How climate information can be made user relevant and usable: the case of the Sectoral Information System of C3S
Samuel Almond (ECMWF, Copernicus)
The European Commission has entrusted ECMWF with the implementation of the Copernicus Climate Change Service (C3S). C3S vision is to provide open access, authoritative, quality-assured climate information to support adaptation and mitigation policies and the development of downstream services in response to a changing climate. At the heart of the C3S infrastructure is the Climate Data Store (CDS), which provides access to an array of climate information; providing access to climate reanalysis, observations, seasonal predictions and climate projections through a unified web-interface. The CDS will also provide a comprehensive set of tools (the CDS toolbox) which enables users to develop customised applications and explore C3S data.
The C3S will also provide a working example of how the data and the tools available through the Climate Data Store (CDS) can be used in specific user relevant contexts. The Sectoral Information System (SIS) component of the C3S engages with a diverse set of users to scope out and to develop user driven, sector specific applications.
The Relative Contributions of ECMWF Deterministic and Ensemble Forecasts in an Automated Consensus Forecasting System
Brett Basarab (Global Weather Corporation)
Global Weather Corporation (GWC) post-processes model output from various national and international weather services to generate a consensus forecast that routinely outperforms the competition. GWC employs a forecast technology known as DICast (Dynamic, Integrated foreCast System) initially developed at the National Center for Atmospheric Research (NCAR) and since adapted and improved at GWC. DICast improves upon component model skill by generating a weighted, bias-corrected consensus forecast that is updated daily to align with recent model performance. In this study, we examine the relative performance of the ECMWF HRES deterministic forecast and ECMWF ensemble mean in several GWC forecast products, including 2m temperature and dewpoint, 10m windspeed, and 80-100m turbine hub-height windspeed (for wind power forecasting applications). The HRES and ensemble mean perform well in GWC's system, determined by the large weighting that both receive across lead times. The ensemble mean is especially valuable at longer lead times when it diverges significantly from the deterministic forecast solution. We hypothesize that the mean, by implicitly taking the consensus of all the ensemble members, is an effective way to add skill to our forecasts. In light of these results, we are encouraged to continue to incorporate ensemble forecasts for applications to both our deterministic and future probabilistic forecast products.
An Ensemble-Based Storm Surge Forecasting System for the Coast of Norway
Haldis Berge (Met Norway)
The Norwegian Meteorological Institute (MET Norway) has established a regional ensemble storm surge forecasting system with 51 members, based on forcing from ECMWF ensemble prediction system. The system receives observations from 23 sea-level stations along the Norwegian coast. The observational system is operated by the Norwegian Mapping Authority and 10-minute data are transferred to MET Norway in near real-time. The data are used to correct the forecast for slowly varying errors and trends in the model sea-level. The data are also used for model verifications. The model predictions are the basis for the forecasters on duty to issue warnings when certain threshold sea-levels are expected to be exceeded. For this purpose, three alert levels has been identified for each coastal stations: yellow when forecast exceeds 1-year return value, amber for the 5-year return value and red for exceedanse of 20-year return value. The latter case is defines an extreme-event in which case warning are send to national and regional authorities in addition to responsible rescue and emergency agencies. The extreme-event forecasts are also broadcasted on national radio and television. To simplify this decision processes, a graphical support tool has been developed. This combines both the ensemble and deterministic forecasts together with the appropriate threshold levels and other relevant information for each single station.
The aim of the presentation is to give an overview of the whole production chain from observations and ensemble model forecasts to the final methods for the issuing of sea-level warnings. In addition some examples of the verification scores for the latest winter season (2017-2018) will be presented.
Project Loon and ECMWF
Robert Carver (Project Loon at X Development)
Project Loon involves operating a network of stratospheric balloons to provide internet to remote areas. Accurate stratospheric wind forecasts from ECMWF are vital for this project. In this talk, we will discuss how Project Loon acquires and processes ECMWF forecasts, and how these forecasts are blended with balloon observations to create the best possible estimate of stratospheric winds for planning. We will also discuss how we use ECMWF reanalyses for planning and interpolations errors we encountered using Climate Data Operators with the ERA5 dataset.
