•Machine learning (ML) models can determine the rainfall flooding threshold as a line. projected in a plane spanned by two principal components, thereby providing a binary result (flood or no ...flood).•Compared to the conventional critical rainfall curve, the proposed models, especially the subspace discriminant analysis, greatly raise the accuracy (ACC) to 96.5% and lowering the false alert rate to 25%.•Rainfall threshold based flood prediction can be executed rapidly and simply, this method allows decision makers time for a high-level assessment of flood risk, providing valuable lead time for citizens in the flood-prone areas to be warned.
Urban pluvial flooding is a threatening natural hazard in urban areas all over the world, especially in recent years given its increasing frequency of occurrence. In order to prevent flood occurrence and mitigate the subsequent aftermath, urban water managers aim to predict precipitation characteristics, including peak intensity, arrival time and duration, so that they can further warn inhabitants in risky areas and take emergency actions when forecasting a pluvial flood. Previous studies that dealt with the prediction of urban pluvial flooding are mainly based on hydrological or hydraulic models, requiring a large volume of data for simulation accuracy. These methods are computationally expensive. Using a rainfall threshold to predict flooding based on a data-driven approach can decrease the computational complexity to a great extent. In order to prepare cities for frequent pluvial flood events – especially in the future climate – this paper uses a rainfall threshold for classifying flood vs. non-flood events, based on machine learning (ML) approaches, applied to a case study of Shenzhen city in China. In doing so, ML models can determine several rainfall threshold lines projected in a plane spanned by two principal components, which provides a binary result (flood or no flood). Compared to the conventional critical rainfall curve, the proposed models, especially the subspace discriminant analysis, can classify flooding and non-flooding by different combinations of multiple-resolution rainfall intensities, greatly raising the accuracy to 96.5% and lowering the false alert rate to 25%. Compared to the conventional model, the critical indices of accuracy and true positive rate (TPR) were 5%-15% higher in ML models. Such models are applicable to other urban catchments as well. The results are expected to be used to assist early warning systems and provide rational information for contingency and emergency planning.
Multi-year observations of aerosol microphysical and optical properties, obtained through ground-based remote sensing at 50 China Aerosol Remote Sensing Network (CARSNET) sites, were used to ...characterize the aerosol climatology for representative remote, rural, and urban areas over China to assess effects on climate. The annual mean effective radii for total particles (ReffT) decreased from north to south and from rural to urban sites, and high total particle volumes were found at the urban sites. The aerosol optical depth at 440 nm (AOD440 nm) increased from remote and rural sites (0.12) to urban sites (0.79), and the extinction Ångström exponent (EAE440–870 nm) increased from 0.71 at the arid and semi-arid sites to 1.15 at the urban sites, presumably due to anthropogenic emissions. Single-scattering albedo (SSA440 nm) ranged from 0.88 to 0.92, indicating slightly to strongly absorbing aerosols. Absorption AOD440 nm values were 0.01 at the remote sites versus 0.07 at the urban sites. The average direct aerosol radiative effect (DARE) at the bottom of atmosphere increased from the sites in the remote areas (−24.40 W m−2) to the urban areas (−103.28 W m−2), indicating increased cooling at the latter. The DARE for the top of the atmosphere increased from −4.79 W m−2 at the remote sites to −30.05 W m−2 at the urban sites, indicating overall cooling effects for the Earth–atmosphere system. A classification method based on SSA440 nm, fine-mode fraction (FMF), and EAE440–870 nm showed that coarse-mode particles (mainly dust) were dominant at the rural sites near the northwestern deserts, while light-absorbing, fine-mode particles were important at most urban sites. This study will be important for understanding aerosol climate effects and regional environmental pollution, and the results will provide useful information for satellite validation and the improvement of climate modelling.
The height of the stable boundary layer is a key parameter in atmospheric transmission and diffusion, air quality, emergency response, wind energy, and numerical weather prediction models. Existing ...methods mainly determine the stable boundary layer height via a threshold or minimum value of the wind speed variance under a low-level jet. Based on multi-meteorological element data from a meteorological gradient observation tower, this paper revealed the limitations of existing methods from the perspective of dynamic and thermal effects. In this paper, it is demonstrated that there were four types of shapes of the wind speed variance profile under the low-level jet and a method for using the shape of the variance profile to retrieve the height of the stable boundary layer was proposed. This method distinguished different types of wind speed variance profiles and solved the problems of the misjudgment and omissions (about 34%) present in existing methods. Our experiment showed that the average absolute error of the proposed method was less than 18 m and the average relative error was less than 9%. The results showed that the proposed inversion method was extended to all kinds of wind field detection equipment for inversion of the stable boundary layer height and has very high universality.
