Anna J. Miller
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Anna J. Miller has a total of 117 co-authors. Here are the top co-authors by number of shared publications:
- Jan Henneberger
Institute for Atmospheric and Climate Science, ETH Zurich, Zurich, Switzerland
45 shared publications - Christopher Fuchs
Institute for Atmospheric and Climate Science, ETH Zurich, Zurich, Switzerland
35 shared publications - Nadja Omanovic
Eidgenössische Technische Hochschule Zürich
35 shared publications - Huiying Zhang
Institute for Atmospheric and Climate Science, ETH Zurich, Zurich, Switzerland
34 shared publications Fabiola Ramelli
32 shared publications- Ulrike Lohmann
Institute for Atmospheric and Climate Science, ETH Zurich, Zurich, Switzerland
31 shared publications
Latest Publications
- Open Access
- 14 October 2026
- Published by American Chemical Society (ACS) in Environmental Science & Technology
- Vol. 60 (42) , 29717-29720
Fluoropolymers across Their Life Cycle: Why “Emissions Reduction” Is Not “Safe”
- 14 October 2026
- Published by American Chemical Society (ACS) in Environmental Science & Technology
- Vol. 60 (42) , 29717-29720
Open AccessFluoropolymers across Their Life Cycle: Why “Emissions Reduction” Is Not “Safe”
- Open Access
- 29 August 2026
- Published by American Chemical Society (ACS) in Environmental Science & Technology
- Vol. 60 (36) , 25947-25948
Correction to “Global Inventory of Fluoropolymer Production Plants and Their Associated PFAS Environmental Contamination”
- 29 August 2026
- Published by American Chemical Society (ACS) in Environmental Science & Technology
- Vol. 60 (36) , 25947-25948
Open AccessCorrection to “Global Inventory of Fluoropolymer Production Plants and Their Associated PFAS Environmental Contamination”
- Open Access
- 28 April 2026
- Published by American Chemical Society (ACS) in ChemRxiv
Inside the Battery Box: A Cross‑Technology Assessment of PFAS Uses, Functions, and Substitution Potential
Understanding where and why PFAS are used in batteries is essential for evaluating their substitution potential, yet detailed information on functions across battery technologies is scarce. We analyzed the more than 5 600 submissions to the Annex XV public consultation on the proposed universal PFAS restriction by the European Union and extracted information describing PFAS uses, associated substances, and stated functions in batteries. A total of 341 records were compiled into a database using our structured Battery Use & Function Framework with support of the Functional Substitution Framework. Among these records, 78 contained sufficient information for qualifying as a potential PFAS use and when accounting for overlaps, these were organized into 36 potential PFAS uses. Each potential PFAS use was then evaluated using supporting evidence from previous use‑mapping studies, interviews with battery stakeholders, teardown-analysis studies, and other targeted academic literature. Across all lines of evidence, two PFAS uses were contradicted, indicating already large-scale availability of PFAS-free alternatives. 16 PFAS uses were confirmed, primarily in cathode components across several battery technologies and chemistries, for which PFAS-free alternatives for all applications were identified in 10 PFAS uses and an additional 5 PFAS uses for some applications within use. Commercially available PFAS‑free alternatives were identified for Li-ion (lithium-ion), Na-ion (sodium‑ion) and lithium metal batteries. The findings indicate that few technical barriers remain for substituting PFASs in most battery components. This mapping provides a detailed foundation for future alternatives assessments, supports essential‑use evaluations, and highlights substitution opportunities and challenges relevant to monitoring and recycling. TOC Reports on PFAS uses in batteries were collected and weighed against supporting evidence. Feasible alternatives were assessed using market deployment as a proxy for technical viability, illustrating a newly developed cross‑sector assessment framework.