Use of direct radiation forecasts to improve the reliability of solar thermal energy
Jose L. Casado-Rubio (AEMET)
Concentrated Solar Power (CSP) is one of the two main techniques employed to harness the energy from the Sun. It has the distinctive advantage of being able to store energy, using it subsequently when needed (for example, at night). PreFlexMS project (www.preflexms.eu) aims to extend the flexibility of CSP to daytime, storing energy when it is sunny and dispatching it when it is cloudy. Forecasts from two meteorological models, the global IFS and the local area model Arome-Harmonie, have been used as inputs to help the CSP plant to decide the optimal schedule to follow.
Both models have been verified against observations from AEMET Radiation Network to estimate their reliability, within the framework of PreFlexMS project. In this talk results will be shown (see attached figures) for the period studied, from March 2015 to February 2018.
Another related project, done jointly with Red Electrica, the Spanish Transmission System Operator (TSO) will be also explained in this presentation. A combination of ECMWF radiation forecasts and Copernicus Atmosphere Monitoring Service (CAMS) aerosol forecasts has been used to improve solar power predictions for days with high aerosol content. Although aerosols don't affect scores significantly when averaged for a long period, it will be shown their huge impact in some events, and how our method can reduce the forecasting error.
Fire forecast: the skill provided by ECMWF ensemble prediction system
Francesca Di Giuseppe (ECMWF)
In the framework of the Copernicus program the European Centre for Medium-range Weather Forecast (ECMWF) on behalf of the Join Research Centre (JRC) is calculating daily fire forecast indices using its medium range ensemble prediction system. ECMWF is also developing open-license software for the manipulation and verification of these data which are made publicly available to guarantee the consistent and reproducible work flow across all users and to aid the use of the products and boost confidence on its quality.
The simultaneous availability of fire danger open-data software tools, poses an unprecedented opportunity to create or advance existing national fire management programs for both European and Worldwide stakeholders. Using one year of operational service in 2017 in this talk we assess the capability of the system globally and analyze in some details three major events that took place in 2017 in Chile, Portugal and California. We also present some suggestions on how products could be tailored to provide information in a probabilistic fashion.
A hands-on session will take place during the course to showcase how to handle ECMWF fire data.
Improvement of numerical model forecasts by using Meteodrone measurements in planetary boundary layer
Martin Fengler (Meteomatics AG)
During the past 2 years Meteomatics drove Meteodrones operationally to collect measure data in the planetary boundary layer up to 1.500 meter height on special places in Switzerland. These data were processed in their own high-resolution numerical model SWISS1k and lead to significantly better predictions compared to the model without drone data - especially in terms of economic-critical events such as thunderstorms, snow and fog situations.
Using this additional data from the lower 1 to 2 km of the planetary boundary layer, it is demonstrably possible to predict more accurately strength and location of meteorological events. The benefit of the forecasts can be seen for the next 24 hours and often up to day ahead. This measurement data also offers a decisive advantage for the calculation of fog and cloud base. And by the way, inversion layers can be identified accurately for measure regions.
The innovation of Meteodrones is groundbreaking. Meteomatics uses a completely new approach and self-developed UAVs (Unmanned Aerial Systems). A Team of young scientists work on perfecting the Meteodrones under special weather conditions like storm and icing conditions. Last year the American Weather Service NOAA has checked the quality of measure data during two project periods.
Using Meteodrones to collect more important data to improve forecast models will be one of the most innovation for next generation of weather forecasts.
Simple access to quality weather data via API
Martin Fengler (Meteomatics AG)
The availability of quality weather data has improved dramatically over the past decade. At the same time the number of big data analytics businesses delivering sector-specific solutions and business insights has also grown dramatically. However timely access to quality weather data, as cut outs that are suited to specific business requirements, delivered in formats that users can simply apply to new and existing in-house systems and models has remained a challenge.
Meteomatics is a commercial weather data provider that is working collaboratively with National Met Services (NMSs), Academia and Scientific communities. We bring together historical, nowcast and forecast weather data from global models such as the ECMWF EC model, satellite operations and station data. By applying in-house modelling and downscaling capabilities, Meteomatics is able to deliver weather data for any lat / long and time series to use in 3rd party models via an industrial scale robust Weather API.
