1. Introduction
In the context of global warming, the intensity and frequency of extreme weather phenomena are on the rise in many regions of the globe, especially extreme precipitation, which has aroused wide academic attention (e.g., Allan and Soden 2008; Lenderink and van Meijgaard 2010; Donat et al. 2013; Sun and Ao 2013; Asadieh and Krakauer 2015). Frequent occurrence of natural disasters such as floods, landslides, and mudslides caused by extreme precipitation has seriously affected social stability, economic development, and human lives (Milly et al. 2002; Pal and Al-Tabbaa 2009). Especially in rapidly urbanizing China, the loss of natural water bodies and the fragmentation of natural water systems greatly diminish the cities’ capability to withstand extreme storms, thus increasing risks of flash floods, which may cause significant damages and fatalities (Xiao et al. 2016; Zheng et al. 2016; Zhou et al. 2017). Therefore, the forecast of extreme precipitation events is currently the focus of meteorological and climate science studies in China.
Atmospheric water vapor is crucial to both weather and the climate system (Kiehl and Trenberth 1997; Trenberth et al. 2005). Abundant water vapor is a prerequisite for extreme precipitation. Furthermore, the spatial–temporal changes of water vapor are closely interrelated with the formation of precipitation system (Zhai and Eskridge 1997), thus studying the correlation between water vapor and precipitation can provide a reference for precipitation forecasts. Lintner et al. (2011) constructed probability distribution functions of observed anomalous tropospheric column water vapor conditions in the western equatorial Pacific Ocean and discussed their relationship with deep convection. Hottovy and Stechmann (2015) presented a linear stochastic model for the dynamics of water vapor and tropical convection. However, the relationship of cause and effect between water vapor and convection is variable. On the mesoscale, rapid water vapor increases result from deep convective activities triggered by moisture convergence and upward motions (Adams et al. 2013, 2017). While on the large-scale, increases in water vapor can cause increases in deep convective activity (Peters and Neelin 2006; Neelin et al. 2009; Holloway and Neelin 2010). Furthermore, the moisture holding capacity of the atmosphere grows exponentially with temperature following the Clausius–Clapeyron (CC) equation that governs saturated vapor pressure (Xiao et al. 2016). Thus, it can be expected that there will be an increase in water vapor according to the CC equation in the context of global warming. Meanwhile, water vapor has strong feedbacks on global warming as the most abundant greenhouse gas in the air, having a substantial effect on the global climate (Held and Soden 2000). Therefore, the accurate knowledge of the correlation between water vapor and precipitation is of great significance for long-term climate analysis as well as extreme precipitation weather forecasting.
Great efforts have been made in recent decades to explore the correlation between atmospheric water vapor and precipitation. Precipitable water vapor (PWV) is the depth of water in a vertical column of unit area of the atmosphere, if all the water vapor in that column were precipitated as liquid water. Here, we use PWV as the parameter to conduct our correlation analysis. Adams et al. (2013) used 3.5 years of global positioning system (GPS) PWV data to derive a new water vapor convergence time scale to characterize the temporal evolution of deep convection over the tropical continental Amazon region. Li and Deng (2013) found that rainfall occurred mainly during high levels of PWV, but highlighted differences in the PWV evolution caused by different weather systems. Benevides et al. (2015) analyzed the temporal behavior of GPS PWV in many case studies of intense precipitation in the Lisbon area of Portugal, coming to a conclusion that such behavior can correlate positively with the probability of precipitation. Sapucci et al. (2018) evaluated the high-frequency GPS PWV estimates for intense rainfall events during the Cloud Processes of the Main Precipitation Systems in Brazil: A Contribution to Cloud-Resolving Modeling and to the GPM (CHUVA) Vale field campaign in Brazil, indicating a sharp increase in the PWV values before intense rainfall events. Barindelli et al. (2018) found a steep decrease in PWV after the rainfall. Despite the relationship between rainfall intensity and PWV varies for each precipitation event, as well as the time interval between the occurrence of the two peaks, the increase in PWV reaching a peak just before the onset of heavy rain has been well documented in many studies (Van Baelen and Penide 2009; Priego et al. 2017), which can help establish warning systems for heavy rainfall combing surface meteorological parameters (Sharifi et al. 2015; Benevides et al. 2019; Huang et al. 2021).
There have also been many studies focusing on PWV and precipitation over China. For example, Cai et al. (2004), Zhang et al. (2012), and Ma et al. (2015) respectively studied the evolution characteristics and correlation of PWV and precipitation in the Tibetan Plateau, north China, and northwest China, demonstrating the indicative role of PWV on precipitation. Dong et al. (2019) evaluated the response of rainfall to temperature variations in China from available moisture and convective instability based on daily radiosonde data, indicating that rainfall intensity and especially rainfall extremes are predominantly controlled by variations in PWV, with the convective available potential energy (CAPE) playing a secondary role. Ayantobo et al. (2021) investigated the spatial–temporal pattern of PWV and precipitation conversion rate (i.e., the percentage of PWV transferred to precipitation) in China based on reanalysis datasets during 1980–2016, results showing that an increase in precipitation is connected with an increase of PWV and a decrease in precipitation is connected with a decrease of PWV, in most regions. China has vast territory with complex terrain and diverse climatic regimes that generally can be divided into five types, including the tropical monsoon type (TM), the subtropical monsoon type (SM), the midlatitude monsoon type (MM), the temperate continental type (TC), and the plateau type (PL) (CMA 1979). However, detailed discussions on distribution characteristics of correlation between PWV and precipitation under different climate types in China remain rare in current studies. Zou et al. (2012) compared the PWV under different climate types based on the GPS data in China from 2000 to 2004, indicating differences in the PWV values of various climate types. The time series curves fluctuate differently, with both amplitude and peak duration varying. However, the actual precipitation and the correlation between the two were not taken into account. In view of the diversity of climate in China, analysis of the distribution characteristics of the correlation between PWV and precipitation in combination with specific climate types may supplement current studies and be of a specific practical significance in China.
It is noteworthy that, currently in research at synoptic and larger time and space scales, radiosonde data with long-term records and good distribution (Bannon and Steel 1960) are the most widely used data in China. However, because of changes in instrument type, processing strategies, observational practice, or other issues, the inhomogeneity problem has seriously hampered radiosonde usage, resulting in biases in long-term climate trends (Zhai and Eskridge 1996). There are artificial downward trend signals for the PWV measurements from radiosonde stations over China, which also affects the long-term trend estimations based on the reanalysis products due to the assimilation of the uncorrected radiosonde data (W. Zhang et al. 2017). There are also many other water vapor measuring techniques, such as infrared radiometers, which require clear-sky conditions (Divakarla et al. 2006), and satellite microwave radiometers whose observation accuracy is limited due to variations in surface conditions (Deeter 2007). In comparison with traditional water vapor measuring methods mentioned above, ground-based GPS has the advantages of high temporal resolution, all-weather operations, low cost, automation, and long-term homogeneity, from which high-quality PWV data can be retrieved (Bevis et al. 1992; Hein 2020). GPS has made a great contribution to studies of extreme weather and climate events around the world (Bonafoni et al. 2019). However, at present, most studies based on GPS PWV in China only focus on cases at a small temporal scale, and few comparisons of long-term trends of PWV and precipitation have been made. The small temporal scale is not large enough to show the overall long-term trend of PWV and precipitation in China, so long GPS PWV time series can be of great significance for long-term trend analysis, in order to take full advantage of its homogeneous measurements.