Black carbon (BC) is an essential climate forcer in the atmosphere. Large uncertainties remain in BC’s radiative forcing estimation by models, partially due to the limited measurements of BC vertical ...distributions near the surface layer. We conducted time-resolved vertical profiling of BC using a 356-m meteorological tower in Shenzhen, China. Five micro-aethalometers were deployed at different heights (2, 50, 100, 200, and 350 m) to explore the temporal dynamics of BC vertical profile in the highly urbanized areas. During the observation period (December 6–15, 2017), the average equivalent BC (eBC) concentrations were 6.6 ± 3.6, 5.4 ± 3.3, 5.9 ± 2.8, 5.2 ± 1.8, and 4.9 ± 1.4 μg m
−3
, from 2 to 350 m, respectively. eBC temporal variations at different heights were well correlated. eBC concentrations generally decreased with height. At all five heights, eBC diurnal variations exhibited a bimodal pattern, with peaks appearing at 09:00–10:00 and 19:00–21:00. The magnitudes of these diurnal peaks decreased with height, and the decrease was more pronounced for the evening peak. eBC episodes were largely initiated by low wind speeds, implying that wind speed played a key role in the observed eBC concentrations. eBC wind-rose analysis suggested that elevated eBC events at different heights originate from different directions, which suggested contributions from local primary emission plumes. Air masses from central China exhibited much higher eBC levels than the other three backward trajectory clusters found herein. The absorption Ångström exponent (AAE
375–880
) showed clear diurnal variations at 350 m and increased slightly with height.
The impacts of weather forecast uncertainties have not been quantified in current air quality forecasting systems. To address this, we developed an efficient 2‐D convolutional neural network‐surface ...ozone ensemble forecast (2DCNN‐SOEF) system using 2‐D convolutional neural network and weather ensemble forecasts, and we applied the system to 216‐hr ozone forecasts in Shenzhen, China. The 2DCNN‐SOEF demonstrated comparable performance to current operating forecast systems and met the air quality level forecast accuracies required by the Chinese authorities up to 144‐hr lead time. Uncertainties in weather forecasts contributed 38%–54% of the ozone forecast errors at 24‐hr lead time and beyond. The 2DCNN‐SOEF enabled an “ozone exceedance probability” metric, which better represented the risks of air pollution given the range of possible weather outcomes. Our ensemble forecast framework can be extended to operationally forecast other meteorology‐dependent environmental risks globally, making it a valuable tool for environmental management.
Plain Language Summary
Weather forecasts are intrinsically uncertain, but the impacts of that uncertainty on air quality forecasts are not explicitly quantified in current air quality forecast systems. We proposed here a surface ozone ensemble forecast system, analogous to modern weather ensemble forecast systems, to represent the probability distribution of forecasted surface ozone concentrations given 30–50 possible future weather outcomes. The computation costs of this surface ozone ensemble forecast system were greatly reduced using deep learning techniques that emphasized the spatial patterns of weather. We showed that the surface ozone ensemble forecast system's accuracy met the Chinese operational requirements. However, half of the ozone forecast error was due to weather forecast uncertainties, which cannot be completely eliminated even with perfect pollutant emission estimates and chemistry models. This weather‐induced innate uncertainty in air quality forecasts should be considered for effective air quality management.
Key Points
We built a deep‐learning surface ozone ensemble forecast system to quantify pollution risks given the range of possible weather outcomes
Deep‐learning models accentuating the spatial patterns of weather effectively represented the ozone‐meteorology relationship
Weather forecast uncertainties contributed 38%–54% of the ozone forecast errors at 24‐hr lead time in Shenzhen
Abstract The recovery of the ozone layer relies on decreasing atmospheric mixing ratios of ozone-depleting substances (ODSs), including chlorofluorocarbons (CFCs). A significant decline in the mixing ...ratio of trichlorofluoromethane (CFC-11 or CCl 3 F ), the second most abundant CFC, has been observed since the mid-1990s. However, a slowdown in the decline after 2012 indicates a rise in emissions, particularly in Eastern Asia. Ground-based observations are lacking in southeastern China, limiting a thorough evaluation of CFC-11 levels and emissions in this region. A new Advanced Global Atmospheric Gases Experiment background station was established at Xichong (XCG), Shenzhen, China, to provide high-frequency continuous in situ observations. The annual mean CFC-11 mixing ratio, recorded from May 2022 to April 2023, is 221.64 ± 2.29 ppt. When compared with a monthly (MHD) or daily (MLO) observation, this value is found to be 0.45% to 5.36% higher than the northern hemispheric background. With the inverse modeling and interspecies correlation method, we estimate CFC-11 emissions in southeastern China between 1.23 ± 0.25 Gg yr −1 and 1.58 ± 0.21 Gg yr −1 , in line with the bottom-up estimation of 1.50 Gg yr −1 . Results indicate that CFC-11 emissions in the Pearl River Delta region have returned to levels before 2010, aligning with regional and global trends. Observations from XCG would compensate for the deficiency of CFC-11 measurements in southeastern China, paving the road for ODS studies in this region and beyond.
