the Creative Commons Attribution 4.0 License.
Special issue: Advanced Global Navigation Satellite Systems tropospheric...
Research article 27 Feb 2019
Research article  27 Feb 2019
Development of timevarying global gridded T_{s}–T_{m} model for precise GPS–PWV retrieval
 ^{1}School of Resources and Environmental Engineering, Anhui University, Hefei, Anhui, China
 ^{2}Anhui Province Key Laboratory of Wetland Ecosystem Protection and Restoration, Anhui University, Hefei, Anhui, China
 ^{3}GNSS Research Center, Wuhan University, Wuhan, Hubei, China
 ^{4}Shenzhen Key Laboratory of Spatial Smart Sensing and Services, College of Civil Engineering, Shenzhen University, Shenzhen, Guangdong, China
 ^{5}Anhui Engineering Research Center for Geographical Information Intelligent Technology, Hefei, Anhui, China
 ^{1}School of Resources and Environmental Engineering, Anhui University, Hefei, Anhui, China
 ^{2}Anhui Province Key Laboratory of Wetland Ecosystem Protection and Restoration, Anhui University, Hefei, Anhui, China
 ^{3}GNSS Research Center, Wuhan University, Wuhan, Hubei, China
 ^{4}Shenzhen Key Laboratory of Spatial Smart Sensing and Services, College of Civil Engineering, Shenzhen University, Shenzhen, Guangdong, China
 ^{5}Anhui Engineering Research Center for Geographical Information Intelligent Technology, Hefei, Anhui, China
Correspondence: Peng Jiang (jiangpeng@ahu.edu.cn)
Hide author detailsCorrespondence: Peng Jiang (jiangpeng@ahu.edu.cn)
Watervaporweighted mean temperature, T_{m}, is the key variable for estimating the mapping factor between GPS zenith wet delay (ZWD) and precipitable water vapor (PWV). For the nearrealtime GPS–PWV retrieval, estimating T_{m} from surface air temperature T_{s} is a widely used method because of its high temporal resolution and fair degree of accuracy. Based on the estimations of T_{m} and T_{s} at each reanalysis grid node of the ERAInterim data, we analyzed the relationship between T_{m} and T_{s} without data smoothing. The analyses demonstrate that the T_{s}–T_{m} relationship has significant spatial and temporal variations. Static and timevarying global gridded T_{s}–T_{m} models were established and evaluated by comparisons with the radiosonde data at 723 radiosonde stations in the Integrated Global Radiosonde Archive (IGRA). Results show that our global gridded T_{s}–T_{m} equations have prominent advantages over the other globally applied models. At over 17 % of the stations, errors larger than 5 K exist in the Bevis equation (Bevis et al., 1992) and in the latituderelated linear model (Y. B. Yao et al., 2014), while these large errors are removed in our timevarying T_{s}–T_{m} models. Multiple statistical tests at the 5 % significance level show that the timevarying global gridded model is superior to the other models at 60.03 % of the radiosonde sites. The secondbest model is the 1^{∘} × 1^{∘} GPT2w model, which is superior at only 12.86 % of the sites. More accurate T_{m} can reduce the contribution of the uncertainty associated with T_{m} to the total uncertainty in GPS–PWV, and the reduction augments with the growth of GPS–PWV. Our theoretical analyses with high PWV and small uncertainty in surface pressure indicate that the uncertainty associated with T_{m} can contribute more than 50 % of the total GPS–PWV uncertainty when using the Bevis equation, and it can decline to less than 25 % when using our timevarying T_{s}–T_{m} model. However, the uncertainty associated with surface pressure dominates the error budget of PWV (more than 75 %) when the surface pressure has an error larger than 5 hPa. GPS–PWV retrievals using different T_{m} estimates were compared at 74 International GNSS Service (IGS) stations. At 74.32 % of the IGS sites, the relative differences of GPS–PWV are within 1 % by applying the static or the timevarying global gridded T_{s}–T_{m} equations, while the Bevis model, the latituderelated model and the GPT2w model perform the same at 37.84 %, 41.89 % and 29.73 % of the sites. Compared with the radiosonde PWV, the error reduction in the GPS–PWV retrieval can be around 1–2 mm when using a more accurate T_{m} parameterization, which accounts for around 30 % of the total GPS–PWV error.
 Article
(16377 KB) 
Supplement
(7869 KB)  BibTeX
 EndNote
Water vapor is an important trace gas and one of the most variable components in the troposphere. The transport, concentration and phase transition of water vapor are directly involved in the atmospheric radiation and hydrological cycle. It plays a key role in many climate changes and weather processes (Adler et al., 2016; Mahoney et al., 2016; Song et al., 2016). However, water vapor has high spatial–temporal variability, and its content is often small within the atmosphere. It is a challenge to measure water vapor content accurately and timely. For decades, several methods have been studied, such as radiosondes and water vapor radiometers, sun photometers and GPS (Campmany et al., 2010; Ciesielski et al., 2010; Liu et al., 2013; PerezRamirez et al., 2014; Li et al., 2016). Compared with the traditional water vapor observations, groundbased GPS water vapor measurement has the advantages of high accuracy, high spatial–temporal resolution, allweather availability and lowcost (Haase et al., 2003; Pacione and Vespe, 2008; Lee et al., 2010; Means, 2013; Lu et al., 2015). Groundbased GPS water vapor products, mainly including precipitable water vapor (PWV), are widely used in many fields such as realtime vapor monitoring, weather and climate research, and numerical weather prediction (NWP) (Van Baelen and Penide, 2009; Karabatic et al., 2011; Rohm et al., 2014; Adams et al., 2017).
GPS observations require some kind of meteorological element to estimate PWV. Zenith hydrostatic delay (ZHD) can be calculated using surface pressure P_{s} with the equation (Ning et al., 2013):
where φ is the latitude, H is the geoid height in meters, and
Then, zenith wet delay (ZWD) is generated by subtracting ZHD from zenith total delay (ZTD). ZTD can be directly estimated from precise GPS data processing. Finally, a conversion factor Q, which is used to map ZWD onto PWV, is determined by the watervaporweighted mean temperature T_{m} over a GPS station. The mapping function from ZWD to PWV is expressed as follows (Bevis et al., 1992):
and Q is computed using following formula:
where ρ_{w} is the density of liquid water, R_{v} is the specific gas constant for water vapor, ${k}_{\mathrm{2}}^{\prime}=(\mathrm{22.1}\pm \mathrm{2.2})$ K mbar^{−1} and ${k}_{\mathrm{3}}=(\mathrm{3.739}\pm \mathrm{0.012})\times {\mathrm{10}}^{\mathrm{5}}$ K^{2} mbar^{−1} are physical constants (Ning et al., 2016). T_{m} is the weighted mean temperature which is defined as a function related to the temperature and water vapor pressure. It can be approximated as the following formula (Bevis et al., 1992):
where e and T represent vapor pressure in hPa and temperature in Kelvin, i denotes the ith level and Δz_{i} is the height difference of ith level. Vapor pressure e is calculated using equation e=e_{s} × RH; RH is the relative humidity, and the saturation vapor pressure e_{s} can be estimated from the temperature observations using a Goff–Gratch formula (Sheng et al., 2013).