Thus, in this study, GPS PWV and the measurements at meteorological stations collocated with the GPS stations are used to analyze the correlation between PWV and precipitation covering the national scale of China and a sizeable period from 1999 to 2015. The specific comparison of correlation under five climate types in China is then discussed, respectively. Additionally, the long-term trend analysis of PWV and precipitation is also conducted by calculating the monthly anomalies and the changes of extremes, respectively. The data and methods used in this paper will be described in section 2. Section 3 will focus on the correlation analysis between PWV and precipitation over China combining climate types and long-term trends, followed by the conclusions given in section 4.
2. Data and method
In this section, we first describe the raw data sources and processing methods. Then the PWV retrieval procedure is briefly introduced. Third, the methods used in correlation analysis and long-term trend analysis are mentioned.
a. GPS PWV data
The 6-hourly GPS PWV data provided by W. Zhang et al. (2017) was directly adopted in this study, which was generated based on the observations of GPS stations from the Crustal Movement Observation Network of China (CMONOC) covering the period from 1 March 1999 to 30 April 2015. The CMONOC is a national geoscience fundamental research facility aimed at monitoring the current intraplate deformation in China, using space geodetic techniques such as GPS (Zhang 2001). The CMONOC project mainly consists of two phases. Phase I was from 1999 to 2010 with about 28 GPS permanent stations over China, and since 2011, CMONOC has entered Phase II with approximately 260 stations built. Considering the need of long-term trend research, 22 stations with good product integrity and a sizeable period of more than 15 years were selected in this study, whose distribution is shown in Fig. 1. The different colors in Fig. 1 represent the different climate types in the area where the station is located.
Geographical distribution of CMONOC GPS stations in China, with different colors corresponding to different climate types. The five representative stations selected in the following analysis are marked with red arrows.
Citation: Journal of Applied Meteorology and Climatology 61, 11; 10.1175/JAMC-D-21-0200.1
The GPS data processing strategy and PWV retrieval method can be briefly introduced as follows where more details can be found in W. Zhang et al. (2017). The three principal variables affecting the accuracy of PWV are zenith total delay (ZTD), station pressure Ps and water vapor–weighted mean temperature Tm. First, the GPS data covering the period from 1999 to 2015 were processed uniformly based on the precise point positioning (PPP) module, using the position and navigation data analysis (PANDA) package (Shi et al. 2008) independently developed by Wuhan University, to get reliable ZTD estimates. The GPS station pressure Ps was acquired through the integrated sources of GPS station meteorological measurements, nearby meteorological station records and ERA-Interim meteorological reanalysis products from the European Centre for Medium-Range Weather Forecasts (ECMWF). Here vertical adjustment was conducted for the pressure from the meteorological station altitude or ERA-Interim surface level to the GPS antenna altitude using the barometric height formula (Torri et al. 2019). The Tm was also obtained using ERA-Interim meteorological reanalysis products. Then we used ZTD, Ps, and Tm to acquire PWV as follows:
b. Precipitation data
The hourly accumulated precipitation observations provided by China Meteorological Administration (CMA) were used in this study with an average annual missing data rate of 0.41%. After checking the extreme values, internal consistency, time consistency, and space consistency, CMA performed quality control on the dataset, and the false precipitation data caused by known input errors were removed by technical means.
Considering that GPS stations are generally not collocated with meteorological stations, the selected meteorological stations are matched with GPS stations on the principle that the horizontal and vertical separation between the two stations are limited within 30 km and 100 m, respectively. According to this principle, 29 meteorological stations are finally matched to 22 GPS stations.
c. Temperature data
Temperature data are used to calculate the thermodynamic expectations of long-term extreme precipitation and PWV changes in the context of global warming. We use the gridded surface air temperature data from the Goddard Institute for Space Studies Surface Temperature Analysis (GISTEMP), with no ocean data included. Considering that temperature change on a regional and local scale does not account for the possibility of remote moisture sources related to large-scale circulation (X. Zhang et al. 2017), we calculate the mean surface temperature change on a global continental scale.
d. Correlation analysis
e. Long-term trend analysis
In the context of global warming, long-term trend analysis of extreme precipitation events based on PWV can help understand the changing regularity. There are various definitions of extreme precipitation events, which may lead to different results of extreme precipitation research (Pendergrass 2018). Many climate studies use quantiles of precipitation to assess the long-term trend changes of heavy precipitation events. However, due to the diversity of climate regions, it is challenging to define proper quantiles of interest. Even if the quantile is calculated for a subset of precipitation events whose precipitation exceeds a certain threshold, changes in precipitation frequency may also interfere with changes in its magnitude and produce misleading results (Schär et al. 2016). Thus, Fischer and Knutti (2016) suggested examining changes in the whole tail of the precipitation distribution rather than for individual quantiles. Guerreiro et al. (2018) extended their work to explore changes in hourly and daily extreme rainfall over the Australian continent during 1966–89 and 1990–2013 with a K-largest method.
The K-largest method is basically equivalent to the traditional method using quantiles. However, it makes up for the shortcomings of the poor signal-to-noise ratio when using an individual quantile. Instead, the entire tail of the data distribution is processed. Unlike fitting a linear trend directly, this method focuses on the segmental study of extreme values. The difference between the data tails of the two periods is taken as the long-term change of extremes, which magnifies the intuitiveness of the changing trend. Additionally, considering the complex spatial patterns of precipitation and PWV changes, the K-largest method takes the spatial average of all stations. Spatial aggregation at continental scales can help reduce the influence of local relative variation patterns and identify a clearer signal in extreme precipitation. Therefore, with reference to this analysis method, we here assess the long-term average changes in the magnitude of extreme hourly precipitation observations and PWV content in China over 1999–2006 and 2007–14. Expected changes of extreme precipitation and PWV are also calculated based on CC scaling to help analyze the association between them, in the hope of obtaining more reliable understanding of extreme weather. The specific process of the K-largest method is introduced as follows:
3. Results and discussion
In this section, we first analyze the geographic distribution characteristics of the PWV and precipitation over China from 1999 to 2015. The correlation analysis between PWV and precipitation is then conducted combining different climate types, followed by discussions on the long-term trend correlation of PWV and precipitation by calculating the monthly anomalies and the changes of extremes, respectively.
a. Basic correlation analysis between precipitation and PWV
1) Correlation distribution characteristics
Considering that the PWV time series has 6-h resolution and the high-frequency noise is large, we conducted the study on a monthly scale. Therefore, the monthly mean value of PWV was calculated for correlation analysis with monthly accumulated precipitation data in the same period.
The geographic distributions of the average monthly PWV and the average monthly accumulated precipitation over China from 1999 to 2015 are shown respectively in Figs. 2a and 2b, from which we can easily find that both PWV and precipitation have the geographic distribution characteristic of a decreasing trend from coast to inland. For PWV, the minimum average value appears in northwestern China (about 5 mm), and the maximum average value in southeastern China (about 45 mm). Similarly, for precipitation, the minimum average value appears in northwestern China (about 10 mm), and the maximum average value appears in south China (about 200 mm). The northwest region is far inland, making it almost unaffected by the monsoon. The low PWV content here may be related to the proximity to mountains and continental source of air. On the other hand, the temperature in southeastern China is relatively high, where abundant water vapor is brought by the summer monsoons from the ocean. Following the Clausius–Clapeyron relation, more water vapor can exist in the atmosphere at higher temperatures so that the PWV content is high. Combining the altitude of the stations in Fig. 2c, it can be seen that the PWV and precipitation measurements at stations with higher altitudes are usually smaller. That is mainly because water vapor rises toward high-altitude regions, during which it condenses into liquid water with the temperature decreasing. Thus, only a small part of water vapor can reach high-altitude regions, leading to less precipitation.
