Effects of Typhoon Fitow residual circulation on the relationship between surface wind field and precipitation in Shanghai Authors: Yuqi Tang, Jishi Zhang, Caijun Yue, Juan Sun, Zhihui Han, Wen Gu, Liping Deng, Yao Yao Published: 2025 ABSTRACT AbstractHeavy precipitation, triggered by the residual circulation from Typhoon Fitow (2013), struck Shanghai with a spatial inhomogeneity distribution. Using multi-physical diagnostic analysis based on automatic weather stations (AWS) observations, the fifth-generation ECMWF re-analyses (ERA5), FY satellite products, and typhoon track data, this study investigates the relationship between the surface wind field and precipitation in Shanghai. The main results are as follows: (1) The Shanghai Central Area (SCA) decelerated the eastward progression of the northerly wind associated with the typhoon’s residual circulation. The low-wind area and wind deflection around the SCA enhanced the wind field convergences in its surrounding regions. (2) The slow-moving convective systems prolonged the duration of rainfall, while the wind deflection around the SCA further strengthened surface wind convergence conditions. (3) Intense surface Q vector convergence persisted between the northerly wind from the residual circulation and the northeasterly maritime wind. Precipitation attenuated rapidly within 2 h following the weakening of the Q vector convergence, suggesting that this parameter served as a reliable indicator for the development of the studied precipitation event. Additionally, surface water vapor advection was channeled by wind deflection around the SCA, leading to water vapor accumulation in both upwind and downwind areas of the SCA.Keywords: Surface wind field, Heavy precipitation, Residual circulation from typhoon, Urban underlying surface, Spatial inhomogeneity distributionSubject terms: Climate sciences, Environmental sciences, Natural hazards Introduction Landfall typhoon circulation usually decays gradually due to surface friction and weakened water vapor supply. The typhoon residual circulation is defined as a weakened typhoon circulation characterized by a central maximum mean wind speed of less than 17.2 m s-1 or an intensity weaker than the tropical storm. However, the residual circulation may still induce widespread, hard-to-forecast heavy precipitation, which received growing attention in recent years. The continued dynamical lift and water vapor transport are critical to the residual circulation rainstorms. The fifth-generation ECMWF Re-Analyses (ERA5) and the Final Operational Global Analysis (FNL) revealed the 3-D vertical wind shear and water vapor transport in the rainstorms caused by typhoon residual circulations1–3. A study by Diao et al.4 using radar and conventional observations indicated that high humidity, strong vertical wind shear, and large storm relative helicity in low-level - maintained by the residual circulation from Typhoon Yagi and Rumbia (2018) - may serve as key physical parameters for tornado formation. Meso- and small-scale topographic features, including the urban underlying surface, can also modulate the intensity of residual circulation precipitation. Urban expansion and urban agglomeration formation have prolonged the retention time and impact of weather systems, and enhanced convergence by weakening surface wind field5,6. Radar and numerical simulations studies have shown that thunderstorms may split, divert, divide, or merge as they pass through cities7,8, while rainfall is also bifurcated by urban buildings in surrounding areas9. The precipitation in downwind areas of cities may be altered by the urbanization process10, even during the typhoon impact period11,12. Urban surface characteristics can significantly impact the movement, stagnation, development, and intensification of typhoon rainstorms13–15. The rainstorms induced by residual circulation from Typhoon Fitow (2013) persisted in impacting the southeastern coastal areas of China from October 5 to 9, 2013. Comprehensive multi-physical diagnostic analyses were conducted for this event based on observations and reanalysis data. The results showed that the intrusion of cold air into residual circulation16,17 and the water vapor transport from Typhoon Danas (2013)18 provided favorable largescale dynamic and water vapor conditions, which jointly contributed to this heavy precipitation. Meso- and small-scale topographic enhancement effect