Simulation of summer climate over East China by convection-permitting regional air-sea coupled model

doi:10.21203/rs.3.rs-4834307/v1

Download PDF

Research Article

Simulation of summer climate over East China by convection-permitting regional air-sea coupled model

https://doi.org/10.21203/rs.3.rs-4834307/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

In this research, the Coupled-Ocean-Atmosphere-Wave-Sediment Transport (COAWST) Model was utilized to study the summer climate over eastern China and its adjacent ocean from 2009 to 2018. The COAWST model, equipped with a convection-permitting resolution of approximately 4km, effectively replicated the patterns of precipitation for both land and ocean, including sub-daily extreme precipitation, and the diurnal cycle. The COAWST model offers accurate simulations of the land-sea contrast in terms of the diurnal cycle of precipitation, showing the peaks of rainfall over land and sea in the afternoon and morning respectively. The model exhibited higher skill in extreme precipitation intensity over the ocean, while it was more skillful in wet frequency over the land. The overestimation of land precipitation can be attributed to stronger water vapor flux convergence and latent heat flux in the model. The inclusion of ocean module affects the ocean condition through fresh water injection, which contribute to the sea surface temperature (SST) bias in the simulation. This article highlights the strong performance of the coupled atmosphere-ocean model in simulating coastal precipitation of eastern China.

coupled atmosphere-ocean model

convection-permitting

extreme precipitation

1. The convection-permitting regional air-sea coupled model effectively simulates the climatological characteristics and extreme precipitation in eastern China.

2. The coupled model shows better skill in simulating extreme hourly precipitation over ocean than over land.

3. Stronger water vapor flux convergence and latent heat fluxes in the model lead to overestimation of precipitation over land.

The East China Coast (ECC) is home to more than 40% of the country’s population, and its offshore and coastal zones are crucial for China’s economy and production (Wang et al. 2020). The ECC is vulnerable to extreme rainfall, with floods triggered by monsoon rainfalls being commonplace (Kundzewicz et al. 2020). Climate variability of the ECC are influenced not only by remote forcing, but also by regional air–sea interaction, which significantly affects short-term climate variations in China. The East Asian monsoon is strengthened by air–sea interaction through a positive feedback process, leading to increased rainfall (Hu et al. 2012). In the East China Sea (ECS), heavy (weak) precipitation over the central part is related to horizontal moisture convergence influenced by strong (weak) SST fronts (Sasaki et al. 2018). Land-Sea breeze plays an important role in the summer sea-air energy transfer in the form of turbulent heat fluxes, affecting the local climate. (Shen et al. 2024). Therefore, careful consideration of air-sea interaction is essential for understanding and predicting local climate and extreme precipitation in the ECC.

High-resolution regional climate models (RCMs) are commonly used to study characteristics and variability of regional climates. Sub-daily and local extreme precipitation, which can strongly impact society, require very high-resolution climate models to provide reliable simulations (Thomassen et al. 2021). RCMs with a grid-spacing of approximately 10–50 km, and Convection-Permitting Models (CPMs) with a grid-spacing of less than 4 km, are two commonly used types of regional models for simulating sub-daily and extreme precipitation. While CPMs are a type of RCM, they differ by explicitly resolving convection.

Studies have shown that CPMs perform better than RCMs in simulating the diurnal cycle of precipitation, and the intensity, frequency and duration of sub-daily rainfall, among other aspects (Caillaud et al. 2021; Dong et al. 2022; Guo et al. 2020). In complex coastal terrain, CPM provide added value compared to coarser resolution models (Belušić et al. 2023). However, CPMs may overestimate onshore winds due to insufficient convergence in coastal regions caused by a decreased temperature difference between land and sea (Lu et al. 2022). The deficiencies in regional atmospheric models are likely related to neglected air-sea interactions and feedback, which can substantially influence the spatial and temporal structure of regional climates, are often neglected in stand-alone model (Ho-Hagemann et al. 2017).

Recent research highlights the importance of coupling different Earth components in regional climate simulations, particularly ocean-atmosphere coupling over East Asia (Giorgi & Gao, 2018; Peng et al. 2023). The coupled model has been reported to enhance the simulation of precipitation variability at both interannual and intraseasonal scales (Dai et al. 2018). By including ocean-atmosphere interaction, the model can reduce the dry bias of atmosphere-only model, which lack adequate moisture transport from the seas to the land (Ho-Hagemann et al. 2017). Interactions of heat, moisture and momentum at the air–sea interface modify low-level atmospheric stability, significantly impacting the intensity of atmospheric convection and precipitation (Rainaud et al. 2017). Previous studies have demonstrated improved simulations of the summer monsoon circulation over East Asia through regional air-sea coupling (Zou & Zhou 2013; Zou et al. 2016). Including are-sea coupling can also reduce bias in ocean convective precipitation simulation by suppressing convective activity through negative feedback mechanisms (Mishra et al. 2022).

In terms of extreme precipitation, coupled models show advantages in many aspects. Coupled models successfully project extreme coastal El Niño events in Peru, where a positive feedback between coastal warming, atmospheric deep convection and coastal winds leads to extreme coastal rainstorms (Peng et al. 2019). Mesoscale Convective Systems are also highly sensitive to sea surface conditions (Rainaud et al. 2017). Coupled simulations indicate that extreme precipitation is likely influenced by low-level jets associated with changes in surface heat fluxes (Berthou et al. 2016; Sauvage et al. 2021). The inclusion of air–sea coupling can improve simulations of circulation anomalies associated with extreme heat (Li et al. 2018). Along the Chinese coast, extreme climate researches also focus on air-sea coupling simulations. Increased oceanic moisture transport through anomalous southerly winds during east Asian monsoon contributes to the growth of precipitation extremes under warming conditions (Liu et al. 2023). Heavy rain triggered by coastal boundary layer jets is reasonably reproduced by the coupled atmosphere-ocean model due to accurate simulation of land-sea thermal contrast (Xu et al. 2023).