Geographic distribution of the (a) average PWV, (b) average precipitation, (c) height of the stations, and (d) correlation coefficients between PWV and precipitation over China from 1999 to 2015.
Citation: Journal of Applied Meteorology and Climatology 61, 11; 10.1175/JAMC-D-21-0200.1
As can be seen from Fig. 2d, the correlation coefficients R between PWV and precipitation at all 22 stations are above 0.5 with a mean value of approximately 0.73, and all of the correlations have passed the significance test at a 95% confidence level. The results show a significant positive correlation between PWV content and precipitation measurements overall, which is consistent with the results of some previous studies (e.g., Benevides et al. 2015; Sapucci et al. 2018; Ayantobo et al. 2021). Based on the conclusion above, the influence of various climate types on the correlation is also taken into account in this work. Figure 2d shows distribution characteristics of the correlation coefficients distinctively related to the climate types. The values of R range from 0.4 to 0.7 at stations located in regions with TC, TM, and SM climate types, which represents a moderate degree of correlation. For the stations located in regions with PL and MM climate types, the values of R are all above 0.7, representing a higher degree of correlation. The average of monthly mean PWV, the average of monthly accumulated precipitation, and the correlation coefficients at all 22 stations divided by climate types are summarized in Table 1.
Average PWV, average monthly precipitation, and correlation coefficients between PWV and precipitation at 22 stations with different climate types in China from 1999 to 2015. The names of the five representative stations are in boldface type.
According to the grade of precipitation intensity issued by CMA, a torrential rain represents daily accumulated precipitation of more than 50 mm. To further explore the indicative effect of PWV on precipitation, we take 50 mm as the daily accumulated precipitation threshold and compare the PWV contents of all stations at the corresponding time. The PWV box-and-whisker plot shown in Fig. 3a depicts the overall distribution of PWV data, from which we can easily identify the outliers. It can be seen that the lower quartile of PWV in torrential rain (>50 mm) is larger than the upper quartile of PWV in other cases (<50 mm), indicating that there is a preference for higher PWV during torrential rain. Thus, PWV content can be used to distinguish whether there is a large amount of precipitation. However, Fig. 3a shows large tails of PWV during a non-torrential-rain period, which can also be found in the specific PWV time series shown in Fig. 3b. There are still many overlapping PWV values between the torrential-rain and less-rain cases. If the threshold is used alone to forecast precipitation, there will be a high false-alarm rate.
The (a) box-and-whisker plot and (b) time series of PWV contents of all stations during cases with torrential rain and less rain, with the threshold of daily accumulated precipitation being 50 mm. The blue and yellow horizontal lines in (b) represent the PWV upper and lower quartile of torrential-rain and less-rain cases, respectively.
Citation: Journal of Applied Meteorology and Climatology 61, 11; 10.1175/JAMC-D-21-0200.1
2) Representative stations correlation
Five representative stations are selected on the basis of the climate type of the region for further study: Urumqi, Xinjiang (URUM); Lhasa, Tibet (LHAZ); Fangshan, Beijing (BJFS); Wuhan, Hubei (WUHN); and Qiongzhong, Hainan (QION) located in regions with TC, PL, MM, SM, and TM climate type, respectively. The geographic distribution of the five selected stations is presented in Fig. 1. Through the correlation analysis given below in turn (Fig. 4), we can easily find that both PWV and monthly precipitation take the year as the cyclical fluctuation with generally the same changing trends and show distinct seasonal variation. There is usually more precipitation in months with higher PWV content, whereas in months with lower PWV content the opposite occurs.
PWV (blue line) and precipitation (red bar) time series and linear regression fitting at five representative stations, (a) URUM, (b) LHAZ, (c) BJFS, (d) WUHN, and (e) QION, in China from 1999 to 2015.
Citation: Journal of Applied Meteorology and Climatology 61, 11; 10.1175/JAMC-D-21-0200.1
(i) URUM—Temperate continental climate
Station URUM is deep inland in northwestern China with TC climate type, always controlled by continental air mass, where the weather maintains dry and rainless all year round. The value of R is approximately 0.59, having passed the significance test with a 95% confident level, which indicates there is a moderate correlation between PWV and precipitation. As is shown in Fig. 4a, the annual precipitation at station URUM is generally concentrated from May to September, and PWV is generally below 10 mm in months without precipitation. The PWV peak is around 20 mm, appearing in July.
The phases of PWV and precipitation are not completely matched, as the station URUM is located in the Tianshan area, where the water vapor transport condition is unfavorable, blocked by the Tianshan mountains. Since PWV reflects the potential for precipitation, only a fraction of which can be converted into actual precipitation, the mean precipitation conversion efficiency of PWV to actual rainfall is measured by taking the ratio of the accumulated precipitation to the total amount of PWV during the same period and at the same location (Robinson 1965; Ye et al. 2014). Cai et al. (2004) pointed out that the average summer precipitation conversion efficiency in China is different in different regions, with the highest value in the Tibetan Plateau and the lowest value in northwest China. This is related to the water vapor divergence and subsidence, making it less conductive to precipitation formation.
(ii) LHAZ—Plateau climate
Station LHAZ is located in the middle of the Tibetan Plateau with high altitude and PL climate type, where the water vapor content generally decreases with the increase of altitude. The value of R is approximately 0.85, having passed the significance test with a 95% confident level, which indicates a high correlation between PWV and precipitation. As shown in Fig. 4b, the annual precipitation at station LHAZ is generally concentrated from May to September with the peak appearing in July or August, and there is almost no precipitation in autumn and winter. Meanwhile, the corresponding PWV content is also the lowest in January (less than 5 mm) and the highest in July or August (around 20 mm).
The results generally show in-phase changes between PWV and precipitation, especially in the rainy season, in accordance with the results of Liang et al. (2020) that PWV increases rapidly before the peak of precipitation, which is likely dominated by mesoscale convective activities above the GPS receiver triggered by low-level convergence. The peak of PWV occurs just before that of precipitation as the local PWV accumulation can stimulate precipitation occurrences. There is more precipitation in summer over the Tibetan Plateau due to the abundant water vapor provided by the warm and moist airflow from the Bay of Bengal in the Indian Ocean. The lower atmosphere always converges and ascends toward the plateau in summer, with strong water vapor convergence, uplift and condensation effect making the precipitation conversion efficiency very high (Zhou et al. 2012).
(iii) BJFS—Midlatitude monsoon climate
Station BJFS is located in the North China Plain with MM climate type, where the seasonal distribution of precipitation is rather uneven. It is cold and dry in winter, while hot and rainy in summer. The value of R is approximately 0.83, having passed the significance test with a 95% confident level, which indicates a high correlation between PWV and precipitation. As shown in Fig. 4c, the annual precipitation at station BJFS is generally concentrated from April to October with the peak appearing in summer, and PWV is generally below 10 mm in months without precipitation. The PWV peak is around 40 mm, generally appearing in July, which indicates the in-phase changes between PWV and precipitation.