was another important factor. Studies indicated that Hangzhou Bay and the surrounding natural underlying surface provided topographic convergence and lifting conditions for the residual circulation rainstorms in Zhejiang province utilizing reanalysis, satellite, radar and conventional observations19,20. Rainfall amount recorded by numerous automatic weather stations (AWSs) in Shanghai surpassed historical records during the residual circulation precipitation event21. However, the underestimation of rainfall intensity, rainfall area, and heavy precipitation duration during this event posed a more severe threat to Shanghai’s safety than the direct impact of the typhoon periphery. Previous studies have forced more on the interactions between large-scale weather systems (e.g., cold air and residual circulation22, as well as their impacts on wind, moisture, and unstable energy23,24. Nevertheless, no dedicated research has been conducted on the correspondence between surface wind fields, precipitation, and the urban underlying surface in Shanghai. Additionally, there is insufficient evidence to explain the potential impact of local meteorological conditions and urban underlying surface on the spatial inhomogeneity distribution of heavy precipitation across Shanghai. Multi-physical diagnostic analysis has been widely applied in study of typhoons and their residual circulation rainstorms. It has been used to analyze the distribution and correspondence of dynamic, moisture, and instability energy conditions. Wind field divergence and the water vapor flux have revealed variations in the dynamic convergence16,24 and water vapor transport25,26, respectively. The Q vector is an effective diagnostic method for analyzing vertical motion in typhoon rainstorm27,28, heavy precipitation29 and other precipitation events30, as it integrates both dynamic and thermodynamic information. In this study, we analyze the spatial and temporal variation characteristics of the surface wind field and heavy precipitation using 129 AWS observations in Shanghai, ERA5 reanalysis, typhoon track data, and FY satellite products. A multi-physical diagnostic analysis of the surface dynamic and moisture conditions was conducted to investigate the correspondence between surface wind field changes and spatial inhomogeneity distribution of heavy precipitation, and to further explore the possible impact of urban underlying surface of Shanghai on the heavy precipitation process. Information about Shanghai, data sources, and the one-layer non-geostrophic Q vector divergence method are presented in Sect. 2. In Sect. 3, first, an overview of the heavy precipitation process is provided based on comprehensive analysis of ERA5 reanalysis, FY satellite products, typhoon track data, AWS observations, and relevant literature. Then, the correspondence between surface wind fields and heavy precipitation is analyzed using AWS observations, including observational and statistical analyses. Finally, a multi-physical diagnostic analysis is conducted using AWS observations and ERA5 reanalysis, with corresponding discussions presented herein. The conclusions are summarized in Sect. 4. Data and methods Study area Shanghai is located between 30.6–31.9°N and 120.6–122.2° E, at the core of the Yangtze River Delta Urban Agglomeration in eastern China. The city is bounded by the Yangtze River to the north, the East China Sea to the east, and Hangzhou Bay to the south. It features a subtropical monsoon climate with four distinct seasons, abundant rainfall, shorter springs and autumns, and longer winters and summers. Shanghai is extremely sensitive to hazardous weather with developed economy, dense population, and high urbanization. It is impacted by 2–3 typhoons annually on average, which often bring strong winds, rainstorms and other severe convective weather events, threatening urban safety and normal operations. This study focuses on surface wind field and precipitation variations around the Shanghai Central Area (SCA). The SCA covers the region within the Shanghai Outer Ring Expressway (Fig. 1, the area enclosed by the red polygon), including the districts of Huangpu, Xuhui, Changning, Jing’an, Putuo, Hongkou, and Yangpu, as well as parts of Pudong New Area, Minhang, Baoshan, and Jiading. Distribution of automatic weather stations (AWSs) in Shanghai; the red polygon indicates the Shanghai Outer Ring Expressway, and the area within the