There are the three main approaches that are used to estimate T_{m}. They have respective advantages and disadvantages when they are applied for different purposes:

The integration of vertical temperature and humidity profiles is believed to be the most accurate method. The profile data can be extracted from radio soundings or NWP data sets (Wang et al., 2016). However, some inconveniences have to be endured. It usually takes a considerable amount of time to acquire the NWP data, which are normally released in a large volume every 6 h. This limits the use of NWP data in the nearrealtime GPS–PWV retrieval. Radiosonde data are another profile data source, but it has low spatial and temporal resolution. At most of the radiosonde sites, sounding balloons are cast daily at 00:00 and 12:00 UTC. Furthermore, a large number of GPS stations are not located close enough to the radio sounding sites. Therefore, such methods are appropriate for climate research or the study of longterm PWV trends, but do not meet the realtime requirements.

Several global empirical models of T_{m} are established based on the analyses of T_{m} time series from NWP data sets or other sources (Yao et al., 2012; Chen et al., 2014; Bohm et al., 2015). T_{m} at any time and any location can be estimated from these models. They are often independent of the current meteorological observations, which are required to be observed together with the GPS data. However, some important real variations, which may be dramatic during some extreme weather events, can be lost without the constraints of current real data (Jiang et al., 2016). Therefore, these modeled estimates are not accurate enough for highprecision meteorological applications, such as providing GPS–PWV estimates for weather prediction.

Many studies indicated that the T_{m} parameter has a relationship with some surface meteorological elements, such as surface air temperature or surface air humidity (Bevis et al., 1992; Y. Yao et al., 2014). These surface meteorological parameters can be measured accurately and rapidly. T_{m} is then estimated using these surface measurements. However, these studies also revealed that the relationships are often weak, except the T_{s}–T_{m} relationship. For example, Bevis et al. (1992) introduced the equation T_{m}=0.72 T_{s}+70.2 [K] after analyzing 8712 radiosonde profiles collected at 13 sites in the US over 2 years. This equation has been widely used in many other studies.
According to Rohm et al. (2014), GPS–ZTD can be estimated very precisely by realtime GPS data processing. This means that T_{m} is one of the key parameters in the nearrealtime GPS–PWV estimation. On the other hand, method (3) is the most suitable method for estimating T_{m} in near realtime because of its balance between timeliness and accuracy. The T_{s}–T_{m} relationship has spatial–temporal variations. Several regional T_{s}–T_{m} equations were established using the profile data over corresponding fields (Wang et al., 2012). However, a T_{s}–T_{m} model without spatial variation is not good enough for a vast field, e.g., the Indian region (Singh et al., 2014). Aside from this, some vast areas have no specific highprecision T_{s}–T_{m} model, for example over the oceans. In general, significant differences exist between oceanic and terrestrial atmospheric properties, especially near the surface layer and within the boundary layer. The change in T_{s} from land to ocean may be very different from that of T_{m}. Therefore it is necessary to model the T_{s}–T_{m} relationship over oceanic regions, since several oceanbased GPS meteorology experiments demonstrated the potential of this technique to retrieve PWV over the broad ocean (Rocken et al., 2005; Kealy et al., 2012). A global gridded T_{s}–T_{m} model has been established by Lan et al. (2016). In this model, the 2.0^{∘} × 2.5^{∘} T_{m} data from GGOS Atmosphere and the 0.75^{∘} × 0.75^{∘} T_{s} data from the European Centre for MediumRange Weather Forecasts (ECMWF) reanalysis data are both smoothed to the resolution of 4^{∘} × 5^{∘}. However, the T_{s}–T_{m} relationship is varying in time (Y. Yao et al., 2014), while the Lan et al. (2016) model is static.
The objective of this study is mainly to (1) develop global gridded T_{s}–T_{m} models without any smoothing of the data, then assess their precision, and (2) study the performances of GPS–PWV retrievals using our T_{s}–T_{m} models. Table 1 lists the main differences between the T_{s} and T_{m} models developed in this study and the other global used T_{m} models. In Sect. 2, the data sources and determining methods of T_{m} are introduced in detail. Then, in section 3 we analyze the T_{s}–T_{m} relationships and their variations on a global scale. Globalgridded T_{s}–T_{m} estimating models in different forms are established and evaluated in Sect. 4. Section 5 assesses the accuracies of different PWV retrievals and Sect. 6 presents conclusions based on our experiments.
As the definition of T_{m} in Eq. (5), the e_{i} parameter in the middle of ith level is calculated by vertical exponential interpolation of the water vapor pressure of its two neighbor measurement points. The temperature is estimated by linear interpolation of the two neighbor temperatures. The integral intervals are from the Earth's surface to the top level of the profile data. The height of the top level depends on the data sources we employed. The essential profile data, including the temperature, height and relative humidity values through the entire atmospheric column, can be obtained from the radiosondes or NWP data sets.
We employed radiosonde data from the Integrated Global Radiosonde Archive (IGRA, ftp://ftp.ncdc.noaa.gov/pub/data/igra, last access: 25 February 2019) to calculate T_{m}. Version 2.0 of the IGRAderived sounding parameters provides pressure, geopotential height, temperature, saturation vapor pressure and relative humidity observations at the observed levels. A bias may be introduced if the integrals were terminated at lower levels (Wang et al., 2005); thus the integrations were performed up to the topmost valid radiosonde data. According to our quality control processes, some radiosonde profile data were rejected. In each profile, the surface observations must be available and the top profile level should not be lower than the standard 300 hPa level. Furthermore, the level number between the surface and the top level should be greater than 10 to avoid vertical profiles that are too sparse. At most of the radio sounding stations, sounding balloons are launched every 12 h, and their ascending paths are assumed to be vertical.
Profile data are usually provided by NWP products at certain vertical levels. The ERAInterim product from ECMWF provides data on a regular 512 longitude by 256 latitude N128 Gaussian grid after the grid transformation performed by the NCAR Data Support Section (DSS). On each grid node of ERAInterim, temperature, relative humidity and geopotential at 37 isobaric levels from 1000 to 1 hPa can be obtained. By dividing the geopotential by constant gravitational acceleration value (g≈9.80655 m s^{−2}), we can determine the geopotential heights of the surface and levels. Data sets are available at 00:00, 06:00, 12:00 and 18:00 UTC every day and have been covering a period from 1979.01 to the present.
In theory, the computation of Eq. (5) should be integrated through the entire atmospheric column, and the geopotential height should be converted to the geometric height. However, water vapor is solely concentrated in the troposphere, and most of it is specifically located within the first 3 km a.s.l. (above sealevel). Moreover, in the two selected data sets, the geopotential heights of top pressure levels are approximately 30–40 km. Geopotential height is very close to geometric height in such height ranges. According to our computation, the relative difference between them is only between 0.1 % and 0.9 %. In fact, the height difference Δz can be replaced by the geopotential height difference Δh in Eq. (5), since the division can almost eliminate the difference between the two different height types. The T_{m} value nearly has no change after such height replacement. For the convenience of the calculation, we directly employed the geopotential height variable. In this paper, we denoted the T_{m} derived from ERAInterim as T_{m_ERAI}.
At each reanalysis grid node, the computation of Eq. (5) always starts from the surface height to the top pressure level. The pressure levels below surface height were rejected. T_{s} is defined as the variable of “temperature at 2 m above ground”, and surface water vapor pressure can be derived from the “2 m dewpoint temperature” variable in ERAInterim. These T_{s} were also used in the regression analyses between T_{s} and T_{m}.
Many studies have indicated the close relationship between T_{s} and T_{m}. However, T_{m}is also found to not be closely related to T_{s} in some regions, e.g., in the Indian zone (Suresh Raju et al., 2007). Using the T_{m} and T_{s} generated from the global gridded reanalysis data, we are able to study the T_{s}–T_{m} relationship in detail.
