“…3 (j)) at 12:00 was also observed relative to 2019 and 2020, respectively. This finding further revealed that the production of secondary inorganic particles was under the NO x -saturated regime during the Level 3 COVID-19 alert, wherein a reduction of NO x emission could decrease the exhaustion of OH and O 3 by the chemical reaction with NO x , resulting in an increase in NO 3 − production ( Lin et al, 2022a ).…”
Section: Discussionmentioning
confidence: 91%
“…3 (b), (f), and (j), in which a significant peak at noon is observed. It was because the homogeneous reaction between gaseous NH 3 and HNO 3 with the aid of sunlight enhanced the secondary particle formation ( Lin et al, 2022a ). Fig.…”
Section: Resultsmentioning
confidence: 99%
“…3 (b), (f), and (j)). In 2019 and 2020, the SO 4 2− concentrations peaked at noon due to the gas to particle partitioning through the photochemical reaction between HSO 3 − and OH − radicals ( Lin et al, 2022a ). After 21:00, an insignificant diurnal pattern of SO 4 2− was observed because the regional accumulation of nonvolatile sulfate was observed to be critical to the diurnal trend of SO 4 2− in Taichung City ( Lin et al, 2022b ).…”
Section: Resultsmentioning
confidence: 99%
“…From June to July 2021, the average O 3 and NO 3 − concentrations increased by 13.2% and 98.2%, respectively, as compared to those during the same period in 2020. The result depicts that the VOC-limited O 3 production regime with a reduction of NO x concentration improved OH radical formations to produce more O 3 and secondary aerosols ( Lin et al, 2022a ).…”
Section: Discussionmentioning
confidence: 97%
“…A significant increase in the NO 3 − concentration in 2021 by 38.2% and 63.1% at 12:00 was also observed relative to 2019 and 2020, respectively. Thus, during the Level 3 COVID-19 alert, the production of secondary particles was under the VOC-limited regime to enhance NO 3 − production ( Leung et al, 2020 ; Lin et al, 2022a ).…”
“…3 (j)) at 12:00 was also observed relative to 2019 and 2020, respectively. This finding further revealed that the production of secondary inorganic particles was under the NO x -saturated regime during the Level 3 COVID-19 alert, wherein a reduction of NO x emission could decrease the exhaustion of OH and O 3 by the chemical reaction with NO x , resulting in an increase in NO 3 − production ( Lin et al, 2022a ).…”
Section: Discussionmentioning
confidence: 91%
“…3 (b), (f), and (j), in which a significant peak at noon is observed. It was because the homogeneous reaction between gaseous NH 3 and HNO 3 with the aid of sunlight enhanced the secondary particle formation ( Lin et al, 2022a ). Fig.…”
Section: Resultsmentioning
confidence: 99%
“…3 (b), (f), and (j)). In 2019 and 2020, the SO 4 2− concentrations peaked at noon due to the gas to particle partitioning through the photochemical reaction between HSO 3 − and OH − radicals ( Lin et al, 2022a ). After 21:00, an insignificant diurnal pattern of SO 4 2− was observed because the regional accumulation of nonvolatile sulfate was observed to be critical to the diurnal trend of SO 4 2− in Taichung City ( Lin et al, 2022b ).…”
Section: Resultsmentioning
confidence: 99%
“…From June to July 2021, the average O 3 and NO 3 − concentrations increased by 13.2% and 98.2%, respectively, as compared to those during the same period in 2020. The result depicts that the VOC-limited O 3 production regime with a reduction of NO x concentration improved OH radical formations to produce more O 3 and secondary aerosols ( Lin et al, 2022a ).…”
Section: Discussionmentioning
confidence: 97%
“…A significant increase in the NO 3 − concentration in 2021 by 38.2% and 63.1% at 12:00 was also observed relative to 2019 and 2020, respectively. Thus, during the Level 3 COVID-19 alert, the production of secondary particles was under the VOC-limited regime to enhance NO 3 − production ( Leung et al, 2020 ; Lin et al, 2022a ).…”
Recently, runoff simulations in small, headwater basins have been improved by methodological advances such as deep learning (DL). Hydrologic routing modules are typically needed to simulate flows in stem rivers downstream of large, heterogeneous basins, but obtaining suitable parameterization for them has previously been difficult. It is unclear if downstream daily discharge contains enough information to constrain spatially-distributed parameterization. Building on recent advances in differentiable modeling principles, here we propose a differentiable, learnable physics-based routing model. It mimics the classical Muskingum-Cunge routing model but embeds a neural network (NN) to provide parameterizations for Manning's roughness coefficient (n) and channel geometries. The embedded NN, which uses (imperfect) DL-simulated runoffs as the forcing data and reach-scale attributes as inputs, was trained solely on downstream hydrographs. Our synthetic experiments show that while channel geometries cannot be identified, we can learn a parameterization scheme for n that captures the overall spatial pattern. Training on short real-world data showed that we could obtain highly accurate routing results for both the training and inner, untrained gages. For larger basins, our results are better than a DL model assuming homogeneity or the sum of runoff from subbasins. The parameterization learned from a short training period gave high performance in other periods, despite significant bias in runoff. This is the first time an interpretable, physics-based model is learned on the river network to infer spatially-distributed parameters. The trained n parameterization can be coupled to traditional runoff models and ported to traditional programming environments.
Air pollution is the one of the most significant environmental risks to health worldwide. An accurate assessment of population exposure would require a continuous distribution of measuring ground-stations, which is not feasible. Therefore, significant efforts are spent in implementing air-quality models. However, a complex scenario emerges, with the spread of many different solutions, and a consequent struggle in comparison, evaluation and replication, hindering the definition of the state-of-art. Accordingly, aim of this scoping review was to analyze the latest scientific research on air-quality modelling, focusing on particulate matter, identifying the most widespread solutions and trying to compare them. The review was mainly focused, but not limited to, machine learning applications. An initial set of 940 results published in 2022 were returned by search engines, 142 of which resulted significant and were analyzed. Three main modelling scopes were identified: correlation analysis, interpolation and forecast. Most of the studies were relevant to east and south-east Asia. The majority of models were multivariate, including (besides ground stations) meteorological information, satellite data, land use and/or topography, and more. 232 different algorithms were tested across studies (either as single-blocks or within ensemble architectures), of which only 60 were tested more than once. A performance comparison showed stronger evidence towards the use of Random Forest modelling, in particular when included in ensemble architectures. However, it must be noticed that results varied significantly according to the experimental set-up, indicating that no overall best solution can be identified, and a case-specific assessment is necessary.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