expressway is defined as The Shanghai Central Area (SCA), and the green line indicates the four-region scheme centered on the People’s Park AWS for the following analysis in Sect. 3.2. To further analyze the relationship between the residual circulation heavy precipitation and Shanghai’s urban underlying surface characteristics, this study divided Shanghai into four regions (Fig. 1, the northeast (R1), the southeast (R2), the southwest (R3), and the northwest (R4)) centered on People’s Square (Fig. 1, green solid line intersection), which is located at the center of the SCA and also servs as Shanghai’s geographic coordinate reference point. The four-region division aims to quantify the impact of northerly winds from residual circulation on the SCA, as well as its upwind region (R1) and downwind region (R3), during the eastward movement of the synoptic system affecting Shanghai (i.e., from R3 and R4 to R1 and R2). Data sources The AWS observations in Shanghai (Fig. 1) were provided by Shanghai Meteorological Information and Technology Support Center. The data includes hourly rainfall, 2 min mean wind speed, 2 min mean wind direction, and dew point temperature, covering the period from 0000 UTC 7 to 2300 UTC 8 October 2013. Strict quality analysis and control were conducted including boundary value check, internal consistency check, temporal and spatial consistency check, and a manual recheck. Detailed information on data quality control and the instruments is available in the study by Sun et al.31. Finally, 129 AWSs were selected for analysis and calculation of physical diagnostic quantities in this study. The intensity and the track data of Typhoon Fitow (2013) can be obtained from the China Meteorological Administration (CMA) with temporal resolution of 6 h, including data of the typhoon center, maximum wind speed and central pressure in the western North Pacific32. The FY-2E satellite infrared (10.3–11.3 μm) bright temperature product is obtained from National Satellite Meteorological Center (NSMC). ERA5 hourly relative humidity and horizontal wind data on pressure levels can be obtained from the European Centre for Medium-Range Weather Forecasts (ECMWF) with a spatial resolution of 0.25° × 0.25° and a temporal resolution of 1 h. The data was used for the analysis of the surface environmental meteorological conditions in Shanghai and surrounding areas from 0000 UTC 7 to 2300 UTC 8 October 2013. One-layer non-geostrophic Q vector divergence In this study, the one-layer non-geostrophic Q vector divergence is calculated to analyze the ascending motion in the precipitation progress at the surface. The advantage of the one-layer non-geostrophic Q vector divergence (Q vector divergence) lies in its requirement of only one layer of meteorological data and can be calculated using actual winds. It was developed by Yue33 based on the quasi-geostrophic Q vector, derived from the primitive equations under the assumptions of quasi-static, adiabatic, frictionless conditions on an f-plane in P-coordinate. By employing the substitution balance approximation (i.e., replacing the vertical shear of actual wind with the vertical shear of geostrophic wind), it was applied to analyze the impact of the landing structure of Typhoon Haitang (2005) on its precipitation. In recent years, the Q vector divergence has been frequently used in the study and analysis of mesoscale and microscale weather systems. For instance, the overlapping region of Q vector forcing and moisture conditions based on surface AWS observations showed good indicative significance for the location of a short-duration heavy rainfall in Shanghai27. Additionally, it accurately captured the surface ascending motion preceding a severe convection with a short life near the coast of Shanghai34. The physical diagnostic quantities are as follows: Here, Qx and Qy are zonal and meridional one-layer non-geostrophic Q vector divergence components (hPa−1 s−1) in the P coordinate, respectively; T and P are one-layer temperature and pressure, respectively; u and v are zonal and meridional wind components (m s−1), respectively; and R = 287.05 (J kg−1 K−1) is the specific gas constant of dry air. Results and discussions Overviews of the precipitation Typhoon Fitow (2013) made landfall on the coast of Fuding, Fujian province, at 1715 UTC 6 October 2013, with a maximum wind speed of 42.0 m s−1 and a minimum pressure of 955 hPa. It weakened to