White Paper One Contributor: Thomas Stocker

Q1. How good are the forecasts today and what are the unresolved challenges in the weather/climate forecasting? (e.g., before looking 10 years ahead, look 10 years back – _what has changed) 

This contribution comes from my perspective of the long time scale relevant for climate and climate services and which uses global models to provide information on future climate conditions and impacts. In the past 20 years I have witnessed the increasing resolution of these models such that today global climate models have a very high degree of realism in certain large-scale variables. However, there are still substantial limitations regarding the representation of modes of variability such as ENSO, AMO, NAO and their different flavors as suggested by observations. These limitations still reduce substantially the value of such models for impact studies and decision making at regional and national level. The major challenges are (list not complete):

1. Hydrological cycle and regional patterns,
2. Interannual and decadal modes of variability,
3. Extreme events and their statistical distribution in time and space,
4. Regime changes and tipping points in atmospheric and ocean circulations, sea ice distribution, 
5. Impacts on terrestrial and marine biological systems.

These are all key to provide information that goes beyond the "classical" weather forecast and forms the basis for information on possible futures in all regions of the world.

An extension of the classical notion of "seamlessness" must achieved in the next decade of model development. Model improvements that are developed, tested and implemented in NWP models must be effectively "transferred" to the global models that are used for projection and that continue to increase their spatial resolution. WMO has to play an important role in accelerating this by encouraging and facilitating work at the boundary of NWP and global model development. 

It is evident that in a globalized world, global model projection must be a global endeavour. However, there is still a strong spirit of "National Weather Services" that remains a stumbling block in translating climate and impact projections from a global scale to the scale that matters for people. Unfortunately, "Climate Services" also has a taken the road of being distinctly national, and this will prove, before long, its major shortcoming. National climate services have an inherent and potentially dangerous limitation that is built into their design. The scientific foundation of projections does not derive from the best possible and most scientifically advanced source accessible to all, in particular the vulnerable countries and regions. However, delivered reliably, robustly, and usefully can only be what has been developed, tested, and standardized before. There is currently neither a process nor an institution in place that would be able to ensure this in the coming 10 years. WMO could take leadership of such an initiative.

Considering the five challenges outlined above, a strong case must be made for an international climate computation facility that disposes of the most powerful computers, is able to produce exascale ensembles, and ensures the implementation of the most up-to-date scientific knowledge into the global model. Production is intimately coupled with development: ECMWF may serve as a blueprint for such a global R&D centre. Recognizing the scale and the urgency of the problem, in view of the financial implications of wrong or uninformed decisions around climate change adaptation and mitigation, the ambition must be to build a global climate change R&D institute similar to CERN in particle physics or ESA and NASA in space science. Close collaboration with developers of computation and storage infrastructure is a must. Data and information must be open access, and a clear emphasis must be given to the most vulnerable and exposed regions and countries.

Up to now this globality was reflected in the joint, and very successful, effort of successive model intercomparisons at which an increasing number of institutions worldwide is participating. The scientific substance for these intercomparison projects is provided on a voluntary basis by research centers scattered around the globe. The collaboration is excellent, but it does lack the consolidated financial and structural basis for a long-term effort. Without diminishing the important achievements of such intercomparisons, and the most valuable information that they have produced up to now, this activity is minute in comparison to the required effort called for here. 

There are several developments converging that strongly suggest that this activity be enabled under a global umbrella with the according global funds:

1. Making available to world's most effective computing power for exascale simulations requires substantial funds that are beyond the capabilities of an individual nation;
2. Consolidation of intellectual and scientific capacity for a global scale problem requires a global institution;
3. Providing information for the more exposed and most vulnerable represents a global responsibility.

This would form the basis for a massive scaling up of the development and computational investment which is direly needed to overcome the challenges outline above. 

Q2. Are we going to eliminate completely the “black swans”, i.e., a future without surprises from extreme weather? 

There is still a large amount of missing physical understanding and relevant observations for such surprises in the physical and biological Earth system. Knowledge on instabilities, irreversible changes, very extreme events and compound events, in the atmosphere, ocean, cryosphere, and terrestrial and marine ecosystems, is only emerging. A similar effort of scientific acceleration as that triggered by the successive IPCC assessments since 1990 should be initiated by an IPCC Special Report on Tipping Points to be prepared in the 7th assessment cycle of the IPCC that will start in 2023 and likely complete by 2030. This would provide the utterly needed consensus on what we know, and more importantly, on what we do not know about low probability-high risk events.

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

It has been evident that in situ, high quality observation systems have decreased in number over the past 20 years in most regions of the world. The most vulnerable areas such as developing countries require well equipped and maintained stations that serve as early warning systems, and as reference stations for climate and weather simulations. Precipitation is particularly difficult and reliable observations are not available in the required spatial density. In contrast to temperature, high time and space resolution is a must if model are to be checked critically against observations. Limitations of model evaluation against observations have become evident already in the Working Group I contribution to the IPCC's 5th Assessment Report in 2013. The situation has become more dramatic as the resolution of the climate models increases, and the discrepancy between "model information" and real measurements continues to grow.

Progress in the above areas covered by Q1 and Q9 hinges on a global observations data base with open data access, quality monitoring, and high spatial resolution.