Previous studies underscore the significance of atmosphere-ocean coupling, but there is still a limited number of coupled simulations for ECC, especially at convection-permitting scale. The present study aims to (1) evaluate the performance of the regional air-sea coupled model in simulating the precipitation over ECC at convection-permitting resolution, (2) explore how ocean–atmosphere coupling impacts extreme precipitation, and (3) characterize the moisture transport and circulation regimes associated with precipitation. In this study, we utilized a convection-permitting-scale ocean-atmosphere coupling simulation for evaluation. A detailed description of the coupled model and the experimental configurations is introduced in section 2. Section 3 discusses the simulated precipitation, including climatology, diurnal cycle, and extreme. The discussions and conclusions are provided in sections 4 and 5, respectively.

2.1 Model description and configuration

The regional air-sea coupling model used in this work is the COAWST model. The COAWST model includes an atmosphere model, the Weather Research and Forecasting (WRF) model; an ocean model, the Regional Ocean Modeling System (ROMS); a wave model, the Simulating Waves Nearshore (SWAN) model; and a sediment-transport model, the Community Sediment Transport Model (CSTM) (Warner et al. 2010). The model coupling toolkit (MCT) is used as the coupler to exchange data fields between different components (Warner et al. 2008). This work uses the coupled atmosphere module (WRF) and ocean module (ROMS).

In the coupled system, the WRF runs at 4 km convection-permitting resolution with 359x295 grids, covering most coastal region in East China (Fig. 1). It includes 35 sigma vertical levels with the model top at 10hPa. The WRF model is driven by the fifth generation of global reanalysis from the Europe Centre for Medium-Range Weather Forecasts (ECMWF) (ERA5, Hersbach et al. 2020), with a temporal resolution of 3 hours and a spatial resolution of 0.25^o. The physical parameterization schemes used in the WRF simulation include the Thompson microphysics scheme (Thompson et al. 2008), the Mellor-Yamada-Janjic TKE scheme (MYJ) planetary boundary layer (PBL) parameterization (Janjic 1994), the RRTMG shortwave and longwave radiation schemes (Iacono et al. 2008), and the NOAH land surface model (Tewari et al. 2004). The convection parameterization is not used at convection-permitting scale.

The ROMS regional ocean model is a three-dimensional numerical model and free surface that solves the Reynolds-Averaged Navier–Stokes equations (RANS). It operates under the Boussinesq and hydrostatic assumptions and employs a split-explicit time stepping algorithm (Shchepetkin & McWilliams 2005). In this study, ROMS uses a 4-km horizontal resolution grid with 31 vertically stretched layers. The vertical coordinate transformation equation is an improved formulation (Shchepetkin & McWilliams, 2005) and the stretching function is Shchepetkin & McWilliams (2009) improved double stretching function. ROMS is driven by Simple Ocean Data Assimilation (SODA) data, which includes current, salinity, and temperature fields at a temporal resolution of 5-day and spatial resolution of 0.5 degrees.

The coupled simulations are initialized on May 20th and integrated until September 1st for each year from 2009 to 2018, with the first 10 days are considered for spin-up.

2.2 Validation data

Two datasets are used to evaluate the COAWST model’s skill in reproducing precipitation. One dataset is the merged satellite gridded precipitation produced by the Integrated Multi-satellite Retrievals for Global Precipitation Measurement (IMERG; Huffman et al. 2015). IMERG provides near real-time estimates of Earth’s precipitation updated every half-hour, with a horizontal grid spacing of 0.1^o. It is particularly valuable over areas lacking ground-based precipitation-measuring instruments, such as oceans and remote areas. This study utilizes gauge-adjusted estimates from the Final runs of IMERG V06. The second dataset consists of the daily gauge observations provided by the National Meteorological Information Center (NMIC) of the China Meteorological Administration (CMA). This dataset is derived from a dense national network of over 2400 gauges and undergoes rigorous quality control procedures (Han et al. 2023).

For sea surface temperature (SST), this study utilizes a gridded dataset with a horizontal resolution of 0.05 degrees from Group for High-Resolution Sea Surface Temperature (GHRSST; Martin et al. 2012; Good et al. 2020). GHRSST is an international organization dedicated to developing and validating satellite-based methods for measuring sea surface temperature (SST) at high spatial resolution.

2.3 Evaluation methods

The climatology indices are assessed using the root mean squared errors (RMSEs), and spatial or temporal correlations (CORR) between the simulations and observations. The RMSE and CORR are calculated by:

$$\:\text{R}\text{M}\text{S}\text{E}=\sqrt{\frac{\sum\:{\left(m-o\right)}^{2}}{N}},$$

$$\:\text{C}\text{O}\text{R}\text{R}=\frac{\sum\:\left({m}_{i}-\stackrel{-}{m}\right)\bullet\:\sum\:\left({o}_{i}-\stackrel{-}{o}\right)}{\sqrt{\sum\:{\left({m}_{i}-\stackrel{-}{m}\right)}^{2}\bullet\:\sum\:{\left({o}_{i}-\stackrel{-}{o}\right)}^{2}}}\:,$$

where $\:\stackrel{-}{m}$ is the averaged value of model, $\:\stackrel{-}{o}$ is the average value of the observation, and N is the number of grid numbers.