Based on the meteorological, ozone (O3), and vertical observation data of 2020, this study sought to evaluate the daily variation in O3, particularly the characteristics of nocturnal ozone pollution. ...We also discuss the effect of local and mesoscale horizontal transport and vertical mixing on the formation of nocturnal O3 pollution. Distinct seasonal characteristics of the daily O3 variation in Shenzhen were identified. In particular, significant nocturnal peaks were found to regularly occur in the winter and spring (November–December and January–April). The monthly average of daily variation had a clear bimodal distribution. During the period, O3 pollution frequently occurred at night, with the maximum hourly O3 concentration reaching 203.5 μg/m3. Nocturnal O3 pollution was closely associated with horizontal transport and vertical mixing. During the study period, the O3 maximum values were recorded on 68 nights, primarily between 23:00 and 03:00, with occasional observation of two peaks. The impact of horizontal transport and vertical mixing on the nocturnal secondary O3 maximum values was elaborated in two case studies, where vertical mixing was mainly associated with low-level jets, with strong wind shear enhancing turbulent mixing and transporting O3 from the upper layers to the surface.
It is well-known that coastal low-level jets (CLLJs) play an essential role in transporting heat, water vapor and pollutants. However, the CLLJ characteristics in the Pearl River Estuary have not ...been deeply revealed due to the lack of long-term observations. Based on the long-term observations from a wind lidar, we analyze the primary climatic characteristics of the CLLJs in the Pearl River Estuary and investigate their relationships with large-scale and local-scale synoptic systems. The results show that the CLLJs mainly appear during the flood season, with the most occurrence in May. The CLLJ occurrence during the flood season is mainly influenced by the large-scale north–south pressure gradient driven by the western Pacific subtropical high and terrestrial low-pressure systems. The occurrence of the CLLJs exhibits a distinct diurnal cycle with two different peaks in different seasons. One peak appears at nighttime, mainly during non-flood seasons. The other appears in the afternoon, mainly during the flood season. In the non-flood seasons, under the influence of cold air, the inertial oscillations triggered by the land–sea thermal contrast lead to CLLJ onset at nighttime in the Pearl River Estuary. During the flood season, the strong near-surface pressure gradient contributes to CLLJ onset in the afternoon, while the topography (blocking and passing) is more conducive to the occurrence of the CLLJs in the Pearl River Estuary. These findings reveal the formation mechanisms of the CLLJs over the Pearl River Estuary, thus providing a basis for further understanding the precipitation in the Pearl River Estuary and the occurrence of the CLLJs in other coastal areas with complex mountain ranges.
A gust front is generated when the cold downdraft from a thunderstorm reaches the surface, causing the warm air mass at the surface to lift. This is always accompanied by gusts and strong winds, and ...results in severe destruction of property and loss of life, especially in city areas. Despite the potentially adverse effects associated with this weather phenomenon, the microphysical structure of gust fronts remains unclear, owing to a lack of high temporal–spatial resolution observations, and because research on the thermodynamic structure of gust fronts is insufficient. This study mainly analysed the atmospheric boundary layer (ABL) structure before, during, and after the passage of two selected gust fronts, using meteorological data obtained from a 356-m tall meteorology tower located in the eastern Pearl River Estuary area. The thermodynamic structure of the ABL was of particular interest. The results show that meteorological factors fluctuated significantly and simultaneously at the 13 observation altitudes from the bottom to the top of the meteorology tower, and significant changes in the vertical distribution of the analysed parameters were observed during the passage of the gust fronts. The ABL structure transformed from an unstable state before the passage of the gust fronts to a stable state after they had passed, and a ground temperature inversion was evident. Moreover, the temperature inversion lasted for several hours after the dissipation of the gust fronts, and occurred intermittently, with the inversion centre shifting up or down and showing multiple intensity centres. The turbulence also showed a strengthening trend during gust front passage. Furthermore, we determined that the thermal and mechanical turbulence affected the maintenance of the temperature inversion. The results of this study will contribute to our knowledge on the microphysics of the ABL of gust fronts, and will further benefit the correction of numerical modelling of severe weather.