<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE article PUBLIC "-//NLM//DTD Journal Publishing with OASIS Tables v3.0 20080202//EN" "journalpub-oasis3.dtd">
<article xmlns:xlink="http://www.w3.org/1999/xlink" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:oasis="http://docs.oasis-open.org/ns/oasis-exchange/table" xml:lang="en" dtd-version="3.0" article-type="research-article">
  <front>
    <journal-meta><journal-id journal-id-type="publisher">AMT</journal-id><journal-title-group>
    <journal-title>Atmospheric Measurement Techniques</journal-title>
    <abbrev-journal-title abbrev-type="publisher">AMT</abbrev-journal-title><abbrev-journal-title abbrev-type="nlm-ta">Atmos. Meas. Tech.</abbrev-journal-title>
  </journal-title-group><issn pub-type="epub">1867-8548</issn><publisher>
    <publisher-name>Copernicus Publications</publisher-name>
    <publisher-loc>Göttingen, Germany</publisher-loc>
  </publisher></journal-meta>
    <article-meta>
      <article-id pub-id-type="doi">10.5194/amt-14-5593-2021</article-id><title-group><article-title>High-temporal-resolution wet delay gradients estimated from multi-GNSS and microwave radiometer observations</article-title><alt-title>Gradients from the GNSS and microwave radiometry</alt-title>
      </title-group><?xmltex \runningtitle{Gradients from the GNSS and microwave radiometry}?><?xmltex \runningauthor{T.~Ning and G.~Elgered}?>
      <contrib-group>
        <contrib contrib-type="author" corresp="yes" rid="aff1">
          <name><surname>Ning</surname><given-names>Tong</given-names></name>
          <email>tong.ning@lm.se</email>
        </contrib>
        <contrib contrib-type="author" corresp="no" rid="aff2">
          <name><surname>Elgered</surname><given-names>Gunnar</given-names></name>
          
        <ext-link>https://orcid.org/0000-0001-5711-0073</ext-link></contrib>
        <aff id="aff1"><label>1</label><institution>Lantmäteriet (the Swedish mapping, cadastral, and land registration authority), 80182, Gävle, Sweden</institution>
        </aff>
        <aff id="aff2"><label>2</label><institution>Department of Space, Earth and Environment, Chalmers University of Technology,<?xmltex \hack{\break}?> Onsala Space Observatory, 43992 Onsala, Sweden</institution>
        </aff>
      </contrib-group>
      <author-notes><corresp id="corr1">Tong Ning (tong.ning@lm.se)</corresp></author-notes><pub-date><day>17</day><month>August</month><year>2021</year></pub-date>
      
      <volume>14</volume>
      <issue>8</issue>
      <fpage>5593</fpage><lpage>5605</lpage>
      <history>
        <date date-type="received"><day>12</day><month>April</month><year>2021</year></date>
           <date date-type="accepted"><day>15</day><month>July</month><year>2021</year></date>
           <date date-type="rev-recd"><day>14</day><month>July</month><year>2021</year></date>
           <date date-type="rev-request"><day>6</day><month>May</month><year>2021</year></date>
      </history>
      <permissions>
        <copyright-statement>Copyright: © 2021 Tong Ning</copyright-statement>
        <copyright-year>2021</copyright-year>
      <license license-type="open-access"><license-p>This work is licensed under the Creative Commons Attribution 4.0 International License. To view a copy of this licence, visit <ext-link ext-link-type="uri" xlink:href="https://creativecommons.org/licenses/by/4.0/">https://creativecommons.org/licenses/by/4.0/</ext-link></license-p></license></permissions><self-uri xlink:href="https://amt.copernicus.org/articles/14/5593/2021/amt-14-5593-2021.html">This article is available from https://amt.copernicus.org/articles/14/5593/2021/amt-14-5593-2021.html</self-uri><self-uri xlink:href="https://amt.copernicus.org/articles/14/5593/2021/amt-14-5593-2021.pdf">The full text article is available as a PDF file from https://amt.copernicus.org/articles/14/5593/2021/amt-14-5593-2021.pdf</self-uri>
      <abstract><title>Abstract</title>
    <p id="d1e98">We have used 1 year of multi-GNSS observations at the Onsala Space Observatory on the Swedish west coast to estimate the linear horizontal
gradients in the wet propagation delay. The estimated gradients are compared to the corresponding ones from a microwave radiometer. We have
investigated different temporal resolutions from 5 <inline-formula><mml:math id="M1" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">min</mml:mi></mml:mrow></mml:math></inline-formula> to 1 d. Relative to the GPS-only solution and using an elevation cutoff angle of
10<inline-formula><mml:math id="M2" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula> and a temporal resolution of 5 <inline-formula><mml:math id="M3" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">min</mml:mi></mml:mrow></mml:math></inline-formula>, the improvement obtained for the solution using GPS, Glonass, and Galileo data is an
increase in the correlation coefficient of 11 % for the east gradient and 20 % for the north gradient. Out of all the different GNSS
solutions, the highest correlation is obtained for the east gradients and a resolution of 2 <inline-formula><mml:math id="M4" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">h</mml:mi></mml:mrow></mml:math></inline-formula>, while the best agreement for the north
gradients is obtained for 6 <inline-formula><mml:math id="M5" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">h</mml:mi></mml:mrow></mml:math></inline-formula>. The choice of temporal resolution is a compromise between getting a high correlation and the possibility of
detecting rapid changes in the gradient. Due to the differences in geometry of the observations, gradients which happen suddenly are either not
captured at all or captured but with much less amplitude by the GNSS data. When a weak constraint is applied in the estimation of process, the GNSS
data have an improved ability to track large gradients, however, at the cost of increased formal errors.</p>
  </abstract>
    </article-meta>
  </front>
<body>
      