We use a list of hourly extreme precipitation indices to characterize the rainfall attributes (Table 1), including frequency, intensity and duration (Li et al. 2022). R10 and R20 are frequency-related indices that measure the frequencies of extreme precipitation exceeding 10 or 20 mm/hr. Intensity-related indices include Rx1hr, Rx3hr, and Rx6hr, which measure the maximum rainfall, and RnnI and RnnpTOT, which are based on extreme thresholds. The precipitation thresholds are calculated using all hourly records gathered during the summer of 2009–2018 to avoid underestimating extreme precipitation in the validation data (Yuan et al. 2020; Kendon et al. 2012). MeLWS and MxLWS are duration-related indices that indicate the average and maximum lengths of consecutive wet hours. Duration is defined as the number of hours between the start and end of an event, with rainfall occurring after a pause of 2 hours or more deemed a separate event (Xiao et al. 2018).

Table 1

Extreme precipitation indices used in the study.
Index	Definition	Units
Rx1hr	Maximum 1 h precipitation	mm
Rx3hr	Maximum 3 h precipitation	mm
Rx6hr	Maximum 6 h precipitation	mm
R10	Numbers of hours with precipitation > = 10mm	hours
R20	Numbers of hours with precipitation > = 20mm	hours
MxLWS	Maximum number of consecutive wet hours	hours
MeLWS	Average number of consecutive wet hours	hours
RnnI	Average intensity of wet hours > = nn% percentile	mm/h
RnnpTOT	Contribution to total precipitation from very wet hours	%

3.1 Climatology

Figure 2 shows the 10-yr (2009–2018) averaged summer mean precipitation and SST from the observations and COAWST simulation. The gauge-observed precipitation decreases from south to north, with a precipitation center in the mid- south (9.0mm/day), and ranges from 5.0 to 8.0 mm/day over most land areas. IMERG observations reveals similar spatial pattern of land precipitation, with precipitation center in the mid- south and maximum precipitation exceeding 7.0 mm/day. Over the ocean, the main precipitation belt is observed east of 125^oE, ranging from 7.0 to 11.0 mm/day, while low precipitation areas are observed over the southern nearshore ocean, with precipitation rates below 5.0 mm/day. The COAWST model successfully reproduces the spatial pattern of summer precipitation over both land and ocean areas, with a spatial CORR higher than 0.61 and an RMSE of approximately 1.50 mm/day. However, it shows wet biases over land, overestimating precipitation by more than 20%, with the maximum relative bias reaching 40% over southeastern coastal regions. The simulated main precipitation belt is located between longitudes 125^oE and 129^oE, with a precipitation rate of about 8.0 mm/day, consistent with observations. The COAWST model also reproduces the distribution of low precipitation over the ocean, but tends to underestimate the precipitation south of 28^oN with a relative dry bias of -20%.

Regarding the distribution of summer mean SST, observations indicate warm water (28°C) in the southeast ocean, with SST decreasing from southeast to northwest. There is a noticeable southward extension of cooler temperatures near the coast. The COAWST model reproduces the spatial pattern of summer mean SST with a CORR of 0.79 and an RMSE of about 1.48 ^oC, although it shows warm biases over most ocean regions, with a warm bias below 1.5 ^oC. Overall, the COAWST model can simulate the distribution of summer mean precipitation and SST over East China.

To evaluate the model’s performance in simulating interannual variation, the temporal RMSE and CORR of summer mean precipitation and SST between the observations and simulations are calculated and shown in Fig. 3. The results show that the COAWST model can effectively simulate the inter-annual variation of summer mean precipitation over East China, with temporal correlation exceeding 0.6 over most areas. The CORR for precipitation over the ocean generally exceeds 0.6, passing the 95% confidence level (CORR = 0.55). However, large RMSEs of precipitation variation, larger than 3.0 mm/day, can be found over coastal land areas and the most eastern ocean. For the SST, the CORR exceeds 0.7 and reaches 0.9 in the eastern ocean, except in some local offshore areas, indicating that the model reproduces the inter-annual variation of summer mean SST. The RMSEs for SST are mainly located in the ocean areas north of 33^oN and along the coastal areas with the RMSEs around 1.8°C.

Figure 4 shows the inter-annual variations of regionally averaged precipitation and SST over ocean and land areas. Over the land, the average summer daily precipitation is about 7 mm/day according to observations, with the lowest recorded precipitation occurring in 2013. The simulated precipitation over land is approximately 1 mm/day higher than the observation each year, with a RMSE of 1.14 mm/day and a CORR of 0.97. The observed precipitation over the ocean has a greater amplitude than over land, with the lowest precipitation occurs in 2013 and the highest in 2015. The COAWST model successfully captures the variability of ocean precipitation, as evidenced by an RMSE of 0.91 mm/day and a CORR of 0.94. The primary deviation occurs during 2014–2015, when the model underestimates precipitation by approximately 2 mm/day. Regarding SST, observations indicate an upward trend of 0.1°C/year during 2009–2018. This trend is also shown in the simulation with a similar slope, though the simulated SST is ~ 1 ^oC warmer. In terms of average precipitation and SST, the COAWST simulation demonstrates its reliability in reproducing the observed distributions.