Within this workshop the API is introduced and demonstrated.
Weather API data enable insights that are not only relevant to industry but across all sectors, both public and private. By way of example, leveraging simplified access to these weather data makes it possible to better inform traditional catastrophe modelling and to write business for new perils and territories. Increasingly Insurance Linked Securities managers and reinsurers are looking to weather linked products such as catastrophe bonds as potential new avenues to put capital to work.
In the field of agriculture, weather risk management solutions are already protecting the crops of farmers across Africa from drought and innovative start-ups around the globe are applying weather data to a variety of models to meet precision farming challenges.
Energy companies, both in the traditional and renewable sectors, are extensively using these solutions to forecast demand, power output, inform energy trading, protect themselves against unfavourable seasons and safeguard revenues. Meanwhile, wind farm operators seek protection against low or excessively strong wind to secure cash flow and underpin their financing.
Marine insurers are combining vessel tracks and crew behaviours in differing weather conditions to influence their view of risk, and Lloyd’s of London are using historical weather data and ship tracks to identify fraudulent marine claims.
Water utilities are enhancing demand and leakage models, better managing system capacity and ensuring regulatory compliance through weather-enabled automation of alarms, catchment modelling and enhanced workforce management.
So, in summary, simple API access to quality weather is extending the understanding of weather risk for a broad range of sectors. The speed of development of new products and services underpinned by quality weather data, indices, benchmarks and parametric triggers is growing rapidly.
Diagnostics at ECMWF
Laura Ferranti (ECMWF)
We look at the potential of early warning for severe cold conditions using data from ECMWF extended range forecasts and from the Subseasonal to Seasonal (S2S) Prediction research project archive. We explore the use of a 2-dimensional phase space based on the leading Empirical Orthogonal Functions (EOFs) of mid-tropospheric flow computed over the Euro-Atlantic region, to study the time evolution of flow patterns associated with high-impact temperature anomalies. We find that the phase space is an effective tool for monitoring predictions of regime transitions at medium and extended ranges. We show that a number of S2S systems have some skill in the prediction of cold spells over Europe, even beyond the medium range. In particular the ECMWF model represents well the observed preferential transition paths. We reveal that the MJO impact on the predictive skill of large-scale flow over Europe is asymmetric. We discuss the predictability of the recent cold spell occurred at the end of February 2018.
How well do we use weather forecasts? Current examples of ‘impact’ forecasting, and ideas for future improvements
Isla Finney (Lake Street Consulting Ltd)
Having spoken at previous end user meetings about examples of ‘impact’ forecasting in the energy and agricultural sectors, the aim here is to take one step further. Using some current examples of ‘impact’ forecasts, we examine what causes their limitations. And then suggest how these might be overcome.
Calibration at ECMWF
Estibaliz Gascon (ECMWF)
Uncertainties exist in numerical weather prediction (NWP) ensemble precipitation forecasts due to combination of the imperfect knowledge of the initial atmospheric conditions, the chaotic nature of the atmosphere and because of the incorrect representation of physical processes by NWP models. As ensemble precipitation forecasts from a global forecast model are sometimes under-spread and not always valid at small scales, numerous post-processing techniques have been developed to statistically correct for biases in NWP output. These techniques usually compare a set of past model forecasts with observations in order to identify systematic relationships that can be used to correct the operational forecast. Additionally, computer resource availability is the main cause of the current limitation imposed on the horizontal resolution of ensemble systems, so another option to improve the forecast can be to increase the number of ensemble members at the expense of reducing their spatial resolution.
As part of ECMWF’s contribution to the EU H2020 IMPREX (Improving PRedictions and management of hydrological Extremes) project, several experiments investigating the performance of non-calibrated and calibrated daily precipitation using two different ensemble forecast resolutions have been undertaken. Firstly, a quantile mapping procedure was applied to calibrate both ensemble systems, using 20 years of reforecasts and EFAS (European Flood Awareness System) 5 km gridded precipitation analysis for Europe; supplemental locations were chosen based on the similarity of precipitation climatology and terrain to increase the sample size. Five different ensemble combinations combining subsets of the 50-member operational ECMWF configuration (18 km grid) and an experimental 200-member low-resolution configuration (28 km grid) were tested. Each combination would have similar computational cost to the current operational ensemble.