- 28 April 2026
- Published by American Chemical Society (ACS) in ChemRxiv
Open AccessInside the Battery Box: A Cross‑Technology Assessment of PFAS Uses, Functions, and Substitution Potential
Understanding where and why PFAS are used in batteries is essential for evaluating their substitution potential, yet detailed information on functions across battery technologies is scarce. We analyzed the more than 5 600 submissions to the Annex XV public consultation on the proposed universal PFAS restriction by the European Union and extracted information describing PFAS uses, associated substances, and stated functions in batteries. A total of 341 records were compiled into a database using our structured Battery Use & Function Framework with support of the Functional Substitution Framework. Among these records, 78 contained sufficient information for qualifying as a potential PFAS use and when accounting for overlaps, these were organized into 36 potential PFAS uses. Each potential PFAS use was then evaluated using supporting evidence from previous use‑mapping studies, interviews with battery stakeholders, teardown-analysis studies, and other targeted academic literature. Across all lines of evidence, two PFAS uses were contradicted, indicating already large-scale availability of PFAS-free alternatives. 16 PFAS uses were confirmed, primarily in cathode components across several battery technologies and chemistries, for which PFAS-free alternatives for all applications were identified in 10 PFAS uses and an additional 5 PFAS uses for some applications within use. Commercially available PFAS‑free alternatives were identified for Li-ion (lithium-ion), Na-ion (sodium‑ion) and lithium metal batteries. The findings indicate that few technical barriers remain for substituting PFASs in most battery components. This mapping provides a detailed foundation for future alternatives assessments, supports essential‑use evaluations, and highlights substitution opportunities and challenges relevant to monitoring and recycling. TOC Reports on PFAS uses in batteries were collected and weighed against supporting evidence. Feasible alternatives were assessed using market deployment as a proxy for technical viability, illustrating a newly developed cross‑sector assessment framework.
- Open Access
- 23 April 2026
- Published by American Chemical Society (ACS) in Environmental Science & Technology
- Vol. 60 (17) , 12708-12711
Allowing Unlimited PFAS Manufacturing Contradicts the Core Intention of the European Union’s PFAS Restriction Proposal
- 23 April 2026
- Published by American Chemical Society (ACS) in Environmental Science & Technology
- Vol. 60 (17) , 12708-12711
Open AccessAllowing Unlimited PFAS Manufacturing Contradicts the Core Intention of the European Union’s PFAS Restriction Proposal
- Open Access
- 14 March 2026
- Published by Copernicus GmbH
Unveiling In-Situ Ice Aggregation: Deep Learning, Causal Discovery, and Physics
Ice aggregation is a fundamental driver of cloud evolution and precipitation formation. However, quantifying its rate in natural environments remains challenging due to the difficulty of tracking particle history in a Lagrangian frame. To address this issue, we use a unique dataset from 21 targeted glaciogenic seeding experiments (CLOUDLAB, Henneberger et al., 2023) conducted in supercooled stratiform clouds ranging from -7.8 °C to -4.7 °C. This experimental design establishes a controlled initial state and advection time (5–10 minutes). Central to our methodology is IceDetectNet (Zhang et al., 2024), a deep learning architecture that applies in situ holographic imagery to detect and classify individual monomers within complex aggregates. Quantifying the number of collisions per aggregate at the monomer level allows us to reconstruct the initial ice crystal number concentration (ICNCt0) directly from downwind observations.To disentangle the microphysical and environmental drivers of aggregation, we implemented a comprehensive analytical workflow that integrated three distinct paradigms: data-driven causal inference, a theoretically derived physical equation, and machine learning regressors. These independent approaches converge on the conclusion that ICNCt0 parameter is governing aggregation, significantly outweighing the influence of temperature, turbulence, or aspect ratio. Our analysis reveals a significant departure from classical collection theory: the aggregation rate exhibits sub-quadratic power-law dependence on initial concentration (mean exponent 0.92; 95% confidence interval CI: 0.88–0.97), contrasting with the traditional quadratic scaling assumed in kinetic