Nowcasting Guidelines – A Summary

An international expert task team on nowcasting has developed the WMO Guidelines for Nowcasting Techniques (WMO, 2017) to initiate a process for an enhanced integrated and seamless WMO Data-processing and Forecasting System (DPFS). Their purpose is to help National Meteorological and Hydrological Services (NMHSs) by providing them with information and knowledge on how to implement a nowcasting system with the resources available to them and an understanding of the current state of science and technology. This article summarizes the guidelines.

Keith Browning first defined nowcasting in 1981 as “the description of the current state of the weather in detail and the prediction of changes that can be expected on a timescale of a few hours.” In 2010, the WMO Working Group on Nowcasting Research defined nowcasting as forecasting with local detail, by any method, over a period from the present to six hours ahead, including a detailed description of the present weather. This is the definition used in this article.

Nowcasting is generally applied to weather that occurs on the mesoscale and local scales over very short time periods. Thus, the emphasis is on the need for rapidly updated, high-resolution observations of, for example, thunderstorms, tornadoes, hail, heavy precipitation, severe wind, visibility (fog) and winter precipitation types. A well-trained nowcasting forecaster requires an integrated display system that contains observations from the various instruments and sensors on the same display with the same grid spacing for each dataset. Types of observation include radar, satellite, lightning networks, surface stations, wind profilers and radiosondes. During periods of high-impact weather, forecasters should be continually monitoring the latest observations via frequently updated integrated displays. In addition to high-resolution observations, numerical weather prediction (NWP) analysis and forecast fields and products from nowcasting systems should be viewable on the same display.


Observations and nowcasting techniques

While surface and upper-air observations are important for nowcasting purposes, only remote-sensing systems can adequately provide high-resolution spatial coverage. Sophisticated nowcasting techniques exist in developed countries where radar systems are mature and robust. However, in less developed countries and remote areas, the required operational radar systems needed for nowcasting are missing.

Weather radar systems are the most important instruments for nowcasting, particularly for convective weather phenomena. However, these instruments are also the most expensive and sophisticated, and are difficult to maintain. Radar systems have advantages over all other observing systems when it comes to nowcasting the phenomena associated with precipitation because they directly observe precipitation particles in three dimensions over a large area with an update rate of a few minutes. At radar ranges of < 60 km, the resolution of the precipitation is < 1 km. This makes it possible to: (a) estimate rainfall rates and amounts; (b) observe the three-dimensional (3-D) structure of a storm, which has proven useful in estimating storm severity; and (c) obtain the movement of storms, which is central to nowcasting. With the addition of Doppler capability, wind can be estimated. This has proven particularly valuable for the issuing of warnings for tornadoes, microbursts and other damaging winds. With the further addition of dual polarization (transmitting and receiving two differently polarized wave forms), it is possible to differentiate precipitation particle type (rain, snow or hail) and to identify non-precipitation echoes such as insects and ground clutter. This is particularly useful for data quality control, identifying precipitation type and improving precipitation estimates.

The Guidelines for Nowcasting Techniques

The Guidelines for Nowcasting Techniques was published by WMO in 2017

Surface weather stations are sparsely located and/or of poor quality in many regions of the world. Existing stations may be sited incorrectly (not according to WMO standards), not well maintained and have limited communications established at the sites for real-time monitoring. In developing countries, the lack of resources to acquire and deploy instrumentation and the lack of training of local weather service staff to properly site, calibrate and maintain the equipment amplify the issues of poor-quality observations. As commercially available meteorological instruments are relatively expensive, instrumentation that fails, or is stolen, is often not replaced. The result is that weather observations in critical regions are not available.

In these data-sparse regions, an international initiative has been established to develop and deploy low-cost weather instrumentation. The goal is to provide technology to weather services in developing countries so they can build, deploy and maintain their own surface observation network. Instrumentation has been designed using innovative new, low-cost technologies such as 3-D printers, small inexpensive computing systems (for example, Raspberry Pi) and wireless communications. If the station gets destroyed or a sensor fails, it can be replaced at low cost using 3-D printer designs. With the advent of the Internet, wireless communications, mobile phone coverage and faster computers, the technology now exists for rapid transmission of surface-station data.

Different nowcasting methods exist that vary from simple extrapolation of radar precipitation echoes or animated loops of clouds observed by satellite, to sophisticated systems that combine output from feature detection and nowcasting algorithms with rapidly updating, integrated displays of observations and NWP output. The transitory nature and smaller scale of some types of weather (for example, tornadoes and microbursts) generally dictate the type of nowcasting techniques that can be applied for severe weather warnings; in this case, simple extrapolation techniques are frequently employed. (For weather phenomena with longer timescales and larger spatial extent, nowcasting systems have been designed to use observations in combination with NWP forecasts to extend the nowcasting guidance out to 6 hours.)

