White Paper One Contributor: Dimitar Ivanov

Contribution

Q9. What will be needed in terms of investment for infrastructure, to realize the vision of future weather and climate forecast? 

(I am not responding to this one at this stage, I guess there will be sufficient material from other responders).

Q10. What would be optimal mechanisms to support developing countries through the advanced future forecasting systems? 

The PPE perspective to this important question could be described as follows.

Currently, there exists a marked inequality in the availability of weather and climate forecast information in the developing countries, and lack of capacity to utilize such information bringing socio-economic benefits to the society. The reasons are well known, but one striking element of this situation is that the massive investments through various development assistant projects, engaging numerous development funding agencies, provided very disappointing results. The impact of these investments estimated in billions of dollars on the local capacity has been very low. Some analysts pointed out that the main reason for such a failure was a model in which the development funds were spent on ‘things’ – automatic stations, telecom systems, computers to run NWP, weather radars (widely non-used after the delivery due to maintenance problems), etc., etc., with little care about the local capacity of the NMHSs to use all this equipment for serving  their countries, and without planning for sustainable operations in the post-project period. A standard part of many of these projects was the idea to install some sort of NWP and run it locally on a PC cluster, without clear idea of the usefulness of such model output. 

There is a need for a paradigm change in the provision of such development assistance, already well understood by many players in the weather enterprise. The main focus of the projects should be on the rapid improvement of the capacity to serve the public and local users thorough utilizing cost-efficient technology and available data from international sources. Of paramount importance is to stabilize the national observing networks, ensure their long-term maintenance for providing standard quality observational data; in addition, these data should be shared internationally – an international obligation of each and every WMO Member which is not fully understood and followed nowadays.   

Another way to help developing countries in places like Africa, for example, is to strengthen the regional centres for weather and climate. The ECMWF model may need to be replicated to certain extent to other regions. This will allow to consolidate human resources and expertise and to optimize the running cost. Such regional approach could be successful only with strong political support, which is a task and challenge for the WMO in cooperation with other relevant international organizations. 

Another big opportunity to be explored and promoted in the future is building partnerships between the NMHSs and the private sector players engaged in the development assistance projects. Private companies can assist in raising the IT capacity, maintaining the infrastructure, and the overall modernization of the systems operated at national and regional level. To do this, perception and culture issues need to be resolved and political will demonstrated from both public and private sides. There is a need for successful pilot and demonstration projects through PPE as a proof of concept. As already demonstrated in other sectors, countries with very low capacity at present may have a historic change to leap frog into the digital era without passing through all stages of development that happened in more developed countries. Some ‘low hanging fruits’ could be picked up by twinning projects in the implementation of APIs to utilize the huge amount of available data for creating local products – e.g., through weather forecast apps for smartphones.

Q11. Human factor – how to develop the needed experts both in NEWP and in the forecast services; how the “weather forecaster” profession will be changed in general? 

The next decade will be a transition period for the profession ‘weather forecaster’ as we know it. The decade of digital transformation will remove from the weather forecasters most of their usual workflow functions, which will be fully automated. One huge challenge for the forecasters will be how to deal with enormous amount of information coming from the observational and forecasting sources. It is already a state-of-the-art to see in the forecasting offices a person sitting behind ten or more screens displaying different types of graphical or alphanumeric information. It is believed that this ‘assimilation process’ happening in the forecaster’s brain will be replaced by artificial intelligence and machine learning – to what extent this will happen in the next decade, is still to wait and see.

The new major role of the weather forecaster as the intelligent human being sitting between the machines and the users (in many cases, also machine), will be to select the right weather data from the 4-dimensional ‘data ocean’ and translate into user-specified (or co-designed) information products that feed the users decision-making process. In some cases, this translation process could be possible without a human middle man, but it is believed that in areas, such as early warning systems for DRR, the human expertise will continue playing key role, in particular, in rapidly changing situations like tropical cyclones, or convective developments. 

With all these disruptive changes in mind, during the next decade the qualification and competence requirements of the weather forecasters need to be revisited and updated, starting from the university programmes and curricula, to the refresher trainings. 

Q12. Based on your views on the impact of these development to the way NMHSs will be organized in the future, your advice to NMHSs – where to put effort and how to avoid unjustified investments? 