<sec id="Ch1.S1" sec-type="intro">
  <label>1</label><title>Introduction</title>
      <p id="d1e151">An accurate modelling of the atmospheric effects on GNSS observations is relevant for both geodetic and meteorological applications, in forecasting as
well as in climate research. In geodetic applications the standard method is to estimate an equivalent zenith propagation delay together with a linear
horizontal gradient. Early results showed an improved repeatability for the estimated coordinates when estimating gradients using GPS data
<xref ref-type="bibr" rid="bib1.bibx2 bib1.bibx15" id="paren.1"/>. A recent study <xref ref-type="bibr" rid="bib1.bibx23" id="paren.2"/> found that estimating gradients with a temporal resolution of
1 <inline-formula><mml:math id="M6" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">h</mml:mi></mml:mrow></mml:math></inline-formula> can achieve even better positioning performance than strategies where gradients are estimated with resolutions of many hours up to 1 d.</p>
      <p id="d1e168">In meteorological applications the zenith total delay (ZTD) and horizontal gradients may be assimilated directly into the forecasting model; see e.g.
<xref ref-type="bibr" rid="bib1.bibx25" id="paren.3"/>. Inferred values of the zenith wet delay (ZWD) and the integrated water vapour (IWV) may be used to study long-term trends
<xref ref-type="bibr" rid="bib1.bibx1" id="paren.4"/>. Linear horizontal gradients estimated from the GNSS have been used to study specific meteorological conditions. For example, on
the island of Corsica, where the physical meaning of gradients in coastal areas with a steep topography was studied <xref ref-type="bibr" rid="bib1.bibx16" id="paren.5"/>, and in
Texas, USA, where significant gradients during Hurricane Harvey were reported <xref ref-type="bibr" rid="bib1.bibx9" id="paren.6"/>.</p>
      <p id="d1e183">The quality of the estimated gradients has been assessed by comparisons to independent measurements, such as using a microwave radiometer (in the
following this instrument is referred to as a water vapour radiometer, WVR), the space geodetic technique of very-long-baseline interferometry (VLBI),
and numerical weather models.</p>
      <p id="d1e186">Such an assessment was carried out by <xref ref-type="bibr" rid="bib1.bibx6" id="text.7"/>, where the GPS-derived gradients were compared with the ones obtained from WVR, VLBI,
and the European Centre for Medium-Range Weather Forecasts (ECMWF)<?pagebreak page5594?> analyses. The results show that the best agreement is obtained when an elevation
cutoff angle of 3<inline-formula><mml:math id="M7" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula> is applied in the GPS data processing in spite of the fact that the radiometer did not observe below 20<inline-formula><mml:math id="M8" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula>. They
also found that a homogeneous and frequent sampling of the sky is a critical parameter for gradient estimation. Using multi-GNSS observations instead
of GPS only, <xref ref-type="bibr" rid="bib1.bibx12" id="text.8"/> found a significant increase in the correlation coefficient from below 0.5 to about 0.6 when compared to the
gradients computed from the ECMWF reanalysis product. The corresponding decrease in the root-mean-square (RMS) difference in the gradients was
25 %–35 % for multi-GNSS processing. The temporal resolutions of such comparisons are to our knowledge so far limited to 1 <inline-formula><mml:math id="M9" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">h</mml:mi></mml:mrow></mml:math></inline-formula> for WVRs
<xref ref-type="bibr" rid="bib1.bibx13" id="paren.9"/>, 2 <inline-formula><mml:math id="M10" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">h</mml:mi></mml:mrow></mml:math></inline-formula> for VLBI <xref ref-type="bibr" rid="bib1.bibx21" id="paren.10"/>, and 1 <inline-formula><mml:math id="M11" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">h</mml:mi></mml:mrow></mml:math></inline-formula> for numerical weather models <xref ref-type="bibr" rid="bib1.bibx11" id="paren.11"/>.</p>
      <p id="d1e248">The aim of this study is to assess the quality of estimated gradients from multi-GNSS observations with temporal resolutions as high as 5 <inline-formula><mml:math id="M12" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">min</mml:mi></mml:mrow></mml:math></inline-formula> using
independent WVR data. Section <xref ref-type="sec" rid="Ch1.S2"/> describes how the gradients are estimated from the GNSS and the WVR data. In
Sect. <xref ref-type="sec" rid="Ch1.S3.SS1"/> we present the gradients estimated for different GNSS constellations and different elevation cutoff angles. These are
thereafter compared to the WVR data, first in Sect. <xref ref-type="sec" rid="Ch1.S3.SS2"/> for the highest temporal resolution of 5 <inline-formula><mml:math id="M13" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">min</mml:mi></mml:mrow></mml:math></inline-formula> and then in
Sect. <xref ref-type="sec" rid="Ch1.S3.SS3"/> over timescales up to 1 d, both for the entire data set and for a specific event of short-lived gradients associated
with rapid changes in the ZWD. In Sect. <xref ref-type="sec" rid="Ch1.S3.SS4"/> we also study the impact of using a weaker constraint for the random walk process of the
GNSS gradient time series. Finally, Sect. <xref ref-type="sec" rid="Ch1.S4"/> gives our conclusions.</p>
</sec>
<sec id="Ch1.S2">
  <label>2</label><title>Data sets</title>
<sec id="Ch1.S2.SS1">
  <label>2.1</label><title>GNSS</title>
      <p id="d1e295">We have analysed 1 year (1 January–31 December 2019) of ground-based GNSS observations acquired from one station (ONS1) located at the Onsala Space
Observatory, on the west coast of Sweden. For comparison purposes we also used GNSS data from June and July 2019 acquired at the collocated station
ONSA. The data processing was carried out using GipsyX v.1.5
(<uri>https://gipsy-oasis.jpl.nasa.gov/gipsy/docs/releaseNotes-GipsyX-1.5.pdf</uri>, last access: 12 August 2021) with the precise point
positioning (PPP) strategy <xref ref-type="bibr" rid="bib1.bibx24" id="paren.12"/>. The input to the processing was ionosphere-free linear combinations formed by acquired GNSS
phase-delay observations, while the output included station coordinates, clock biases, and tropospheric parameters. The final multi-GNSS orbit and
clock products used were provided by the Center for Orbit Determination in Europe (CODE) <xref ref-type="bibr" rid="bib1.bibx18" id="paren.13"/>. An ocean tide loading correction using the
FES2004 model <xref ref-type="bibr" rid="bib1.bibx14" id="paren.14"/> was applied, while no atmospheric pressure loading corrections were used. The absolute calibration of the phase centre
variations (PCVs) for all antennas (from the file igs14_2101.atx) was implemented <xref ref-type="bibr" rid="bib1.bibx20" id="paren.15"/>. We used the Vienna Mapping Function 1 (VMF1)
<xref ref-type="bibr" rid="bib1.bibx3" id="paren.16"/> to map the zenith delay, and the gradient mapping function was the one presented by <xref ref-type="bibr" rid="bib1.bibx2" id="text.17"/>.</p>
      <p id="d1e320">The ZTD and linear horizontal delay gradients were estimated every 5 <inline-formula><mml:math id="M14" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">min</mml:mi></mml:mrow></mml:math></inline-formula> using a random walk model with a standard deviation (SD) of 10 and
0.3 <inline-formula><mml:math id="M15" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">mm</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:msqrt><mml:mi mathvariant="normal">h</mml:mi></mml:msqrt><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>, respectively. No weighting was applied. The SD value used for the ZTD is given by
<xref ref-type="bibr" rid="bib1.bibx10" id="text.18"/>, where they found a temporal variability in the wet delay, derived from 71 d of microwave radiometer measurements, varying
in the interval 3–22 <inline-formula><mml:math id="M16" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">mm</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:msqrt><mml:mi mathvariant="normal">h</mml:mi></mml:msqrt><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>. Because our focus here is on high temporal resolution and because we expect wet gradients to sometimes
be short-lived, we also use a weaker constraint (in Sect. <xref ref-type="sec" rid="Ch1.S3.SS4"/>) of 1.0 <inline-formula><mml:math id="M17" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">mm</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:msqrt><mml:mi mathvariant="normal">h</mml:mi></mml:msqrt><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> for the SD in the random walk model
for the gradients.</p>
      <p id="d1e391">The zenith hydrostatic delay (ZHD) was calculated using ground pressure measurements <xref ref-type="bibr" rid="bib1.bibx19" id="paren.19"/>, and thereafter the ZWD was obtained by
subtracting the ZHD from the ZTD. The gradients estimated from the GNSS data are total gradients, and they were interpreted as the sum of hydrostatic
and wet components.</p>
      <p id="d1e397">The model used for the gradient estimation is presented by <xref ref-type="bibr" rid="bib1.bibx2" id="text.20"/>:
            <disp-formula id="Ch1.E1" content-type="numbered"><label>1</label><mml:math id="M18" display="block"><mml:mrow><mml:mi>S</mml:mi><mml:mo>(</mml:mo><mml:mi mathvariant="italic">ϵ</mml:mi><mml:mo>,</mml:mo><mml:mi mathvariant="italic">ϕ</mml:mi><mml:mo>)</mml:mo><mml:mo>=</mml:mo><mml:mi>m</mml:mi><mml:mo>(</mml:mo><mml:mi mathvariant="italic">ϵ</mml:mi><mml:mo>)</mml:mo><mml:mfenced open="[" close="]"><mml:mrow><mml:mi>Z</mml:mi><mml:mo>+</mml:mo><mml:mtext>cot</mml:mtext><mml:mo>(</mml:mo><mml:mi mathvariant="italic">ϵ</mml:mi><mml:mo>)</mml:mo><mml:mfenced open="(" close=")"><mml:mrow><mml:msub><mml:mi>G</mml:mi><mml:mi mathvariant="normal">n</mml:mi></mml:msub><mml:mi>cos⁡</mml:mi><mml:mo>(</mml:mo><mml:mi mathvariant="italic">ϕ</mml:mi><mml:mo>)</mml:mo><mml:mo>+</mml:mo><mml:msub><mml:mi>G</mml:mi><mml:mi mathvariant="normal">e</mml:mi></mml:msub><mml:mi>sin⁡</mml:mi><mml:mo>(</mml:mo><mml:mi mathvariant="italic">ϕ</mml:mi><mml:mo>)</mml:mo></mml:mrow></mml:mfenced></mml:mrow></mml:mfenced><mml:mo>,</mml:mo></mml:mrow></mml:math></disp-formula>
          where <inline-formula><mml:math id="M19" display="inline"><mml:mrow><mml:mi>S</mml:mi><mml:mo>(</mml:mo><mml:mi mathvariant="italic">ϵ</mml:mi><mml:mo>,</mml:mo><mml:mi mathvariant="italic">ϕ</mml:mi><mml:mo>)</mml:mo></mml:mrow></mml:math></inline-formula> is a slant delay for a certain elevation and azimuth angle; <inline-formula><mml:math id="M20" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula> and <inline-formula><mml:math id="M21" display="inline"><mml:mrow><mml:mi>m</mml:mi><mml:mo>(</mml:mo><mml:mi mathvariant="italic">ϵ</mml:mi><mml:mo>)</mml:mo></mml:mrow></mml:math></inline-formula> are the zenith delay and the elevation
mapping function; and <inline-formula><mml:math id="M22" display="inline"><mml:mrow><mml:msub><mml:mi>G</mml:mi><mml:mi mathvariant="normal">n</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M23" display="inline"><mml:mrow><mml:msub><mml:mi>G</mml:mi><mml:mi mathvariant="normal">e</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> are the north and east horizontal gradient, respectively.</p>