The verification of the five ensemble combinations (calibrated and raw) was undertaken with daily EFAS precipitation across Europe for June, July and August in 2016 at 1, 3, 5, 7 and 10 days lead time. The CRPS, ROC, reliability, Brier Score, Quantile Score and Relative economic value were evaluated for different 24-hour precipitation thresholds. The verification shows that the most skilful combination is 40 ensemble members from the operational configuration and 40 from the low-resolution configuration. These results suggest that that this set-up combines the advantage of the high resolution forecast system with an improved the representation of the forecast distribution, especially useful for longer lead times. Finally, for all the lead times and combinations, the calibrated forecast increases not only the reliability but also the resolution of the raw forecast.
Breaking the barriers: eLearning at ECMWF
Anna Ghelli (ECMWF)
ECMWF training courses have changed over the years in response to the needs of its Member and Co-operating States. Practical activities, real-world case studies and webinars on demand are some of the features that have been introduced along side the traditional courses.
Collaborations with the wider training community have inspired the latest development at ECMWF, that is the production a set of eLearning modules. The self-contained modules, which can be used as stand-alones or as part of blended courses (eLearning combined with face to face), follow a learner-centred approach to allow for different knowledge levels and learning styles. The modules are created using an instructional design methodology as set out by Mayes & de Freitas (2004), which assumes that information is processed through two channels (auditory and visual) of limited capacity. In order to minimise overload, the learner filters, selects, organises and integrates the information presented in the resources. These assumptions and learning principles will guide the creation of our eLearning modules.
Verification news from ECMWF
Thomas Haiden (ECMWF)
An update is given on the evolution of HRES and ENS skill as seen in headline and supplementary scores. The increase in skill relative to ERA5 (replacing ERA-Interim as a reference) is shown for various surface parameters. New developments in verification at ECMWF are presented, such as two additional headline scores for 2-m temperature in the medium and extended range, the inclusion of observation error in ensemble verification, and the benefit of high-density observations from ECMWF Member and Cooperating States in precipitation verification.
Using ECMWF data for impact based warnings
Elin Björk Jónasdóttir (Icelandic Met Office)
During the winter of 2017-2018 several unusual rain events caused flooding in Iceland. The most severe of the flooding events happened in September and were directly linked to an atmospheric river connected to three major hurricanes in the Atlantic basin. Three consecutive rain events caused catastrophic flooding and land slides, causing damages to roads, farm land and live stock as well as major bridges along the national highway. Several other events occurred during the winter during intense thawing periods causing river flooding, ground flooding and in some cases landslides and avalanches. Most of these events were well forecasted, with fairly accurate severity and forecasted impact. At IMO we have several tools at our disposal but the visualisation of ECMWF data such as SOT and EFI for precipitation as well as ENS and determinist runs for mean sea level pressure, precipitation and temperature were key ingredients in successful forecasts.
The development of new products by forecasters – Example: Probability of Thunder algorithm
Paavo Korpela (Finnish Meteorological Institute)
Often very good ideas grow up from the basic everyday work. While forecasters continuously face varying weather situations and utilize numerous different forecasting tools, sometimes arises need for a certain derived forecasting parameter that doesn’t exist. FMI’s meteorological workstation (SmartMet) includes programmability that enables forecasters to build up scripts and visualize them on-the-go. Scripts can utilize several data sources including NWP models, radar, conventional surface observations, and also location data regarding weather-related emergency calls, lightning detections and citizen observations. This capability motivates and inspires forecasters to visualize, test and develop new algorithms that help weather monitoring, forecasting and decision-making process. Well-designed algorithms can be implemented into the automatic FMI post-processing routines and even into customer products. One of the recent post-processed algorithms is probability of thunder (POT), which is now demonstrated as an example of using ECMWF model data.