To gain further insights into the seasonal evolution of rainfall systems during the summer, latitudinal-temporal cross sections of precipitation over land and ocean are shown in Fig. 5. The observations indicate that the main rain belts propagating from 26^oN to 36 ^oN from 11 Jun to 21 July over both land and ocean area. These systems are characterized by relatively high precipitation intensity, often surpassing 10 mm/day. The rain belt over the ocean is more expansive than that over land, covering a broader swath of latitude on any given day. The simulation captures the evolution of the main rain belts during summer for both land and ocean. However, the simulation shows more frequent precipitation near 20 Jun and after 21 July over land area, resulting in a broader belt with daily rainfall exceeding 14 mm. It also underestimates the precipitation intensity and coverage from 1 July to 11 July over the ocean.

3.2 Diurnal precipitation

In the coastal region, the interaction between land and sea can significantly influence local diurnal variation. Figure 6 shows the spatial distributions of diurnal peak timing for summer hourly precipitation amount (PA), frequency (PF), and intensity (PI). The IMERG observations reveal a distinct dominance of afternoon precipitation over land areas, whereas morning precipitation prevails over oceanic regions. The peak time of PA over land typically occurs between 1500 and 1800 LST. In contrast, over oceanic regions, the peak time of PA is predominantly observed from 0600 to 1000 LST, except for several local parts where the peak time shifts to the nocturnal period. The COAWST model can reproduce the spatial distribution of the diurnal peak in precipitation amount, depicting a prevalence of afternoon rainfall over land and morning precipitation over ocean. The simulated diurnal peak of PA occurs roughly one hour earlier than observed over land, and the early morning peak over ocean area is less pronounced in the simulation.

The peaking time of PF shows a similar spatial pattern to that of PA in the observation, with the afternoon peak over land and the morning peak over ocean. It should be noted that the average diurnal peak of PF occurs at 2200–2300 LST over the offshore area. The COAWST model simulates the peak time of PF one hour earlier over both land and ocean areas. For the southern offshore area, the simulated peak time is at 1700–1900 LST which is about 4 hours earlier than observed.

The distribution of the diurnal peak of PI differs significantly from PA and PF. The peaking time of PI is in the morning over the Yangtze River basin and most ocean areas, while midnight precipitation occurs over land areas north of 33 ^oN and the southern coastal region. The model excels at reproducing the diurnal peak timing of PI over most ocean regions and the Yangtze-River basin. However, it simulates the peak of PI at about 1400–1800 LST over southern coastal regions and land areas north of 33 ^oN, which is about 8 hours earlier than the observation.

Figure 7 displays the regionally averaged normalized diurnal variation of PA, PF and PI for both land and ocean. Observations clearly show a peak of PA in land-based rainfall during the late afternoon around 1600 LST, and an oceanic peak in the early morning around approximately 0700 LST. The COAWST model is capable of simulating the diurnal peaks of PA over both land and ocean areas, though the simulated land rainfall peak occurs around 1500 LST, one hour earlier than the observation. Additionally, the model simulates a larger amplitude of PA over land than the observation. The simulated diurnal variation over the ocean commendably reproduces the observed cycle, in both the peaking time and amplitude. From the observations, PF reaches its peak around 1500 LST, and it shows low amplitude over the ocean. The COAWST model accurately captures the peaking time of PF at 1500 LST with a higher amplitude over land, and skillfully simulates the low variability and morning peak over ocean. For PI, the diurnal amplitude is similar over both land and ocean. The simulated fluctuation of PI over land closely matches the observation in terms of peak time and amplitude. However, the coupled model does not capture the peak time of PI over ocean.

The duration of rainfall events significantly impacts their diurnal peaks, with short-duration rainfall typically peaking in the late afternoon and long-duration rainfall predominantly featuring nocturnal peaks (Yu et al. 2014). Figure 8 illustrates the diurnal distribution of precipitation events of different durations for the PA, PF, and PI. Based on IMERG observations, most precipitation events last no more than 7 hours, with the maximum PA over land starting at 1200 LST and lasting for 2–5 hours. The diurnal PA over the ocean usually lasts 3–5 hours from the observation. In the COAWST simulation, precipitation events are of shorter duration. Most PA over land occurs between 1000 and 2000 LST, with a duration of 1–3 hours, while the ocean PA is mainly nocturnal, lasting for 2–3 hours. Regarding the observed precipitation frequency (PF), the most common rainfall events last for 1–3 hours. The COAWST model also simulates a high frequency of short-duration events. Regarding PI, observations indicate a high proportion of precipitation lasting for around 5 hours, and the model reveals that the heaviest rainfall events last for around 4 hours.

