This study investigates the sensitivity of the GNSSseg segmentation method to change in: GNSS data processing method, length of time series (17 to 25 years), auxiliary data used in the integrated water vapor (IWV) conversion, and reference time series used in the segmentation (ERA-Interim versus ERA5). Two GNSS data sets (IGS repro1 and CODE REPRO2015), representative of the first and second IGS reprocessing, were compared. Significant differences were found in the number and positions of detected change-points due to different a priori ZHD models, antenna/radome calibrations, and mapping functions. The more recent models used in the CODE solution have reduced noise and allow the segmentation to detect smaller offsets. Similarly, the more recent reanalysis ERA5 has reduced representativeness errors, improved quality compared to ERA-Interim, and achieves higher sensitivity of the segmentation. Only 45–50% of the detected change-points are similar between the two GNSS data sets or between the two reanalyses, compared to 70–80% when the length of the time series or the auxiliary data are changed. About 35% of the change-points are validated with respect to metadata. The uncertainty in the homogenized trends is estimated to be around 0.01–0.02 kg m${}^{-2}$ year${}^{-1}$.

Background: Climate mitigation policies just with air quality improvements can deliver substantial co-benefits through reductions in greenhouse gas emissions and air pollutant concentrations. Mitigation strategies range from sufficiency to technology-driven transitions, yet linked co-benefits remain poorly understood. This study evaluates PM _2.5 and NO ₂ exposure changes and associated co-benefits using France’s energy-transition scenarios. Methods: Emission projections were incorporated to the CHIMERE chemistry-transport model to estimate PM _2.5 and NO ₂ concentrations for 2030 and 2050. Health impacts were assessed using disease-specific cessation-lag assumptions relative to 2019, covering premature mortality, morbidity, DALYs, and economic benefits across nine outcomes (hypertension, lung cancer, ischaemic heart disease, stroke, COPD, type-2 diabetes, acute lower respiratory infections, and asthma in children and adults). Findings: Population exposure in continental France is projected to decline by 23% and 45% for PM _2.5 and NO ₂ by 2030, rising to 40% and 70% by 2050. Health gains are substantial and broadly consistent across all four scenarios, with modest differences between sufficiency-oriented and technology-driven pathways. Under delayed-impact assumptions, which accounts for the latency between exposure reduction and health response, avoided premature deaths reached 8,800-9,300 for PM _2.5 and 11,000-12,800 for NO ₂ in 2030, rising to 21,300-22,100 and 24,500-26,200 by 2050. Avoided morbidity cases grew from 51,100–55,200 in 2030 to 84,000-87,800 by 2050, with total DALYs averted increasing from 278,000-310,000 to 427,000-450,000 over the same period. Economic benefits scaled accordingly, direct medical cost savings reached €1.0-1.1 billion/year by 2050, with intangible cost savings of €41-43 billion with PM _2.5 and €36-39 billion with NO ₂ reductions. Conclusion: Net-zero transition delivers substantial, progressive health and economic co-benefits that are robust across diverse policy pathways. Rather than sectoral composition, commitment to decarbonisation itself drives air quality improvements and population health gains. These findings support integrating co-benefits into climate policy frameworks to strengthen the evidence base for ambitious mitigation action. Highlights: Net-zero policies can deliver large air-quality co-benefits; here, chemistry-transport modeling coupled with emission projections quantifies these benefits across four contrasting French decarbonisation pathways for 2030 and 2050. Health and economic co-benefits are large and robust across all four contrasting scenarios, with modest differences between sufficiency-oriented and technology-driven pathways. By 2050, PM _2.5 and NO ₂ concentrations decline by 38-40% and 70-75%, averting 20,000-22,000 and 24,500-26,200 attributable premature deaths annually. Avoided morbidity reaches 84,000-87,800 cases/year and DALYs averted reach 427,000-450,000/year in 2050, growing substantially from 2030 estimates. Monetization of health impacts results in, direct medical savings of €1 billion/year in 2050 and intangible cost savings of €43 billion (PM _2.5 ) and €38 billion (NO ₂ ) per year.