a tropical depression at 0100 UTC 7, was discontinued by the National Meteorological Centre of China at 0300 UTC 7 the same day. Its last recorded was documented at 0600 UTC 7(26.7°N, 117.8°E) - approximately 570 km from Shanghai (Fig. 2(a) and (b), purple line). Water vapor transport within the easterly jet associated with Typhoon Danas (2013), combined with the ascending motion induced by cold air intrusion extended the influence of the residual circulation from Typhoon Fitow (2013). Rainstorms triggered by this residual circulation impacted Shanghai from west to east between 7 and 8 October, posing a serious threat to the city’s safety and normal operation24. The rainstorm showed a spatial inhomogeneity distribution across Shanghai (Fig. 2c). Several AWSs recorded total rainfall exceeding 200.0 mm in the SCA and its downwind areas (the southern side of the SCA), with one station even surpassing the decade extreme daily rainfal21 - significantly higher than that in surrounding areas. Overview of the precipitation process. Distribution of (a,b) the infrared (10.3–11.3 μm) bright temperature (shading; units: K) from FY-2E satellite, and (c) the total rainfall (shaded scatter; units: mm) from AWS observations; the purple line in (a) and (b) indicates the track of Typhoon Fitow (2013). From October 7 to 8, Typhoon Fitow (2013) began to dissipate after landing, and the water vapor transport by its original southeasterly jet weakened (Fig. 3a). Meanwhile, the water vapor transport formed by the easterly jet of Typhoon Danas (2013) persisted. The strong water vapor flux area southwest of Shanghai coincided with the junction of Typhoon Fitow (2013) and the cold air, with an intensity comparable to that near the center of Typhoon Danas (2013) at the same time (Fig. 3a,b). As the residual circulation of Typhoon Fitow (2013) further dissipated and Typhoon Danas (2013) continued moving northward away from Shanghai, the water vapor transport by the easterly jet and local water vapor convergence weakened (Fig. 3c), and the heavy precipitation in Shanghai also abated accordingly. After Typhoon Fitow (2013) landed, the water vapor during the phase when its residual circulation combined with the cold air was mainly provided by the peripheral circulation of Typhoon Danas (2013). The sustained strong water vapor transport by the easterly jet was conducive to the occurrence of heavy precipitation in Shanghai. This result is consistent with the analyses of water vapor flux budget in the heavy precipitation area of Typhoon Fitow (2013) conducted by Shen et al.16, Qin et al.18, Zhou et al.19, and Cao et al.24 The vertically integrated moisture flux (vector; units: kg m2 s−1) and its divergence (shading; units: kg m−2 hPa−1 s−1) from the fifth-generation ECMWF re-analyses (ERA5). The spatial inhomogeneity distribution of heavy precipitation from the residual circulation might be caused by the urban effects of Shanghai near the surface. The ERA5 reanalysis revealed that the wind field in the lower layer (1000 to 900 hPa) changed as the residual circulation pushing eastward (Fig. 4a, b), and the same to those at 10 m (Fig. 4c,d). The northeasterly wind (NE wind, 22.5° − 67.5°) from sea prevailed during the night of the 7th, while the northerly wind (N wind, 337.5° − 360°, and 0° − 22.5°) from the residual circulation predominated in the early morning of the 8th. The high humidity condition at surface and lower layer covered nearly the whole city. Overview of the wind and rainfall during this precipitation. The wind and humidity field from ERA5: (a,b) The average wind (barb; units: m s−1) and the relative humidity (shading; units: %) from 1000 to 900 hPa, (c,d) the wind at 10 m (barb; units: m s−1) and the relative humidity at 2 m (shading; units: %), and (e,f) the wind at 10 m (barb and shading; units: m s−1) in Shanghai. (g) The variation of surface wind direction and average hourly rainfall (bar; units: mm) in Shanghai from AWS observations. During the occurrence of this heavy precipitation event, the 10 m horizontal wind reanalysis data show a low-wind area at the surface near Shanghai (Fig. 4e,f). However, the orientation of the wind speed isopleths was similar to that of the coastline, and the low-wind area was located in the southwest corner of Shanghai, the downwind area of Shanghai on that day. This indicates that the surface wind speed reduction over Shanghai in the ERA5 reanalysis