Station BJFS shows distinct characteristics of MM climate type, greatly influenced by monsoons (Huang et al. 2003). In summer, the MM region is dominated by the summer monsoon under the influence of subtropical high over the western Pacific Ocean, with relatively heavy precipitation and high PWV content. In winter, the winter monsoon prevails, affected by the Siberian continental cold high, resulting in relatively small amount of precipitation and low PWV content.
(iv) WUHN—Subtropical monsoon climate
Station WUHN is located in the midlower reaches of the Yangtze River with SM climate type, where the winter is mild with little rain, and the summer is hot and rainy. The value of R is approximately 0.59, having passed the significance test with a 95% confident level, which indicates a moderate correlation between PWV and precipitation. As is shown in Fig. 4d, the annual precipitation at station WUHN is generally concentrated from April to October when the corresponding PWV is also relatively large, and PWV is generally below 30 mm in months without precipitation. The average PWV peak is around 55 mm, appearing in July or August, whereas the precipitation peak usually appears in June, and then there is a significant decline in precipitation.
There is a phase deviation between PWV and precipitation peaks, mainly due to the typical rainy season in the midlower reaches of Yangtze River, well known as the mei-yu season, approximately beginning in June and lasting until early July (Ni and Zhou 2004). During this period, because of the encounter and long-term stalemate of the warm and cold air masses coming from the tropical ocean in the south and continent in the north, respectively, a quasi-stationary front is formed at the intersection, bringing continuously heavy precipitation. This heavy precipitation is greatly affected by the interannual variability of the large-scale circulation. Besides, Chen et al. (2019) pointed out that the water vapor flux in the mid reaches of Yangtze River is mainly characterized by divergence, and the primary source of precipitation is not the transportation of water vapor. In other words, abundant water vapor content is only one of the necessary conditions for heavy precipitation, which cannot determine the actual precipitation independently.
(v) QION—Tropical monsoon climate
Station QION is located in the middle of Hainan island with TM climate type, where the precipitation is abundant and the temperature is high all year. Hainan Island is frequently and seriously affected by tropical cyclones occurring on the tropical ocean, which is the main water source. The value of R is approximately 0.55, having passed the significance test with a 95% confident level, which indicates a moderate correlation between PWV and precipitation. As is shown in Fig. 4e, the precipitation is less from November to April of the following year and more from May to October, with peaks generally concentrating from August to October when the corresponding PWV is also relatively large, with the peak of around 55 mm.
The phases of PWV and precipitation are not completely matched, which is probably because of the complicated topography of Hainan Island, with the middle terrain being higher and the surroundings being lower. Station QION is located in the mountainous area in the middle of Hainan Island, where the topography blocking plays the role of forced uplifting, leading to convective weather and much precipitation on the windward slope of the central and eastern regions (Liu et al. 2011; Houze 2012). According to Li et al. (2015), although there is abundant water vapor all year round in Hainan Island, the spatial distribution of PWV is uneven, affected by terrain and land–sea location. Because of the high altitude and noticeable land–sea difference, the PWV value in the middle of Hainan Island is relatively lower. Thus, PWV and precipitation in this region have different distribution characteristics.
In our work, the relationship between PWV and precipitation is with considered linear correlation, while there are also studies using a nonlinear relationship. Bretherton et al. (2004) used a nonlinear least squares method to fit the relationship between precipitation and column relative humidity (CRH) by an exponential over the tropical oceans. The average precipitation is only weakly sensitive to the CRH up until a threshold, beyond which the precipitation rapidly increases (Peters and Neelin 2006). The rapid increase in precipitation beyond the critical CRH value is more pronounced for stratiform rain, while convective rain displays a weaker nonlinear relationship (Ahmed and Schumacher 2015). The lower free-tropospheric water vapor has a strong effect on deep convection (Tompkins 2001; Derbyshire et al. 2004). Even though the environment tends to be stable on the large-scale, water vapor will contribute to conditional instability through entrainment, which can result in deep convection (Holloway and Neelin 2009). The exponential curve has been mainly used in the studies of the tropical atmosphere (Bergemann and Jakob 2016; Schiro et al. 2016), while most of China is located in the north temperate zone, with a few areas in the south are close to the tropical ocean. For example, Fig. 4e also shows weak exponential trend at station QION. Because of the lower-tropospheric temperature and hence lower saturation vapor pressure at midlatitude regions, precipitation will increase at lower PWV values (Neelin et al. 2009). Thus, it is expected that the relationship between PWV and precipitation remains nonlinear for midlatitude regions. However, the nonlinear trend is not very significant in our results. Actually, there are many factors affecting the relationship between PWV and precipitation, such as topography and seasonality. In summer when conditional instability plays a major role, PWV can cause convective activities (Peters and Neelin 2006). While in winter when baroclinic instability dominates, high PWV results from moisture transported by synoptic cyclones, no longer being the direct cause of precipitation (Catto et al. 2019). We study here mainly on an interannual scale, and the calculation of monthly average value of PWV and monthly accumulated precipitation may have smoothed the nonlinear relationship between PWV and precipitation. Besides, the moisture precipitation relationship is somewhat dependent on the datasets used to derive CRH and precipitation, and the linear fitting is more applicable for our data. In further study, we expect to use PWV data with higher temporal resolution to investigate the PWV and precipitation characteristics on smaller time scales.
b. Long-term trend analysis of precipitation and PWV
Global warming leads to the concentration of moisture in warmer atmosphere that in turn increases the intensity of extreme precipitation events (Zhou et al. 2017). Water vapor has strong feedbacks on global warming as the most abundant greenhouse gas in the air, whose variation strongly correlates with that of precipitation (Held and Soden 2000). Given the great harm and possible growing trend of extreme hourly precipitation in cities, exploring the relationship between PWV and precipitation developing trends can help provide a reference for long-term precipitation forecasting. On the basis of using monthly data directly to study the consistency of the actual fluctuations of PWV and precipitation time series as above, we further use monthly anomalies to show the consistency of the developing trends of the two and then focus on the change of extremes.
Monthly anomalies, which measure the changes in PWV and precipitation relative to the same period in history, can indicate the developing trends of PWV and precipitation in a specific place. Therefore, correlation analysis on monthly anomalies is conducted to help show the indicative significance of PWV for the long-term change of precipitation. First, the monthly mean PWV anomaly time series and monthly accumulated precipitation anomaly time series are obtained by removing the mean value of the corresponding month over the years from the raw time series. The correlation coefficients R of PWV and precipitation anomalies are then calculated at each station, which can show whether the monthly variations of the two are consistent. As is shown in Fig. 5, the R values at most stations in the study region range from 0.2 to 0.6 and have passed the significance test at a 95% confidence level, indicating a certain positive correlation between PWV and precipitation monthly anomaly time series.
Geographic distribution of the correlation coefficient R between PWV and precipitation monthly anomalies over China from 1999 to 2015.