      <?xmltex \floatpos{t}?><fig id="Ch1.F1" specific-use="star"><?xmltex \currentcnt{1}?><?xmltex \def\figurename{Figure}?><label>Figure 1</label><caption><p id="d1e543">Observations of the sky acquired from ONS1 for 0 to 24 <inline-formula><mml:math id="M24" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">h</mml:mi></mml:mrow></mml:math></inline-formula> on 24 July 2019 for GPS <bold>(a)</bold>, Glonass <bold>(b)</bold>, and Galileo <bold>(c)</bold>.</p></caption>
          <?xmltex \igopts{width=497.923228pt}?><graphic xlink:href="https://amt.copernicus.org/articles/14/5593/2021/amt-14-5593-2021-f01.png"/>

        </fig>

      <p id="d1e569">In order to compare to the wet component inferred by the WVR, we subtracted the hydrostatic component computed from the reanalysis product of the
ECMWF, ERA5, from the total gradient to get the GNSS wet gradient. The hydrostatic gradients at
the site are much less variable compared to the wet ones, and especially for timescales of minutes to hours <xref ref-type="bibr" rid="bib1.bibx6" id="paren.21"/>. The gradients
were calculated from ERA5 by vertical integration of the horizontal refractivity gradients times the height. The profile closest to the site was used
together with one profile to the east and one profile to the north to calculate the refractivity gradient profiles. The temporal resolution of the
ERA5 gradients is 1 <inline-formula><mml:math id="M25" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">h</mml:mi></mml:mrow></mml:math></inline-formula>, and we therefore interpolated them to a 5 <inline-formula><mml:math id="M26" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">min</mml:mi></mml:mrow></mml:math></inline-formula> resolution in order to perform the subtraction for the GNSS
data. For the same reason, we did not use wet gradients from ERA5 because we want to study the gradients with a temporal resolution down to
5 <inline-formula><mml:math id="M27" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">min</mml:mi></mml:mrow></mml:math></inline-formula>.</p>

      <?xmltex \floatpos{t}?><fig id="Ch1.F2" specific-use="star"><?xmltex \currentcnt{2}?><?xmltex \def\figurename{Figure}?><label>Figure 2</label><caption><p id="d1e601">Number of daily observations for each GNSS constellation acquired by ONS1 applying three different elevation cutoff angles (3, 10, and 15<inline-formula><mml:math id="M28" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula>).</p></caption>
          <?xmltex \igopts{width=355.659449pt}?><graphic xlink:href="https://amt.copernicus.org/articles/14/5593/2021/amt-14-5593-2021-f02.png"/>

        </fig>

      <p id="d1e619">The data processing was run for three different elevation cutoff angles (3, 10, and 15<inline-formula><mml:math id="M29" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula>). For each elevation cutoff angle we used four
different combinations of GNSS constellations in the processing: GPS only (G), GPS <inline-formula><mml:math id="M30" display="inline"><mml:mo>+</mml:mo></mml:math></inline-formula> Glonass (GR), GPS <inline-formula><mml:math id="M31" display="inline"><mml:mo>+</mml:mo></mml:math></inline-formula> Galileo (GE), and
GPS <inline-formula><mml:math id="M32" display="inline"><mml:mo>+</mml:mo></mml:math></inline-formula> Glonass <inline-formula><mml:math id="M33" display="inline"><mml:mo>+</mml:mo></mml:math></inline-formula> Galileo (GRE). Due to limitations in the receiver capacity, not all BeiDou<?pagebreak page5595?> observations were recorded. Therefore we decided
not to include BeiDou data in our GNSS data processing.</p>
      <p id="d1e659">An example of the sky coverage of the observations for different GNSS constellations, applying an elevation cutoff angle of 3<inline-formula><mml:math id="M34" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula>, is shown in
Fig. <xref ref-type="fig" rid="Ch1.F1"/> for the ONS1 station. The three systems show a similar geometry for the observations. At this latitude a significant
part of the sky, just north of the zenith direction, is never sampled. The Glonass satellite orbits have a higher inclination angle, implying that a
smaller part of the sky is not sampled compared to GPS and Galileo. As a consequence there are more observations from Glonass to the north, especially
below the elevation angle of 20<inline-formula><mml:math id="M35" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula>. It is therefore interesting to study how this difference will affect the quality of the estimated gradients.</p>
      <p id="d1e683">Figure <xref ref-type="fig" rid="Ch1.F2"/> depicts the number of daily observations given for each GNSS constellation and obtained for each elevation cutoff
angle. Some days have fewer or no data due to receiver problems, especially from 13 to 15 July, where no Galileo observations were recorded. Averaged
over the year, for the GPS system, the number of observations drops about 11 % and 23 % when the elevation cutoff angle changes from 3 to 10
and 15<inline-formula><mml:math id="M36" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula>, respectively. For the Galileo system, the corresponding decrease in the number of observations is 10 % and 22 %,
respectively, while for Glonass, the corresponding values are 8 % and 17 %, respectively.</p>