We first carried out a linear regression analysis on 4 years of T_{s} and T_{m} data generated from the radiosonde data and the global gridded ERAInterim data sets, with data covering the period January 2009 to December 2012. The analysis results are shown in Fig. 1. Although the two data sets have different temporal resolutions (12 h for the radiosonde data and 6 h for the ERAInterim data) and spatial resolutions, both analyses agree well with each other. This is expected because the radiosonde data have been assimilated into the ERAInterim products. Our analyses also indicate that the T_{s}=T_{m} correlation coefficient is generally related to the latitude. The same conclusion has been drawn in other studies (Y. B. Yao et al., 2014). Significant positive correlation coefficients can be found at middle and high latitudes and reach a maximum in the polar regions. The correlation coefficients drop dramatically at low latitudes. This is because T_{m} is stable there, showing independency of the other parameters. To study the variations of T_{s} and T_{m}, we illustrated the denary logarithm values of their standard deviations in Fig. 2. It is evident that T_{m} varies to a lesser degree than T_{s} at low latitudes. Aside from the latituderelated features, there are obvious differences of the T_{s}–T_{m} correlation coefficients between land and ocean. We even found that negative correlation coefficients over certain oceans, e.g., lowlatitude western Pacific, Bay of Bengal or Arabian Sea (see Fig. 1). Unreliable regression analysis results may be derived when the T_{s} and T_{m}data both have small variations. In Fig. 3, scatter plots of T_{s} and T_{m} from ERAInterim at two locations 0.35^{∘} N 180.00^{∘} E and 70.53^{∘} N 180.00^{∘} E are given. As the blue dots show, the T_{s}–T_{m} relationship is weak in the areas near the equator, because the entire variation ranges of T_{s} and T_{m} are both within 10 K. This results in a meaningless linear regression (see the magenta line). The T_{s}–T_{m} correlation coefficient is only −0.0893 there. Other than the large spatial variations, studies have revealed that the T_{s}–T_{m} relationship also has temporal variations (Wang et al., 2005). Therefore, a good T_{s}–T_{m} model should take both the spatial and temporal variations into consideration, and this is the main aim in the following sections.
Since the T_{s}–T_{m} relationship has large spatial variations, a global gridded T_{s}–T_{m} model is preferred for precise GPS–PWV estimations. In this section, a static global gridded model and a timevarying global gridded model are established and assessed.
4.1 Static globalgridded T_{s}–T_{m} model
A linear formula ${T}_{\mathrm{m}}=a{T}_{\mathrm{s}}+b$ for the relation between T_{m} and T_{s} has been adopted in many studies. Based on the T_{s} and T_{m} products from the ERAInterim data covering the years 2009 to 2012, we performed linear fittings of T_{m} versus T_{s} on each grid point. Then, the slope constant (a), the intercept constant (b) and the fitting root mean square error (RMSE) of each linear expression were calculated and contoured in Fig. 4. The a and b values are related to the latitude as well as the underlying surface (e.g., land, ocean). In the middle–high latitudes over the Northern Hemisphere, constant a value varies from 0.6 to 0.8, and constant b is approximately 100–50 over most of the continents. The constants in the Bevis equation are within these value ranges. Constant a is smaller (approximately 0.5–0.7) over land at the middle–high latitudes over the Southern Hemisphere. In particular, there are abrupt changes in the values of constants a and b from land to ocean at the middle–high latitudes due to the different feature variations of T_{s} and T_{m} (see Fig. 2). At low latitudes, the a value is smaller than over the other regions, because of the low variations of T_{s} and T_{m}. The fitting RMSEs are within 2–4 K over the middle–high latitude lands, and lower values are obtained over the oceans or at low latitudes. The reason for the low RMSE around the equator is the smaller fluctuation of T_{m}. Meanwhile, there is no RMSE larger than 4.5 K in the results of our model. As we did not perform any spatial or temporal smoothing of the data during the data processing, both the precision and resolution of our static model are better than other models (e.g., Lan et al., 2016).
4.2 Timevarying globalgridded T_{s}–T_{m} model
The time variation in the T_{s}–T_{m} relationship should also be considered in a precise T_{s}–T_{m} model. Therefore, a timevarying equation is applied for T_{s}–T_{m} regression at each grid node:
where “doy” represents the observed day of year and “h” is the observed hour in UTC time; (m_{1}, m_{2}), (n_{1}, n_{2}) and (p_{1}, p_{2}) are fitting coefficients. These equations can reflect the amplitudes of annual, semiannual and diurnal variations in our T_{s}–T_{m} models.
Our new regression model found similar values for the coefficients a and b (of its static term) as for the static model in Sect. 4.1, except for some differences over the oceans. In Fig. 5, besides these constants a and b, we illustrate the amplitudes of annual, semiannual and diurnal terms. We can see that there are large annual variations (amplitude > 5 K) in the vast regions from Tibet to northern Africa, and in some places of the Siberia and Chile. Large diurnal variations (amplitude > 3 K) mainly occur over the midlatitude lands such as northeastern Asia or North America. Semiannual variations, however, are small in most areas except some highlatitude areas (amplitude > 3 K). All variations are smaller over the oceans due to the slower temperature changes over water than over land. The estimated T_{m} RMSE is also contoured in Fig. 5, and we can see that the RMSE dropped significantly in the regions with large annual or diurnal variations.
4.3 Assessments of T_{s}–T_{m} models
To further assess the precision of the T_{s}–T_{m} models using other independent data sources, we generated T_{m} and T_{s} from the radiosonde data at 723 radiosonde stations in the year 2016. These data are not assimilated into the 2009–2012 ERAInterim data sets. As a result, we can regard them to be independent of our model. At each radiosonde site, different T_{s}–T_{m} models were employed to calculate T_{m}. In addition, we also estimated T_{m} using the 1^{∘} × 1^{∘} GPT2w model (Bohm et al., 2015), which is a global gridded T_{m} empirical model independent of the surface meteorological observation data. Then, these calculated T_{m} will be evaluated by being compared with the integrated T_{m} of radiosondes (denoted as T_{m_RS}) twice a day.
The model estimations of T_{m} are denoted by T_{m_Bevis}, T_{m_LatR}, T_{m_static}, T_{m_varying} and T_{m_GPT2w} for the Bevis equation, the latituderelated model, our static global gridded model, timevarying global gridded model and the GPT2w model. When the global gridded models are employed, the radiosonde station may not be located at a grid node. Therefore, we interpolated the coefficients in the T_{s}–T_{m} equations from the neighboring grids to the radiosonde sites. The interpolation formula is expressed as follows (Jade and Vijayan, 2008):
C_{site} and ${C}_{\mathrm{site}}^{i}$ represent the coefficients in T_{s}–T_{m} equations at the radiosonde site location and its neighboring grids, respectively. w^{i} are the interpolation coefficients, which are determined using the equation
where R=6378.17 km is the mean radius of the earth, λ is the scale factor which equals one in our study, and ψ^{i} is the angular distance between the ith grid node and the station's position. ψ^{i} are computed using following formula (with latitude φ and longitude θ):
Considering the fact that the reanalysis grids are definite and every radiosonde site is in situ, we can compute the interpolation coefficients in Eq. (7) for all of the radiosonde stations. Then, these coefficients are stored as constants to avoid reduplicating the calculation.
