysis”, URL: https://climate.copernicus.eu/climate-reanalysis
20) NCEI, “Numerical Weather Prediction”, URL: https://www.ncei.noaa.gov/products/weather-climate-models/numerical-weather-prediction
21) Gunter's Space Page, “GOES N, O, P, Q / EWS-G 1, 2”, URL: https://space.skyrocket.de/doc_sdat/goes-n.htm
22) Gunter's Space Page, “SMS 1, 2 / GOES 1, 2, 3”, URL: https://space.skyrocket.de/doc_sdat/goes-a_sms.htm
23) NOAA, “GOES Performance Status”, URL: https://www.ospo.noaa.gov/operations/goes/status.html
24) CGMS, “Status of Himawari-8/9 and Their Follow-On Satellite Himawari-10”, URL: https://www.cgms-info.org/Agendas/GetWpFile.ashx?wid=e7f585f8-ada6-4930-95a5-027d96e3d5b9&aid=f6137863-9bc0-4498-92f0-df716bff6a85
25) Eumetsat, “Meteosat Third Generation”, URL: https://www.eumetsat.int/meteosat-third-generation
26) WMO Oscar, “Himawari 4th Generation”, URL: https://space.oscar.wmo.int/satelliteprogrammes/view/himawari_4th_generation
27) NOAA, “Joint Polar Satellite Mission”, URL: https://www.nesdis.noaa.gov/our-satellites/currently-flying/joint-polar-satellite-system
28) ESA, “MetOp Second Generation”, URL: https://www.esa.int/Applications/Observing_the_Earth/Meteorological_missions/MetOp_Second_Generation
29) NOAA, “Polar Operational Environmental Satellites (POES)”, URL: https://www.nesdis.noaa.gov/our-satellites/currently-flying/polar-operational-environmental-satellites-poes
30) NASA, “POES”, URL: https://www.nasa.gov/goddard/earth-science-projects-division/poes/
31) WMO Oscar, “Feng-Yun - 4”, URL: https://space.oscar.wmo.int/satelliteprogrammes/view/fy_4
32) NASA, “The Global Precipitation Measurement Mission (GPM)”, URL: https://gpm.nasa.gov/missions/GPM
33) ESA, “Aeolus”, URL: https://www.esa.int/Applications/Observing_the_Earth/FutureEO/Aeolus