3.3 Extreme precipitation

To further reveal the regional characteristics of sub-daily and extreme precipitation, various hourly extreme precipitation indices are assessed. Figure 9 shows the probability density functions of the sub-daily extreme precipitation indices from IMERG and the simulation. The observed annual maxima indices (Rx1hr, Rx3hr, and Rx6hr) indicate relatively low intensity over the land. The primary frequency of observed Rx1hr has an intensity between 20–40 mm/h over the land, while the highest frequency of Rx1hr over the ocean exceeds that over land, ranging from 30–60 mm/h. The COAWST model clearly overestimate Rx1hr over both land and ocean areas. The simulated terrestrial Rx1hr has intensities of up to 60–80 mm/h at the highest frequencies, while the simulated marine precipitation shows intensities of about 50–70 mm/h at the highest frequencies. Observations suggest that the frequency distribution of Rx3hr is similar to that of Rx1hr, with the most common precipitation intensities ranging from 50–70 mm/3h for Rx3hr over land and approximately 100 mm/3h over the ocean. The COAWST model overestimates the intensity of Rx3hr, with the most common intensity ranging between 100–120 mm/3h over both land and ocean. Over land, observations show that the maximum frequency of Rx6hr is about 7%, with rainfall ranging from 80 to 90 mm. Over the ocean, the highest frequency of Rx6hr ranges from 100 to 120 mm/6h, occurring about 4% of the time. The model roughly simulates the distribution of Rx6hr over land and sea, but the simulation shows higher precipitation intensity (120–140 mm/6h) than the observations. Regarding the duration of continuous rainfall, observations indicate that the frequency of MxLWS over land peaks at around 20 hours, while over the ocean it reaches about 30 hours. However, the COAWST model significantly underestimates the duration of precipitation especially over ocean, with the highest frequency of MxLWS being less than 20 hours. The frequency distribution of MeLWS is similar to that of MxLWS. Observations show that the average duration mainly falls within the range of 4–6 hours, with longer durations occurring over the sea. The simulated results show that MeLWS is generally within 5 hours over both land and ocean.

The precipitation threshold is used to evaluate the model's performance in simulating hourly extreme precipitation. The percentile-based thresholds, which include all hours in the research period regardless of whether they are wet or dry, containing all hours in the research period (Kendon et al. 2012), are illustrated in Fig. 10. As shown in the IMERG data, there is hardly any hourly precipitation below the 95% threshold, with the most extreme precipitation over land reaching nearly 30 mm/h at the 99.99% threshold. The COAWST model tends to overestimate land rainfall, showing a steep increase at the 99.99% threshold, reaching around 50 mm/h. Over the ocean, the model skillfully replicates the characteristics of the hourly extreme precipitation threshold, with the 99.99% threshold near 50 mm/h.

To assess the frequency of hourly extreme precipitation the indices, R10 and R20 are utilized and shown in Fig. 11. From the observations, R10 over land averages between 6–10 hours during summer, with the southwestern region being the epicenter of high frequency, reaching beyond 14 hours. Over the ocean, there is a distinct gradient increase from west to east, and R10 exceeds 18 hours in the eastern maritime areas. The coupled model accurately captures the oceanic gradient of R1, but clearly overestimates the frequency particularly over land. The simulated R10 exceeds 16 hours over land which is about 10 hours longer than the observation. The observed R20 is less than 2 hours over land, with the high values exceeding 7 hours mainly located over the eastern ocean areas. The COAWST model simulates the R20 over ocean well, with the maximum R20 (> 7 hours) in the eastern ocean regions. However, it clearly overestimates the R20 over land, ranging from 5 to 10 hours, indicating a wet bias of approximately 4 to 6 hours.

We consider 99.9% as the extreme threshold (R999I and R999pTOT) to investigate the intensity and volume of very extreme rainfall (Fig. 12). The observed R999I clearly shows high intensity over ocean with the maximum intensities larger than 55 mm/h located in the northern and eastern parts of the ocean, and low intensity over land, which is less than 25 mm/h over most land areas. The COAWST simulation shows a considerable overestimation of more than 60% is found over land, whereas the simulated R999I is relatively underestimated over the ocean, with a bias of under 20%. For R999pTOT, it is observed that precipitation events exceeding the 99.9% threshold contribute 6–10% in most regions over land. Over the ocean, the R999pTOT is higher than 16% in the nearshore areas. The coupled model accurately represents the R999pTOT over the northwestern land and nearshore ocean, but it overestimates the R999pTOT by more than 5% over the southern ocean and western land area.

As short-duration precipitation events are the most common and extreme rainfall contributes significantly to the total amount, thus the relationship between extreme precipitation and its duration is also assessed. Figure 13 illustrates the frequency distribution of extreme precipitation events with different durations across different thresholds. Observations indicate that the majority of extreme precipitation events last between 1 and 3 hours. For all precipitation thresholds, the proportion of 1-hr events exceeds 20% over land. Extreme precipitation events lasting 2 hours also account for more than 20% at most thresholds. Precipitation durations for the 99.9% and above thresholds are mainly less than 5 hours. The proportion of precipitation events with longer durations increases with lower thresholds. From the observation, extreme precipitation events over the ocean tend to have longer durations than those over land, but 1-hour precipitation is still predominant. The COAWST model can simulate the significant proportion of 1-hour rainfall events well both over land and ocean. But it overestimates the frequency of 1-hour precipitation to some extent. For precipitation lasting 2–3 hours, the model significantly underestimates the frequency of precipitation events at thresholds of 99% and above. The model also underestimates the frequency of precipitation events lasting 3 hours or longer at each threshold, but to a lesser extent. Despite the model's tendency to simulate shorter-duration rainfall events compared to the observation, COAWST accurately reflects that ocean precipitation events last longer than those over land.

