Search

• ERA5 describes moderate-intensity rainfall, but fails for hourly/daily precipitation. • ERA5-HRES outperforms ERA5-Land in detecting rainfall magnitude and rainfall events. • ERA5 performs best in coastal cities and mountainous vegetation areas during the wet and dry season, respectively. • ERA5 is insufficient in reproducing precipitation in areas with high urbanization level. Precipitation is an important component of the global water cycle. A comprehensive understanding of precipitation characteristics, especially that in urban areas, is essential for practical applications in hydrology. ERA5, a state-of-the-art reanalysis product with wide temporal coverage as well as fine spatial and temporal resolution, has greater potential to reproduce precipitation than many other products. However, how ERA5 precipitation performs has not been well investigated in urban areas. Therefore, we evaluate and compare the ability of two ERA5 products, ERA5-Land and ERA5-HRES, in the Guangdong-Hong Kong-Macao Greater Bay Area (GBA) using observations from over 3000 rain gauges in a high-density network during 2018. The results show that ERA5 products generally characterize moderate-intensity rain. Due to the different causes of seasonal precipitation as well as the combined influence of coastal effects and topography, the best performance of ERA5 products is found in the southern coastal plains during the wet season and in the northern mountainous vegetation areas during the dry season. Meanwhile, it is challenging for ERA5 products to model precipitation in areas with higher urbanization levels. We also find that ERA5-Land data with finer spatial resolution fails to deliver any preferable results than ERA5-HRES. ERA5 products needs further improvement in the GBA for providing accurate precipitation patterns. This paper gives a detailed insight into the performance of ERA5 products, which offers a valuable reference for its probable applicability in the coastal urban areas. [ABSTRACT FROM AUTHOR]

The Guangdong-Hong Kong-Macao Greater Bay Area (GBA) is the world's largest bay area in terms of land area and population, which has been increasingly suffering from weather and climate extremes under global warming. It is thus desired to produce reliable high-resolution climate information at a regional scale in order to enhance resilience to climate change over the GBA. For the first time, this study develops the multi-decadal nested-grid climate projections at a convection-permitting scale for the GBA, and assesses the abilities of the Weather Research and Forecasting (WRF) model with 36-, 12- and 4 km resolutions in representing precipitation, temperature and their extremes. Our findings indicate the added value of the convection-permitting WRF model for simulating the spring and summertime precipitation as well as extreme heavy rainfall events with daily amounts larger than 30 mm over the GBA. Increasing the spatial resolution of the WRF model does not necessarily lead to a significant improvement on temperature simulations. In addition, our findings reveal that the GBA is expected to have an increasing number of heavy and extreme heavy rainfall events by the end of the twenty-first century. Moreover, the GBA is projected to have a large temperature change across different seasons, and an enhanced warming will appear in autumn. The GBA is also expected to have more summer days with longer durations, thereby leading to an increasing risk of heatwaves and heat stress. [ABSTRACT FROM AUTHOR]

The recent intensification of extreme rainfall events has the potential to aggravate the frequency and intensity of floods in urban areas and signals the urgent need for accurate and high spatiotemporal resolution precipitation data for flood monitoring and warning to alleviate the associated damages. In this context, this paper aims to statistically assess the latest Integrated Multisatellite Retrievals from Global Precipitation Measurement (IMERG) multisatellite constellation and radar quantitative precipitation estimation (QPE) products against dense rain gauge network observations under multiple temporal scales (hourly and 3-hourly) and three spatial scales (grid cumulative, regional, and grid) over the Guangdong-Hong Kong-Macao Greater Bay Area (GBA) in China. The results indicate that (1) at the grid cumulative scale, QPE significantly outperforms IMERG products. The correlation coefficients of QPE (0.83 ≤ CORR ≤ 0.85) are much higher than those of IMERG products (0.35 ≤ CORR ≤ 0.56), the estimation of QPE is generally significantly underestimated (−0.23 ≤ BIAS ≤ -0.16), and the BIAS of IMERG products is larger than that of QPE (−0.24 ≤ BIAS ≤ 0.23). (2) At the regional scale, the accuracy of IMERG products has been significantly improved, with relatively high correlation coefficients (CORR≥0.8) and decreased root mean squared error (RMSE) values (1.18 mm ≤ RMSE ≤ 0.22 mm), and Taylor diagrams visually show that QPE is closer to the gauge observations than are IMERG products. (3) The evaluation results at the grid scale show that both IMERG products and QPE can accurately reflect the spatial distribution of precipitation. Among these results, the spatial distribution of QPE is the most consistent with the actual rainfall, although an obvious underestimation of rainfall occurs in areas with strong rainfall. (4) In terms of detection indicators, the three IMERG products are superior to QPE at capturing rainfall events and have more advantages over QPE in the estimation of instantaneous precipitation, especially for light rain events. However, QPE has better detection effects for rainfall events under rainstorm and heavy rainstorm events, indicating the hydrologic utility of this product. Among the three IMERG products, the IMERG Early Run (ER) product has the advantages of a short lag time and high accuracy, and it has the potential to be applied in flood forecasting. Future studies should improve IMERG algorithms and promote the application of QPE in urban rainstorm and flood research. • The quality of the IMERG products and radar quantitative precipitation estimation (QPE) is compared over the Guangdong-Hong Kong-Macao Greater Bay Area (GBA) in China. • The dense rain gauge network observations are used for evaluation in the GBA. • The evaluation of multisource precipitation products under different temporal and spatial scales has different results and effects. [ABSTRACT FROM AUTHOR]