The concept of having one agency designated by the government of each country (or WMO Member) as ‘the Meteorological Service’, at present known as NMHS, is central for the whole WMO construct (see WMO Convention, Article 3, for example). Considering that it represents a thinking from the first half of the 20th century, it is natural that this concept is being currently evaluated and updated to represent the new realities of the weather enterprise.

With regard to the future weather and climate forecasting, the NMHSs should continue to play a central role as national agencies funded through the public budget to conduct meteorological, hydrological and climate activities. However, their specific role and responsibility should be determined based on the public needs, technological development, resources, and trends towards regionalization and globalization of a number of essential activities. It is foreseen that the NMHS in the future will retain several major functions through public funding – the operation and maintenance of the basic observing network, the collection and archiving of the national observations, the interface to the regional/global data exchange systems coordinated by the WMO (and other international players – ICAO, IMO, ECMWF, etc.), and serving the government and the public with essential services such as early warnings. The individual portfolio of each NMHS will be determined by the mandates given to it by the national government, so, there will be a great diversity in these portfolios; the part not mandated by the government will represent the market area where products and services will be provided through competitive arrangements by stakeholders from all eligible sectors (national legal frameworks will play a great role to make this possible).

With regard to the core business are of forecasting, the classical NMHS model included always a strong organizational element as a ‘forecasting office’, and in the majority of cases, it has been supported by in-house NWP operations. The huge cost of operating a proper NWP model at a national level has led over the years to a ‘cascading’ model whit a global, regional and national component with the deeper layers of this model aimed at bringing finer resolution to account for the sub-grid local effects. While the regional/local models run by the NMHSs are massively cheaper than the global models operated by, e.g., ECMWF, they still require significant resources of the NMHS – both human and financial. 

One of the most significant changes over the coming decade would be in changing the cascading model and eliminating the need for the so-called limited area models at NMHS level mostly due to the great improvement of the global and regional models operated by the dedicated centres. With the resolution of the global models already below 10 km and moving towards 1 km grid size, little improvement can be achieved by the NMHSs by running local models. Through the open data policy, all the high resolution global and regional NWP data will be made available to the NMHSs which will allow them to focus on the other side of the weather forecasting – namely, generating value for the users by the translation of the NWP into impact-based actionable forecasts, warnings and advisories. The advice to NMHS’s decision-makers should be to give up on their ambition to run inferior quality NWP models ‘in house’ but focus the human development and the financial resources in the development of platforms and interfaces to serve users with end products based on the high-quality NWP data from global and regional models. In addition, the NMHS should provide important verification and feed-back information to the regional and global centres that would support their further improvement.   

(PPE: present and future)

Q13. Quality of forecasts from different sources – how to validate and inform the user? 

For many years, the only source of scientifically based weather forecasts was the national meteorological institution (met office, agency, service, institute, etc.). All the successes and failures of the weather forecasts were attributed to that agency and the user had almost no other choice for getting forecast information. The current multi-stakeholder enterprise provides a completely different landscape with many opportunities to receive forecast information for the same location from multiple sources. With the open data availability of observational data, gridded NWP data at global and regional scale, the possibility to run finer resolution NWP models at affordable cost, and the explosion of opportunities in delivering products through mobile apps, there are now many players – big, medium and small, forming a very dynamic part of the enterprise. It has been estimated that the number of weather apps for smartphones and other mobile devices ranges between 10 and 20 thousands. These apps are used by all sorts of users - from individuals to corporate users, and are provided on free or paid basis. It should be realized, however, that the basic input information used by the plethora of weather apps is quite limited and in many cases, the apps present just ‘repackaging’ of the information from the available sources – trying to attract users by better graphical interface, possibility for localization/customization and other non-essential ‘goodies’. More recently, the competitive advantage is also sought by utilization of AI and machine learning to improve the forecast products for locations and use cases. 

A big question about the quality assurance (QA) exists in this situation and it is linked to the fact that people sometimes need to take important decisions based on the forecasts from those apps (e.g., in some extremes sports, farming, sailing, etc.). The question is how the QA information could be attributed to the individual sources, how to be presented to users, and finally, who should do such quality assessment in an authoritative manner. 