Citation: Journal of Applied Meteorology and Climatology 61, 11; 10.1175/JAMC-D-21-0200.1
Stations LHAZ and QION are selected as representative stations, whose PWV and precipitation monthly anomaly time series can be seen in Fig. 6. We can easily find that PWV and precipitation monthly anomalies still show some relationships after removing the seasonal variation. Months with positive PWV anomalies are usually accompanied by positive precipitation anomalies, which means higher PWV content relative to historical period corresponds to higher precipitation to a certain extent. The law is also valid for negative anomalies. That is in line with the study results of Shi et al. (2020) over the China–Indochina Peninsula region, and our result expands to the national scale of China. In other words, PWV monthly anomalies can be used to measure the degree of drought in a certain place during a certain period of time, and then help show the long-term changing trend of precipitation. Nevertheless, different from monthly data, the correlation of monthly anomalies does not show distinctive distribution characteristics related to climate zones.
PWV (blue line) and precipitation (red bar) monthly anomaly time series at stations (a) LHAZ and (b) QION from 1999 to 2015.
Citation: Journal of Applied Meteorology and Climatology 61, 11; 10.1175/JAMC-D-21-0200.1
To further compare the overall long-term trends of extreme precipitation and PWV content in the study region, the hourly precipitation data and 6-hourly PWV data in China from 1999 to 2014 are divided into two 8-yr periods, and the average changes in the magnitude of the extremes are calculated. Long-term changes of precipitation and PWV extremes in China are shown in Fig. 7a, that is, the spatial average difference between the periods of 1999–2006 and 2007–14 at each K-largest group. The labels in the figure represent different K-largest groups. For example, when the label is “K1–K25” the plot shows the average change of the 1st–20th maxima, which can be considered to be the data with the highest “extreme degree.” We also calculated the expected changes in precipitation and PWV based on CC scaling using the change in global mean surface temperature (0.30°C) between the 1999–2006 and 2007–14 periods, shown as the dashed line and dot–dashed line in Fig. 7a. Figures 7b and 7c show the ranges of precipitation and PWV values associated with each K-largest group, respectively.
(a) Long-term changes with different extreme degree in the magnitude of hourly precipitation (blue line) and PWV (red line), and expected changes based on CC scaling (dashed line and dot–dashed line). Ranges of (b) precipitation. (c) PWV values associated with each K-largest group. All changes are shown as spatial means of all stations for each K-largest group across China during 1999–2006 and 2007–14.
Citation: Journal of Applied Meteorology and Climatology 61, 11; 10.1175/JAMC-D-21-0200.1
It can be seen that the long-term changes of precipitation corresponding to different extreme degrees are all positive, with more significant increase at the highest magnitudes (K1–K25). That means the magnitude of extreme hourly precipitation shows an increasing trend on the whole, with the precipitation events with higher extreme degree showing more significant changes. Besides, the increases of extreme precipitation far exceed the expected changes from CC scaling, showing approximately 5 times the expected rate of increase (CC × 5). That may be due to the condensation and latent heat release determined by more extreme precipitation, leading to stronger cloud updrafts feeding back again onto the precipitation formation (Lenderink and van Meijhaard 2010). The latent heat release associated with extreme precipitation is important in maintaining the vertical coupling of low-level convergence and upper-level divergence near the storm center (Juneng et al. 2007). The growth rate of the atmospheric instability will also increase due to the latent heat release (Emanuel et al. 1987), which may strengthen atmospheric circulation and bring a positive feedback for extreme precipitation, thus amplifying the most extreme precipitation events (Pendergrass 2018). Our results are in line with those of Guerreiro et al. (2018) for hourly precipitation observations over Australia, while here we also find that the magnitude of extreme hourly PWV shows similar performance. The increase of PWV is more significant at the highest magnitudes, and the expected changes from CC scaling are relatively gentle. The increases of extreme PWV are overall in line with a sub-CC scaling rate of increase (CC × 0.4), except for the substantially increasing part at the highest magnitudes (K1–K25), which becomes slightly above CC scaling. O’Gorman and Schneider (2009) indicate that the CC scaling of precipitation extremes cannot be easily explained with any basic physical principle, and identified several factors causing deviations from CC scaling, such as changes in the moist-adiabatic lapse rate. In summary, the magnitudes of both extreme precipitation and extreme PWV show an overall long-term upward trend as the climate change causes a shift of environments to moister and warmer. The extreme precipitation–temperature scaling rate of changes can reach superCC scaling, while that of the extreme PWV-temperature is sub-CC overall.
It is noteworthy that the number of the used stations may have effect on the results. Guerreiro et al. (2018) tested the sensitivity of the extreme precipitation changes to the number of gauges and noted that the continental mean changes for precipitation extremes vary widely when using fewer gauges (different values depending on which gauges are chosen). Because of the short time of CMONOC development in China, currently only limited stations can provide relatively long-term data for PWV retrieval. Insufficient Global Navigational Satellite System (GNSS)-based PWV data may lead to inaccurate long-term extreme precipitation–temperature scaling rate, but still show the increasing trend. Despite this, we subdivided the stations into north or south according to climate types to assess whether there is a different pattern of behavior between the midlatitude climate of the north of China and the tropical and subtropical south of China (Fig. 8).
Long-term changes in the magnitude of extreme precipitation (blue line) and PWV (red line) with expected changes based on CC scaling (dashed line and dot–dashed line) in the (a) north and (b) south of China.
Citation: Journal of Applied Meteorology and Climatology 61, 11; 10.1175/JAMC-D-21-0200.1
Results indicate that the overall changes of extreme precipitation and extreme PWV in the north are lower than those in the south, but the extreme precipitation at the highest magnitudes (K1–K25) in the north show more intense increase. Both northern and southern changes in precipitation are well above expectations from CC scaling, with the scaling rate being higher in the midlatitude north (around CC × 9) and relatively lower in the tropical and subtropical south (around CC × 3). That is similar to the results of some existing studies, reporting on various rates of precipitation extremes changes from negative in the tropics and subtropics (Hardwick-Jones et al. 2010) to superCC scaling rates in midlatitudes (Prein et al. 2017). In contrast, the scaling rate of the extreme PWV increases is generally lower in the north (around CC × 0.3) than that in the south (around CC × 0.5). That is because of the smaller absolute increase in precipitation and larger absolute increase in PWV amount in the tropical and subtropical south. In conclusion, though the data duration and the number of the stations may bring uncertainties, the long-term trend analysis based on CC-scaling aids an understanding of the relationship between extreme precipitation and PWV under global climate change.
4. Conclusions
The correlation analysis between PWV and precipitation will contribute to the long-term extreme weather research in the context of global warming. GPS-derived PWV data covering the period from 1999 to 2015 and the measurements at meteorological stations over China were used to conduct the study. Results show that both PWV and precipitation have the geographic distribution characteristic of a decreasing trend from coast to inland. Furthermore, there is a significant positive correlation between monthly PWV mean content and precipitation measurements with a mean correlation coefficient R of 0.73 at all stations.
Given the diverse climate in China, the comparison of correlation among different climate types was then discussed, suggesting that the distribution characteristics of the correlation coefficients are distinctively related to different climate types. The values of R range from 0.4 to 0.7 at stations located in regions with TC, TM, and SM climate types, which represents a moderate degree of correlation. While for the stations located in regions with PL and MM climate types, the values of R range from 0.7 to 1.0, representing a higher degree of correlation. In this work, we chose the simple Pearson correlation coefficients to study the linear relationship between PWV and precipitation, while the correlation of the two may be more nonlinear in some regions, which requires further correlation analysis with more complicated and various methods.