collection kernels. We hypothesize that this scaling involves aggregation among smaller crystals, where subsequent diffusional growth masks the boundaries between monomers, making early collisions difficult to detect. Furthermore, benchmarking eleven machine learning architectures against the physically derived formulation revealed a clear trade-off. While CatBoost's gradient boosting ensembles achieved higher statistical accuracy (R² = 0.87), the theoretical model showed greater robustness and generalizability in sensitivity testing. This multi-perspective framework uses a combination of experimental atmospheric physics and AI-driven interpretation to demonstrate how data-driven plasticity and physically-based stability complement each other. It provides a practical approach to understanding complex microphysical processes. Reference:Henneberger J, Ramelli F, Spirig R, et al. Seeding of supercooled low stratus clouds with a UAV to study microphysical ice processes: an introduction to the CLOUDLAB project[J]. Bulletin of the American Meteorological Society, 2023, 104(11): E1962-E1979. https://doi.org/10.1175/BAMS-D-22-0178.1Zhang H, Li X, Ramelli F, et al. IceDetectNet: a rotated object detection algorithm for classifying components of aggregated ice crystals with a multi-label classification scheme[J]. Atmospheric Measurement Techniques, 2024, 17(24): 7109-7128. https://doi.org/10.5194/amt-17-7109-2024
- 14 March 2026
- Published by Copernicus GmbH
Open AccessUnveiling In-Situ Ice Aggregation: Deep Learning, Causal Discovery, and Physics
Ice aggregation is a fundamental driver of cloud evolution and precipitation formation. However, quantifying its rate in natural environments remains challenging due to the difficulty of tracking particle history in a Lagrangian frame. To address this issue, we use a unique dataset from 21 targeted glaciogenic seeding experiments (CLOUDLAB, Henneberger et al., 2023) conducted in supercooled stratiform clouds ranging from -7.8 °C to -4.7 °C. This experimental design establishes a controlled initial state and advection time (5–10 minutes). Central to our methodology is IceDetectNet (Zhang et al., 2024), a deep learning architecture that applies in situ holographic imagery to detect and classify individual monomers within complex aggregates. Quantifying the number of collisions per aggregate at the monomer level allows us to reconstruct the initial ice crystal number concentration (ICNCt0) directly from downwind observations.To disentangle the microphysical and environmental drivers of aggregation, we implemented a comprehensive analytical workflow that integrated three distinct paradigms: data-driven causal inference, a theoretically derived physical equation, and machine learning regressors. These independent approaches converge on the conclusion that ICNCt0 parameter is governing aggregation, significantly outweighing the influence of temperature, turbulence, or aspect ratio. Our analysis reveals a significant departure from classical collection theory: the aggregation rate exhibits sub-quadratic power-law dependence on initial concentration (mean exponent 0.92; 95% confidence interval CI: 0.88–0.97), contrasting with the traditional quadratic scaling assumed in kinetic collection kernels. We hypothesize that this scaling involves aggregation among smaller crystals, where subsequent diffusional growth masks the boundaries between monomers, making early collisions difficult to detect. Furthermore, benchmarking eleven machine learning architectures against the physically derived formulation revealed a clear trade-off. While CatBoost's gradient boosting ensembles achieved higher statistical accuracy (R² = 0.87), the theoretical model showed greater robustness and generalizability in sensitivity testing. This multi-perspective framework uses a combination of experimental atmospheric physics and AI-driven interpretation to demonstrate how data-driven plasticity and physically-based stability complement each other. It provides a practical approach to understanding complex microphysical processes. Reference:Henneberger J, Ramelli F, Spirig R, et al. Seeding of supercooled low stratus clouds with a UAV to study microphysical ice processes: an introduction to the CLOUDLAB project[J]. Bulletin of the American Meteorological Society, 2023, 104(11): E1962-E1979. https://doi.org/10.1175/BAMS-D-22-0178.1Zhang H, Li X, Ramelli F, et al. IceDetectNet: a rotated object detection algorithm for classifying components of aggregated ice crystals with a multi-label classification scheme[J]. Atmospheric Measurement Techniques, 2024, 17(24): 7109-7128. https://doi.org/10.5194/amt-17-7109-2024