      <?xmltex \floatpos{t}?><fig id="Ch1.F3" specific-use="star"><?xmltex \currentcnt{3}?><?xmltex \def\figurename{Figure}?><label>Figure 3</label><caption><p id="d1e699">The water vapour radiometer (WVR) and GNSS stations (ONS1 and ONSA) at the Onsala Space Observatory.</p></caption>
          <?xmltex \igopts{width=398.338583pt}?><graphic xlink:href="https://amt.copernicus.org/articles/14/5593/2021/amt-14-5593-2021-f03.jpg"/>

        </fig>

      <?xmltex \floatpos{t}?><fig id="Ch1.F4"><?xmltex \currentcnt{4}?><?xmltex \def\figurename{Figure}?><label>Figure 4</label><caption><p id="d1e710">One cycle of the WVR observations consists of 52 observations. The cycle is repeated every 5 <inline-formula><mml:math id="M37" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">min</mml:mi></mml:mrow></mml:math></inline-formula>. Note that the zenith point is only measured once, the first time it passes over zenith.</p></caption>
          <?xmltex \igopts{width=176.407087pt}?><graphic xlink:href="https://amt.copernicus.org/articles/14/5593/2021/amt-14-5593-2021-f04.png"/>

        </fig>

</sec>
<sec id="Ch1.S2.SS2">
  <label>2.2</label><title>Water vapour radiometer (WVR)</title>
      <p id="d1e735">The WVR, shown in Fig. <xref ref-type="fig" rid="Ch1.F3"/>, is located close to the GNSS stations, 9 <inline-formula><mml:math id="M38" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">m</mml:mi></mml:mrow></mml:math></inline-formula> from ONSA and 59 <inline-formula><mml:math id="M39" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">m</mml:mi></mml:mrow></mml:math></inline-formula> from ONS1, with height differences of less than 2 <inline-formula><mml:math id="M40" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">m</mml:mi></mml:mrow></mml:math></inline-formula>. The WVR was designed in order to provide independent estimates of the wet propagation delays for space geodetic applications. It measures the sky brightness temperature on and off the water vapour emission line at 22 <inline-formula><mml:math id="M41" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">GHz</mml:mi></mml:mrow></mml:math></inline-formula>. More detailed specifications are given by <xref ref-type="bibr" rid="bib1.bibx6" id="text.22"/>.</p>
      <p id="d1e776">Starting in January 2019 the observations were scheduled in 5 <inline-formula><mml:math id="M42" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">min</mml:mi></mml:mrow></mml:math></inline-formula> long cycles with the ambition to sample the whole atmosphere at elevation
angles above 25<inline-formula><mml:math id="M43" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula>, which is illustrated in Fig. <xref ref-type="fig" rid="Ch1.F4"/>. Data were acquired throughout 2019, except from<?pagebreak page5596?> mid-August to early
October because of a failure caused by a thunderstorm.</p>
      <p id="d1e798">A four-parameter model was used to estimate the mean ZWD, a linear trend in the ZWD, and east and north linear horizontal gradients over 5 <inline-formula><mml:math id="M44" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">min</mml:mi></mml:mrow></mml:math></inline-formula>
<xref ref-type="bibr" rid="bib1.bibx4" id="paren.23"/>. Before the model was applied, all observations during rain and with a liquid water content larger than 0.7 <inline-formula><mml:math id="M45" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">mm</mml:mi></mml:mrow></mml:math></inline-formula> were
removed. Thereafter, the quality of the data was assessed through manual editing. Gain jumps occurring sometimes at the beginning of a 5 <inline-formula><mml:math id="M46" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">min</mml:mi></mml:mrow></mml:math></inline-formula>
cycle were frequent during the whole year. The jumps were later found to be caused by vibrations when the mechanical waveguide switch was activated at
the beginning of each 5 <inline-formula><mml:math id="M47" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">min</mml:mi></mml:mrow></mml:math></inline-formula> period. When such a jump was identified, one or several complete 5 <inline-formula><mml:math id="M48" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">min</mml:mi></mml:mrow></mml:math></inline-formula> cycles were removed. The jumps were
identified by viewing the ZWD during each day. The temporal resolution is then sufficient to identify a 5 <inline-formula><mml:math id="M49" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">min</mml:mi></mml:mrow></mml:math></inline-formula> long group of data that is
discontinuous with the adjacent 5 <inline-formula><mml:math id="M50" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">min</mml:mi></mml:mrow></mml:math></inline-formula> periods. Thereby we eliminated the possibility that the large gain jumps could have an impact on the
estimated ZWD trends and gradients for the 5 <inline-formula><mml:math id="M51" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">min</mml:mi></mml:mrow></mml:math></inline-formula> cycle since it is not synchronized with the estimation period. Smaller gain jumps may still
have degraded the accuracy of the estimated gradients. Thereafter, when the model was applied, we required that at least 40 of the 52 observations
were available in the 5 <inline-formula><mml:math id="M52" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">min</mml:mi></mml:mrow></mml:math></inline-formula> cycle period.</p>
      <p id="d1e877">There are 105 120 possible 5 <inline-formula><mml:math id="M53" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">min</mml:mi></mml:mrow></mml:math></inline-formula> data points in 1 year. After removing data acquired during rain, data that were unstable because of the
jumps, and all 5 <inline-formula><mml:math id="M54" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">min</mml:mi></mml:mrow></mml:math></inline-formula> periods where there were less than 40 (of the 52 scheduled) observations (typically caused by large liquid water
contents), we ended up with 56 612 data points. There were 14 236 periods of 5 <inline-formula><mml:math id="M55" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">min</mml:mi></mml:mrow></mml:math></inline-formula> during the time when the WVR was in the lab, meaning that
the gradients estimated from 62 % of the time when the WVR was operated were used to compare to the ones from the GNSS. An overall reason for applying
this strict editing was that the primary goal was to have accurate gradients from the WVR rather than a statistical characterization of the specific
atmospheric conditions at the site.</p>

      <?xmltex \floatpos{t}?><fig id="Ch1.F5" specific-use="star"><?xmltex \currentcnt{5}?><?xmltex \def\figurename{Figure}?><label>Figure 5</label><caption><p id="d1e907">The time series of the ZWD estimated from the WVR data for each 5 <inline-formula><mml:math id="M56" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">min</mml:mi></mml:mrow></mml:math></inline-formula> cycle. The radiometer was repaired from mid-August to early October.</p></caption>
          <?xmltex \igopts{width=341.433071pt}?><graphic xlink:href="https://amt.copernicus.org/articles/14/5593/2021/amt-14-5593-2021-f05.png"/>