Taking T_{m_RS} as the reference values, we calculated the biases and RMSEs of T_{m_Bevis}, T_{m_LatR}, T_{m_static}, T_{m_varying} and T_{m_GPT2w} at each radiosonde site. The results are illustrated in Fig. 6. Obviously, in many regions, the Bevis equation has bad precision with the absolute bias and RMSE both larger than 5 K. T_{m_LatR} can reduce the estimated biases in many areas, but the RMSEs remain large. Large biases still exist at quite a few radiosonde stations, e.g., in Africa or western Asia. T_{m_static} and T_{m_GPT2w} remove the large T_{m} biases at most of the radiosonde stations. T_{m_varying} performs significantly better over the world, especially in the Middle East, North America, Siberia, etc.
Detailed statistics of the distributions of the bias and RMSE using different models are shown in Fig. 7 and Table 2. At over 97.37 % of the radiosonde stations, the biases of T_{m_varying} are within −3–3 K. Large positive biases (>3 K) nearly disappear in T_{m_varying}. In contrast, there are significant large biases in T_{m_Bevis} and T_{m_LatR}. Improvements in RMSE are more evident. The RMSEs of T_{m_varying} are smaller than 4 K at over 91 % of the radiosonde sites, while few sites (<1 %) have RMSEs larger than 5 K. This is clearly better than the other models. In T_{m_Bevis} and T_{m_LatR}, there are more than 17 % of the radiosonde sites with RMSEs larger than 5 K. The overall performance of T_{m_GPT2w} is very close to T_{m_Bevis}, except that its absolute bias is smaller than the other T_{s}–T_{m} models.
To identify the superior T_{m} estimation model at each radiosonde site, we employed the following statistical tests under the assumption of a normal distribution of the estimated T_{m} error:

First, Brown–Forsythe tests (Brown and Forsythe, 1974) of the equality of variances were carried out at each site for estimating the T_{m} errors from two different models, e.g., model A and B. The purpose of this step is to determine whether there are significant differences in the variances of the T_{m} results. If the test rejects the null hypothesis at a 5 % significance level and the errors of model A and B have the same variance, the model with the smaller sample variance is regarded as the better one. However, if the test does not reject the homogeneity of variances, analysis of variance (ANOVA) is performed in the next step.

ANOVA is a technique used to analyze the differences among group means (Hogg and Ledolter, 1987). It evaluates the null hypothesis that the samples all have the same mean against the alternative that the means are not the same. If the null hypothesis is rejected at a 5 % significance level, the T_{m} sample with smaller absolute mean value is believed to be better. Otherwise, we think that the two models perform almost as well at this radiosonde site.

After multiple tests and comparisons, the best model at each radiosonde station may be identified. However, at some sites no superior model can be confirmed. All the models are believed to have equivalent performance.
Finally, we counted the number of sites at which each T_{m} model performed the best. The results are given in Table 3. The timevarying global gridded model is superior to the others at 434 radiosonde stations (60.03 % of all sites), while the secondbest estimation, T_{m_GPT2w}, is superior at only 12.86 % of the sites.
In Fig. 8 the T_{m} series at the IGRA station no. 62378 (29.86^{∘} N 31.34^{∘} E in Egypt) are given. We can see that large negative biases ($<\mathrm{5}$ K) between T_{m_Bevis} (or T_{m_LatR}) and T_{m_RS} exist. T_{m_static} performs only slightly better from July to October. However, T_{m_varying} and T_{m_GPT2w} can eliminate most of the seasonal errors. Different properties of T_{m} series appear at another IGRA station no. 40841 (30.25^{∘} N 56.97^{∘} E in Iran). Some observation data are missing, but we can still see that there are large positive differences (>5 K) between T_{m_Bevis} (or T_{m_LatR}) and T_{m_RS} throughout the year. The biases of T_{m_static} are much smaller, but some large errors still appear in many months. The T_{m_varying}, however, performs as well as the T_{m} calculated from the radiosonde data, with small biases, and captures the variations well. The time series ofT_{m_GPT2w} are smoother and cannot capture the fluctuations of the T_{m} time series, causing an accuracy worse than T_{m_varying}.
On the other hand, even T_{m_varying} have large differences from T_{m_RS} at a few IGRA stations. This can be explained by the fact that our fitting analyses are based on the T_{m} values derived from ERAInterim profiles. The quality of ERAInterim data can be very poor in the regions with sparse observation data (Itterly et al., 2018).
GPS–PWV has different error sources with different properties. It is complicated to evaluate the GPS–PWV uncertainty here due to the lack of collaborated additional independent techniques that monitor water vapor at the GPS site.
5.1 Theoretical analysis of the GPS–PWV uncertainty
Comprehensive research on the uncertainty in GPS–PWV has been carried out by Ning et al. (2016). The uncertainties in ZTD, ZHD and conversion factor Q have been studied in detail. The total uncertainty in GPS–PWV is as follows:
where σ_{PWV}, σ_{ZTD}, ${\mathit{\sigma}}_{{P}_{\mathrm{s}}}$ and σ_{Q} are, respectively, the uncertainties of GPS–PWV, the ZTD estimation, the P_{s} observations and the conversion factor Q. σ_{c}=0.0015 denotes the uncertainty in constant C=2.2767 in Eq. (1), PWV is the value of GPS–PWV, and
where ${\mathit{\sigma}}_{{k}_{\mathrm{3}}}=\mathrm{0.012}\times {\mathrm{10}}^{\mathrm{5}}$ K^{2} hPa^{−1}, ${\mathit{\sigma}}_{{k}_{\mathrm{2}}^{\prime}}=\mathrm{2.2}$ K hPa^{−1} and ${\mathit{\sigma}}_{{T}_{\mathrm{m}}}$, respectively, denote the uncertainties of k_{3}, ${k}_{\mathrm{2}}^{\prime}$ and T_{m} in Eq. (4). The variation in σ_{Q} with the values of T_{m} and ${\mathit{\sigma}}_{{T}_{\mathrm{m}}}$ is depicted in Fig. 9. Assuming that T_{m} is 280 K, we find that the σ_{Q} increases by over 60 % (from 0.069 to 0.112) as the ${\mathit{\sigma}}_{{T}_{\mathrm{m}}}$ rises from 3.0 to 5.0 K. However, the σ_{Q} is less sensitive to the value of T_{m}. The σ_{Q} rises only by 17.96 % (about from 0.061 to 0.075) as the value of T_{m} drops from 300 to 270 K with ${\mathit{\sigma}}_{{T}_{\mathrm{m}}}=\mathrm{3.0}$ K.
Ning et al. (2016) assumed the T_{m} were obtained from NWP models so the uncertainty in T_{m} was set to be small (${\mathit{\sigma}}_{{T}_{\mathrm{m}}}=\mathrm{1.1}$ K). However, as shown in Sect. 4.3, the uncertainties of T_{m} from different T_{m} models are significantly larger at the radiosonde stations. For each radiosonde station, we calculated the mean value of T_{m} and assigned the ${\mathit{\sigma}}_{{T}_{\mathrm{m}}}$ with the RMSEs of T_{m} given in Fig. 6. Then we obtained the σ_{Q} in Eq. (11). Our statistics indicate that the σ_{Q} using our varying T_{s}–T_{m} model decreases on average by 19.26 %, 17.77 %, 7.79 % and 18.67 % with respect to the σ_{Q}, respectively, using T_{m_Bevis}, T_{m_LatR}, T_{m_static} and T_{m_GPT2w}. For example, at the IGRA station no. 42724 (22.88^{∘} N 91.25^{∘} E in India), σ_{Q} drops by 53 % from 0.141 of the T_{m_Bevis} to 0.066 of the T_{m_varying}.