The goal was to create a simple but physically well-reasoned thunderstorm forecasting tool that indicates also probability. The best way to approach the likelihood of thunder is to assess the availability of three necessary ingredients; moisture, instability and lift. This approach is used also in the POT algorithm. Conceptual-wise both instability and moisture are represented by most unstable CAPE integrated between temperatures -10....-40C. This parameter is associated with the electrification of convective clouds. Lift, and especially the sufficiency of it, is difficult to determine from the ECMWF model output. However, model precipitation is a result of activation of convective parametrization schemes, which is a sign of sufficient lifting. Hence, the lifting term was considered to be well enough represented simply by model precipitation. POT algorithm is based on these two parameters, in which probabilistic characteristics are generated with approximated thresholds and dependency functions. Also ensemble forecasts have been tested showing promising results.
Visualisation products using COSMO-LEPS: recent upgrades at Arpae-SIMC
Andrea Montani (Arpae-SIMC)
COSMO-LEPS is the limited-area ensemble prediction system of the COSMO consortium, running on an operational basis since 2002.
The system, implemented and maintained by Arpae-SIMC, runs at ECMWF as a member-state time-critical application and provides operational ensemble forecasts at 7 km of horizontal resolution over Central and Southern Europe in the short and medium range (up to day 5).
In this contribution, we will show the most recent developments of the system in terms of product generation to assist the activity of Civil Protection Agencies for the generation of alerts. Particular attention will be paid to the visual representation of probabilistic products for wind, precipitation and fog forecasting.
The planned upgrades, including a multi-physics approach for the representation of moist convection and an increase in horizontal resolution to 5 km, will be also presented and their benefits on the probabilistic prediction of surface fields will be quantified over a 2-month verification period as well as for individual case studies.
Finally, the blending of COSMO-LEPS products with those of the recently-developed convection-permitting ensemble system COSMO-I2-EPS will be discussed.
Forecasting the "Beast from the East" and Storm Emma
Ken Mylne (Met Office)
The end of February and first days of March 2018 saw the most severe outbreak of cold winter weather in the UK for several years, as well as over much of western Europe, compounded by Storm Emma which came from the south and generated exceptional blizzards and freezing rain in the south of the country. This series of events presented a major challenge to forecast systems and operational meteorologists alike. The cold outbreak had been well predicted in advance, firstly from long-range predictions of a Sudden Stratospheric Warming increasing the risk of blocking, and then two weeks ahead from the Decider system. Decider clusters ECMWF and other global ensembles according to weather regimes, and identified very high probabilities of cold easterly flows up to two weeks ahead. In fact the real cold air came a few days later than first anticipated but by the end of the week before it was very clear that a major winter outbreak was coming. Forecasting throughout the week exploited a combination of ECMWF and Met Office models, and a few others including NCEP, to provide multi-model ensemble guidance at all time ranges. Key to successful forecasting under such extreme conditions was a strong interaction between Science teams and operational meteorologists to analyse model guidance and provide new diagnostics, resulting in exceptionally good guidance and warnings throughout the week, including an unprecedented simultaneous two Red warnings in different parts of the UK. This talk will focus particularly on the use of ECMWF data in supporting the early forecasting of the event, and the benefits of a multi-model ensemble approach.
ECMWF@UEF2018: past, present and future
Florian Pappenberger (ECMWF)
ECMWF has made a noticeable step towards fulfilling our Strategy to 2025, with the implementation of Cycle 45r1 in June bringing seamless coupling to a dynamical 3- dimensional ocean and sea-ice model in all ECMWF forecasts. In the past year, coupling processes have progressed in line with the Earth system approach, options for higher resolution ensembles are being investigated, new headline scores are proposed focussing on ensembles and extremes, and Scalability is delivering its share of the efficiency gains required to remain sustainable. Particular highlights include:
Very significant progress has been made in the representation of physical processes, leading to a reduction in radiation biases and an improved representation of precipitation near land/water boundaries.
Recent changes in the Ensemble of Data Assimilation (EDA), which plays a key role in providing input to 4D‐Var, contributed to better scores in Cycle 43r3. Cycle 45r1 is improving analyses notably through a better use of radiosondes, whilst changes addressing the systematic short-wave radiation biases in the storm tracks and over the southern oceans, as well as changes addressing the longstanding precipitation issues along coastlines, improve forecasts.
