Periodic Reporting for period 1 - ISSUL (Improving Subseasonal and Seasonal sUmmer forecast over southern Europe through machine Learning)
Berichtszeitraum: 2022-02-01 bis 2024-01-31
Short- and medium-range weather forecast systems show a predictability limit of about ten days after which the forecast degrades strongly. However, reliable predictions on longer timescales are needed to prepare and better protect citizens and any economic sector sensitive to weather and climate against the occurrence of extreme events. Therefore, in the recent years, the development of forecast model on subseasonal to seasonal (S2S) timescales has become the focus of an intense research work from the scientific community. However, despite the large number of research studies, S2S forecast models still show a limited skill in summer over Europe. This is especially the case for southern Europe, that has generally received less attention, even though it is highly vulnerable to high-impact summer heatwaves, and very sensitive to climate change. Therefore, this project aims at improving and better understanding summer S2S predictability of heatwaves in southern Europe. To do that, a fairly new approach in the field of weather and extremes prediction is employed. Specifically, the S2S forecast model is entirely based on two machine learning algorithms: an optimization algorithm, which aims to select optimal predictors from a pool of candidate drivers, and a regression algorithm (linear or nonlinear) to predict the heatwave occurrence.
The overall objectives of this project are:
1. The identification and evaluation of the optimal predictors of heatwave frequency and strength on S2S timescales for the three sub-regions
2. The prediction of the heatwave frequency and strength on S2S timescales over the three areas
3. The evaluation of the S2S forecast system.
The main conclusions of the action are the following:
1. The machine learning forecast model based on the coupling of an optimization algorithm with a regression algorithm skilfully predicts regional monthly mean temperature conditions over southern Europe with one month lead time.
2. Regional monthly heatwave intensity in southern Europe is more difficult to predict one month in advance than the mean temperature conditions.
3. Dynamical connections are evidenced between recurrent predictors and the targets, which give confidence in the model architecture.
The overall objectives of this project are:
1. The identification and evaluation of the optimal predictors of heatwave frequency and strength on S2S timescales for the three sub-regions
2. The prediction of the heatwave frequency and strength on S2S timescales over the three areas
3. The evaluation of the S2S forecast system.
The main conclusions of the action are the following:
1. The machine learning forecast model based on the coupling of an optimization algorithm with a regression algorithm skilfully predicts regional monthly mean temperature conditions over southern Europe with one month lead time.
2. Regional monthly heatwave intensity in southern Europe is more difficult to predict one month in advance than the mean temperature conditions.
3. Dynamical connections are evidenced between recurrent predictors and the targets, which give confidence in the model architecture.
Summary of the different tasks performed in ISSUL:
1. Two targets of different complexity (see black curves in the attached figure), which describe regional summer temperature conditions in Iberia, were computed. This was done to test the sensitivity of the machine learning based forecast model to the complexity of the targets.
2. A pool of predictors containing 121 potential drivers was created. This pool of predictor is composed of 120 PCs (20 leading PCs of the 6 predictor variables) plus one more predictor: the day of the summer that we aim to predict. The predictor variables are: the sea surface temperature (SST), the 2m-temperature over Europe (T2M), the soil moisture (SM) over Europe, the outgoing-longwave radiation (OLR), the sea ice cover (SIC) in the Northern Hemisphere and the snow cover (SC) in the Northern Hemisphere.
3. The optimization algorithm was coupled to a regression algorithm. This machine learning model was trained then trained using the training data. Finally, a prediction of the test set part of the target was performed. These tasks are part of Objective 2
4. Once a satisfying skill was obtained for the targets, the dynamical link between the set of optimal predictors selected by the optimization algorithm and the target was analysed. This task is part of Objective 1.
Main results achieved in this project:
1. A machine learning prediction model based on the coupling of an optimization algorithm with a regression algorithm skilfully predicts regional monthly mean temperature conditions over southern Europe with one month lead time. This is the main scientific result of the project and its main technological achievement. This work further demonstrated that a machine learning prediction model of Iberian monthly 2m maximum temperature outperforms a simple statistical model based on the climatology for a lead time of at least one month (see figure).
2. Regional monthly heatwave intensity in southern Europe is more difficult to predict one month in advance than the mean temperature conditions. The model did not achieve a skilful prediction for the regional heatwave intensity index. Although the result can be considered negative, it also illustrates that extreme events are more difficult to predict. This outcome was somewhat anticipated, given that extreme events are inherently rare, resulting in an unbalanced training set (fewer days with a heatwave than non-heatwave days) to train the model.
3. Confidence in the model: dynamical connections between recurrent predictors and the target. ISSUL demonstrates that there are recurrent sets of optimal predictors (i.e. preferred combinations of skilful predictors’ states), and hence windows of opportunity to further improve the predictability of summer high temperatures and extremes. Second, the most commonly selected predictors have impacts on regional temperature in the next month, indicating that the model is not an artefact and that the drivers have true predictive signals. More specifically, the calendar day and the first PC of the SIC (or SST) allows the model to learn the climatology and trend in the data. The fourth PC of the OLR appears to be associated with cooler 2m maximum temperatures via a wave train developing in the tropical Atlantic during its positive phase and a weaker anomaly dipole during its negative phase related to summer NAO-like negative phase. All these results contribute to uncover new predictors of monthly 2m maximum temperatures over southern Europe and to understand the sources of its skill, thus enhancing trust in our model's performance.