As can be seen from the previous section, the simulation results indicate an overestimation of summer precipitation amount and intensity over land, and a relatively small deviation over the ocean. The water vapor flux divergence (WVFD), which is closely related to precipitation, is shown in Fig. 14. From ERA5, it is clear that significant water vapor convergence occurs in the summer, with stronger convergence in the southern part of the land and the central part of the ocean. The southern coastal land and its adjacent sea exhibit water vapor divergence. The COAWST simulation shows a spatial distribution of water vapor convergence similar to ERA5, with stronger convergence over land, particularly in the southern regions. The WVFD varies significantly over land in the simulation, indicating increased local air ascending and descending motions leading to the formation of robust convective activity. The difference in wind speed suggests stronger divergence in the eastern ocean. The overestimated southerly wind in the air-sea model has also been identified by Liu et al. (2023). This bias may be related to excessively powerful Western Pacific subtropical high and local circulation errors.

In the ocean-atmosphere coupled model, heat exchange between components is achieved through heat flux (Gentemann et al. 2020). The intensity and distribution of precipitation are also influenced by the latent heat flux. The distribution of latent heat fluxes in the study area during summer is shown in Fig. 15. The largest latent heat fluxes are observed over the northern land and southeastern sea, exceeding 100 W/m². Latent heat fluxes in the northwestern region of the ocean are below 60 W/m². The model generally overestimates the upward latent heat flux by more than 20 W/m² in most areas of the study region, except for a clear underestimation over the southern offshore ocean. In coastal areas, the latent heat flux in the model is overestimated. This overestimation may be due to several factors. Higher wind speeds in the model promote water vapor transport and evaporation, leading to an increase in the latent heat flux (Drake et al. 2022). Tidal movements may cause turbulent movements at the sea surface as well as alternating wet and dry conditions along the seashore, increasing water evaporation and gas exchange (Geng et al. 2016).

Regarding ocean currents and salinity, the simulated surface current speed and salinity are higher than the observation (Fig. 16). Higher salinity and flow velocity generally lead to a greater release of latent heat (Mor et al. 2018). Under high salinity conditions, seawater has a relatively low heat storage capacity. This means that temperature changes will be more pronounced compared to freshwater under similar heating or cooling conditions. The increased turbulence caused by high current speeds can enhance the mixing of surface water, resulting in more heat being transferred to the surface and released into the atmosphere as latent heat. The high salinity in the simulation may be associated with several factors. Faster current carrying high-salinity seawater to low-salinity areas, a process that may not be accurately captured by the model. The absence of freshwater input from estuaries in the model could contribute to maintaining higher salinity levels. Additionally, the dry bias of ocean precipitation in the model hinders the dilution of salinity. Overall, these factors indicate that while the model captures some aspects of ocean currents and salinity, discrepancies exist compared to observations, particularly in terms of current speeds and salinity levels.

In this study, we conducted 10-year air-sea coupling numerical simulations at convection-permitting scale during summer to investigate diurnal cycle and extreme precipitation simulation performance over the coastal region in East China. The results indicate that the air-sea coupling model effectively simulated the spatiotemporal distribution of precipitation in the study region, showing a strong correlation with the observations both over the sea and land. The model also replicates the differences in precipitation patterns between land and sea.

The COAWST model effectively captures the land-sea contrast in precipitation between land and sea in diurnal variation. Typically, land areas experience peak rainfall in the afternoon, while the ocean exhibits a morning peak. The frequency of precipitation plays a more significant role in land-based rainfall, while precipitation intensity is a key factor for oceanic precipitation. This difference is primarily influenced by the prevalence of short or long-duration rainfall events (Zhao et al. 2020). Short-duration rainfall tends to peak in the late afternoon, while long-duration rainfall peaks during the night. However, the coupled model shows some common modeling inaccuracies, such as a wet bias and early afternoon peak times. These inaccuracies may be associated with the higher sensible heat flux of the CPM, which enhances water vapor convergence (Lu et al. 2022).

The COAWST model performs well in simulating extreme precipitation intensity over the ocean, but exhibits some wet bias over land. In terms of wet frequency, model simulations for land and ocean are similar to observations, with a slight overestimation. In addition, the model accurately captures the duration of extreme rainfall events for 1-hour events, but underestimates the duration for longer events. Overall, the model tends to overestimate the intensity of extreme precipitation over land with shorter durations while performing relatively better in simulating extreme precipitation over the ocean.

The significant water vapor convergence in the southern part of the land and the central part of the ocean explains the precipitation biases in the simulations. The excessive formation of convective activity related to extreme hourly precipitation may be caused by intensified upward and downward air motions over the land. The overestimated southerly wind in the air-sea model favor water vapor transport and evaporation, leading to an increase in latent heat fluxes. The coupled model overestimates heat fluxes, especially near the coast and in the northwestern ocean region. Furthermore, the simulated surface current speed and salinity are higher than the observed values, which can promote heat mixing and increase latent heat release.

The impact of air-sea interaction on extreme precipitation in coastal regions is incredibly complex. The combined effects of land-sea breeze, topography, and small-scale hydrological processes constitute complex dynamics that influence coastal precipitation. Future studies should explore the combined influences of these various elements on short-duration precipitation extremes in coastal regions.

Compliance with ethical standards

We declare that we have no conflict of interest.

Funding:

This study was funded by the National Key Research and Development Program of China (grant no. 2023YFF0805404).

Authors’ contributions:

All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by Zhiyu Jiang, Jianping Tang. Shuguang Wang, Guangtao Dong, and Shuya Wang helped perform the analysis with constructive discussions. The first draft of the manuscript was written by Zhiyu Jiang and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Acknowledgments

This study was funded by the National Key Research and Development Program of China (grant no. 2023YFF0805404).