A cost–loss value analysis indicates that the precipitation type (rain and snow) probabilities are useful for decision-making for a broad range of cases.
Substantial progress has been made by coupling the complete IFS moist physical parametrizations to the prospective IFS-Finite Volume Model (FVM) dynamical core. The results so far bode well for the future with the new approach giving us enhanced flexibility to be able to take advantage of evolving HPC architectures.
New Continuous Data Assimilation is allowing the assimilation of observations up to 30 minutes before cut-off time, as well as the start of the 4D-Var analysis earlier for an identical delivery time. Experimental results for this process to be implemented as part of IFS Cycle 46r1 show consistent improvements, translating into a 2 to 3-hour gain in forecast skill.
The new seasonal system, SEAS5, partially supported by Copernicus, was successfully implemented last November, providing significantly improved El Niño forecasts.
Diagnostics and developments linked to the Arctic area have been the focus of various activities, in collaboration with Member States in the context of the Year of Polar Prediction.
In Copernicus, three chemical modules are now included in CAMS (CB05/BASCOE from KNMI, MOZART from MPI and MOCAGE from Météo‐France), ERA‐I now figures prominently for the first time on the state of climate presented by WMO at COP23 in Bonn, and the global flood predictions from GloFAS now extend to the seasonal timescales. The highly-anticipated release of the Climate Data Store (CDS) marks the marks the transition of C3S into its operational phase. Other notable Copernicus activities include the release of CAMS reanalysis, GloFAS also becoming fully operational, and the new FIRE project delivering its first forecasts.
eLearning modules have been made available, to improve the efficiency of the training provided by the Centre to its Member States.
A new user guide to ECMWF forecasts was released on 14 May. This includes improved descriptions of the most popular parameters and will make ECMWF forecasts easier to use. Also on the users’ front, a new vision for data services has been developed, based upon the concept of providing Member States and other customers with ECMWF NWP products via an on-site cloud named ecCloud.
Identification of weather factors affecting the number of refugees and migrants following the central Mediterranean route
Thomas Petroliagkis (Joint Research Center)
Every summer since 2014 has seen an increase in the number of refugees and migrants fleeing war, violence, persecution, or poverty to cross the Mediterranean Sea to Italy and Malta following the Central Mediterranean Route (CMR), which re-emerged as the world’s deadliest maritime migration route in 2015 and again in 2016 and 2017. The CMR is heavily influenced by a variety of interconnected factors, including immigration policies and border control in North Africa and Europe; shifts in the main countries of origin of refugees and migrants; changing social, political, economic, and environmental conditions in countries of origin, transit, and destination; the adaptability of smuggling networks and the prevailing weather and seasonal patterns. Refugees and migrants are increasingly using the CMR again, which has become the main route after the effectively closing of the Greece’s northern borders and the Turkey/Greece route in Kos/Lesvos islands as a result of the EU/Turkey agreements (2016). Based on the results so far, it seems highly probable that a certain number of meteorological parameters and weather factors affect significantly the number of monthly arrivals over CMR areas especially during the later years (2013 to 2017). Above all, a seasonal pattern can be seen in the monthly number of arrivals peaking up during the warm months and decreasing during the cold months of the year. On close investigation, a distinct aspect of this pattern relating to air temperature was identified and documented especially during the later years.
In non-extreme mode, although a coherence (correlation) between the number of refugee arrivals and air temperature (T2M) values does not seem to exist during the first years (until 2013), a significant (positive) correlation was found between the monthly mean of daily mean air temperature and monthly number of arrivals while a similar behaviour exists for the maximum and minimum air temperature. Further, a coherence between the number of refugee arrivals and significant wave height (SWH) values does not seem to exist during the first years (until 2013), whereas a significant (negative) correlation has found between the monthly mean of daily mean significant wave height and the monthly number of arrival for the later years. An almost identical (negative) type and strength of coherence was found to exist between arrivals and maximum wave height (HMAX), wind speed (WS) and wind speed gusts (FG6). For sea surface temperature (SST), results showed that in zero lag mode significant (positive) correlation has found between the monthly mean of daily maximum SST and monthly number of arrival for the later years. Correlations become distinctly higher in one-month lag mode for the later years. In addition, for total (large-scale plus convective) precipitation (PRECI), results revealed that a significant (negative) correlation has found between the monthly mean of daily-accumulated PRECI and monthly number of arrival for the later years. Same wise, for total snowfall (SNOW), results revealed that a considerable (negative) correlation has found between the monthly mean of daily-accumulated SNOW and monthly number of arrival for the later years.