The long-term trend analysis of PWV and precipitation was also conducted, which may contribute to better extreme precipitation research and adaptation to climate change. The correlation coefficients of monthly anomalies in the study region generally ranging from 0.2 to 0.6 do not show distinctive distribution characteristics related to climate types, yet there is still a certain positive correlation between PWV and precipitation anomalies. That is, months with higher PWV content correspond to higher precipitation, to a certain extent, when compared with the same time period. Furthermore, the overall long-term trends of extremes indicate that extremes of both precipitation and PWV are on an upward trend with the most extreme events showing more significant changes. The extreme precipitation–temperature scaling rate of changes can reach superCC scaling, while that of the extreme PWV–temperature is sub-CC overall, and there exist regional differences in the specific changing factor values.
It is noteworthy that, although PWV content has a significant correlation with precipitation, the physical mechanism of PWV and precipitation is complicated, associated with deep convective activities triggered by moisture convergence, and precipitation cannot be determined simply by the PWV value in the short term. Thus, PWV content is just one of the necessary conditions for precipitation instead of the decisive factor. Besides, this work mainly focuses on the long-term correlation and trends of PWV and precipitation on an interannual scale to show the relationship between PWV and extreme precipitation. However, for the short-term precipitation forecast, it is still necessary to pay attention to the increase/decrease trend and amplitude of PWV, combined with other surface meteorological parameters such as the variation in atmospheric pressure. Considering that the monsoon climate in China is dominant, which brings frequent and strong convective precipitation, we expect to study the water vapor–precipitation relationship in the context of monsoon movement on a seasonal scale in future work.
Acknowledgments.
This work was supported by the National Natural Science Foundation of China (41961144015; 42174027); Key Research and Development Program of Guangxi Zhuang Autonomous Region, China (2020AB44004); and the Fundamental Research Funds for the Central Universities (2042022kf1198). The authors thank the China Meteorological Administration for providing precipitation data and ECMWF for providing reanalysis products. The authors also thank David K. Adams and another anonymous reviewer for reviewing this work and providing useful inspiration and guidance. The authors declare no conflict of interest.
Data availability statement.
Precipitation data used during this study are openly available from the China Meteorological Administration (http://data.cma.cn/en), and GPS PWV data for this research are included in W. Zhang et al. (2017). Global surface temperature data provided by NASA Goddard Institute for Space Studies can be accessed online (https://data.giss.nasa.gov/gistemp/).
REFERENCES
Adams, D. K., S. I. Gutman, K. L. Holub, and D. S. Pereira, 2013: GNSS observations of deep convective time scales in the Amazon. Geophys. Res. Lett., 40, 2818–2823, https://doi.org/10.1002/grl.50573.
Adams, D. K., H. M. J. Barbosa, and K. P. Gaitán De Los Ríos, 2017: A spatiotemporal water vapor–deep convection correlation metric derived from the Amazon dense GNSS meteorological network. Mon. Wea. Rev., 145, 279–288, https://doi.org/10.1175/MWR-D-16-0140.1.
Ahmed, F., and C. Schumacher, 2015: Convective and stratiform components of the precipitation-moisture relationship. Geophys. Res. Lett., 42, 10 453–10 462, https://doi.org/10.1002/2015GL066957.
Allan, R. P., and B. J. Soden, 2008: Atmospheric warming and the amplification of precipitation extremes. Science, 321, 1481–1484, https://doi.org/10.1126/science.1160787.
Asadieh, B., and N. Y. Krakauer, 2015: Global trends in extreme precipitation: Climate models versus observations. Hydrol. Earth Syst. Sci., 19, 877–891, https://doi.org/10.5194/hess-19-877-2015.
Ayantobo, O. O., J. Wei, B. Kang, T. Li, and G. Wang, 2021: Spatial and temporal characteristics of atmospheric water vapour content and its relationship with precipitation conversion in China during 1980–2016. Int. J. Climatol., 41, 1747–1766, https://doi.org/10.1002/joc.6928.
Bannon, J. K., and L. P. Steel, 1960: Averaged water vapor contents of the air. U.K. Meteorological Office Geophysical Memo. 102, 38 pp.
Barindelli, S., E. Realini, G. Venuti, A. Fermi, and A. Gatti, 2018: Detection of water vapor time variations associated with heavy rain in northern Italy by geodetic and low-cost GNSS receivers. Earth Planets Space, 70, 28, https://doi.org/10.1186/s40623-018-0795-7.
Benevides, P., J. Catalao, and P. M. A. Miranda, 2015: On the inclusion of GPS precipitable water vapour in the nowcasting of rainfall. Nat. Hazards Earth Syst. Sci., 15, 2605–2616, https://doi.org/10.5194/nhess-15-2605-2015.
Benevides, P., J. Catalao, and G. Nico, 2019: Neural network approach to forecast hourly intense rainfall using GNSS precipitable water vapor and meteorological sensors. Remote Sens., 11, 966, https://doi.org/10.3390/rs11080966.
Bergemann, M., and C. Jakob, 2016: How important is tropospheric humidity for coastal rainfall in the tropics? Geophys. Res. Lett., 43, 5860–5868, https://doi.org/10.1002/2016GL069255.
Bevis, M., S. Businger, T. A. Herring, C. Rocken, R. A. Anthes, and R. H. Ware, 1992: GPS meteorology: Remote sensing of the atmospheric water vapor using the global positioning system. J. Geophys. Res., 97, 15 787–15 801, https://doi.org/10.1029/92JD01517.
Bonafoni, S., R. Biondi, H. Brenot, and R. Anthes, 2019: Radio occultation and ground-based GNSS products for observing, understanding and predicting extreme events: A review. Atmos. Res., 230, 104624, https://doi.org/10.1016/j.atmosres.2019.104624.
Bretherton, C. S., M. E. Peters, and L. E. Back, 2004: Relationships between water vapor path and precipitation over the tropical oceans. J. Climate, 17, 1517–1528, https://doi.org/10.1175/1520-0442(2004)017<1517:RBWVPA>2.0.CO;2.
Cai, Y., Z. Qian, T. Wu, X. Liang, and M. Song, 2004: Distribution, changes of atmospheric precipitable water over Qinghai-Xizang Plateau and its surroundings and their changeable precipitation climate (in Chinese). J. Plateau Meteor., 23 (1), 1–10.
Catto, J. L., and Coauthors, 2019: The future of midlatitude cyclones. Curr. Climate Change Rep., 5, 407–420, https://doi.org/10.1007/s40641-019-00149-4.
Chen, T., T. Ao, and X. Li, 2019: Temporal and spatial variation of air water resources in the Yangtze River basin in recent seventy years. China Rural Water Hydropower, 5, 6–11.
CMA, 1979: Atlas of Climate of the People’s Republic of China. SinoMaps Press, 226 pp.
Deeter, M. N., 2007: A new satellite retrieval method for precipitable water vapor over land and ocean. Geophys. Res. Lett., 34, L02815, https://doi.org/10.1029/2006GL028019.
Derbyshire, S. H., I. Beau, P. Bechtold, J.-Y. Grandpeix, J.-M. Piriou, J.-L. Redelsperger, and P. M. M. Soares, 2004: Sensitivity of moist convection to environmental humidity. Quart. J. Roy. Meteor. Soc., 130, 3055–3079, https://doi.org/10.1256/qj.03.130.