data was mainly caused by the land-sea contrast and did not reflect the characteristics of Shanghai’s urban underlying surface (Fig. 1, with the SCA located in the north-central part of Shanghai), resulting in an underestimation of its barrier effect. Therefore, it is necessary to use the dense surface AWS observations established in Shanghai to analyze the impact of the urban underlying surface on the distribution of meteorological elements such as the surface wind field, and further discuss the possible reasons for the spatial inhomogeneity distribution of precipitation associated with the residual circulation from Typhoon Fitow (2013). The precipitation from the night of 7 October to the early morning of 8 October 2013 was significantly stronger than during other periods of the precipitation using 129 AWSs observations, which is consistent with the results of Cao et al.24. The average hourly rainfall reached heavy to hard rain levels (≥ 7.0 mm) and persisted for approximately 9 h (1900 UTC 7 to 0300 UTC 8 October). Therefore, the period from 1700 UTC 7 to 0400 UTC 8 October is defined as the study period in this paper (Fig. 4g, hereafter referred to as the “study period”). Variations in the number of AWSs recording different wind directions reflected changes in the surface wind field. The surface prevailing wind gradually changed from NE wind and easterly wind (E wind, 67.5° − 112.5°) to N wind in after 1800 UTC 7. Concurrently, the precipitation rapidly intensified to heavy - hard rainfall levels, with the average hourly rainfall peaking at 16.0 mm (2200 to 2300 UTC 7). After 0200 UTC 8, the precipitation weakened rapidly as NE wind gradually regained dominance, and then the heavy precipitation process ultimately concluded. It is evident that the heavy precipitation in Shanghai was closely associated with changs in the surface prevailing wind during the study period. In addition, the surface temperature in Shanghai decreased gradually from west to east under the influence of residual circulation precipitation. The temperature gradient across the city was insignificant, so the urban heat island (UHI) effect on precipitation was nearly negligible. This study focuses on the evolutionary characteristics of the surface wind field under the influence of Shanghai’s urban underlying surface, as well as the impact of the surface wind field on the spatial inhomogeneity distribution of heavy precipitation in Shanghai. Correspondence between surface wind fields and heavy precipitation Surface wind in Shanghai changed from NE wind to N wind as the movement of residual circulation from Typhoon Fitow (2013), and precipitation associated with the residual circulation was distributed along the N wind front (Fig. 5). Before 1700 UTC 7, NE wind remained dominant over Shanghai’s surface, after which N wind from the residual circulation gradually advanced eastward. At approximately 1800 UTC 7, N wind began to prevail across the SCA and the surrounding areas, accompanied by an increase in the number of AWSs recording heavy rain (HR, 7.0–14.9 mm h−1). Then the eastward advance of N wind slowed over the subsequent 4 h (2000 to 2300 UTC 7), during which time the SCA and its downwind areas experienced severe rainstorms (SR, 40.0–49.9 mm h−1) - the peak phase of the heavy precipitation process. At this stage, the surface wind field around the SCA was more disorganized. N wind resumed their eastward advance by 0000 UTC 8, and heavy precipitation impacted eastern Shanghai with a more uniform distribution compared to the peak period. After N wind became predominant across the whole city at approximately 0100 UTC 8, precipitation weakened rapidly, marking the end of the heavy precipitation process. The hourly 2 min mean wind (vector; units: m s−1) and the hourly rainfall (shaded scatter; units: mm) at the surface from AWS observations. Wind deflects and detours before merging in the downwind area when it passes through urban areas35. N wind from the residual circulation underwent significant deflection as it passed the SCA (Fig. 6), with a concomitant strengthening of the surface wind field convergence (denser streamlines). Wind deflection even impacted the distribution of N wind front during 1800 to 1900 UTC 7 – its northern part moved to the east of the SCA, while its northern part remained to the west of the SCA. The surface wind speeds also represented a spatial inhomogeneity distribution at