Water vapor plays a key role in the Earth's climate as a dominant greenhouse gas. It is also the most efficient actor of heat transfer from the surface to the atmosphere and from low to high latitudes which shapes the global atmospheric circulation and weather systems. Monitoring and understanding the spatial and temporal variability and changes of water vapor are thus of crucial importance.This work aims at computing decadal trends of total column Integrated Water Vapour (IWV) from a global network of ground-based GNSS observations. Although GNSS observations are available with high accuracy in all weather conditions, it has been shown that, over long periods of time, changes in instrumentation, in station location and environment, and in processing methods can introduce spurious shifts in the IWV time series and bias trend estimates. Homogenization is a crucial step to detect and correct such non-climatic signals.We have developed a relative homogenization method which involves three steps.Segmentation. First, change-points are detected from the difference series (GNSS – reference) with the help of the GNSSseg segmentation package (Quarello et al., 2022). The method uses a difference series in order to cancel out the common climatic variations. It also accounts for changes in the variance on fixed intervals (monthly) and a periodic bias (annual) due to representativeness differences between GNSS and the reference (in our case the ERA5 reanalysis). Because the change-points detected in the difference series could be either due to GNSS or to the reference (ERA5), the next step is the attribution. Attribution. Second, the detected change-points are attributed to either GNSS or to the reference (ERA5) using a statistical test based on linear regression and a predictive rule based on the Random Forest learning algorithm (Nguyen et al., 2023). This step requires additional neighbors stations (at least one). Correction. The last step is the correction. Here the initial GNSS series is corrected only for the shifts which are attributed to the GNSS in the second step. We will present results of the homogenization procedure applied to a global network of GNSS stations and discuss the impact of homogenization on linear trend estimates for stations that have more than 20 years of observations.Quarello et al., 2022, https://doi.org/10.3390/rs14143379Nguyen et al., 2023, https://hal-obspm.ccsd.cnrs.fr/IGN-ENSG/hal-04014145v

Background: Climate change impacts are increasing, while fine particulate matter (PM2.5) concentration is declining. Most studies of climate change health impacts have ignored the urban heat island phenomenon.Methods: We developed a microsimulation-based health impact assessment approach relying on fine-scale climate models incorporating the urban heat island and on atmospheric pollution models based on planned policies. Impacts were expressed as changes in life expectancy and attributable mortality in an urban area (Grenoble, France), between the early and mid-century, and were stratified by social deprivation.Findings: Climate change resulted in a 2-month (95\% CI, 2–4) decline in life expectancy by 2050 compared to 2016. The fraction of deaths above age 30 attributable to non-optimal temperatures increased from 8.1 to 9.9%. This impact was stronger in socially deprived areas. It was underestimated by one third when the urban heat island was ignored. The regional PM2.5 mitigation plan was expected to reduce the mortality fraction attributable to PM2.5 down to 6% by 2050, yielding a 7-month (95\% CI, 5–9) life expectancy gain. The combined impacts of temperature and PM2.5 changes corresponded to a larger life expectancy gain in the least deprived neighborhoods (average gain, 5-month; 95\% CI, 2–7). Interpretation: By 2050, non-optimal temperatures may become a larger driver of mortality than PM2.5 in this urban area; climate change will partly offset the health gains expected from air pollution mitigation strategies and will more strongly affect the most deprived populations in this urban area.

In recent years, the detection and correction of the non-natural irregularities in the long climatic records, so-called homogenization, has been studied. This work is motivated by the problem of identification of origins of the breakpoints in the segmentation of difference series (difference between a candidate series and a reference series). Several segmentation methods have been developed for the difference series, but many of them assume that the reference series is homogenous. However, the homogeneity of the reference series, in reality, is uncertain and unproven. In our study, we applied the segmentation method GNSSseg (Quarello et al., 2020) on the difference between the Integrated water vapour estimates of the CODE REPRO2015 GNSS data set and the ERA5 reanalysis. About 36.5% of change points can be validated from the GPS metadata, and the origins of the remaining 64.5% are questionable (Nguyen et al., 2021). The ambiguity can be leveraged when there is at least one nearby GPS station with respect to which the candidate series can be compared. The proposed method uses weighted t-tests combining the candidate GPS and ERA series and their homologues (denoted GPS' and ERA') from each nearby station. If sufficient consistency emerges from the six tests for all the nearby stations, a decision can be made whether the breakpoint detected in the candidate GPS-ERA series is due to GPS or, alternatively, to ERA. For each quadruplet (GPS, ERA, GPS', ERA'), six t-tests are performed, and the outcomes are combined. In a set of 81 globally distributed GNSS time series spanning more than 25 years, 56 series have at least one nearby station, where 171 breakpoints are detected in segmentation, in which 136 breakpoints are attributed to the GPS. Among those, 94 breakpoints have consistent results between all the nearby stations. GPS-related breakpoints are used for the correction of the mean shift in the difference series. The impact of the breakpoint correction on the GNSS IWV trend estimates is then evaluated. Quarello A, Bock O, & Lebarbier E. (2020). A new segmentation method for the homogenisation of GNSS-derived IWV time-series. arXiv: Methodology.Nguyen KN, Quarello A, Bock O, Lebarbier E. Sensitivity of Change-Point Detection and Trend Estimates to GNSS IWV Time Series Properties. Atmosphere. 2021; 12(9):1102. https://doi.org/10.3390/atmos12091102