Divakarla, M. G., C. D. Barnet, M. D. Goldberg, L. M. McMillin, E. Maddy, W. Wolf, L. Zhou, and X. Liu, 2006: Validation of atmospheric infrared sounder temperature and water vapor retrievals with matched radiosonde measurements and forecasts. J. Geophys. Res., 111, D09S15, https://doi.org/10.1029/2005JD006116.
Donat, M. G., and Coauthors, 2013: Updated analyses of temperature and precipitation extreme indices since the beginning of the twentieth century: The HadEX2 dataset. J. Geophys. Res. Atmos., 118, 2098–2118, https://doi.org/10.1002/jgrd.50150.
Dong, W. D., Y. Lin, J. S. Wright, Y. Xie, X. Yin, and J. Guo, 2019: Precipitable water and CAPE dependence of rainfall intensities in China. Climate Dyn., 52, 3357–3368, https://doi.org/10.1007/s00382-018-4327-8.
Emanuel, K. A., M. Fantini, and A. J. Thorpe, 1987: Baroclinic instability in an environment of small stability to slantwise moist convection. Part I: Two-dimensional models. J. Atmos. Sci., 44, 1559–1573, https://doi.org/10.1175/1520-0469(1987)044<1559:BIIAEO>2.0.CO;2.
Fischer, E. M., and R. Knutti, 2016: Observed heavy precipitation increase confirms theory and early models. Nat. Climate Change, 6, 986–991, https://doi.org/10.1038/nclimate3110.
Guerreiro, S. B., H. J. Fowler, R. Barbero, S. Westra, G. Lenderink, S. Blenkinsop, E. Lewis, and X. F. Li, 2018: Detection of continental-scale intensification of hourly rainfall extremes. Nat. Climate Change, 8, 803–807, https://doi.org/10.1038/s41558-018-0245-3.
Hardwick-Jones, R., S. Westra, and A. Sharma, 2010: Observed relationships between extreme sub‐daily precipitation, surface temperature, and relative humidity. Geophys. Res. Lett., 37, L22805, https://doi.org/10.1029/2010GL045081.
Hein, G. W., 2020: Status, perspectives and trends of satellite navigation. Satell. Navig., 1, 22, https://doi.org/10.1186/s43020-020-00023-x.
Held, I. M., and B. J. Soden, 2000: Water vapor feedback and global warming. Annu. Rev. Energy Environ., 25, 441–475, https://doi.org/10.1146/annurev.energy.25.1.441.
Holloway, C. E., and J. D. Neelin, 2009: Moisture vertical structure, column water vapor, and tropical deep convection. J. Atmos. Sci., 66, 1665–1683, https://doi.org/10.1175/2008JAS2806.1.
Holloway, C. E., and J. D. Neelin, 2010: Temporal relations of column water vapor and tropical precipitation. J. Atmos. Sci., 67, 1091–1105, https://doi.org/10.1175/2009JAS3284.1.
Hottovy, S., and S. N. Stechmann, 2015: A spatiotemporal stochastic model for tropical precipitation and water vapor dynamics. J. Atmos. Sci., 72, 4721–4738, https://doi.org/10.1175/JAS-D-15-0119.1.
Houze, R. A., 2012: Orographic effects on precipitating clouds. Rev. Geophys., 50, RG1001, https://doi.org/10.1029/2011RG000365.
Huang, L., Z. Mo, S. Xie, L. Liu, J. Chen, C. Kang, and S. Wang, 2021: Spatiotemporal characteristics of GNSS-derived precipitable water vapor during heavy rainfall events in Guilin, China. Satell. Navig., 2, 13, https://doi.org/10.1186/s43020-021-00046-y.
Huang, R., L. Zhou, and W. Chen, 2003: The progresses of recent studies on the variabilities of the East Asian monsoon and their causes. Adv. Atmos. Sci., 20, 55–69, https://doi.org/10.1007/BF03342050.
Juneng, L., F. T. Tangang, and C. J. C. Reason, 2007: Numerical case study of an extreme rainfall event during 9–11 December 2004 over the east coast of Peninsular Malaysia. Meteor. Atmos. Phys., 98, 81–98, https://doi.org/10.1007/s00703-006-0236-1.
Kiehl, J. T., and K. E. Trenberth, 1997: Earth’s annual global mean energy budget. Bull. Amer. Meteor. Soc., 78, 197–208, https://doi.org/10.1175/1520-0477(1997)078<0197:EAGMEB>2.0.CO;2.
Lenderink, G., and E. van Meijgaard, 2010: Linking increases in hourly precipitation extremes to atmospheric temperature and moisture changes. Environ. Res. Lett., 5, 025208, https://doi.org/10.1088/1748-9326/5/2/025208.
Li, G., and J. Deng, 2013: Atmospheric water monitoring by using ground-based GPS during heavy rains produced by TPV and SWV. Adv. Meteor., 13, 793957, https://doi.org/10.1155/2013/793957.
Li, G., C. Li, J. Ao, Y. He, and Y. Huang, 2015: Research on characteristics of spatiotemporal distribution of precipitable water in Hainan Island. J. Nat. Disasters, 24, 129–138, https://doi.org/10.13577/j.jnd.2015.0515.
Liang, H., Y. Zhang, L. Cao, and Y. Cao, 2020: Temporal relations between precipitable water vapour and precipitation during wet seasons based on nearly two decades of data from the Lhasa River valley, Tibetan Plateau. Int. J. Climatol., 40, 1656–1668, https://doi.org/10.1002/joc.6293.
Lintner, B. R., C. E. Holloway, and J. D. Neelin, 2011: Column water vapor statistics and their relationship to deep convection, vertical and horizontal circulation, and moisture structure at Nauru. J. Climate, 24, 5454–5466, https://doi.org/10.1175/JCLI-D-10-05015.1.
Liu, S., J. Zhang, Z. He, D. Cai, and G. Tian, 2011: Estimate of topography influences on precipitation distribution during typhoon process in Hainan Island. J. Nat. Disasters, 20, 196–199.
Ma, X., K. Shang, J. Li, J. Wang, Y. Kang, and S. Wang, 2015: Spatial and temporal changes of atmospheric precipitable water in the eastern part of Northwest China from 1981 to 2010. J. Desert Res., 35, 448–455.
Milly, P. C. D., R. T. Wetherald, K. A. Dunne, and T. L. Delworth, 2002: Increasing risk of great floods in a changing climate. Nature, 415, 514–517, https://doi.org/10.1038/415514a.
Neelin, J. D., O. Peters, and K. Hales, 2009: The transition to strong convection. J. Atmos. Sci., 66, 2367–2384, https://doi.org/10.1175/2009JAS2962.1.
Ni, Y., and X. Zhou, 2004: Study for formation mechanism of heavy rainfall within the Meiyu front along the middle and downstream of Yangtze River and theories and methods of their detection and prediction. Acta Meteor. Sin., 62, 647–662, https://doi.org/10.11676/qxxb2004.063.
O’Gorman, P. A., and T. Schneider, 2009: Scaling of precipitation extremes over a wide range of climates simulated with an idealized GCM. J. Climate, 22, 5676–5685, https://doi.org/10.1175/2009JCLI2701.1.
Pal, I., and A. Al-Tabbaa, 2009: Trends in seasonal precipitation extremes—An indicator of ‘climate change’ in Kerala, India. J. Hydrol., 367, 62–69, https://doi.org/10.1016/j.jhydrol.2008.12.025.