this time. A low wind speed center (below 2.0 m s−1) formed over the SCA and its downwind areas, which enhanced and sustained the wind speed gradient until N wind prevailed across the entire city. The wind speed gradient further strengthened the surface wind speed convergence between the stronger N wind to the west (from typhoon’s residual circulation) and the weaker NE wind to the east from 1700 to 2300 UTC 7. Collectively, enhance of the surface wind deflection, convergence, and wind speed gradient near the SCA provided favorable local dynamic uplift conditions for the intensification and persistence of precipitation7,8. The 2 min mean wind (shading and streamline; units: m s−1) at surface from AWS observations. A statistical analysis was conducted using observations from 129 AWSs in Shanghai to quantitatively examine the relationship between the surface wind field and heavy precipitation distribution. For this study, Shanghai was divided into four regions: R1, R2, R3, and R4 (Fig. 1). As illustrated in Fig. 7, most areas of Shanghai were dominated by NE and E wind before 1800 UTC 7. Subsequently, N wind first became predominant in R1, R3, and R4 under the influence of N wind front from the residual circulation, accompanied by enhanced precipitation. N wind prevailed in R1, R3, and R4 between 1800 and 2300 UTC 7, while E wind remained in R2 (Fig. 5). From 1800 UTC 7 onward, precipitation in R1, R3, and R4 exceeded the HR; by 2100 UTC 7, precipitation in R1 and R3 reached rainstorm (RS, 15.0 mm h−1− 39.9 mm h−1) and persisted for approximately 3 h (2100 to 2300 UTC 7). Rainfall intensified further in R1 and R3 at 2200 UTC 7, coinciding with a wind direction change to northwesterly wind (NW wind, 292.5° − 337.5°) in R1 due to surface wind deflection around the SCA (Fig. 6). N wind became predominant in R2 at 0000 UTC 8, and precipitation there intensified to RS, while precipitation in R1 and R3 began to weaken rapidly. By approximately 0400 UTC 8, the heavy precipitation process had essentially concluded, with precipitation reduced to light to moderate (LM, 1.6 mm h−1-6.9 mm h−1). Under the influence of residual circulation of Typhoon Fitow (2013), N wind rapidly prevailed in R1, R3, and R4, with R2 experiencing N wind about 6 h later. This indicates that the SCA can also exert a barrier effect on N wind front associated with residual circulation, and further confirms that when N wind passed the SCA, it deflected and detoured at surface (evidenced by the temporary occurrence of NW wind in R1). The SCA and its upwind /downwind areas (R1 and R3) experienced the RSs for approximately 3 h for the stagnation of N wind front, exhibiting greater intensity and longer duration compared to other regions of Shanghai. This is consistent with the findings of Lorenz et al.8 and Qi et al.13. Correspondence between the surface wind direction and the heavy precipitation. The surface prevailing wind and average rainfall are determined by statistical analysis of AWS observations in four regions shown in Fig. 1. The surface prevailing wind (defined as the most frequent directional sector per hour) is obtained from statistics of the 2-minute mean wind direction recorded hourly in each region, and the average rainfall is mean value of rainfall recorded hourly in each region, respectively. The light gray shading indicates HRs, and the dark gray shading indicates RSs and above. Multi-physical diagnostic analysis This section presents a multi-physical diagnostic analysis of the spatiotemporal variations in dynamic and moisture conditions using AWSs observations and ERA5 reanalysis during the study period, with a specific focus on their correspondence to the distribution of hourly rainfall exceeding RS. Q vector divergence During the study period, Q vector convergence areas on the basis of ERA5 existed at 700 hPa near Shanghai (Fig. 8), indicating that Shanghai was located within the ascending motion area of the residual circulation from Typhoon Fitow (2013) and possessed favorable low-level dynamic uplift conditions for the occurrence of heavy precipitation. However, the distribution of Q vector convergence at 700 hPa is insufficient to explain why the precipitation intensity in SCA was significantly higher than that in other regions of Shanghai. The Q vector divergence (shading; units: ×10–10 m−1 s−1) at 700 hPa based on ERA5. In contrast, the Q vector convergence using AWS observations during