The uncertainty in Q will be propagated to the total uncertainty in GPS–PWV according to Eq. (10). We obtained the contributions of the different terms in Eq. (10) to the total GPS–PWV uncertainty. The contribution of one term is measured by the percentage it accounts for the total σ_{PWV}. The percentages are computed using the formulas
where p_{ZTD}, ${p}_{{P}_{\mathrm{s}}}$, p_{C} and p_{Q}, respectively, indicate the contributions of the uncertainties associated with ZTD, P_{s}, constant C and factor Q to the total σ_{PWV}. Following the summaries of Ning et al. (2016), we assumed that σ_{ZTD}=4 mm and σ_{C}=0.0015. T_{m} identically equals 280 K since the σ_{Q} is less sensitive to the value of T_{m} with respect to the ${\mathit{\sigma}}_{{T}_{\mathrm{m}}}$. Table 4 gives five sets of typical values which are assigned to the ${\mathit{\sigma}}_{{P}_{\mathrm{s}}}$, ${\mathit{\sigma}}_{{T}_{\mathrm{m}}}$, P_{s} and PWV in Eqs. (10)–(12).
The ${\mathit{\sigma}}_{{P}_{\mathrm{s}}}$ equals 0.2 hPa in Ning et al. (2016); however we enlarged its typical value to 0.5 hPa in consideration of the possible worse performance of the surface barometers. In Fig. 10, we illustrated the contributions of the terms in Eq. (12) based on the assumptions (a)–(e) in Table 4. Some feature variations of the contributions of different terms can be found from the comparisons between different subplots:

No significant difference exists between Fig. 10a and b. Because of the small value of σ_{c} in Eq. (10), the σ_{PWV} is not sensitive to the value of P_{s}. Meanwhile, the uncertainty associated with σ_{c} contributes less than 10 % of the σ_{PWV}.

With the typical values in Table 4 (values a and b), a reduction of ${\mathit{\sigma}}_{{T}_{\mathrm{m}}}$ can reduce the p_{Q} significantly. For example, in Fig. 10a, the p_{Q} accounts for 69.54 % with ${\mathit{\sigma}}_{{T}_{\mathrm{m}}}=\mathrm{6}$ K, and it declines to 38.19 % with ${\mathit{\sigma}}_{{T}_{\mathrm{m}}}=\mathrm{3}$ K.

As Fig. 10c shows, the uncertainty associated with σ_{ZTD} accounts for the main part of σ_{PWV} when the values of PWV and ${\mathit{\sigma}}_{{P}_{\mathrm{s}}}$ are not high. With the typical values in Table 4 (value c), the p_{ZTD} can be up to 74.21 % with ${\mathit{\sigma}}_{{T}_{\mathrm{m}}}=\mathrm{3}$ K. The p_{Q}, however, can drop from 26.76 % to 9.00 % as the ${\mathit{\sigma}}_{{T}_{\mathrm{m}}}$ decreases from 6 to 3 K. Although the p_{Q} is not large under this situation, a smaller ${\mathit{\sigma}}_{{T}_{\mathrm{m}}}$can still reduce the contribution of σ_{Q} to the σ_{PWV}.

The uncertainty associated with ${\mathit{\sigma}}_{{P}_{\mathrm{s}}}$ dominates the error budget of PWV when the ${\mathit{\sigma}}_{{P}_{\mathrm{s}}}$ is large. In Fig. 10d and e, the ${p}_{{P}_{\mathrm{s}}}$ is over 80 % with ${\mathit{\sigma}}_{{T}_{\mathrm{m}}}<\mathrm{3}$ K and ${\mathit{\sigma}}_{{P}_{\mathrm{s}}}=\mathrm{5}$ hPa. In Fig. 10d, the p_{Q} increases from 7.55 % to 23.19 % as the ${\mathit{\sigma}}_{{T}_{\mathrm{m}}}$ rises from 3 to 6 K. However, in Fig. 10e, the p_{Q} only grows from 1.29 % to 4.61 % with the same variation in ${\mathit{\sigma}}_{{T}_{\mathrm{m}}}$.
Theoretical analyses of σ_{PWV} were also carried out at two representative stations. At the IGRA station no. 42971 (20.25^{∘} N 85.83^{∘} E, in India), the mean value of PWV is 53.88 mm. The RMSEs of T_{m_Bevis}, T_{m_LatR}, T_{m_static}, T_{m_varying} and T_{m_GPT2w} are 4.30, 3.15, 2.41, 1.93 and 1.97 K. The ${\mathit{\sigma}}_{{T}_{\mathrm{m}}}$ in Eq. (11) was replaced by the calculated RMSEs, and the p_{ZTD}, ${p}_{{P}_{\mathrm{s}}}$, p_{C} and p_{Q} were generated with two typical values, 0.5 and 5 hPa, assigned to the ${\mathit{\sigma}}_{{P}_{\mathrm{s}}}$. With ${\mathit{\sigma}}_{{P}_{\mathrm{s}}}=\mathrm{0.5}$ hPa, the p_{C} accounts for around 7 %, while the ${p}_{{P}_{\mathrm{s}}}$ accounts for around 4 % of the total σ_{PWV}. By using different T_{m} estimations, the variations of p_{C} and ${p}_{{P}_{\mathrm{s}}}$ are both within 4 %. However, the p_{Q} varies more evidently. It accounts for averages of 55.69 %, 40.77 %, 30.70 %, 23.53 % and 24.11 % of the σ_{PWV} with the estimations of T_{m_Bevis}, T_{m_LatR}, T_{m_static}, T_{m_varying} and T_{m_GPT2w}, respectively. The p_{ZTD} rises with the reduction of p_{Q}, e.g., from 36.23 % of T_{m_Bevis} to 62.53 % of T_{m_varying}. On the other hand, with ${\mathit{\sigma}}_{{P}_{\mathrm{s}}}=\mathrm{5}$ hPa, the ${p}_{{P}_{\mathrm{s}}}$ accounts for more than 75 % of the σ_{PWV}, while the p_{Q} decreases from 14.21 % of T_{m_Bevis} to 3.9 % of T_{m_varying}.
At another representative station, the IGRA station no. 50557 (49.17^{∘} N 125.22^{∘} E, in northeastern China), the mean PWV is only 12.17 mm. The RMSEs of T_{m_Bevis}, T_{m_LatR}, T_{m_static}, T_{m_varying} and T_{m_GPT2w} are 5.16, 3.94, 3.54, 2.99 and 5.10 K. We can see that the accuracy of T_{m} has been improved significantly. However, because of the low average value of PWV, the p_{ZTD} averagely contributes over 73.5 % of the σ_{PWV}, while the p_{Q} averagely contributes less than 10.5 % assuming ${\mathit{\sigma}}_{{P}_{\mathrm{s}}}=\mathrm{0.5}$ hPa and less than 1.5 % assuming ${\mathit{\sigma}}_{{P}_{\mathrm{s}}}=\mathrm{5}$ hPa. But such a discussion only concerns the average values. In fact, even at this station there are still some high values of PWV, for example at 12:00 UTC 22 July 2016, the PWV reached 48 mm. For the observations with high PWV, the improvement in the accuracy of T_{m} can still exert a significant positive impact on the reduction of p_{Q}.