Pendergrass, A. G., 2018: What precipitation is extreme? Science, 360, 1072–1073, https://doi.org/10.1126/science.aat1871.
Peters, O., and J. D. Neelin, 2006: Critical phenomena in atmospheric precipitation. Nat. Phys., 2, 393–396, https://doi.org/10.1038/nphys314.
Prein, A. F., R. M. Rasmussen, K. Ikeda, C. Liu, M. P. Clark, and G. J. Holland, 2017: The future intensification of hourly precipitation extremes. Nat. Climate Change, 7, 48–52, https://doi.org/10.1038/nclimate3168.
Priego, E., J. Jones, M. J. Porres, and A. Seco, 2017: Monitoring water vapour with GNSS during a heavy rainfall event in the Spanish Mediterranean area. Geomatics Nat. Hazards Risk, 8, 282–294, https://doi.org/10.1080/19475705.2016.1201150.
Robinson, P. J., 1965: Sellers, W. D. 1965: Physical climatology. Chicago: University of Chicago Press, 272 pp. Prog. Phys. Geogr., 33, 130–132, https://doi.org/10.1177/0309133308096757.
Saastamoinen, J., 1972: Atmospheric correction for the troposphere and stratosphere in radio ranging of satellites. The Use of Artificial Satellites for Geodesy, Geophys. Monogr., Vol. 15, Amer. Geophys. Union, 247–251, https://doi.org/10.1029/GM015p0247.
Sapucci, L., L. Machado, E. Souza, and T. Campos, 2018: Global positioning system precipitable water vapour (GPS-PWV) jumps before intense rain events: A potential application to nowcasting. Meteor. Appl., 26, 49–63, https://doi.org/10.1002/met.1735.
Schär, C., and Coauthors, 2016: Percentile indices for assessing changes in heavy precipitation events. Climatic Change, 137, 201–216, https://doi.org/10.1007/s10584-016-1669-2.
Schiro, K. A., J. D. Neelin, D. K. Adams, and B. R. Lintner, 2016: Deep convection and column water vapor over tropical land versus tropical ocean: A comparison between the Amazon and the tropical western Pacific. J. Atmos. Sci., 73, 4043–4063, https://doi.org/10.1175/JAS-D-16-0119.1.
Sharifi, M. A., A. S. Khaniani, and M. Joghataei, 2015: Comparison of GPS precipitable water vapor and meteorological parameters during rainfalls in Tehran. Meteor. Atmos. Phys., 127, 701–710, https://doi.org/10.1007/s00703-015-0383-3.
Shi, C., Q. Zhao, J. Geng, Y. Lou, M. Ge, and J. Liu, 2008: Recent development of PANDA software in GNSS data processing. Proc. SPIE, 7285, 72851S, https://doi.org/10.1117/12.816261.
Shi, C., W. Zhang, Y. Cao, Y. Lou, H. Liang, and L. Fan, 2020: Atmospheric water vapor climatological characteristics over Indo–China region based on Beidou/GNSS and relationships with precipitation. Acta Geod. Cartogr. Sin., 49, 1112–1119, https://doi.org/10.11947/j.AGCS.2020.20200339.
Sun, J., and J. Ao, 2013: Changes in precipitation and extreme precipitation in a warming environment in China. Chin. Sci. Bull., 58, 1395–1401, https://doi.org/10.1007/s11434-012-5542-z.
Tompkins, A. M., 2001: Organization of tropical convection in low vertical wind shears: The role of water vapor. J. Atmos. Sci., 58, 529–545, https://doi.org/10.1175/1520-0469(2001)058<0529:OOTCIL>2.0.CO;2.
Torri, G., D. K. Adams, H. Wang, and Z. Kuang, 2019: On the diurnal cycle of GPS-derived precipitable water vapor over Sumatra. J. Atmos. Sci., 76, 3529–3552, https://doi.org/10.1175/JAS-D-19-0094.1.
Trenberth, K. E., J. Fasullo, and L. Smith, 2005: Trends and variability in column-integrated atmospheric water vapor. Climate Dyn., 24, 741–758, https://doi.org/10.1007/s00382-005-0017-4.
Van Baelen, J., and G. Penide, 2009: Study of water vapor vertical variability and possible cloud formation with a small network of GPS stations. Geophys. Res. Lett., 36, L02804, https://doi.org/10.1029/2008GL036148.
Xiao, C., P. Wu, L. Zhang, and L. Song, 2016: Robust increase in extreme summer rainfall intensity during the past four decades observed in China. Sci. Rep., 6, 38506, https://doi.org/10.1038/srep38506.
Ye, H., E. J. Fetzer, S. Wong, A. Behrangi, E. T. Olsen, J. Cohen, B. H. Lambrigtsen, and L. Chen, 2014: Impact of increased water vapor on precipitation efficiency over Northern Eurasia. Geophys. Res. Lett., 41, 2941–2947, https://doi.org/10.1002/2014GL059830.
Zhai, P., and R. E. Eskridge, 1996: Analysis of inhomogeneities in radiosonde temperature and humidity time series. J. Climate, 9, 884–894, https://doi.org/10.1175/1520-0442(1996)009<0884:AOIIRT>2.0.CO;2.
Zhai, P., and R. E. Eskridge, 1997: Atmospheric water vapor over China. J. Climate, 10, 2643–2652, https://doi.org/10.1175/1520-0442(1997)010<2643:AWVOC>2.0.CO;2.
Zhang, B., J. Han, J. Chen, and P. Liu, 2012: Analysis of correlation between water vapor content and precipitation over North China. J. Arid Meteor., 30, 207–214.
Zhang, W., Y. Lou, J. S. Haase, R. Zhang, G. Zheng, J. Huang, C. Shi, and J. Liu, 2017: The use of ground-based GPS precipitable water measurements over China to assess radiosonde and ERA-Interim moisture trends and errors from 1999 to 2015. J. Climate, 30, 7643–7667, https://doi.org/10.1175/JCLI-D-16-0591.1.
Zhang, X., F. W. Zwiers, G. Li, H. Wan, and A. J. Cannon, 2017: Complexity in estimating past and future extreme short-duration rainfall. Nat. Geosci., 10, 255–259, https://doi.org/10.1038/ngeo2911.
Zhang, Z., 2001: The Crustal Movement Observation Network of China (in Chinese). China Basic Sci., 3, 45–49.
Zheng, Z., J. Gao, Z. Ma, Z. Wang, X. Yang, X. Luo, J. Thierry, and G. Fu, 2016: Urban flooding in China: Main causes and policy recommendation. Hydrol. Processes, 30, 1149–1152, https://doi.org/10.1002/hyp.10717.
Zhou, S., P. Wu, C. Wang, and J. Han, 2012: Spatial distribution of atmospheric water vapor and its relationship with precipitation in summer over the Tibetan Plateau. J. Geogr. Sci., 22, 795–809, https://doi.org/10.1007/s11442-012-0964-8.
Zhou, X., Z. Bai, and Y. Yang, 2017: Linking trends in urban extreme rainfall to urban flooding in China. Int. J. Climatol., 37, 4586–4593, https://doi.org/10.1002/joc.5107.
Zou, Y., Y. Wang, D. Wen, and Y. Su, 2012: A comparative study of GPS precipitable water using different climate types (in Chinese). J. Wuhan Univ., 37, 573–581, https://doi.org/10.13203/j.whugis2012.05.019.