Periodic Reporting for period 2 - CAFE (Climate Advanced Forecasting of sub-seasonal Extremes)
Reporting period: 2021-03-01 to 2023-02-28
Climate extremes such as heat waves or tropical storms have huge social and economic impact, and they are increasing due to climate change.
Reliable forecasting of such extreme events at the sub-seasonal time scale (from 10 days to 3 months) is of great importance, allowing for early warnings and adequate mitigation strategies, reducing damage and saving human lives.
The World Meteorological Organization (WMO), the World Weather Research Programme and World Climate Research Programme acknowledge the importance of sub-seasonal forecasts by running the Sub-seasonal to Seasonal Prediction Project.
Delivering reliable and timely information on extremes is an enormous challenge due to the complexity of the atmosphere-ocean coupled system, which requires an interdisciplinary effort of meteorologists, climate scientists, mathematicians, statisticians, and non-linear physicists.
Advancing the understanding of phenomena (Madden-Julian Oscillation, planetary waves, and atmospheric blocking) which are believed to act as sources of predictability of extremes at this scale is fundamental, as well as translating this knowledge into tools for prediction.
The ultimate goal of the project "CAFE" (Climate Advanced Forecasting of sub-seasonal Extremes) is to train a new generation of interdisciplinary scientists to move forward this area of knowledge, in order to advance three scientific objectives, during the project and beyond:
1. Understanding the phenomena behind the sub-seasonal variability (blockings, Madden-Julian Oscillation, Rossby Wave packets, etc.).
2. Improving the detection and analysis of weather extremes and their possible increase with global warming.
3. Developing a diverse range of tools to improve forecasting of extremes at the sub-seasonal scale.
Reliable forecasting of such extreme events at the sub-seasonal time scale (from 10 days to 3 months) is of great importance, allowing for early warnings and adequate mitigation strategies, reducing damage and saving human lives.
The World Meteorological Organization (WMO), the World Weather Research Programme and World Climate Research Programme acknowledge the importance of sub-seasonal forecasts by running the Sub-seasonal to Seasonal Prediction Project.
Delivering reliable and timely information on extremes is an enormous challenge due to the complexity of the atmosphere-ocean coupled system, which requires an interdisciplinary effort of meteorologists, climate scientists, mathematicians, statisticians, and non-linear physicists.
Advancing the understanding of phenomena (Madden-Julian Oscillation, planetary waves, and atmospheric blocking) which are believed to act as sources of predictability of extremes at this scale is fundamental, as well as translating this knowledge into tools for prediction.
The ultimate goal of the project "CAFE" (Climate Advanced Forecasting of sub-seasonal Extremes) is to train a new generation of interdisciplinary scientists to move forward this area of knowledge, in order to advance three scientific objectives, during the project and beyond:
1. Understanding the phenomena behind the sub-seasonal variability (blockings, Madden-Julian Oscillation, Rossby Wave packets, etc.).
2. Improving the detection and analysis of weather extremes and their possible increase with global warming.
3. Developing a diverse range of tools to improve forecasting of extremes at the sub-seasonal scale.
CAFE most significant results in the period 2019-2021, organized per scientific objective, are the following:
-Understanding the relation between Rossby Wave Packets (RWPs) and the large-scale environment through statistics:
Influence of ENSO and the Southern Annular Mode (SAM) on RWPs in the Southern-Hemisphere extratropics during austral summer, in particular for long-lived RWPs.
which opens an opportunity to better forecast extreme events during these conditions.
-Statistical characetrization of the Madden Julian (MJ) Oscillation events:
For size and duration of MJ events, a power law has been found to describe both the correlations and the tails of the distributions, which could affect the limits of predictability of the phenomenon.
-Development of diagnosis tools for identification and tracking of the MJ oscillation, blocking, and some other oceanic structures:
Coherent Lagrangian structures have been used to identify circulation patterns during atmospheric blocking. Harmonic closeness centrality identifies the location of the jet stream and its meanders.
-Classification of daily weather patterns (synoptic circulations) over Europe into eleven main categories. The evaluation of past and future trends in the seasonal frequencies of the weather types shows
that most of the trends arise from the SSP5-8.5 scenario in the summer. Results prove weather types associated to the most extreme rainfall.
-Correlation of El Niño-Southern Oscillation (ENSO) with extreme temperatures and extreme precipitation at a global scale have been performed, using diverse analysis methods (composition; correlation analysis, Granger causality).
-Circulation patterns are associated to heatwaves by means of an empirical orthogonal function analysis and a clustering algorithm.
Anti-cyclonic structures over certain European areas are found to occur together with cyclonic structures over neighbouring areas, pointing towards a blocking effect triggering the associated heatwaves.
-Development of machine-learning algorithms for the analysis of climatological phenomena, tested for MJO events, and obtaining similar skill to conventional methods.
-Assessment of the response of extreme weather events for different levels of stabilized global warming and comparison with their response to internal modes of climate variability such as ENSO or MJO.
climate networks linking different spatial points by means of correlations have been constructed, restricted to relative short windows (10 to 14 days). Several metrics characterizing the networks are revealing distinct signatures of tropical cyclones, in particular high clustering and low degree along the tropical-cyclone track.
-Development of a method to find maximally correlated structures between two spatial fields, with many potential applications. When applied to sea surface temperature and land precipitation trying to identify extremes, it yields very well-known climatological patterns as the successive resulting modes, among them the El Niño-Southern Oscillation (ENSO) and the Pacific Decadal Oscillation (PDO).