It is worth mentioning that the uncertainty in ZHD may be underestimated in some situations. There are two reasons for this. Firstly, the calculation of ZHD assumes that the water vapor does not contribute to the mass of the atmosphere. The ZHD error introduced by this assumption is often negligible. But in some very wet regions, the mass of water vapor could produce significant errors in the ZHD calculation. Secondly, and more importantly, the error of P_{s} in Eq. (1) can sometimes be very large. Small ${\mathit{\sigma}}_{{P}_{\mathrm{s}}}$ is reasonable when the surface barometer is calibrated routinely and equipped together with the GPS antenna. However, if there were significant height difference between the GPS antenna and the barometer, the error for ZHD would increase significantly. Snajdrova et al. (2006) found that 10 m of height difference approximately causes a difference of 3 mm in the ZHD. On the other hand, P_{s} can be generated from NWP data if there are no nearby barometers at GPS site. The error of P_{s} could be very large using this method (Means and Cayan, 2013; Jiang et al., 2016). In these cases, the GPS–PWV error reduction will be very limited due to the more precise T_{m} estimation.
5.2 Impact of real T_{m} estimation
To study the impact of T_{m} on the real GPS–PWV retrieval, we first downloaded GPS ZTD products (Byun and BarSever, 2009) at 74 IGS sites in the year 2016 from the NASA Crustal Dynamics Data Information System (CDDIS) ftp address (ftp://cddis.gsfc.nasa.gov/pub/gps/products/troposphere/zpd, last access: 25 February 2019). These selected GPS sites were equipped with meteorological sensors so that the surface pressure and temperature measurements could also be obtained. ZHD was calculated using Eq. (1). It is subtracted from ZTD to obtain ZWD. Then, T_{m} was generated with six approaches: the first five T_{m} series were T_{m_Bevis}, T_{m_LatR}, T_{m_static}, T_{m_varying} and T_{m_GPT2w}. The sixth T_{m} was integrated from the ERAInterim profiles and interpolated to each GPS site (Jiang et al., 2016; Wang et al., 2016). Finally, the GPS–PWV was generated from the ZWD and the six different T_{m} estimates leading to over 100 compared points for each GPS–PWV series. We denoted these GPS–PWV sets as PWV_{BTm}, PWV_{LTm}, PWV_{STm}, PWV_{VTm}, PWV_{GTm} and PWV_{ETm}. The only difference between these GPS and PWV estimations is the T_{m} estimation model; therefore, the impact of other errors is excluded.
The T_{m} from ERAInterim is believed to be the most accurate among our T_{m} estimates at the selected GPS sites. We therefore took the PWV_{ETm} as reference values to assess the other PWV. The relative RMSEs of PWV_{BTm}, PWV_{LTm}, PWV_{STm}, PWV_{VTm} and PWV_{GTm} at these selected stations were calculated and are illustrated in Fig. 11. Detailed statistics are given in Table 5. The mean relative error of all sites drops from 1.18 % of the PWV_{BTm} to 0.91 % of the PWV_{VTm}. PWV_{VTm} has the minimum mean relative errors at 51.35 % of the sites, while PWV_{STm} is superior at 27.03 % of the sites. PWV_{STm} and PWV_{VTm} obtain relative RMSEs smaller than 1.0 % at 55 sites, while only 28 sites of PWV_{BTm}, 31 sites of PWV_{LTm} and 22 sites of PWV_{GTm} perform similarly. For example, at the ALIC site (23.67^{∘} S 133.89^{∘} E, in Australia), with a mean PWV of approximately 23 mm, the relative RMSE dropped from 1.97 % of PWV_{BTm} to 1.10 % of PWV_{VTm}. The time series of the relative differences of PWV_{BTm}, PWV_{LTm}, PWV_{STm}, PWV_{VTm} and PWV_{GTm} are given in Fig. 12. We found that some relative RMSEs could be reduced by more than 2 % from PWV_{BTm} to PWV_{VTm}. Obviously, PWV_{BTm} and PWV_{LTm} have larger relative errors throughout the year, while the PWV differences are significantly larger only in the summer season (when the PWV values are highest). Apparently, the T_{m} variations in summer are not modeled well by either the Bevis model and the latituderelated model. PWV_{STm} eliminate those large differences but still retain some residual errors, which are removed by more than 0.5 mm in PWV_{VTm}. PWV_{GTm} has some large errors during the period from May to July. All of these results demonstrate that our timevarying model has a precision advantage.
5.3 Comparisons between GPS–PWV and radiosonde PWV
Among our selected 74 IGS sites, there are only 11 sites located within 5 km to a nearby IGRA radiosonde station. At these common stations, we generated PWV from the radiosonde data (PWV_{RS}) by adjusting the sounding profiles to the heights of IGS sites. It is worth noting that a geoid undulation correction should be carried out on each IGS site geoid height (Jiang et al., 2016). Then, we compared PWV_{BTm}, PWV_{LTm}, PWV_{STm}, PWV_{VTm}, PWV_{GTm} and PWV_{ETm} with PWV_{RS}. Figure 13 shows the statistics. The RMSEs of GPS–PWV are approximately 1–5 mm. Comparisons indicate that the RMSEs of different GPS–PWV retrievals are very close (differences < 0.2 mm) regardless of the applied T_{m} sources at most of the selected sites. This means that other errors (e.g., ZTD estimation errors or sounding sensors errors) instead of the T_{m} make up the bulk of the differences between the GPS–PWV and the radiosonde PWV. Actually, each sounding does not represent the vertical sounding centered at the radiosonde site because of the complex path of the balloon. And GPS–PWV represents the averaged value of the water vapor zenithal projection from all the slant signal paths during the observation period. Such differences can introduce significant uncertainty into our comparisons. However, we still found obvious gaps between PWV at the NRIL station (88.36^{∘} N 69.36^{∘} E, 4.1 km away from the IGRA station no. 23078 in Russia). The RMSE decreases from 2.29 mm of PWV_{BTm} to 1.84 mm of PWV_{VTm} and 1.42 mm of PWV_{ETm}. As shown in Fig. 14, the large PWV differences appear mainly from May to September. During those five months, the mean GPS–PWV difference to PWV_{RS} decreases by over 30 % from 2.52 mm of PWV_{BTm} to 1.67 mm of PWV_{VTm}, and the reductions of GPS–PWV error are mainly around 1–2 mm. This is attributed to the wetter atmosphere in these months. As indicated by the uncertainty analysis in Sect. 5.1, the improvement in the accuracy of T_{m} can be translated into more error reduction in the GPS–PWV retrieval with higher values of PWV.
We developed two global gridded T_{s}–T_{m} models, which are, respectively, static and timevarying with a spatial resolution of 0.75^{∘} × 0.75^{∘}. The models are established by analyzing the ERAInterim reanalysis data sets covering the years 2009–2012, which indicated the significant spatial–temporal variations in T_{s}–T_{m} relationship as well as the radiosondes covering the same period. The annual, semiannual and diurnal variations in T_{s}–T_{m} relationship are considered in the timevarying model. The timevarying global gridded T_{s}–T_{m} model has a significant global precision advantage over the other globally applied models, including the Bevis equation, the latituderelated model and the GPT2w model. The average RMSE of T_{m} reduces by approximately 1 K. At over 90 % of the radiosonde sites, our timevarying model has RMSE smaller than 4 K, while the RMSEs larger than 5 K nearly disappear. On the other hand, in the Bevis model or in the latituderelated model, there are more than 17 % of the radiosonde sites with RMSEs larger than 5 K. Multiple statistical tests at the 5 % significance level identified the significant superiority of our varying model at more than 60 % of the radiosonde sites. Analyses at the specific stations demonstrate that the errors larger than 5 K in the estimated T_{m} series can be eliminated by our varying T_{s}–T_{m} model.