- Open Access
- 14 March 2026
- Published by Copernicus GmbH
A Shape-Aware Mass–Diameter Parameterization for Ice Crystals Constrained by Glaciogenic Seeding Experiments
Accurate estimates of ice crystal mass are essential for reducing uncertainties in cloud radiative forcing and precipitation forecasts. Ice crystal mass can be derived from imaging cloud probes using power-law mass–diameter (m-D) relationships. However, these often do not account for variability in crystal habit, leading to significant biases in ice water content (IWC) retrievals and complicating the comparison between in-situ observations and numerical models. To address this, we developed a shape-aware m-D parameterization by explicitly incorporating the aspect ratio (AR) into a power-law framework. The parametrization is fitted using data from the CLOUDLAB campaigns, which use glaciogenic seeding to induce ice formation in supercooled stratus clouds. This experimental setup allows for two total water content (TWC) conservation assumptions: (i) the temporal stability of the stratus clouds allows to use the TWC of the unseeded cloud as a reliable baseline for the seeded section., and (ii) the introduction of ice crystals via seeding does not alter the TWC, even as the Wegener–Bergeron–Findeisen process redistributes mass from the liquid to the ice phase. Using data from 20 seeding experiments, we optimized the m-D parameters by minimizing the difference between seeded and the baseline TWC using a loss function based on the Wasserstein distance to ensure that the probability distribution of the derived TWC match the observed variability of the background cloud. The resulting parameterization m = 0.0487 D2.045 / AR2 aligns well with existing m-D relationships but predicts lower ice crystal masses for high aspect ratios. When applied to the CLOUDLAB data, the formula successfully removes systematic overestimations in ice mass across various temperatures and growth stages. While riming and aggregation were only weakly present in our dataset, they did not lead to significant deviations. This study provides a shape-aware m-D formulation suitable for bulk microphysics schemes and demonstrates a robust, data-driven framework for constraining cloud parameters using field measurement from cloud seeding experiments.
- 14 March 2026
- Published by Copernicus GmbH
Open AccessA Shape-Aware Mass–Diameter Parameterization for Ice Crystals Constrained by Glaciogenic Seeding Experiments
Accurate estimates of ice crystal mass are essential for reducing uncertainties in cloud radiative forcing and precipitation forecasts. Ice crystal mass can be derived from imaging cloud probes using power-law mass–diameter (m-D) relationships. However, these often do not account for variability in crystal habit, leading to significant biases in ice water content (IWC) retrievals and complicating the comparison between in-situ observations and numerical models. To address this, we developed a shape-aware m-D parameterization by explicitly incorporating the aspect ratio (AR) into a power-law framework. The parametrization is fitted using data from the CLOUDLAB campaigns, which use glaciogenic seeding to induce ice formation in supercooled stratus clouds. This experimental setup allows for two total water content (TWC) conservation assumptions: (i) the temporal stability of the stratus clouds allows to use the TWC of the unseeded cloud as a reliable baseline for the seeded section., and (ii) the introduction of ice crystals via seeding does not alter the TWC, even as the Wegener–Bergeron–Findeisen process redistributes mass from the liquid to the ice phase. Using data from 20 seeding experiments, we optimized the m-D parameters by minimizing the difference between seeded and the baseline TWC using a loss function based on the Wasserstein distance to ensure that the probability distribution of the derived TWC match the observed variability of the background cloud. The resulting parameterization m = 0.0487 D2.045 / AR2 aligns well with existing m-D relationships but predicts lower ice crystal masses for high aspect ratios. When applied to the CLOUDLAB data, the formula successfully removes systematic overestimations in ice mass across various temperatures and growth stages. While riming and aggregation were only weakly present in our dataset, they did not lead to significant deviations. This study provides a shape-aware m-D formulation suitable for bulk microphysics schemes and demonstrates a robust, data-driven framework for constraining cloud parameters using field measurement from cloud seeding experiments.