the same period exhibited more refined spatiotemporal variation characteristics (Fig. 9). It was primarily concentrated near the N wind front associated with the residual circulation. The Q vector convergence center (with a magnitude stronger than − 0.3 × 10–11 hPa−1s−3) persisted around the SCA and its downwind area until 1900 UTC 7, showing good correspondence with the surface wind convergences. It provided favorable surface conditions of water vapor and dynamic uplift for heavy precipitation before 2300 UTC 7. Subsequently, the Q vector convergence area began to weaken and fragment, leading to a mismatch between surface dynamic uplift, wind field convergence, and water vapor flux convergence near the SCA. From 0000 UTC 8 onward, Q vector convergence rapidly diminished across most areas of Shanghai, and the heavy precipitation weakened to below the level of RS within the following 2 h. The N wind front, triggered by the N wind from the residual circulation, was the primary factor promoting surface dynamic uplift23, and contributed significantly to the pronounced Q vector convergence. The spatiotemporal distribution of Q vector convergence showed good correspondence with the heavy precipitation, serving as a reliable indicator for development of the heavy precipitation in the next 2 h during the study period28,36. The Q vector divergence (shading; units: ×10–11 hPa−1 s−3) and hourly rainfall (shaded scatter; units: mm) stronger than the RS level at the surface based on AWS observations. Water vapor flux and its divergence The distribution of surface water vapor flux and its divergence over Shanghai was highly consistent with that of surface wind field during the study period (Fig. 10). The water vapor flux divergence was directly impacted by surface wind field: the first water vapor flux convergence area was mainly distributed near N wind front at approximately 1800 UTC 7, while the second one was located along the northern riverside of Shanghai. Even after the first convergence area moved away with N wind front, water vapor flux convergence persisted over the southeastern edge and downwind area of the SCA. The second convergence area enhanced over the upwind area of the SCA (stronger than − 6 × 10−7 g cm−2 hPa−1 s−1) from 2000 UTC 7 to 0100 UTC 8. The SCA weakened the water vapor transport across itself, leading to moisture accumulation in the upwind area and the moisture merging in the downwind area13,37. Although the intensity of this convergence was weaker than low-level water vapor convergence associated with synoptic-scale weather systems (e.g., low-level jet)25, it was comparable to that observed in local severe rainstorm26,38, and rainstorm induced by typhoon inverted troughs36. Under the influence of the SCA, the spatiotemporal distributions of surface water vapor flux convergence and wind field convergence have been changed; however, they still show a strong correspondence, providing favorable surface water vapor conditions for the heavy precipitation. The water vapor flux (vectors; units: ×10− 3 g cm−1 hPa−1 s−1) and its divergence (shading; units: ×10− 7 g cm−2 hPa−1 s−1), and hourly rainfall (shaded scatter; units: mm) stronger than the RS level at the surface from AWS observations. Conclusions An extremely heavy precipitation occurred in Shanghai, characterized by spatial inhomogeneity distribution, under the influence of residual circulation from Typhoon Fitow (2013) during 1700 UTC 7 to 0400 UTC 8 October 2013. This study analyzed the spatiotemporal correspondence between the surface wind field and heavy precipitation using AWS observations, FY satellite product, typhoon track data, and ERA5 reanalysis. Additionally, diagnostic analysis and mechanism investigations were also presented based on multi-physical quantities, including Q vector divergence, water vapor flux, and water vapor flux divergence. The main conclusions are summarized as follows: The wind field from ERA5 reanalysis revealed that under the influence of residual circulation from Typhoon FItow (2013) at the low-level in Shanghai has changed from NE wind to N wind, with favorable largescale water vapor transport conditions at that time. Wind observations from AWSs further indicated that the SCA impeded the eastward progression of N wind, which deflected and detoured around the SCA as it approached. The low-wind area (below 2 m s−1) over the SCA and the wind deflection in its upwind and