More precise T_{m} estimations can decrease by around 20 % of the uncertainty in the conversion factor Q, which maps GPS–ZWD to GPS–PWV, and the reduction can be even more than 50 % at some stations. The contribution of the uncertainty associated with Q to the total GPS–PWV uncertainty also declines when using a more precise T_{m} model. The reduction is related to the value of PWV and the uncertainty in the surface pressure. With GPS–PWV higher than 50 mm, the uncertainty associated with Q contributes more than 55 % of the uncertainty in GPS–PWV when using the Bevis equation and less than 25 % when using our varying T_{s}–T_{m} model, assuming the ZTD and the surface pressure are measured accurately with the uncertainties of 4 mm and 0.5 hPa, respectively. However, the uncertainty in ZTD or in surface pressure would dominate the error budget of GPS–PWV (>70 %) if the value of GPS–PWV were small or the uncertainty in surface pressure were large. In these cases, the uncertainty associated with Q only contributes around 10 % of the GPS–PWV uncertainty or even smaller. Taking the GPS–PWV, using ERAInterim T_{m} estimates at 74 IGS sites as the references, we found that the GPS–PWV using our timevarying T_{s}–T_{m} model obtained the minimum mean relative error at 51.35 % of the sites, while the GPS–PWV using the static gridded T_{s}–T_{m} model is superior at only 27.03 % of the sites. The differences between GPS–PWV and radiosonde PWV are approximately 1–5 mm. And our varying T_{s}–T_{m} model can reduce the error in the GPS–PWV retrieval by 30 % (around 1–2 mm) with respect to the Bevis equation.
According to our experiments, we are confident that the timevarying global gridded T_{s}–T_{m} models presented here will help us to retrieve GPS PWV more precisely and to study the precise PWV variations in high temporal resolution. The Matlab array file consisting of the global gridded coefficients in our model, as well as codes for interpolating coefficients at any given location, is provided in the supplement of this study.

Radiosonde data: ftp://ftp.ncdc.noaa.gov/pub/data/igra (IGRA radiosonde data, 2019);

ERAInterim project: https://doi.org/10.5065/D6CR5RD9 (European Centre for MediumRange Weather Forecasts, 2019);

GPSZTD product: ftp://cddis.gsfc.nasa.gov/pub/gps/products/troposphere/zpd (CDDIS GPSZTD Product, 2019).
The supplement related to this article is available online at: https://doi.org/10.5194/amt1212332019supplement.
PJ, SY and YW conceived and designed the experiments; PJ, YL performed the experiments and analyzed the results; and DC and YL processed the data. All authors contributed to the writing of the paper.
The authors declare that they have no conflict of interest.
This study is supported by the National Natural Science Foundation of China
(no. 41604028), the Anhui Provincial Natural Science Foundation (no. 1708085QD83),
and the Doctoral Research Startup Funds Projects of Anhui University (no. J01001966).
The authors thank the European Centre for MediumRange Weather Forecasts for
providing the ERAInterim data set. We also thank the National Centers for
Environmental Information for the IGRA data sets and International GNSS Service
for the GNSS troposphere products.
Edited by: Roeland Van Malderen
Reviewed by: David Adams and four anonymous referees
Adams, D. K., Barbosa, H. M. J., and De Los Rios, K. P. G.: A Spatiotemporal Water VaporDeep Convection Correlation Metric Derived from the Amazon Dense GNSS Meteorological Network, Mon. Weather Rev., 145, 279–288, https://doi.org/10.1175/mwrd160140.1, 2017.
Adler, B., Kalthoff, N., Kohler, M., Handwerker, J., Wieser, A., Corsmeier, U., Kottmeier, C., Lambert, D., and Bock, O.: The variability of water vapour and preconvective conditions over the mountainous island of Corsica, Q. J. Roy. Meteorol. Soc., 142, 335–346, https://doi.org/10.1002/qj.2545, 2016.
Bevis, M., Businger, S., Herring, T. A., Rocken, C., Anthes, R. A., and Ware, R. H.: GPS meteorology: Remote sensing of atmospheric water vapor using the global positioning system, J. Geophys. Res.Atmos., 97, 15787–15801, https://doi.org/10.1029/92JD01517, 1992.
Bohm, J., Moller, G., Schindelegger, M., Pain, G., and Weber, R.: Development of an improved empirical model for slant delays in the troposphere (GPT2w), GPS Solutions, 19, 433–441, https://doi.org/10.1007/s1029101404037, 2015.
Brown, M. B. and Forsythe, A. B.: Robust Tests for the Equality of Variances, Publ. Am. Stat. Assoc., 69, 364–367, 1974.
Byun, S. H. and BarSever, Y. E.: A new type of troposphere zenith path delay product of the international GNSS service, J. Geodesy, 83, 367–373, https://doi.org/10.1007/s0019000802888, 2009.
Campmany, E., Bech, J., RodriguezMarcos, J., Sola, Y., and Lorente, J.: A comparison of total precipitable water measurements from radiosonde and sunphotometers, Atmosp. Res., 97, 385–392, https://doi.org/10.1016/j.atmosres.2010.04.016, 2010.
CDDIS GPSZTD Product: ftp://cddis.gsfc.nasa.gov/pub/gps/products/troposphere/zpd, last access: 26 February 2019.
Chen, P., Yao, W. Q., and Zhu, X. J.: Realization of global empirical model for mapping zenith wet delays onto precipitable water using NCEP reanalysis data, Geophys. J. Int., 198, 1748–1757, https://doi.org/10.1093/gji/ggu223, 2014.
Ciesielski, P. E., Chang, W. M., Huang, S. C., Johnson, R. H., Jou, B. J. D., Lee, W. C., Lin, P. H., Liu, C. H., and Wang, J. H.: QualityControlled UpperAir Sounding Dataset for TiMREX/SoWMEX: Development and Corrections, J. Atmos. Ocean. Tech., 27, 1802–1821, https://doi.org/10.1175/2010jtecha1481.1, 2010.
European Centre for MediumRange Weather Forecasts: ERAInterim Project, Research Data Archive at the National Center for Atmospheric Research, Computational and Information Systems Laboratory, https://doi.org/10.5065/D6CR5RD9, 2019.
Haase, J., Ge, M., Vedel, H., and Calais, E.: Accuracy and variability of GPS tropospheric delay measurements of water vapor in the western Mediterranean, J. Appl. Meteorol., 42, 1547–1568, https://doi.org/10.1175/15200450(2003)042<1547:AAVOGT>2.0.CO;2, 2003.
Hogg, R. V. and Ledolter, J.: Enginering Statistics, Macmillan, New York, 1987.
IGRA radiosonde data: https://www1.ncdc.noaa.gov/pub/data/igra/, last access: 26 Februray 2019.
Itterly, K. F., Taylor, P. C., and Dodson, J. B.: Sensitivity of the Amazonian Convective Diurnal Cycle to Its Environment in Observations and Reanalysis, J. Geophys. Res.Atmos., 23, 12621–12646, https://doi.org/10.1029/2018JD029251, 2018.
Jade, S. and Vijayan, M. S. M.: GPSbased atmospheric precipitable water vapor estimation using meteorological parameters interpolated from NCEP global reanalysis data, J. Geophys. Res.Atmos., 113, D03106, https://doi.org/10.1029/2007jd008758, 2008.
Jiang, P., Ye, S. R., Chen, D. Z., Liu, Y. Y., and Xia, P. F.: Retrieving Precipitable Water Vapor Data Using GPS Zenith Delays and Global Reanalysis Data in China, Remote Sensing, 8, 389, https://doi.org/10.3390/rs8050389, 2016.
Karabatic, A., Weber, R., and Haiden, T.: Near realtime estimation of tropospheric water vapour content from ground based GNSS data and its potential contribution to weather nowcasting in Austria, Adv. Space Res., 47, 1691–1703, https://doi.org/10.1016/j.asr.2010.10.028, 2011.