An expert system to nowcast thunderstorms may use: (a) satellite to monitor cumulus cloud lines (convergence lines) and cumulus cloud growth; (b) radar to identify thunderstorms, their intensity and motion as well as boundary-layer convergence-line location and motion; (c) lightning and lightning tendency to fill in locations of thunderstorms and their evolution not observed by radar; (d) upper-air temperature, moisture and winds to obtain vertical wind-shear and stability profiles to estimate potential storm type; and (e) surface stations to monitor potential changes in atmospheric stability. Limited observations are available in many locations around the world, and thus the ability to nowcast specific phenomena varies significantly.

Quantitative nowcasts of rainfall are based on extrapolation of the observed rainfall field forward in time. The advection that is used for the nowcast is based on the apparent motion that has been analysed using the latest radar images. Typically, the advection is estimated using a cross-correlation or optical flow technique that uses tiles of 20–40 km in size. The technique that is used to advect the image forward in time usually requires an estimate of the advection at each pixel in the field. Some form of interpolation is required to distribute the advection estimates from the tiles onto the entire image. Extrapolation algorithms first identify storms as objects in the current radar scan and then track the motion of the storm by identifying the same object in successive scans. This approach is called cell tracking and is suitable for identifying and tracking severe convective storms. Typically, the data are used as input to systems that generate warnings for the hazards that are associated with severe convection: large hail, damaging wind, heavy rain and lightning.

Automated tools and systems are highly desirable because of the very short time periods associated with nowcasting. Automated nowcasting systems are available worldwide for the extrapolation of precipitation and severe storms. Other nowcasting tools are used for quickly detecting trends in storm characteristics, such as intensity, size and movement. Many of these tools have proven particularly successful for warning purposes, for example, for microbursts, mesocyclones and heavy rain accumulations. Users of nowcasting systems are often in need of real-time information for their downstream applications and for a quick assessment of the current weather situation. Consequently, the available computation time is limited, in particular for meteorological parameters with a high update frequency such as precipitation.

In contrast to NWP models, which feature a high level of sophistication and comprehensive physics, resulting in long computation times, nowcasting systems must be kept comparatively simple and often exhibit heuristic approaches. In this context, heuristic means that limitations in a method (for example, uncertainties, inaccuracies and limited applicability) are accepted with regard to an otherwise disproportionate expenditure of time or resources. These nowcasting methods or tools have different limitations depending on the regions they are designed for or the main goal they are used for.

The most accurate warnings of severe convective weather require human involvement. First, the nowcaster should examine the synoptic pattern and NWP forecasts. Based on the nowcaster’s knowledge of the local climatology and conceptual models of severe storm evolution for the local area, the nowcaster should decide if severe weather is likely for the day. Second, the nowcaster should conduct an analysis of the latest local sounding, if available, for vertical wind shear and stability and likely changes that may occur during the day. Based on this analysis, the nowcaster should estimate the type of storms that might occur, such as supercells, multicells, single cells and squall lines. Third, if rapidly updating satellite and/or radar are available, the nowcaster should continually monitor for boundary-layer convergence lines – locations where storms are likely to first develop. Once storms develop, and if a radar is available, the primary focus should be to look for features or signatures that might signal imminent severe weather. These signatures include high reflectivity, velocity rotation couplets, divergence velocity couplets, and bow and flare echoes. The nowcaster should then use automated extrapolation techniques to pinpoint future locations of severe weather.


Quality of nowcasting information and training

The methodology and metrics of nowcast verification should be carefully chosen to produce information that is meaningful to the user. A two-way dialogue is necessary to ensure that users obtain the information they require. As a basic guide for developing nowcast verification, those responsible should: (a) understand the needs of users interested in nowcast verification; (b) identify verification methods and attributes that can answer the questions of interest; (c) select metrics and graphics that appropriately measure and represent the attributes; (d) identify and collect a representative matched set of forecasts and observations; (e) compute verification metrics using, for example, free available verification tools and packages such as Model Evaluation Tools and the R verification package; (f) depict the verification results in ways that are meaningful to users; and (g) make nowcast quality assessments on a routine basis to provide ongoing information about nowcast performance. (For more details on verification of nowcasts, see Forecast Verification for the African Severe Weather Forecasting Demonstration Projects (WMO, 2014) and the WMO Joint Working Group on Forecast Verification Research website at

