Search

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 outcomesDeep‐learning models accentuating the spatial patterns of weather effectively represented the ozone‐meteorology relationshipWeather forecast uncertainties contributed 38%–54% of the ozone forecast errors at 24‐hr lead time in Shenzhen [ABSTRACT FROM AUTHOR]

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. [ABSTRACT FROM AUTHOR]

A number of squall line passages over the Pearl River Delta (PRD) are documented for the first time based on the wind data collected by a 356 m-high meteorological tower with high temporal resolution (10 Hz). The mean wind and turbulence characteristics are studied based on a sample of six cases. The mean wind profile is consistent with the international standard for wind engineering, but the turbulence intensity profile has quite significant deviations. Moreover, the energy spectrum of the fluctuating wind is found to be consistent with the −5/3 law, as expected in the inertial subrange. For the purpose of wind gust nowcasting, the performance of a nowcasting algorithm based on upper air wind and thermodynamic profiles is examined using the limited dataset. The results highlight that the mean wind gust estimates are sufficient to nowcast the surface wind gust as measured by the anemometers, but for an extreme squall line case, the maximum wind gust estimates would be useful. The information in this paper is believed to be useful for wind engineering applications and weather nowcasting for wind gusts associated with subtropical squall lines, and, to the best of the authors' knowledge, this is the first study of its kind for the PRD region. [ABSTRACT FROM AUTHOR]

This study investigates the performance of two commonly used planetary boundary layer (PBL) schemes, the local turbulence closure (Mellor-Yamada-Janjić: MYJ) and the nonlocal turbulence closure (Yonsei University: YSU), over the tropical coastal region of Shenzhen, China, using the Weather Research and Forecasting (WRF) model. The simulation results are evaluated by comparing them to various measurements. Additionally, we also investigate the performance of the PBL schemes over different land cover types and their ability to reproduce the physical relationships between several pairs of interrelated near-surface atmospheric parameters. The results show that both PBL schemes underestimate temperature and PBL height and overestimate wind speed and turbulent fluxes throughout the whole day, but this can be improved by increasing horizontal resolution. We find that the overestimation of wind speed and turbulent heat fluxes is mainly caused by the strong mechanical mixing in the model. Some improvements are required such as more accurate surface morphology parameters and momentum transport coefficients. Overall, the YSU scheme outperforms the MYJ scheme at all horizontal resolutions over complex terrain in terms of near-surface variables, PBL vertical structures, and the internal relations between turbulent heat fluxes and atmospheric instability. However, the difference between the two PBL schemes, varying over different land covers and with time periods and altitudes within the PBL, is generally greater over the land than over the sea for thermal-related variables (especially during daytime), while for variables associated with water vapor is the opposite. This highlights the importance of the underlying surface for the performance of PBL schemes and the need to select appropriate PBL schemes for simulating the evolution of PBL processes over different land covers. • The YSU scheme generally performs better than MYJ scheme over complex terrain. • Increasing model resolution improves the overall model performance in urban area. • Performance of the PBL schemes differs with land cover types. • Turbulent fluxes-instability relationships provide insight in PBL parameterizations. [ABSTRACT FROM AUTHOR]