        </fig>

      <p id="d1e924">Finally it is noted that the WVR estimates are completely independent of the corresponding estimates from the GNSS data. The study does not need to
assume that the WVR gradients are more accurate compared to the GNSS ones. The main advantage of the WVR gradients is that they are independent, and by
comparing these to the gradients from different GNSS solutions we can assess the different GNSS processing methods. Furthermore, since we want to
study the agreement with as high temporal resolution as possible, we do not apply constraints to the individual 5 <inline-formula><mml:math id="M57" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">min</mml:mi></mml:mrow></mml:math></inline-formula> gradients in order to
have them independent from adjacent estimates in terms of the atmospheric signals.</p>
      <p id="d1e935">The estimated ZWD from the WVR data is shown in Fig. <xref ref-type="fig" rid="Ch1.F5"/>. The seasonal dependence is clearly visible as well as a large short-term
variability. Because of the gain jumps this data set is not optimum for ZWD comparisons on an absolute scale, but our focus is on gradients.</p>
</sec>
</sec>
<?pagebreak page5597?><sec id="Ch1.S3">
  <label>3</label><title>Results</title>
<sec id="Ch1.S3.SS1">
  <label>3.1</label><title>Estimated gradients and their formal errors</title>
      <p id="d1e956">Before carrying out comparisons of the gradients estimated from different GNSS solutions with the WVR gradients, we investigate the characteristics of
the input data. Table <xref ref-type="table" rid="Ch1.T1"/> summarizes the statistics of estimated gradients and their corresponding formal errors. A few features
of the data are worth noting. The mean of and the variability (standard deviation) in the estimated gradient amplitude increase with increasing elevation
cutoff angle together with its mean formal error. When the elevation cutoff angle increases, a smaller number of observations were included for the
gradient estimation. For the GNSS, the geometry of the satellite constellation is also deteriorated for a larger elevation cutoff angle. As a result, the
formal error in the estimated gradient increases as well as the variability. In addition, when using a lower elevation, the larger volume sensed by
the GNSS introduces an averaging effect that reduces the mean amplitude of the estimated gradients (see <xref ref-type="bibr" rid="bib1.bibx6" id="altparen.24"/>).</p>