Nowcasting needs to generate scientifically correct products and also needs to explain to the customer how to use these products. Thus, continuous training in meteorology and in communication is necessary for successful nowcasting. Trainers must rely on high scientific expertise and social and pedagogical competencies (see WMO (2013) for more details). Skilled forecasters and trainers in many meteorological disciplines are necessary, which is usually impossible for one training institute. Thus, international cooperation is needed to deliver training in nowcasting that fulfils the requirements according to WMO. Several organizations provide resources that are useful for those conducting nowcasting training. Below are four key examples to consider:

  1. WMO: the link leads to learning resources of aeronautical meteorology that can be also used for nowcasting training. The link contains resources for instructors and training managers. Additional assistance in education and training, also in relation to nowcasting, is given at
  2. European Organisation for the Exploitation of Meteorological Satellites (EUMETSAT): the link leads to training courses and a training library.
  3. European Meteorological Training (EUMETRAIN) ( an international training project, sponsored by EUMETSAT, that offers training material (manuals, interactive training modules, case studies, online teaching and courses) and training support in the field of satellite meteorology.
  4. Cooperative Program for Operational Meteorology, Education and Training MetEd (COMET MetEd) ( online training resources are available for all fields of meteorology, hydrology and climatology, and for different target groups.


Applications of nowcasting

An important aspect in nowcasting is the early detection of situations conducive to warnings and the rapid dissemination of this information to communities. Adequate application of nowcasting products and correct response to warnings could significantly contribute to the optimization of protective measures and the reduction of losses due to disasters. Population and economic growth have increased the number of lives at risk during severe weather and the financial impacts resulting from severe storms. Meanwhile, modernization of societies means that meteorological information is now indispensable to people’s daily lives. Compared to the more specific and narrower requirements for nowcasting products in specialized areas (such as aviation, road, hydrology and marine applications), requirements of general-public nowcasting bring a wider variety of expectations on nowcasting products. These requirements are for severe weather phenomena and also on different kinds of elements that are relevant to safety, health, daily life, tourism and entertainment.

Some aviation nowcasting systems, such as the Convective Nowcasting Oceanic system, blends polar orbital and GEO satellite observations with global model forecasts to produce 0–2 hour nowcasting of thunderstorm hazards (turbulence, icing and lighting) along aviation routes over the oceans. The accumulation of ice on taxiways, runways and aircraft due to freezing precipitation also greatly affects the safety and efficiency of aviation. The Weather Support to De-icing Decision-Making system, developed in the United States by the National Center for Atmospheric Research (NCAR) Research Applications Laboratory, is a real-time, operational system run at airports that provides aviation users with current and short-term forecasts of weather conditions, including the liquid-equivalent snowfall rate during winter storms. The system combines radar reflectivity information with precipitation rate from a network of surface precipitation gauges and uses a cross-correlation tracking algorithm to produce 60 minute nowcasting of precipitation rate on the ground. The system also provides alert information when icing conditions (freezing drizzle, freezing rain, freezing fog and frost) are observed. Displays at the operational facilities of the airport show the detection and nowcasting products, giving airport operations personnel the ability to observe and monitor changing weather conditions in real time.



While they were working on the Nowcasting Guidelines, the expert task team developed the following recommendations for NMHSs interested in building or developing nowcasting capability:

  • Contact the WMO DPFS division to be put in contact with experts for assistance
  • Engage end users to identify and prioritize their needs and requirements related to high-impact weather warnings
  • Assess all the available observations in terms of high data quality, timely transmission to a central point, data display and storage, and address deficiencies
  • Identify, in consultation with experts, the gaps in observations, infrastructures and available resources, and determine feasible nowcasting solutions to address the high-priority needs of end users
  • Develop a plan for an efficient seamless nowcasting system that integrates observations, automated nowcasting techniques and models that are all displayed on a common workstation; this plan should include collaboration with neighbouring countries to share data and model products
  • Develop a plan to ensure long-term technical support, training and expertise to keep equipment, hardware and software updated, calibrated and operational
  • Develop a plan that ensures sustainable nowcasting techniques, continuous forecaster training on all aspects of nowcasting processes and, where appropriate, takes advantage of WMO Severe Weather Forecasting Demonstration Project training workshops and available material
  • Verify the quality of nowcasting products according to the weather phenomena and users’ requirements
  • Ensure that forecasters play a vital role in the nowcasting processes, despite the availability of automatic nowcasting techniques.

The Nowcasting Guidelines contain much more information than be covered in a short article. We encourage all NMHS to use the publication to develop their nowcasting capability.