downwind areas enhanced surface wind field convergences. This further reinforced maintenance of surface dynamic disturbances and the spatial inhomogeneity distribution of heavy precipitation. Statistical analysis using AWS observations verified that the SCA slowed the eastward progression of N wind front and its associated convective systems, prolonged the residence time of the convective systems in the SCA and its upwind (R1) and downwind (R3) areas, and consequently enhanced the impact of heavy precipitation in surrounding areas. The enhanced surface wind field convergence resulting from wind deflection over the SCA further contributed to the spatial inhomogeneity distribution of heavy precipitation in the SCA, R1, and R3. Multi-physical diagnostic analysis was conducted using AWSs observations and ERA5 reanalysis to further investigate the physical mechanisms underlying the spatial inhomogeneity distribution and evolution of heavy precipitation. Strong surface Q vector convergence was mainly promoted by the wind field convergence between N wind from the residual circulation and NE wind from the sea, as well as that shown at 700 hPa. Therefore, surface Q vector convergence maintained simultaneously around the SCA when N wind persisted in this area. The Q vector convergence decayed rapidly after N wind predominated the entire city, and the heavy precipitation weakened within 2 h. The surface Q vector convergence can serve as a reliable indicator for development of heavy precipitation. Additionally, surface water vapor transport is also mainly influenced by the surface wind field, resulting in moisture retention in the upwind area of SCA and water vapor convergence in its downwind area. The analysis of surface wind fields based on AWS observations provides valuable insights into investigating the impacts of Shanghai urban underlying surface on the local wind field and water vapor transport under influence of typhoon residual circulation, as well as how this further contributed to spatial inhomogeneity distribution of the heavy precipitation. Future research should focus on the 3-D analysis, quantitative assessment, and numerical simulation of the urbanization effect on convective systems associated with typhoon residual circulation and other synoptic systems. Acknowledgements The authors benefited from the discussion with Xu Tong, Ao Xiangyu of Shanghai Typhoon Institute, China Meteorological Administration, and Ma Bingxin, Hu Yan of Shanghai Ecological Forecasting & Remote Sensing Center. Author contributions Y. Tang, J. Zhang, and C. Yue designed conceptualization, methodology andwrote the manuscript; Y. Tang, J. Sun, and Z. Han performed AWSs data processing; Y. Tang, J. Zhang, and W. Gu performed ERA5 data processing; Y. Tang, andY. Yao performed intensity and the track data and FY-2E satellite data processing; C. Yue, L. Deng, and Y. Tang reviewed the manuscript. Funding This work was supported by the National Natural Science Foundation of China [grant numbers 41875059, 41875071]; the Shanghai Natural Science Foundation [grant number 22ZR1456100, 21ZR1457700]. Data availability The intensity and the track data of Typhoon Fitow (2013) are available from the China Meteorological Administration (https://tcdata.typhoon.org.cn). The FY-2E satellite infrared bright temperature products are available from National Satellite Meteorological Center (https://satellite.nsmc.org.cn/). The ERA5 hourly relative humidity and horizontal wind data on pressure levels are available from the European Centre for Medium-Range Weather Forecasts (https://cds-beta.climate.copernicus.eu/datasets). The AWS datasets analyzed during the current study are available from the corresponding author upon reasonable request. Declarations Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement The intensity and the track data of Typhoon Fitow (2013) are available from the China Meteorological Administration (https://tcdata.typhoon.org.cn). The FY-2E satellite infrared bright temperature products are available from National Satellite Meteorological Center (https://satellite.nsmc.org.cn/). The ERA5 hourly relative humidity and horizontal wind data on pressure levels are available from the European Centre for Medium-Range Weather Forecasts (https://cds-beta.climate.copernicus.eu/datasets). The AWS datasets analyzed during the current study are available from the corresponding author upon reasonable request.