- Open Access
- 13 March 2026
- Published by Copernicus GmbH
Ice-nucleating particle depletion in the wintertime boundary layer in the pre-Alpine region during stratus cloud conditions
The remote-sensing equipment of LACROS (Leipzig Aerosol and Cloud Remote Observations System) was installed in Eriswil, Switzerland during the winter campaign of 2023 / 2024. We utilize a big dataset of in situ (especially ice-nucleating particle (INP) sampler; low volume sampler) and remote-sensing (especially cloud radar and Raman lidar) equipment. In addition, INP measurements were available as well in Hohenpeißenberg, Germany and Melpitz, Germany.We evaluate the regional variability of the number concentration of INPs between the two pre-Alpine central-European sites of Eriswil and Hohenpeißenberg, supported by INP measurements from Melpitz during the winter months of 2024. The aim of the study is to spatially and temporally evaluate INP availability and removal within the planetary boundary layer (PBL) during Bise situations because reasons for the lack of ice and precipitation in the supercooled clouds observed over the Swiss Plateau remain unclear and may be caused by the lack of INPs. Target scenario of the study were situations when northeasterly winds (so-called Bise winds) prevailed and layers of stratus clouds formed at the top of the PBL at temperatures down to −10 °C. In these situations, it is expected that INPs are depleted along the transport path.We will present our main insights from our measurements:1) During the cold-Bise (cloud minimum temperatures as low as −10 °C) and warm-Bise (cloud minimum temperatures above 0 °C), almost no INP contrast was found between Hohenpeißenberg and Eriswil if both were within the PBL. Also, the INP concentration was overall found to be much lower during the cold-Bise than during the later warm-Bise situation.2) When the Hohenpeißenberg site was located in the free troposphere during the cold-Bise situation, INP concentrations were much higher compared to Eriswil (still within the PBL) but similar to cloud-free Melpitz. These observations led to the conclusion that during cold-Bise situations the INP reservoir within the PBL is depleted, likely by the presence of supercooled stratus. The inversion-capped wintertime PBL, especially during periods of widespread snow cover, is apparently not capable to replenish the INP reservoir from the free troposphere.3) INP observations of around 10^−3 L^−1 at Hohenpeißenberg, when this site was above the PBL were on a similar order as the ice crystal number concentrations (ICNC) observed during the same period at Eriswil. This supports the hypothesis that INPs are entrained from the free troposphere via turbulence and afterwards immediately removed as they interact with the Bise cloud layer, leading to reduced availability of INPs downwind. It must be noted that an ICNC concentration which is higher than the observed INP concentration can in principle also be a result of secondary ice formation processes. Nevertheless, secondary ice formation processes generally lead to orders of magnitudes of increase in ICNC, which was, besides occasional peaks in the ICNC, not observed in the average ICNC values during the investigated time periods.
- 13 March 2026
- Published by Copernicus GmbH
Open AccessIce-nucleating particle depletion in the wintertime boundary layer in the pre-Alpine region during stratus cloud conditions
The remote-sensing equipment of LACROS (Leipzig Aerosol and Cloud Remote Observations System) was installed in Eriswil, Switzerland during the winter campaign of 2023 / 2024. We utilize a big dataset of in situ (especially ice-nucleating particle (INP) sampler; low volume sampler) and remote-sensing (especially cloud radar and Raman lidar) equipment. In addition, INP measurements were available as well in Hohenpeißenberg, Germany and Melpitz, Germany.We evaluate the regional variability of the number concentration of INPs between the two pre-Alpine central-European sites of Eriswil and Hohenpeißenberg, supported by INP measurements from Melpitz during the winter months of 2024. The aim of the study is to spatially and temporally evaluate INP availability and removal within the planetary boundary layer (PBL) during Bise situations because reasons for the lack of ice and precipitation in the supercooled clouds observed over the Swiss Plateau remain unclear and may be caused by the lack of INPs. Target scenario of the study were situations when northeasterly winds (so-called Bise winds) prevailed and layers of stratus clouds formed at the top of the PBL at temperatures down to −10 °C. In these situations, it is expected that INPs are depleted along the transport path.We will present our main insights from our measurements:1) During the cold-Bise (cloud minimum temperatures as low as −10 °C) and warm-Bise (cloud minimum temperatures above 0 °C), almost no INP contrast was found between Hohenpeißenberg and Eriswil if both were within the PBL. Also, the INP concentration was overall found to be much lower during the cold-Bise than during the later warm-Bise situation.2) When the Hohenpeißenberg site was located in the free troposphere during the cold-Bise situation, INP concentrations were much higher compared to Eriswil (still within the PBL) but similar to cloud-free Melpitz. These observations led to the conclusion that during cold-Bise situations the INP reservoir within the PBL is depleted, likely by the presence of supercooled stratus. The inversion-capped wintertime PBL, especially during periods of widespread snow cover, is apparently not capable to replenish the INP reservoir from the free troposphere.3) INP observations of around 10^−3 L^−1 at Hohenpeißenberg, when this site was above the PBL were on a similar order as the ice crystal number concentrations (ICNC) observed during the same period at Eriswil. This supports the hypothesis that INPs are entrained from the free troposphere via turbulence and afterwards immediately removed as they interact with the Bise cloud layer, leading to reduced availability of INPs downwind. It must be noted that an ICNC concentration which is higher than the observed INP concentration can in principle also be a result of secondary ice formation processes. Nevertheless, secondary ice formation processes generally lead to orders of magnitudes of increase in ICNC, which was, besides occasional peaks in the ICNC, not observed in the average ICNC values during the investigated time periods.