<?xmltex \floatpos{t}?><table-wrap id="Ch1.T1" specific-use="star"><?xmltex \currentcnt{1}?><label>Table 1</label><caption><p id="d1e967">The mean and standard deviations (SDs) of the east and the north gradients and the gradient amplitude, together with the mean and SD of their 1<inline-formula><mml:math id="M58" display="inline"><mml:mi mathvariant="italic">σ</mml:mi></mml:math></inline-formula> formal errors.</p></caption><oasis:table frame="topbot"><oasis:tgroup cols="13">
     <oasis:colspec colnum="1" colname="col1" align="left"/>
     <oasis:colspec colnum="2" colname="col2" align="right"/>
     <oasis:colspec colnum="3" colname="col3" align="right" colsep="1"/>
     <oasis:colspec colnum="4" colname="col4" align="right"/>
     <oasis:colspec colnum="5" colname="col5" align="right" colsep="1"/>
     <oasis:colspec colnum="6" colname="col6" align="right"/>
     <oasis:colspec colnum="7" colname="col7" align="right" colsep="1"/>
     <oasis:colspec colnum="8" colname="col8" align="right"/>
     <oasis:colspec colnum="9" colname="col9" align="right" colsep="1"/>
     <oasis:colspec colnum="10" colname="col10" align="right"/>
     <oasis:colspec colnum="11" colname="col11" align="right" colsep="1"/>
     <oasis:colspec colnum="12" colname="col12" align="right"/>
     <oasis:colspec colnum="13" colname="col13" align="right"/>
     <oasis:thead>
       <oasis:row>
         <oasis:entry colname="col1">GNSS constellation</oasis:entry>
         <oasis:entry rowsep="1" namest="col2" nameend="col5" align="center" colsep="1">East, <inline-formula><mml:math id="M66" display="inline"><mml:mrow><mml:msub><mml:mi>G</mml:mi><mml:mi mathvariant="normal">e</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula></oasis:entry>
         <oasis:entry rowsep="1" namest="col6" nameend="col9" align="center" colsep="1">North, <inline-formula><mml:math id="M67" display="inline"><mml:mrow><mml:msub><mml:mi>G</mml:mi><mml:mi mathvariant="normal">n</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula></oasis:entry>
         <oasis:entry rowsep="1" namest="col10" nameend="col13" align="center">Amplitude, <inline-formula><mml:math id="M68" display="inline"><mml:msqrt><mml:mrow><mml:msubsup><mml:mi>G</mml:mi><mml:mi mathvariant="normal">e</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msubsup><mml:mo>+</mml:mo><mml:msubsup><mml:mi>G</mml:mi><mml:mi mathvariant="normal">n</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msubsup></mml:mrow></mml:msqrt></mml:math></inline-formula></oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">with elevation</oasis:entry>
         <oasis:entry rowsep="1" namest="col2" nameend="col3" align="center" colsep="1">Estimated </oasis:entry>
         <oasis:entry rowsep="1" namest="col4" nameend="col5" align="center" colsep="1">Formal error </oasis:entry>
         <oasis:entry rowsep="1" namest="col6" nameend="col7" align="center" colsep="1">Estimated </oasis:entry>
         <oasis:entry rowsep="1" namest="col8" nameend="col9" align="center" colsep="1">Formal error </oasis:entry>
         <oasis:entry rowsep="1" namest="col10" nameend="col11" align="center" colsep="1">Estimated </oasis:entry>
         <oasis:entry rowsep="1" namest="col12" nameend="col13" align="center">Formal error </oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">cutoff angle</oasis:entry>
         <oasis:entry colname="col2">Mean</oasis:entry>
         <oasis:entry colname="col3">SD</oasis:entry>
         <oasis:entry colname="col4">Mean</oasis:entry>
         <oasis:entry colname="col5">SD</oasis:entry>
         <oasis:entry colname="col6">Mean</oasis:entry>
         <oasis:entry colname="col7">SD</oasis:entry>
         <oasis:entry colname="col8">Mean</oasis:entry>
         <oasis:entry colname="col9">SD</oasis:entry>
         <oasis:entry colname="col10">Mean</oasis:entry>
         <oasis:entry colname="col11">SD</oasis:entry>
         <oasis:entry colname="col12">Mean</oasis:entry>
         <oasis:entry colname="col13">SD</oasis:entry>
       </oasis:row>
       <oasis:row rowsep="1">
         <oasis:entry colname="col1">(c.f. the WVR)</oasis:entry>
         <oasis:entry colname="col2">(mm)</oasis:entry>
         <oasis:entry colname="col3">(mm)</oasis:entry>
         <oasis:entry colname="col4">(mm)</oasis:entry>
         <oasis:entry colname="col5">(mm)</oasis:entry>
         <oasis:entry colname="col6">(mm)</oasis:entry>
         <oasis:entry colname="col7">(mm)</oasis:entry>
         <oasis:entry colname="col8">(mm)</oasis:entry>
         <oasis:entry colname="col9">(mm)</oasis:entry>
         <oasis:entry colname="col10">(mm)</oasis:entry>
         <oasis:entry colname="col11">(mm)</oasis:entry>
         <oasis:entry colname="col12">(mm)</oasis:entry>
         <oasis:entry colname="col13">(mm)</oasis:entry>
       </oasis:row>
     </oasis:thead>
     <oasis:tbody>
       <oasis:row>
         <oasis:entry colname="col1">G(PS)      3<inline-formula><mml:math id="M69" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col2">0.00</oasis:entry>
         <oasis:entry colname="col3">0.43</oasis:entry>
         <oasis:entry colname="col4">0.14</oasis:entry>
         <oasis:entry colname="col5">0.04</oasis:entry>
         <oasis:entry colname="col6">0.09</oasis:entry>
         <oasis:entry colname="col7">0.40</oasis:entry>
         <oasis:entry colname="col8">0.14</oasis:entry>
         <oasis:entry colname="col9">0.04</oasis:entry>
         <oasis:entry colname="col10">0.48</oasis:entry>
         <oasis:entry colname="col11">0.36</oasis:entry>
         <oasis:entry colname="col12">0.20</oasis:entry>
         <oasis:entry colname="col13">0.05</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">GE<inline-formula><mml:math id="M70" display="inline"><mml:msup><mml:mi/><mml:mi mathvariant="normal">a</mml:mi></mml:msup></mml:math></inline-formula>   3<inline-formula><mml:math id="M71" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col2"><inline-formula><mml:math id="M72" display="inline"><mml:mo>-</mml:mo></mml:math></inline-formula>0.00</oasis:entry>
         <oasis:entry colname="col3">0.45</oasis:entry>
         <oasis:entry colname="col4">0.10</oasis:entry>
         <oasis:entry colname="col5">0.03</oasis:entry>
         <oasis:entry colname="col6">0.10</oasis:entry>
         <oasis:entry colname="col7">0.41</oasis:entry>
         <oasis:entry colname="col8">0.10</oasis:entry>
         <oasis:entry colname="col9">0.03</oasis:entry>
         <oasis:entry colname="col10">0.49</oasis:entry>
         <oasis:entry colname="col11">0.38</oasis:entry>
         <oasis:entry colname="col12">0.15</oasis:entry>
         <oasis:entry colname="col13">0.03</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">GR<inline-formula><mml:math id="M73" display="inline"><mml:msup><mml:mi/><mml:mi mathvariant="normal">b</mml:mi></mml:msup></mml:math></inline-formula>   3<inline-formula><mml:math id="M74" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col2"><inline-formula><mml:math id="M75" display="inline"><mml:mo>-</mml:mo></mml:math></inline-formula>0.00</oasis:entry>
         <oasis:entry colname="col3">0.45</oasis:entry>
         <oasis:entry colname="col4">0.11</oasis:entry>
         <oasis:entry colname="col5">0.03</oasis:entry>
         <oasis:entry colname="col6">0.10</oasis:entry>
         <oasis:entry colname="col7">0.40</oasis:entry>
         <oasis:entry colname="col8">0.10</oasis:entry>
         <oasis:entry colname="col9">0.02</oasis:entry>
         <oasis:entry colname="col10">0.48</oasis:entry>
         <oasis:entry colname="col11">0.37</oasis:entry>
         <oasis:entry colname="col12">0.15</oasis:entry>
         <oasis:entry colname="col13">0.03</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">GRE<inline-formula><mml:math id="M76" display="inline"><mml:msup><mml:mi/><mml:mi mathvariant="normal">c</mml:mi></mml:msup></mml:math></inline-formula>  3<inline-formula><mml:math id="M77" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col2"><inline-formula><mml:math id="M78" display="inline"><mml:mo>-</mml:mo></mml:math></inline-formula>0.00</oasis:entry>
         <oasis:entry colname="col3">0.45</oasis:entry>
         <oasis:entry colname="col4">0.09</oasis:entry>
         <oasis:entry colname="col5">0.02</oasis:entry>
         <oasis:entry colname="col6">0.10</oasis:entry>
         <oasis:entry colname="col7">0.41</oasis:entry>
         <oasis:entry colname="col8">0.09</oasis:entry>
         <oasis:entry colname="col9">0.02</oasis:entry>
         <oasis:entry colname="col10">0.49</oasis:entry>
         <oasis:entry colname="col11">0.38</oasis:entry>
         <oasis:entry colname="col12">0.13</oasis:entry>
         <oasis:entry colname="col13">0.03</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">G(PS)     10<inline-formula><mml:math id="M79" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col2"><inline-formula><mml:math id="M80" display="inline"><mml:mo>-</mml:mo></mml:math></inline-formula>0.05</oasis:entry>
         <oasis:entry colname="col3">0.48</oasis:entry>
         <oasis:entry colname="col4">0.22</oasis:entry>
         <oasis:entry colname="col5">0.03</oasis:entry>
         <oasis:entry colname="col6">0.04</oasis:entry>
         <oasis:entry colname="col7">0.48</oasis:entry>
         <oasis:entry colname="col8">0.22</oasis:entry>
         <oasis:entry colname="col9">0.03</oasis:entry>
         <oasis:entry colname="col10">0.56</oasis:entry>
         <oasis:entry colname="col11">0.38</oasis:entry>
         <oasis:entry colname="col12">0.31</oasis:entry>
         <oasis:entry colname="col13">0.05</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">GE<inline-formula><mml:math id="M81" display="inline"><mml:msup><mml:mi/><mml:mi mathvariant="normal">a</mml:mi></mml:msup></mml:math></inline-formula>  10<inline-formula><mml:math id="M82" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col2"><inline-formula><mml:math id="M83" display="inline"><mml:mo>-</mml:mo></mml:math></inline-formula>0.02</oasis:entry>
         <oasis:entry colname="col3">0.52</oasis:entry>
         <oasis:entry colname="col4">0.17</oasis:entry>
         <oasis:entry colname="col5">0.03</oasis:entry>
         <oasis:entry colname="col6">0.04</oasis:entry>
         <oasis:entry colname="col7">0.50</oasis:entry>
         <oasis:entry colname="col8">0.18</oasis:entry>
         <oasis:entry colname="col9">0.03</oasis:entry>
         <oasis:entry colname="col10">0.58</oasis:entry>
         <oasis:entry colname="col11">0.42</oasis:entry>
         <oasis:entry colname="col12">0.25</oasis:entry>
         <oasis:entry colname="col13">0.04</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">GR<inline-formula><mml:math id="M84" display="inline"><mml:msup><mml:mi/><mml:mi mathvariant="normal">b</mml:mi></mml:msup></mml:math></inline-formula>  10<inline-formula><mml:math id="M85" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col2"><inline-formula><mml:math id="M86" display="inline"><mml:mo>-</mml:mo></mml:math></inline-formula>0.02</oasis:entry>
         <oasis:entry colname="col3">0.51</oasis:entry>
         <oasis:entry colname="col4">0.17</oasis:entry>
         <oasis:entry colname="col5">0.02</oasis:entry>
         <oasis:entry colname="col6">0.05</oasis:entry>
         <oasis:entry colname="col7">0.49</oasis:entry>
         <oasis:entry colname="col8">0.17</oasis:entry>
         <oasis:entry colname="col9">0.02</oasis:entry>
         <oasis:entry colname="col10">0.58</oasis:entry>
         <oasis:entry colname="col11">0.41</oasis:entry>
         <oasis:entry colname="col12">0.25</oasis:entry>
         <oasis:entry colname="col13">0.03</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">GRE<inline-formula><mml:math id="M87" display="inline"><mml:msup><mml:mi/><mml:mi mathvariant="normal">c</mml:mi></mml:msup></mml:math></inline-formula> 10<inline-formula><mml:math id="M88" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col2"><inline-formula><mml:math id="M89" display="inline"><mml:mo>-</mml:mo></mml:math></inline-formula>0.02</oasis:entry>
         <oasis:entry colname="col3">0.53</oasis:entry>
         <oasis:entry colname="col4">0.15</oasis:entry>
         <oasis:entry colname="col5">0.02</oasis:entry>
         <oasis:entry colname="col6">0.03</oasis:entry>
         <oasis:entry colname="col7">0.51</oasis:entry>
         <oasis:entry colname="col8">0.15</oasis:entry>
         <oasis:entry colname="col9">0.02</oasis:entry>
         <oasis:entry colname="col10">0.59</oasis:entry>
         <oasis:entry colname="col11">0.44</oasis:entry>
         <oasis:entry colname="col12">0.22</oasis:entry>
         <oasis:entry colname="col13">0.03</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">G(PS)     15<inline-formula><mml:math id="M90" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col2"><inline-formula><mml:math id="M91" display="inline"><mml:mo>-</mml:mo></mml:math></inline-formula>0.06</oasis:entry>
         <oasis:entry colname="col3">0.48</oasis:entry>
         <oasis:entry colname="col4">0.33</oasis:entry>
         <oasis:entry colname="col5">0.04</oasis:entry>
         <oasis:entry colname="col6"><inline-formula><mml:math id="M92" display="inline"><mml:mo>-</mml:mo></mml:math></inline-formula>0.03</oasis:entry>
         <oasis:entry colname="col7">0.60</oasis:entry>
         <oasis:entry colname="col8">0.33</oasis:entry>
         <oasis:entry colname="col9">0.05</oasis:entry>
         <oasis:entry colname="col10">0.66</oasis:entry>
         <oasis:entry colname="col11">0.40</oasis:entry>
         <oasis:entry colname="col12">0.47</oasis:entry>
         <oasis:entry colname="col13">0.06</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">GE<inline-formula><mml:math id="M93" display="inline"><mml:msup><mml:mi/><mml:mi mathvariant="normal">a</mml:mi></mml:msup></mml:math></inline-formula>  15<inline-formula><mml:math id="M94" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col2"><inline-formula><mml:math id="M95" display="inline"><mml:mo>-</mml:mo></mml:math></inline-formula>0.02</oasis:entry>
         <oasis:entry colname="col3">0.51</oasis:entry>
         <oasis:entry colname="col4">0.26</oasis:entry>
         <oasis:entry colname="col5">0.04</oasis:entry>
         <oasis:entry colname="col6"><inline-formula><mml:math id="M96" display="inline"><mml:mo>-</mml:mo></mml:math></inline-formula>0.07</oasis:entry>
         <oasis:entry colname="col7">0.61</oasis:entry>
         <oasis:entry colname="col8">0.27</oasis:entry>
         <oasis:entry colname="col9">0.04</oasis:entry>
         <oasis:entry colname="col10">0.68</oasis:entry>
         <oasis:entry colname="col11">0.41</oasis:entry>
         <oasis:entry colname="col12">0.38</oasis:entry>
         <oasis:entry colname="col13">0.05</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">GR<inline-formula><mml:math id="M97" display="inline"><mml:msup><mml:mi/><mml:mi mathvariant="normal">b</mml:mi></mml:msup></mml:math></inline-formula>  15<inline-formula><mml:math id="M98" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col2">0.00</oasis:entry>
         <oasis:entry colname="col3">0.50</oasis:entry>
         <oasis:entry colname="col4">0.26</oasis:entry>
         <oasis:entry colname="col5">0.03</oasis:entry>
         <oasis:entry colname="col6"><inline-formula><mml:math id="M99" display="inline"><mml:mo>-</mml:mo></mml:math></inline-formula>0.05</oasis:entry>
         <oasis:entry colname="col7">0.60</oasis:entry>
         <oasis:entry colname="col8">0.26</oasis:entry>
         <oasis:entry colname="col9">0.03</oasis:entry>
         <oasis:entry colname="col10">0.67</oasis:entry>
         <oasis:entry colname="col11">0.40</oasis:entry>
         <oasis:entry colname="col12">0.37</oasis:entry>
         <oasis:entry colname="col13">0.04</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">GRE<inline-formula><mml:math id="M100" display="inline"><mml:msup><mml:mi/><mml:mi mathvariant="normal">c</mml:mi></mml:msup></mml:math></inline-formula> 15<inline-formula><mml:math id="M101" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col2">0.00</oasis:entry>
         <oasis:entry colname="col3">0.53</oasis:entry>
         <oasis:entry colname="col4">0.23</oasis:entry>
         <oasis:entry colname="col5">0.03</oasis:entry>
         <oasis:entry colname="col6"><inline-formula><mml:math id="M102" display="inline"><mml:mo>-</mml:mo></mml:math></inline-formula>0.09</oasis:entry>
         <oasis:entry colname="col7">0.61</oasis:entry>
         <oasis:entry colname="col8">0.23</oasis:entry>
         <oasis:entry colname="col9">0.03</oasis:entry>
         <oasis:entry colname="col10">0.69</oasis:entry>
         <oasis:entry colname="col11">0.43</oasis:entry>
         <oasis:entry colname="col12">0.33</oasis:entry>
         <oasis:entry colname="col13">0.04</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">WVR</oasis:entry>
         <oasis:entry colname="col2">0.12</oasis:entry>
         <oasis:entry colname="col3">1.00</oasis:entry>
         <oasis:entry colname="col4">0.09</oasis:entry>
         <oasis:entry colname="col5">0.08</oasis:entry>
         <oasis:entry colname="col6">0.19</oasis:entry>
         <oasis:entry colname="col7">0.80</oasis:entry>
         <oasis:entry colname="col8">0.08</oasis:entry>
         <oasis:entry colname="col9">0.07</oasis:entry>
         <oasis:entry colname="col10">0.99</oasis:entry>
         <oasis:entry colname="col11">0.84</oasis:entry>
         <oasis:entry colname="col12">0.12</oasis:entry>
         <oasis:entry colname="col13">0.11</oasis:entry>
       </oasis:row>
     </oasis:tbody>
   </oasis:tgroup></oasis:table><table-wrap-foot><p id="d1e977"><inline-formula><mml:math id="M59" display="inline"><mml:msup><mml:mi/><mml:mi mathvariant="normal">a</mml:mi></mml:msup></mml:math></inline-formula> GPS <inline-formula><mml:math id="M60" display="inline"><mml:mo>+</mml:mo></mml:math></inline-formula> Galileo, <inline-formula><mml:math id="M61" display="inline"><mml:msup><mml:mi/><mml:mi mathvariant="normal">b</mml:mi></mml:msup></mml:math></inline-formula> GPS <inline-formula><mml:math id="M62" display="inline"><mml:mo>+</mml:mo></mml:math></inline-formula> Glonass, <inline-formula><mml:math id="M63" display="inline"><mml:msup><mml:mi/><mml:mi mathvariant="normal">c</mml:mi></mml:msup></mml:math></inline-formula> GPS <inline-formula><mml:math id="M64" display="inline"><mml:mo>+</mml:mo></mml:math></inline-formula> Glonass <inline-formula><mml:math id="M65" display="inline"><mml:mo>+</mml:mo></mml:math></inline-formula> Galileo.</p></table-wrap-foot></table-wrap>