The task team of international nowcasting experts :

  • Yong Wang (Zentralanstalt für Meteorologie und Geophysik, Vienna)
  • Wilfried Jacobs (Deutscher WetterDienst, Offenbach, Germany)
  • Larisa Nikitina (Russian Federal Service for Hydrometeorology and Environmental Monitoring, Moscow)
  • Rita Roberts (NCAR, Boulder, United States)
  • Jianjie Wang (China Meteorological Administration, Beijing)
  • Jim Wilson (NCAR, Boulder, United States)

With contributions and coordination from the WMO Secretariat (Abdoulaye Harou, Estelle De Coning and Paul Joe)



Bailey, M.E., G.A. Isaac, N. Driedger and J. Reid, 2009: Comparison of nowcasting methods in the context of high-impact weather events for the Canadian Airport Nowcasting Project. International Symposium on Nowcasting and Very Short Range Forecasting, Whistler, British Columbia, 30 August–4 September.

Bowler, N.E., C.E. Pierce and A. Seed, 2004: Development of a precipitation nowcasting algorithm based upon optical flow techniques. Journal of Hydrology, 288(1):74–91.

———, 2006: STEPS: A probabilistic precipitation forecasting scheme which merges an extrapolation nowcast with downscaled NWP. Quarterly Journal of the Royal Meteorological Society, 132:2127–2155.

Dixon, M. and G. Wiener, 1993: TITAN: Thunderstorm identification, tracking, analysis and nowcasting – a radar-based methodology. Journal of Atmospheric and Oceanic Technology, 10:785–797.

Germann, U. and I. Zawadzki, 2002: Scale-dependence of the predictability of precipitation from continental radar images. Part I: Description of the methodology. Monthly Weather Review, 130:2859–2873.Hering, A.M., C. Morel, G. Galli, S. Senesi, P. Ambrosetti and M. Boscacci, 2004: Nowcasting thunderstorms in the Alpine region using a radar-based adaptive thresholding scheme. Proceedings of the Third European Conference on Radar Meteorology, Visby, Sweden, 6–10 September.

James, P., B. Reichert and D. Heizenreder, 2015: NowCastMIX – optimized automatic warnings from continuously monitored nowcasting systems based on fuzzy-logic evaluations of storm attributes. Presented at the Eighth European Conference on Severe Storms, Wiener Neustadt, Austria, 14–18 September.

Johnson, J.T., P.L. MacKeen, A. Witt, E.D. Mitchell, G.J. Stumpf, M.D. Eilts and K.W. Thomas, 1998: The storm cell identification and tracking algorithm: an enhanced WSR-88D algorithm. Weather and Forecasting, 13:263–276.

Jung, S.H. and G. Lee, 2015: Radar‐based cell tracking with fuzzy logic approach. Meteorological Applications, 22(4):716–730.

Li, P.W. and E.S. Lai, 2004: Short-range quantitative precipitation forecasting in Hong Kong. Journal of Hydrology, 288(1):189–209.

Li, L., W. Schmid and J. Joss, 1995: Nowcasting of motion and growth of precipitation with radar over a complex orography. Journal of Applied Meteorology, 34:1286–1300.

Mueller, C., T. Saxen, R. Roberts, J. Wilson, T. Betancourt, S. Dettling, N. Oien and J. Yee, 2003: NCAR Auto-Nowcast System. Weather and Forecasting, 18:545–561.

Reichert, B.K., 2009: AutoWARN – Automatische Unterstützung der Herausgabe von Unwetterwarnungen (AutoWARN – Automatic support for issuing weather warnings). promet, 35(1/2):98–103.

Reinhart, R.E. and E.T. Garvey, 1978: Three-dimensional storm motion detection by convective weather radar. Nature, 273:287–289.

Ruzanski, E., V. Chandrasekar and Y. Wang, 2011: The CASA nowcasting system. Journal of Atmospheric and Oceanic Technology, 28(5):640–655.

World Meteorological Organization, 2013: Guidelines for Trainers in Meteorological, Hydrological and Climate Services (WMO-No. 1114). Geneva.

———, 2014: Forecast Verification for the African Severe Weather Forecasting Demonstration Projects (WMO-No. 1132). Geneva.

———, 2017: Guidelines for Nowcasting Techniques (WMO-No. 1198). Geneva.



Franziska Schmid, Central Institution for Meteorology and Geodynamics (ZAMG), Austria

Yong Wang, Central Institution for Meteorology and Geodynamics (ZAMG), Austria

Abdoulaye Harou, WMO Secretariat 

Share this page