One recent good example of QA is the establishment of a mandatory requirement for a Quality Management System (QMS) for all providers of aeronautical meteorological forecasts and services – this has been imposed as an international requirement by the ICAO, WMO and the European Commission, and includes the need for those providers to obtain an ISO 9001:2015 certificate in order to be considered as legitimate sources. While such approach was applicable for the highly regulated aeronautical sector, it will be difficult to come to at least quasi-mandatory international regulations for the forecast products in other sectors. Nevertheless, the WMO should take a leadership in establishing QA principles and practices that would allow objective evaluation of providers of weather forecasts and related services. The newly established WMO Services Commission should take this task as part of its work programme for the coming decade. WMO should also engage the private sector in the development of the procedures and practices for the QA evaluation in order to consider all use cases and ensure buy-in from the private developers. For instance, there have been already attempts by the private sector to make assessment of the forecast accuracy of various providers (e.g., the ForecastWatch company in the USA). 

Within the span of the next 10 years, it could be envisaged that by 2025, the WMO, in partnership with private sector and academia, should come to an agreeable methodology for QA validation, recognition and attribution of respective visible signs of quality to the various providers from public, private and other sectors. This would allow in the second half of the decade to promote and apply this methodology, thus facilitating the use of high quality sources, in particular, in areas where forecasts are used for critical decision-making. 

Q14. How do you foresee the evolving roles of public, private and academic stakeholders in the future forecasting enterprise? 

An improved coordination and collaboration between the three main sectors on the forecasting enterprise will accelerate the achievement of the vision for 2030. Each sector will continue to play its most relevant role: academic sector – furthering the fundamental science and research, as well as the applied science and research related to the Earth system modeling; public sector – the continuous development of the observing systems, NWP infrastructure and operations, including the WMO-coordinated system of global, regional and national centres and facilities; private sector – developments in several key sub-sectors, like computing and telecom, will lead to new opportunities to run forecasting models with incredible resolution and speed; the weather companies on their side will be working on applying AI and machine learning for providing advanced forecast services to the meet the growing demand trends of the weather and climate services market (WCSM).

The success of the enterprise as a whole will depend on a clear understanding of the inherent interdependence of the three sectors and creating such forms of dialogue and collaboration that would overcome eventual competition and frictions. WMO will play an important role in setting appropriate policies on data sharing that provide for mutually beneficial outcomes. Private sector itself should also adopt an open approach to sharing data and know-how. 

One particular development should be sought through the WMO Commission of Infrastructure which provides an intergovernmental possibility for coordination of plan and investments. WMO should explore the organizational models that ensure more synthetic integration of the private sector into its work, in particular on technology development and standardization, for example, the ITU model in which private sector entities form a tier of ‘sectoral members’ to the organization.

Q15. Ethics: What to do and what to avoid in the relationship between the stakeholders? 

The current and projected growth of the demand for weather and climate services, the core of which are the weather forecasts and climate predictions, expands the opportunities for stakeholders to get their share in the market. Based on their different funding and business models, private and public sector may address these new opportunities from different angles, i.e., the private sector stakeholders will try to monetize those opportunities as soon as possible, while some public sector stakeholders may try to expand their portfolios. Some of the public agencies have already adopted a particle commercialization of their services, thus the profit driver is also present for them. There exist a potential for conflict of business interests and disruption that would have a negative effects not only to individual stakeholders involved but to the whole enterprise.  

While understanding that markets logic and self-regulating mechanisms will apply to the weather enterprise, the part of the enterprise that serves the public interests by providing public good type of services, should avoid such potential disruptions. The well understood global societal risks of weather extremes and climate change require a unified response from the enterprise as a whole. This necessitates the development of a culture and ethics relevant to the overall mission of the enterprise expressed in the WMO Vision 2030 – for a more resilient world.

A good starting point for cultivating such culture and ethic resulting in concrete actions is the WMO Geneva Declaration 2019: Building Community for Weather, Climate and Water Actions. This Declaration represents the high-level WMO policy for an inclusive enterprise contributing to the global sustainable development agenda. It promotes the UN Global Compact principles and will help build a responsible and committed multi-sector forecasting enterprise. The WMO role in this regard will include the promotion of ethical behavior and a recent accord together with the HMEI for the development of a joint ‘code of ethics’ should demonstrate its advantages.