      <p id="d1e2073">As indicated by column 9 in Table <xref ref-type="table" rid="Ch1.T1"/>, the gradient amplitudes estimated by the WVR (0.99 <inline-formula><mml:math id="M103" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">mm</mml:mi></mml:mrow></mml:math></inline-formula>) are approximately twice as
large as the GNSS gradient amplitudes at 3<inline-formula><mml:math id="M104" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula> cutoff angle, i.e. 0.49 <inline-formula><mml:math id="M105" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">mm</mml:mi></mml:mrow></mml:math></inline-formula>, for the GRE solution, but they decrease to around 50 % as
large for the cutoff angle of 15<inline-formula><mml:math id="M106" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula>, i.e. 0.69 <inline-formula><mml:math id="M107" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">mm</mml:mi></mml:mrow></mml:math></inline-formula>, for the GRE solution. The GNSS gradient amplitudes are about twice as large as their
formal errors. The WVR gradient amplitudes are about 8 times larger than their mean formal errors, but the variability in the WVR formal errors is
significantly larger than those from the GNSS. This is due to a varying uncertainty in the measured sky brightness temperatures. These variations are
taken into account in the following comparisons by using the formal errors in the GNSS and the WVR gradients when calculating the weighted
root-mean-square (WRMS) differences and correlations. It should also be pointed out that the uncertainty (here represented by the formal error) in the
WVR gradients is scaled, meaning that if the true wet delays in the different directions have deviations from the linear gradient model the
uncertainties increase. Such deviations are common, e.g. during convection processes, and the assumption of linear changes in the wet refractivity in
a layered atmosphere will not be accurate. The formal error in the gradient given by GipsyX is, however, not scaled. Therefore, these uncertainties
are likely smaller than realistic values.</p>
</sec>
<sec id="Ch1.S3.SS2">
  <label>3.2</label><title>Comparison of gradients from the GNSS and WVR</title>
      <p id="d1e2129">We first carry out comparisons of the gradients estimated from the different GNSS constellations and using the three different elevation cutoff
angles, presented in Table <xref ref-type="table" rid="Ch1.T1"/>, with the WVR gradients. Even though the gradients estimated from both the GNSS and the WVR data
have a temporal resolution of 5 <inline-formula><mml:math id="M108" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">min</mml:mi></mml:mrow></mml:math></inline-formula>, the estimates are not centered at exactly the same time epochs. It is therefore necessary to synchronize
the time series to compare the gradients. We first present results with the highest available temporal resolution of 5 <inline-formula><mml:math id="M109" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">min</mml:mi></mml:mrow></mml:math></inline-formula> where the WVR
gradients were interpolated to the epochs in the GNSS time series using the temporal Gaussian filter, as described by <xref ref-type="bibr" rid="bib1.bibx17" id="text.25"/>, with a full
width at half maximum (FWHM) of <inline-formula><mml:math id="M110" display="inline"><mml:mo>±</mml:mo></mml:math></inline-formula> 2.5 <inline-formula><mml:math id="M111" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">min</mml:mi></mml:mrow></mml:math></inline-formula>.</p>