Availability of data and material

The station observations used in this work are available at: http://data.cma.cn/en. The ERA5 dataset used in this work is available at: https://www.ecmwf.int/en/forecasts/datasets/reanalysis-datasets/era5. The IMERG dataset used in this work is available at: https://gpm.nasa.gov/data/imerg.

Belušić A, Belušić D, Telišman M, et al (2023) Evaluation of the near-surface wind field over the Adriatic region: local wind characteristics in the convection-permitting model ensemble. Clim Dyn 1-18.
Berthou S, Mailler S, Drobinski P, et al (2016) Influence of submonthly air–sea coupling on heavy precipitation events in the Western Mediterranean basin. Q J R Meteorol Soc 142: 453-471.
Caillaud C, Somot S, Alias A, et al (2021) Modelling Mediterranean heavy precipitation events at climate scale: an object-oriented evaluation of the CNRM-AROME convection-permitting regional climate model. Clim Dyn 56: 1717-1752.
Dai Y, Li H, Sun L (2018) The Simulation of East Asian Summer Monsoon Precipitation With a Regional Ocean‐Atmosphere Coupled Model. J Geophys Res Atmos 123(20): 11-362.
Dong G, Jiang Z, Wang Y, et al (2022) Evaluation of extreme precipitation in the Yangtze River Delta Region of China using a 1.5 km mesh convection permitting regional climate model. Clim Dyn 59(7-8): 2257-2273.
Drake SA, Rupp DE, Thomas CK, et al (2022) Increasing Daytime Stability Enhances Downslope Moisture Transport in the Subcanopy of an Even‐Aged Conifer Forest in Western Oregon, USA. J Geophys Res Atmos 127(9): e2021JD036042.
Giorgi F, Gao XJ (2018) Regional Earth system modeling: Review and future directions. Atmos Oceanic Sci Lett 11: 189–197.
Geng X, Boufadel MC, Jackson NL (2016). Evidence of salt accumulation in beach intertidal zone due to evaporation. Sci Rep 6(1): 31486.
Gentemann CL, Clayson CA, Brown S, et al (2020) FluxSat: measuring the ocean–atmosphere turbulent exchange of heat and moisture from space. Remote Sens 12(11): 1796.
Good S, Fiedler E, Mao C, et al (2020) The current configuration of the OSTIA system for operational production of foundation sea surface temperature and ice concentration analyses. Remote Sens 12(4): 720.
Guo Z, Fang J, Sun X, et al (2020) Decadal long convection-permitting regional climate simulations over eastern China: evaluation of diurnal cycle of precipitation. Clim Dyn 54(3): 1329-1349.
Han J, Miao C, Gou J, et al (2023) A new daily gridded precipitation dataset for the Chinese mainland based on gauge observations. Earth Syst Sci Data 15(7): 3147-3161.
Hersbach H, Bell B, Berrisford P, et al (2020) The ERA5 global reanalysis. Q J R Meteorol Soc 146(730): 1999-2049.
Hu Y, Zhong Z, Liu X, et al (2012) Influence of air–sea interaction on the simulation of East Asian summer monsoon: a case study. Dyn Atmos Oceans 53: 1-16.
Ho-Hagemann H T M, Gröger M, Rockel B, et al (2017) Effects of air-sea coupling over the North Sea and the Baltic Sea on simulated summer precipitation over Central Europe. Clim Dyn, 49: 3851-3876.
Huffman GJ, Bolvin DT, Braithwaite D, et al (2015) NASA global precipitation measurement (GPM) integrated multi-satellite retrievals for GPM (IMERG). Algorithm theoretical basis document (ATBD) version 4(26).
Iacono MJ, Delamere JS, Mlawer EJ, et al (2008) Radiative forcing by long‐lived greenhouse gases: Calculations with the AER radiative transfer models. J Geophys Res Atmos 113(D13).
Janjić ZI (1994) The step-mountain eta coordinate model: Further developments of the convection, viscous sublayer, and turbulence closure schemes. Mon Weather Rev 122(5): 927-945.
Kendon EJ, Roberts NM, Senior CA, et al (2012) Realism of rainfall in a very high-resolution regional climate model. J Climate 25(17): 5791-5806.
Kundzewicz ZW, Huang J, Pinskwar I, et al (2020) Climate variability and floods in China-A review. Earth-Sci Rev 211: 103434.
Li D, Zou L, Zhou T (2018) Regional air–sea coupled model simulation for two types of extreme heat in North China. Clim Dyn 50: 2107-2120.
Li X, Zhang K, Bao H, et al (2022) Climatology and changes in hourly precipitation extremes over China during 1970–2018. Sci Total Environ 839: 156297.
Liu Z, Lee SS, Nellikkattil AB, et al (2023) The East Asian summer monsoon response to global warming in a high resolution coupled model: mean and extremes. Asia-Pac J Atmos Sci 59(1): 29-45.
Lu X, Wang Y, Qiu Y (2022) A convection‐permitting numerical study of diurnal cycles of pre‐summer rainfall over southern China. Quart J Roy Meteor Soc 148(749): 3677-3693.
Martin M, Dash P, Ignatov A, et al. (2012) Group for High Resolution Sea Surface temperature (GHRSST) analysis fields inter-comparisons. Part 1: A GHRSST multi-product ensemble (GMPE). Deep-sea Res Pt II 77: 21-30.
Mishra AK, Kumar P, Dubey AK, et al (2022) Impact of air–sea coupling on the simulation of Indian summer monsoon using a high-resolution Regional Earth System Model over CORDEX-SA. Clim Dyn 59(9-10): 3013-3033.