The results of the project were disseminated in three presentations in international conferences and one scientific paper (in preparation). No website has been developed for the project.
1. Two targets of different complexity (see black curves in the attached figure), which describe regional summer temperature conditions in Iberia, were computed. This was done to test the sensitivity of the machine learning based forecast model to the complexity of the targets.
2. A pool of predictors containing 121 potential drivers was created. This pool of predictor is composed of 120 PCs (20 leading PCs of the 6 predictor variables) plus one more predictor: the day of the summer that we aim to predict. The predictor variables are: the sea surface temperature (SST), the 2m-temperature over Europe (T2M), the soil moisture (SM) over Europe, the outgoing-longwave radiation (OLR), the sea ice cover (SIC) in the Northern Hemisphere and the snow cover (SC) in the Northern Hemisphere.
3. The optimization algorithm was coupled to a regression algorithm. This machine learning model was trained then trained using the training data. Finally, a prediction of the test set part of the target was performed. These tasks are part of Objective 2
4. Once a satisfying skill was obtained for the targets, the dynamical link between the set of optimal predictors selected by the optimization algorithm and the target was analysed. This task is part of Objective 1.
Main results achieved in this project:
1. A machine learning prediction model based on the coupling of an optimization algorithm with a regression algorithm skilfully predicts regional monthly mean temperature conditions over southern Europe with one month lead time. This is the main scientific result of the project and its main technological achievement. This work further demonstrated that a machine learning prediction model of Iberian monthly 2m maximum temperature outperforms a simple statistical model based on the climatology for a lead time of at least one month (see figure).
2. Regional monthly heatwave intensity in southern Europe is more difficult to predict one month in advance than the mean temperature conditions. The model did not achieve a skilful prediction for the regional heatwave intensity index. Although the result can be considered negative, it also illustrates that extreme events are more difficult to predict. This outcome was somewhat anticipated, given that extreme events are inherently rare, resulting in an unbalanced training set (fewer days with a heatwave than non-heatwave days) to train the model.
3. Confidence in the model: dynamical connections between recurrent predictors and the target. ISSUL demonstrates that there are recurrent sets of optimal predictors (i.e. preferred combinations of skilful predictors’ states), and hence windows of opportunity to further improve the predictability of summer high temperatures and extremes. Second, the most commonly selected predictors have impacts on regional temperature in the next month, indicating that the model is not an artefact and that the drivers have true predictive signals. More specifically, the calendar day and the first PC of the SIC (or SST) allows the model to learn the climatology and trend in the data. The fourth PC of the OLR appears to be associated with cooler 2m maximum temperatures via a wave train developing in the tropical Atlantic during its positive phase and a weaker anomaly dipole during its negative phase related to summer NAO-like negative phase. All these results contribute to uncover new predictors of monthly 2m maximum temperatures over southern Europe and to understand the sources of its skill, thus enhancing trust in our model's performance.
The results of the project were disseminated in three presentations in international conferences and one scientific paper (in preparation). No website has been developed for the project.
The primary objective of the ISSUL project was to improve the prediction of heatwaves in southern Europe on S2S scales, thereby enhancing readiness for extreme summer conditions among citizens. This topic holds significant value across multiple societal domains including health, agriculture, economy, and water resources, as well as for climate services.
The methodology used in the project ISSUL was the most innovative aspect of the project. At the time of the submission of the proposal, the use of shallow learning for S2S prediction of extremes was a fairly new approach. Therefore, the results of the project ISSUL are also of great interest for the scientific community as they confirm the potential skill of machine learning algorithms for S2S predictions. Indeed, the machine learning based forecast model was able to predict a simple target representing regional temperatures in southern Europe. In addition, the project demonstrated the reliability of its predictions as it consistently identified the same optimal sets of predictors with tangible physical influences on 2m maximum temperature.
Finally, the ISSUL project contributes to the “Research and innovation” objective of the European Commission through innovation activities on the use of machine learning to tackle the challenge of summer temperature predictions on S2S timescales, which is in the forefront of weather and climate prediction research.
The methodology used in the project ISSUL was the most innovative aspect of the project. At the time of the submission of the proposal, the use of shallow learning for S2S prediction of extremes was a fairly new approach. Therefore, the results of the project ISSUL are also of great interest for the scientific community as they confirm the potential skill of machine learning algorithms for S2S predictions. Indeed, the machine learning based forecast model was able to predict a simple target representing regional temperatures in southern Europe. In addition, the project demonstrated the reliability of its predictions as it consistently identified the same optimal sets of predictors with tangible physical influences on 2m maximum temperature.
Finally, the ISSUL project contributes to the “Research and innovation” objective of the European Commission through innovation activities on the use of machine learning to tackle the challenge of summer temperature predictions on S2S timescales, which is in the forefront of weather and climate prediction research.