Homogénéisation et estimation des tendances des observations GNSS de la vapeur d'eau et des réanalyses atmosphériques : application à l'analyse du changement climatique: La vapeur d'eau joue un rôle important dans le bilan énergétique et le cycle hydrologique de la Terre. Avec le réchauffement climatique, la vapeur d'eau dans l'atmosphère tend à augmenter, mais la répartition géographique des tendances de vapeur d'eau dans les simulations de modèles climatiques et les observations reste incertaine. Les estimations de la vapeur d'eau intégrée (IWV) fournies par les systèmes mondiaux de navigation par satellites (GNSS) constituent une nouvelle source de données qui peut servir à vérifier les tendances simulées et observées de vapeur d'eau. Des études antérieures ont fait état d'inhomogénéités dans les données IWV du GNSS dues à des changements d'instrumentation des stations et de méthodes de traitement des données, et ont recommandé leur homogénéisation pour les applications climatiques. Cette étude utilise une méthode de segmentation statistique pour détecter les changements abrupts dans la moyenne (ou ruptures) des séries de différences IWV entre les données GNSS et celles de réanalyse. Les séries de différences sont utilisées pour atténuer la forte variabilité temporelle inhérente aux séries IWV, qui gêne la détection de petites ruptures sur les séries brutes. Cette étude utilise le package R GNSSseg développé par Quarello (2020). Le premier objectif de cette thèse est d'étudier la sensibilité de la segmentation et des estimations de tendances aux propriétés des données d'entrée. Ces dernières sont composées de deux ensembles de données GNSS (IGS repro1 et CODE REPRO2015) et de deux produits de réanalyse (ERA-Interim et ERA5). Il s'avère que seuls 45 à 50% des ruptures détectées sont similaires lors du changement des données d'entrée. La modification de la longueur des séries ou des données auxiliaires utilisées dans le traitement des données GNSS IWV a un impact plus faible, avec 70-80% de détections similaires. Les produits GNSS et de réanalyse plus récents présentent des différences moindres et permettent la détection de ruptures plus faibles. Les estimations des tendances sont sensibles au nombre et à la position des ruptures. Par conséquent, valider l'origine des ruptures avant d'estimer les tendances est crucial. Nous avons constaté que seuls 35 % sont expliqués par les métadonnées, le 65% restant vient de changements non documentés du GNSS ou de la réanalyse. Le deuxième objectif de cette thèse est de développer une méthode de classification automatique, opérée en post-traitement, pour distinguer les points de changement du GNSS et de la réanalyse. La méthode "d'attribution" proposée combine les données GNSS et de réanalyse d'une station principale avec des données similaires de stations voisines. Chaque paire station principale-station voisine comprend quatre séries de base à partir desquelles six séries de différences sont formées. Ensuite, pour chaque rupture détectée dans la série principale, un test de significativité est appliqué à la moyenne avant et après rupture, et une règle statistique prédictive est construite pour attribuer les points de changement testés au GNSS ou à la réanalyse. En présence de plusieurs stations voisines, une solution combinée est proposée. L'aspect original de la méthode d'attribution développée est l'utilisation d'une méthode d'inférence généralisée des moindres carrés, tenant compte de l'hétéroscédasticité et de l'autocorrélation dans les données, et d'un classificateur d'apprentissage automatique. La principale nouveauté réside dans la manière dont la règle prédictive est construite, en utilisant les résultats des tests obtenus à partir des données réelles avec une procédure de rééchantillonnage. Les résultats de la prédiction dépendent de la stratégie de rééchantillonnage (équilibrés ou déséquilibrés) et du niveau de significativité du test. Lorsque la méthode est appliquée aux données GNSS CODE REPRO2015 et à la réanalyse ERA5, 62% des ruptures testées sont attribuées au GNSS, 19% à la réanalyse et 10% à des changements communs aux deux.