- Open Access
- 3 March 2026
- Published by Copernicus GmbH in Atmospheric Chemistry and Physics
- Vol. 26 (4) , 3223-3236
Ice-nucleating particle depletion in the wintertime boundary layer in the pre-Alpine region during stratus cloud conditions
This study evaluates the regional variability of the number concentration of ice-nucleating particles (INPs) between the two pre-Alpine central-European sites of Eriswil, Switzerland, and Hohenpeißenberg, Germany, supported by INP measurements from Melpitz, Germany, during the winter months of 2024. The aim of the study is to spatially and temporally evaluate INP availability and removal within the planetary boundary layer (PBL) during Bise situations because reasons for the lack of ice and precipitation in the supercooled clouds observed over the Swiss Plateau remain unclear and may be caused by the lack of INPs. Target scenario of the study were situations when northeasterly winds (so-called Bise winds) prevailed and layers of stratus clouds formed at the top of the PBL at temperatures down to −10 °C. In these situations, it is expected that INPs are depleted along the transport path. The main insights from INP measurements were: first, during the cold-Bise (cloud minimum temperatures as low as −10 °C) and warm-Bise (cloud minimum temperatures above 0 °C), almost no INP contrast was found between Hohenpeißenberg and Eriswil if both were within the PBL. Also, the INP concentration was overall found to be much lower during the cold-Bise than during the later warm-Bise situation. Second, when the Hohenpeißenberg site was located in the free troposphere during the cold-Bise situation, INP concentrations were much higher compared to Eriswil (still within the PBL) but similar to cloud-free Melpitz. These observations led to the conclusion that during cold-Bise situations the INP reservoir within the PBL is depleted, likely by the presence of supercooled stratus. The inversion-capped wintertime PBL, especially during periods of widespread snow cover, is apparently not capable to replenish the INP reservoir from the free troposphere.
- 3 March 2026
- Published by Copernicus GmbH in Atmospheric Chemistry and Physics
- Vol. 26 (4) , 3223-3236
Open AccessIce-nucleating particle depletion in the wintertime boundary layer in the pre-Alpine region during stratus cloud conditions
This study evaluates the regional variability of the number concentration of ice-nucleating particles (INPs) between the two pre-Alpine central-European sites of Eriswil, Switzerland, and Hohenpeißenberg, Germany, supported by INP measurements from Melpitz, Germany, during the winter months of 2024. The aim of the study is to spatially and temporally evaluate INP availability and removal within the planetary boundary layer (PBL) during Bise situations because reasons for the lack of ice and precipitation in the supercooled clouds observed over the Swiss Plateau remain unclear and may be caused by the lack of INPs. Target scenario of the study were situations when northeasterly winds (so-called Bise winds) prevailed and layers of stratus clouds formed at the top of the PBL at temperatures down to −10 °C. In these situations, it is expected that INPs are depleted along the transport path. The main insights from INP measurements were: first, during the cold-Bise (cloud minimum temperatures as low as −10 °C) and warm-Bise (cloud minimum temperatures above 0 °C), almost no INP contrast was found between Hohenpeißenberg and Eriswil if both were within the PBL. Also, the INP concentration was overall found to be much lower during the cold-Bise than during the later warm-Bise situation. Second, when the Hohenpeißenberg site was located in the free troposphere during the cold-Bise situation, INP concentrations were much higher compared to Eriswil (still within the PBL) but similar to cloud-free Melpitz. These observations led to the conclusion that during cold-Bise situations the INP reservoir within the PBL is depleted, likely by the presence of supercooled stratus. The inversion-capped wintertime PBL, especially during periods of widespread snow cover, is apparently not capable to replenish the INP reservoir from the free troposphere.