-Extreme precipitation events (EPEs) in the Mediterranean region have been studied, paying attention to atmospheric variability in the lower and middle troposphere, using an empirical orthogonal function approach plus a clustering algorithm.
Both spatial (winter and autumn being the seasons of highest EPE frequency for the eastern and western Mediterranean, respectively) and temporal clusterings ( 20 % of the EPEs occurring up to 1 week after a preceding EPE) have been found.
12. Systematic quantification of the predictability potential of a stochastic weather generator of analogues of atmospheric circulation:
Development of a stochastic weather generator (SWG), based on the use of analogues, using as a variable the geopotential height at 500 hPa, related to large-scale circulations.
A pilot test to forecast precipitation in several European cities has delivered results with satisfactory skill up to 20 days in advance.
-Understanding the relation between Rossby Wave Packets (RWPs) and the large-scale environment through statistics:
Influence of ENSO and the Southern Annular Mode (SAM) on RWPs in the Southern-Hemisphere extratropics during austral summer, in particular for long-lived RWPs.
which opens an opportunity to better forecast extreme events during these conditions.
-Statistical characetrization of the Madden Julian (MJ) Oscillation events:
For size and duration of MJ events, a power law has been found to describe both the correlations and the tails of the distributions, which could affect the limits of predictability of the phenomenon.
-Development of diagnosis tools for identification and tracking of the MJ oscillation, blocking, and some other oceanic structures:
Coherent Lagrangian structures have been used to identify circulation patterns during atmospheric blocking. Harmonic closeness centrality identifies the location of the jet stream and its meanders.
-Classification of daily weather patterns (synoptic circulations) over Europe into eleven main categories. The evaluation of past and future trends in the seasonal frequencies of the weather types shows
that most of the trends arise from the SSP5-8.5 scenario in the summer. Results prove weather types associated to the most extreme rainfall.
-Correlation of El Niño-Southern Oscillation (ENSO) with extreme temperatures and extreme precipitation at a global scale have been performed, using diverse analysis methods (composition; correlation analysis, Granger causality).
-Circulation patterns are associated to heatwaves by means of an empirical orthogonal function analysis and a clustering algorithm.
Anti-cyclonic structures over certain European areas are found to occur together with cyclonic structures over neighbouring areas, pointing towards a blocking effect triggering the associated heatwaves.
-Development of machine-learning algorithms for the analysis of climatological phenomena, tested for MJO events, and obtaining similar skill to conventional methods.
-Assessment of the response of extreme weather events for different levels of stabilized global warming and comparison with their response to internal modes of climate variability such as ENSO or MJO.
climate networks linking different spatial points by means of correlations have been constructed, restricted to relative short windows (10 to 14 days). Several metrics characterizing the networks are revealing distinct signatures of tropical cyclones, in particular high clustering and low degree along the tropical-cyclone track.
-Development of a method to find maximally correlated structures between two spatial fields, with many potential applications. When applied to sea surface temperature and land precipitation trying to identify extremes, it yields very well-known climatological patterns as the successive resulting modes, among them the El Niño-Southern Oscillation (ENSO) and the Pacific Decadal Oscillation (PDO).
-Extreme precipitation events (EPEs) in the Mediterranean region have been studied, paying attention to atmospheric variability in the lower and middle troposphere, using an empirical orthogonal function approach plus a clustering algorithm.
Both spatial (winter and autumn being the seasons of highest EPE frequency for the eastern and western Mediterranean, respectively) and temporal clusterings ( 20 % of the EPEs occurring up to 1 week after a preceding EPE) have been found.
12. Systematic quantification of the predictability potential of a stochastic weather generator of analogues of atmospheric circulation:
Development of a stochastic weather generator (SWG), based on the use of analogues, using as a variable the geopotential height at 500 hPa, related to large-scale circulations.
A pilot test to forecast precipitation in several European cities has delivered results with satisfactory skill up to 20 days in advance.
The progress beyond the state of the art achieved by CAFE (ongoing and expected by the end of project) includes:
- Determination of the maximum amount of predictability due to the existence of Rossby wave packets.
- New techniques to identify and track the development of atmospheric and oceanic coherent structures.
- Correlations between damage-relevant regional distribution properties of meteorological variables and El Niño – Southern Oscillation (ENSO).
and linking hazard forecast to expected losses and loss variabilities in relation to ENSO.
- Evaluation of the current ability of sub-seasonal forecasts in predicting the risk of heat waves.
- Possible tool to extract risk of heatwaves from ECMWF forecasts
- Correlations between precipitation on land and diverse oceanic variables measured in advance.
- Adaptation of stochastic weather generator to use ECMWF forecasts in order to obtain downscaled forecasts.
- Determination of the maximum amount of predictability due to the existence of Rossby wave packets.
- New techniques to identify and track the development of atmospheric and oceanic coherent structures.
- Correlations between damage-relevant regional distribution properties of meteorological variables and El Niño – Southern Oscillation (ENSO).
and linking hazard forecast to expected losses and loss variabilities in relation to ENSO.
- Evaluation of the current ability of sub-seasonal forecasts in predicting the risk of heat waves.
- Possible tool to extract risk of heatwaves from ECMWF forecasts
- Correlations between precipitation on land and diverse oceanic variables measured in advance.
- Adaptation of stochastic weather generator to use ECMWF forecasts in order to obtain downscaled forecasts.