Kealy, J., Foster, J., and Businger, S.: GPS meteorology: An investigation of oceanbased precipitable water estimates, J. Geophys. Res.Atmos., 117, D17303, https://doi.org/10.1029/2011jd017422, 2012.
Lan, Z., Zhang, B., and Geng, Y.: Establishment and analysis of global gridded T_{m}–T_{s} relationship model, Geodesy Geodynam., 7, 101–107, https://doi.org/10.1016/j.geog.2016.02.001, 2016.
Lee, J., Park, J.U., Cho, J., Baek, J., and Kim, H. W.: A characteristic analysis of fog using GPSderived integrated water vapour, Meteorol. Appl., 17, 463–473, https://doi.org/10.1002/met.182, 2010.
Li, X., Zhang, L., Cao, X. J., Quan, J. N., Wang, T. H., Liang, J. N., and Shi, J. S.: Retrieval of precipitable water vapor using MFRSR and comparison with other multisensors over the semiarid area of northwest China, Atmos. Res., 172, 83–94, https://doi.org/10.1016/j.atmosres.2015.12.015, 2016.
Liu, Z. Z., Li, M., Zhong, W. K., and Wong, M. S.: An approach to evaluate the absolute accuracy of WVR water vapor measurements inferred from multiple water vapor techniques, J. Geodynam., 72, 86–94, https://doi.org/10.1016/j.jog.2013.09.002, 2013.
Lu, C. X., Li, X. X., Nilsson, T., Ning, T., Heinkelmann, R., Ge, M. R., Glaser, S., and Schuh, H.: Realtime retrieval of precipitable water vapor from GPS and BeiDou observations, J. Geodesy, 89, 843–856, https://doi.org/10.1007/s0019001508180, 2015.
Mahoney, K., Jackson, D. L., Neiman, P., Hughes, M., Darby, L., Wick, G., White, A., Sukovich, E., and Cifelli, R.: Understanding the role of atmospheric rivers in heavy precipitation in the Southeast US, Mon. Weather Rev., 144, 1617–1632, https://doi.org/10.1175/MWRD150279.1, 2016.
Means, J. D.: GPS Precipitable Water as a Diagnostic of the North American Monsoon in California and Nevada, J. Climate, 26, 1432–1444, https://doi.org/10.1175/jclid1200185.1, 2013.
Means, J. D. and Cayan, D.: Precipitable Water from GPS Zenith Delays Using North American Regional Reanalysis Meteorology, J. Atmos. Ocean. Tech., 30, 485–495, https://doi.org/10.1175/jtechd1200064.1, 2013.
Ning, T., Elgered, G., Willen, U., and Johansson, J. M.: Evaluation of the atmospheric water vapor content in a regional climate model using groundbased GPS measurements, J. Geophys. Res.Atmos., 118, 329–339, https://doi.org/10.1029/2012jd018053, 2013.
Ning, T., Wang, J., Elgered, G., Dick, G., Wickert, J., Bradke, M., Sommer, M., Querel, R., and Smale, D.: The uncertainty of the atmospheric integrated water vapour estimated from GNSS observations, Atmos. Meas. Tech., 9, 79–92, https://doi.org/10.5194/amt9792016, 2016.
Pacione, R. and Vespe, F.: Comparative studies for the assessment of the quality of nearrealtime GPSderived atmospheric parameters, J. Atmos. Ocean. Tech., 25, 701–714, https://doi.org/10.1175/2007jtecha935.1, 2008.
PerezRamirez, D., Whiteman, D. N., Smirnov, A., Lyamani, H., Holben, B. N., Pinker, R., Andrade, M., and AladosArboledas, L.: Evaluation of AERONET precipitable water vapor versus microwave radiometry, GPS, and radiosondes at ARM sites, J. Geophys. Res.Atmos., 119, 9596–9613, https://doi.org/10.1002/2014jd021730, 2014.
Rocken, C., Johnson, J., Van Hove, T., and Iwabuchi, T.: Atmospheric water vapor and geoid measurements in the open ocean with GPS, Geophys. Res. Lett., 32, L12813, https://doi.org/10.1029/2005gl022573, 2005.
Rohm, W., Yuan, Y. B., Biadeglgne, B., Zhang, K. F., and Le Marshall, J.: Groundbased GNSS ZTD/IWV estimation system for numerical weather prediction in challenging weather conditions, Atmos. Res., 138, 414–426, https://doi.org/10.1016/j.atmosres.2013.11.026, 2014.
Sheng, P., Mao, J., Li, J., Ge, Z., Zhang, A., Sang, J., Pan, N., and Zhang, H.: Atmospheric Physics, 2nd Edn., Peking University Press, Beijing, 2013.
Singh, D., Ghosh, J. K., and Kashyap, D.: Weighted mean temperature model for extra tropical region of India, J. Atmos. Sol.Terr. Phys., 107, 48–53, https://doi.org/10.1016/j.jastp.2013.10.016, 2014.
Snajdrova, K., Boehm, J., Willis, P., Haas, R., and Schuh, H.: Multitechnique comparison of tropospheric zenith delays derived during the CONT02 campaign, J. Geodesy, 79, 613–623, https://doi.org/10.1007/s001900050010z, 2006.
Song, J. J., Wang, Y., and Tang, J. P.: A Hiatus of the Greenhouse Effect, Sci. Rep., 6, 9, https://doi.org/10.1038/srep33315, 2016.
Suresh Raju, C., Saha, K., Thampi, B. V., and Parameswaran, K.: Empirical model for mean temperature for Indian zone and estimation of precipitable water vapor from ground based GPS measurements, Ann. Geophys., 25, 1935–1948, https://doi.org/10.5194/angeo2519352007, 2007.
Van Baelen, J. and Penide, G.: 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, 2009.
Wang, J. H., Zhang, L. Y., and Dai, A.: Global estimates of watervaporweighted mean temperature of the atmosphere for GPS applications, J. Geophys. Res.Atmos., 110, D21101, https://doi.org/10.1029/2005jd006215, 2005.
Wang, X. M., Zhang, K. F., Wu, S. Q., Fan, S. J., and Cheng, Y. Y.: Water vaporweighted mean temperature and its impact on the determination of precipitable water vapor and its linear trend, J. Geophys. Res.Atmos., 121, 833–852, https://doi.org/10.1002/2015jd024181, 2016.
Wang, X. Y., Song, L. C., and Cao, Y. C.: Analysis of the weighted mean temperature of china based on sounding and ECMWF reanalysis data, Acta Meteorol. Sin., 26, 642–652, https://doi.org/10.1007/s1335101205082, 2012.
Yao, Y., Zhang, B., Xu, C., and Yan, F.: Improved one/multiparameter models that consider seasonal and geographic variations for estimating weighted mean temperature in groundbased GPS meteorology, J. Geodesy, 88, 273–282, https://doi.org/10.1007/s0019001306846, 2014.
Yao, Y. B., Zhu, S., and Yue, S. Q.: A globally applicable, seasonspecific model for estimating the weighted mean temperature of the atmosphere, J. Geodesy, 86, 1125–1135, https://doi.org/10.1007/s0019001205681, 2012.
Yao, Y. B., Zhang, B., Xu, C. Q., and Chen, J. J.: Analysis of the global T(m)–T(s) correlation and establishment of the latituderelated linear model, Chinese Sci. Bull., 59, 2340–2347, https://doi.org/10.1007/s1143401402759, 2014.