In extreme mode, not clear results of coherence were found for the extreme forecast indices of wind speed at 10 meters (WSI10), wind speed gusts at 10 meters (FGI10) and air temperature at 2 meters (TI2) although all of them (as non-extreme parameters) were found to produce distinctly high (absolute) values of correlation. On the contrary, considerably high (negative) values of coherence were found for both extreme forecast indices of precipitation (TPI) and snowfall (SFI).
This study has focused mainly over monthly data so far, so, even if results are quite promising and strengthening the possibility of tailored weather (factor) anomalies, the forecast horizon of one month is mostly pointing to planning rather than actually operational activities performed by (Italian) Coastguard and FRONTEX. Based on this, we are already planning to focus and investigate over daily arrivals data referring to CMR or even to the other two routes (Eastern & Western). A potential coherence between daily / sub-weekly / weekly / half monthly / arrivals and prevailing weather conditions could be the basis of designing and producing new tailored operational products with various forecast horizons spanning the short (T+0 to T+48 hours), early medium-range (T+72 to T+120 hours), medium-range (T+144 to T+240 hours), late medium-range (T+264 to T+360 hours) besides the monthly one (T+720 hours). Such products should be prepared (designed) in full collaboration with (Italian) Coastguard and FRONTEX while their validation, verification, calibration and optimisation before final implementation should be the next steps to follow.
GloFAS / EFAS
Christel Prudhomme (ECMWF)
The European Flood Awareness System (EFAS) and the Global Flood Awareness System (GloFAS) are part of the Early Warning Services of the Copernicus Emergency Management Service, funded by the European commission. They provide operationally trans-boundary probabilistic flood forecasts for up to 30 days ahead, and overviews of upcoming wet and dry conditions up to 4 months in advance. In Europe, the service is accessible to national authorities, whilst outside Europe, GloFAS forecasts are freely available to all users. This presentation will give an overview of the two services and their products, and examples of possible applications.
ECMWF product development
David Richardson (ECMWF)
The presentation will review forecast product development activities at ECMWF over the past year, in response to user requests and feedback. New forecast output fields including lightning activity, vertically integrated water-vapour transport and maximum CAPE will be presented. Enhancements to ecCharts include the addition of extended-range forecasts, with a number of interactive functionalities. Recent work to evaluate the ability to predict severe events in the extended range will be reported. A major new ecCharts feature, the ability to plot vertical profiles of HRES and ENS forecasts, will be introduced. Recent changes to the ECMWF web site will be covered, and the new substantially revised edition of the User Guide to ECMWF Forecast Products will be highlighted.
On the causes of systematic biases in near-surface weather parameters in the ECMWF forecasting system
Irina Sandu (ECMWF)
ECMWF has started an internal project entitled ‘Understanding uncertainties in surface-atmosphere exchange’. The main objective of this project is to better understand the causes of biases in
near-surface weather parameters, such as temperature, humidity and winds, in the Integrated Forecast System in order to come up with suggestions of how they could be reduced. The representation of these
parameters is governed by many processes, such as clouds, radiation, turbulent mixing, surface processes or the strength of the land-atmosphere coupling, which makes error attribution difficult.
Although the forecasts of weather parameters have improved over the years, along with improvements in large-scale forecast skill, many unanswered questions remain regarding systematic and persistent features of these biases, like for example the underestimation of the diurnal cycle of near-surface temperatures over land. The main findings of this project so far will be summarized in this talk.