Mor Z, Assouline S, Tanny J, et al (2018) Effect of water surface salinity on evaporation: The case of a diluted buoyant plume over the Dead Sea. Water Resour Res 54(3): 1460-1475.
Peng J, Li K, Dan L, et al (2023) Sea–air coupling leads to a decrease in precipitation in East Asia under present day conditions that is partially alleviated in future simulations. NPJ Clim Atmos Sci 6(1): 174.
Peng Q, Xie SP, Wang D, et al (2019) Coupled ocean-atmosphere dynamics of the 2017 extreme coastal El Niño. Nat commun 10(1): 298.
Rainaud R, Brossier CL, Ducrocq V, et al (2017) High‐resolution air–sea coupling impact on two heavy precipitation events in the Western Mediterranean. Quart J Roy Meteor Soc 143(707): 2448-2462.
Sasaki YN, Yamada Y (2018) Atmospheric response to interannual variability of sea surface temperature front in the East China Sea in early summer. Clim Dyn 51(7): 2509-2522.
Sauvage C, Lebeaupin Brossier C, Bouin MN (2021). Towards kilometer-scale ocean–atmosphere–wave coupled forecast: a case study on a Mediterranean heavy precipitation event. Atmos Chem Phys 21(15): 11857-11887.
Shen L, Zhao C, Xu C, et al (2024) Effects of sea land breeze on air‐sea turbulent heat fluxes in different seasons using platform observation in East China Sea. J Geophys Res Atmos 129(5): e2023JD040001.
Shchepetkin AF, McWilliams JC (2005) The regional oceanic modeling system (ROMS): a split-explicit, free-surface, topography-following-coordinate oceanic model. Ocean Model 9(4): 347-404.
Shchepetkin AF, McWilliams JC (2009) Correction and commentary for “Ocean forecasting in terrain-following coordinates: Formulation and skill assessment of the regional ocean modeling system” by Haidvogel et al, J Comp Phys 227: 3595–3624. J Comput Phys 228(24): 8985-9000.
Tewari M, Chen F, Wang W, et al (2004) Implementation and verification of the unified NOAH land surface model in the WRF model. In 20th conference on weather analysis and forecasting/16th conference on numerical weather prediction (Vol. 1115, No. 6, pp. 2165-2170).
Thomassen ED, Kendon EJ, Sørup HJ, et al (2021) Differences in representation of extreme precipitation events in two high resolution models. Clim Dyn 57(11-12): 3029-3043.
Thompson G, Field PR, Rasmussen RM, Hall WD (2008) Explicit forecasts of winter precipitation using an improved bulk microphysics scheme. Part II: Implementation of a new snow parameterization. Mon Weather Rev 136(12): 5095-5115.
Wang K, Feng G, Zhang H, et al (2020) Climate change characteristics and adaptation in the offshore East China Sea from 1979 to 2017. J Coast Res 99(SI): 54-59.
Warner JC, Armstrong B, He R, et al. (2010) Development of a coupled ocean–atmosphere–wave–sediment transport (COAWST) modeling system. Ocean Model 35(3): 230-244.
Xiao C, Yuan W, Yu R (2018) Diurnal cycle of rainfall in amount, frequency, intensity, duration, and the seasonality over the UK. Int J Climatol 38(13): 4967-4978.
Xu J, Zhi X, Sein DV, et al (2023) Predictability of coastal boundary layer jets in South China using atmosphere–ocean coupling. J Geophy Res Atmos 128(16): e2023JD039184.
Yu R, Li J, Chen H, et al (2014) Progress in studies of the precipitation diurnal variation over contiguous China. J Meteorol Res 28(5): 877-902.
Yuan Y, Zhai P, Chen Y, et al (2021) Hourly extreme precipitation changes under the influences of regional and urbanization effects in Beijing. Int J Climatol 41(2): 1179-1189.
Zhao Y, Huang A, Kan M, et al (2020) Characteristics of hourly extreme precipitation along the Yangtze River Basin, China during warm season. Sci Rep 10(1): 5613.
Zou L, Zhou T, et al (2016) Dynamical downscaling of historical climate over CORDEX East Asia domain: A comparison of regional ocean‐atmosphere coupled model to stand‐alone RCM simulations. J Geophy Res Atmos 121(4): 1442-1458.
Zou L, Zhou T (2013) Can a regional ocean-atmosphere coupled model improve the simulation of the interannual variability of the western North Pacific summer monsoon? J Climate 26(7): 2353–2367

Download PDF

Editorial decision: Major Revision
07 Oct, 2024
Reviewers agreed at journal
26 Aug, 2024
Reviewers invited by journal
22 Aug, 2024
Editor assigned by journal
08 Aug, 2024
First submitted to journal
07 Aug, 2024

You are reading this latest preprint version

Simulation of summer climate over East China by convection-permitting regional air-sea coupled model

Status:

Version 1

Abstract

Figures

Key Points

1. Introduction

2. Model, data and methods

2.1 Model description and configuration

2.2 Validation data

2.3 Evaluation methods

3. Results

3.1 Climatology

3.2 Diurnal precipitation

3.3 Extreme precipitation

4. Discussion

5. Conclusions

Declarations

Compliance with ethical standards

Funding:

Authors’ contributions:

Acknowledgments

Availability of data and material

References

Status:

Version 1