- Open Access
- 28 January 2026
- Published by Copernicus GmbH in Atmospheric Chemistry and Physics
- Vol. 26 (2) , 1459-1481
Inferring the controlling factors of ice aggregation from targeted cloud seeding experiments
Ice aggregation in clouds plays a crucial role in cloud development and precipitation formation. Despite the significance of ice aggregation, direct in situ quantification of aggregation rates in natural clouds has been challenging due to the difficulty of tracking ice crystals. Here, we present in situ measurements of ice aggregation rates in persistent supercooled stratiform clouds. Using novel glaciogenic seeding experiments (CLOUDLAB), ice crystals are nucleated upwind and subsequently measured downwind after a known advection time in cloud, allowing us to estimate their age. A deep-learning-based detection algorithm (IceDetectNet) counts the individual monomers of aggregates to derive the initial ice crystal number concentration (ICNC ). We considered several factors that may influence ice aggregation, including ICNC , temperature, ice crystal size, aspect ratio, and turbulence. Among these, ICNC was found to be the dominant factor controlling aggregation rates by three independent approaches: causal inference, a physical equation, and machine learning models. We report, however, a subquadratic dependence of the aggregation rate on ICNC (mean exponent ∼ 0.92; 95 % CI: 0.88–0.97), in contrast to theoretical expectations (quadratic dependence). One possible explanation is that aggregation may also involve smaller ice crystals, but this remains hypothetical. To predict aggregation rates, we evaluated 11 machine learning models and a physically based formulation. CatBoost achieved the best statistical performance, while the physical model proved more robust in sensitivity tests. These findings provide new insights into the microphysical and environmental controls of ice aggregation and establish a robust methodological foundation for studying aggregation processes in natural clouds.- 28 January 2026
- Published by Copernicus GmbH in Atmospheric Chemistry and Physics
- Vol. 26 (2) , 1459-1481
Open AccessInferring the controlling factors of ice aggregation from targeted cloud seeding experiments
Ice aggregation in clouds plays a crucial role in cloud development and precipitation formation. Despite the significance of ice aggregation, direct in situ quantification of aggregation rates in natural clouds has been challenging due to the difficulty of tracking ice crystals. Here, we present in situ measurements of ice aggregation rates in persistent supercooled stratiform clouds. Using novel glaciogenic seeding experiments (CLOUDLAB), ice crystals are nucleated upwind and subsequently measured downwind after a known advection time in cloud, allowing us to estimate their age. A deep-learning-based detection algorithm (IceDetectNet) counts the individual monomers of aggregates to derive the initial ice crystal number concentration (ICNC ). We considered several factors that may influence ice aggregation, including ICNC , temperature, ice crystal size, aspect ratio, and turbulence. Among these, ICNC was found to be the dominant factor controlling aggregation rates by three independent approaches: causal inference, a physical equation, and machine learning models. We report, however, a subquadratic dependence of the aggregation rate on ICNC (mean exponent ∼ 0.92; 95 % CI: 0.88–0.97), in contrast to theoretical expectations (quadratic dependence). One possible explanation is that aggregation may also involve smaller ice crystals, but this remains hypothetical. To predict aggregation rates, we evaluated 11 machine learning models and a physically based formulation. CatBoost achieved the best statistical performance, while the physical model proved more robust in sensitivity tests. These findings provide new insights into the microphysical and environmental controls of ice aggregation and establish a robust methodological foundation for studying aggregation processes in natural clouds. - Open Access
- 1 January 2026
- Published by Royal Society of Chemistry (RSC) in Environmental Science: Processes & Impacts
Inside the battery box: a cross-technology assessment of PFAS uses, functions, and substitution potential
Reports on PFAS uses in batteries were collected and weighed against supporting evidence. Feasible alternatives were assessed using market deployment as a proxy for technical viability, illustrating a new cross-sector assessment framework.
- 1 January 2026
- Published by Royal Society of Chemistry (RSC) in Environmental Science: Processes & Impacts
Open AccessInside the battery box: a cross-technology assessment of PFAS uses, functions, and substitution potential
Reports on PFAS uses in batteries were collected and weighed against supporting evidence. Feasible alternatives were assessed using market deployment as a proxy for technical viability, illustrating a new cross-sector assessment framework.