Exploring Geoscience with AR/VR Technologies
Tim Scheitlin (NCAR)
This contribution is invited (by Dr. Anna Ghelli). This would be an overview presentation regarding efforts at the National Center for Atmospheric Research (NCAR) in Boulder, Colorado, USA, to develop augmented and virtual reality tools for exploring geoscience datasets. Specifically, the talk would focus on two mobile applications, Meteo AR and Meteo VR, that were developed to help make weather, climate, and space science more accessible and engaging to the general public. The talk would be followed by a hands on demonstration where the audience could download the apps and try them on their own mobile devices. Note - the demo would require that the venue have wifi access for the participants.
Imaginative Insights or Flights of (Forecast) Fantasy?
Leonard Smith (LSE CATS)
Embracing the possibility that weather prediction will never provide accountable probabilities frees us to do much more insightful things with the models we have today, things much more relevant and more useful for the consumers of weather information. This discussion will be framed in the context of anticipatory disaster risk reduction and the energy sector; it is trivially generalise to a host of other applications in the insurance sector, agricultural sector, satellite launches and elsewhere.
First we illustrate the value (and need) for easing practitioners’ ability to explore bespoke forecast evaluation statistics. The aim here is to make it easier to judge the strengths and weaknesses of a forecast for particular targets at a particular (reference class of) locations; this will significantly aid those tasked with deciding whether today's best available forecast system is adequate for their purpose. It is useful to keep in mind the value of getting even the best available forecast off the table when it does not inform decision of the day. Second, letting go of the drive for "fantastic objects" (for example probability forecasts, useable as such) frees physical scientists to reorient resource allocation toward providing more valuable forecast information. Examples from the Just Enough Decisive Information (JEDI) framework, including the use of sculpted ensembles, illustrate why this approach is attractive to the consumers of forecasts and conceivable given the resources of an evolving Global Weather Enterprise (GWE). Accepting that the probabilities issued will never reflect the true probability of future conditions can significantly increase the value of our forecasts and the speed with which our understanding of the Earth System increases. While physical scientists may imagine things beyond the limits of mathematical possibility, we will soon see more forecast products for the heavens and the earth than are dreamt of in our philosophy.
ECMWF forecasts for African small-scale farmers
Fiona van der Burgt (Weather Impact BV)
In drought-prone Sub-Saharan Africa every rain drop counts; around 96% of agriculture is rain-fed. Weather affects almost every aspect of agricultural business, from determining the time of planting and harvests to efficient planning of irrigation.
Weather Impact provides African farmers with local weather forecasts based on output of the ECMWF ensemble model. From the terabytes of model output data a mobile-text message of maximum 160 characters is formulated. Text messaging allows to send localised forecasts through a technology that is widely available in rural Africa. Based on the forecasts, farmers are supported in determining the best time to plant, optimise the usage of fertiliser or protect their crops if hazardous weather is coming up.
We present several case studies from Ethiopia, Kenya and South-Africa. For Ethiopia we validated the performance of the ECMWF model output in collaboration with the National Meteorological Agency. With this work we aim at supporting African farmers to increase food security and strengthen their resilience to climate change.
Turbulence resolving weather forecasting : applications and operational aspects
Remco Verzijlbergh (Whiffle Weather Finecasting Ltd)
Despite advances in computing power, currently operational NWP models are still far away from directly resolving small-scale phenomena like turbulence and small boundary layer clouds. Whiffle Ltd. is a spin-off from Delft University of Technology that is running, to our best knowledge, the first operational large-eddy simulation (LES) based NWP model in the world. One of the main motivations of using LES in the context of operational weather forecasting is that it overcomes some of the difficulties in the parameterization of turbulence and boundary layer clouds.
In addition to an improved representation of the parameterized atmospheric processes named above, a high resolution LES model also has the benefit of capturing phenomena on scales that are important in many economic sectors. For wind energy applications, dispersion of pollutants or small inland water wind forecasts, to name a few application areas, a resolution below 100m brings potential benefits because it allows for directly resolving wind turbines, buildings or (high) vegetation. Facilitated by the emergence of high-resolution datasets of e.g. land-cover, land-use and elevation, new applications of local weather forecasting are arising.
Whiffle has been running its atmospheric LES model coupled to the ECMWF deterministic high resolution forecasts as well as the recently released ERA5 reanalysis. In this talk we discuss a number of practical applications and operational aspects of LES based weather forecasting.