<?xmltex \floatpos{t}?><table-wrap id="Ch1.T2" specific-use="star"><?xmltex \currentcnt{2}?><label>Table 2</label><caption><p id="d1e2172">The WRMS differences and correlations of the east and north gradients, obtained from different satellite constellations, relative to the WVR data.</p></caption><oasis:table frame="topbot"><oasis:tgroup cols="9">
     <oasis:colspec colnum="1" colname="col1" align="left"/>
     <oasis:colspec colnum="2" colname="col2" align="right"/>
     <oasis:colspec colnum="3" colname="col3" align="right" colsep="1"/>
     <oasis:colspec colnum="4" colname="col4" align="right"/>
     <oasis:colspec colnum="5" colname="col5" align="right" colsep="1"/>
     <oasis:colspec colnum="6" colname="col6" align="right"/>
     <oasis:colspec colnum="7" colname="col7" align="right" colsep="1"/>
     <oasis:colspec colnum="8" colname="col8" align="right"/>
     <oasis:colspec colnum="9" colname="col9" align="right"/>
     <oasis:thead>
       <oasis:row>
         <oasis:entry colname="col1">GNSS solution</oasis:entry>
         <oasis:entry rowsep="1" namest="col2" nameend="col3" align="center" colsep="1">G(PS) </oasis:entry>
         <oasis:entry rowsep="1" namest="col4" nameend="col5" align="center" colsep="1">GE<inline-formula><mml:math id="M119" display="inline"><mml:msup><mml:mi/><mml:mi mathvariant="normal">a</mml:mi></mml:msup></mml:math></inline-formula></oasis:entry>
         <oasis:entry rowsep="1" namest="col6" nameend="col7" align="center" colsep="1">GR<inline-formula><mml:math id="M120" display="inline"><mml:msup><mml:mi/><mml:mi mathvariant="normal">b</mml:mi></mml:msup></mml:math></inline-formula></oasis:entry>
         <oasis:entry rowsep="1" namest="col8" nameend="col9" align="center">GRE<inline-formula><mml:math id="M121" display="inline"><mml:msup><mml:mi/><mml:mi mathvariant="normal">c</mml:mi></mml:msup></mml:math></inline-formula></oasis:entry>
       </oasis:row>
       <oasis:row rowsep="1">
         <oasis:entry colname="col1"><inline-formula><mml:math id="M122" display="inline"><mml:mo>+</mml:mo></mml:math></inline-formula> cutoff angle</oasis:entry>
         <oasis:entry colname="col2">East</oasis:entry>
         <oasis:entry colname="col3">North</oasis:entry>
         <oasis:entry colname="col4">East</oasis:entry>
         <oasis:entry colname="col5">North</oasis:entry>
         <oasis:entry colname="col6">East</oasis:entry>
         <oasis:entry colname="col7">North</oasis:entry>
         <oasis:entry colname="col8">East</oasis:entry>
         <oasis:entry colname="col9">North</oasis:entry>
       </oasis:row>
     </oasis:thead>
     <oasis:tbody>
       <oasis:row rowsep="1">
         <oasis:entry colname="col1">WRMS differences</oasis:entry>
         <oasis:entry colname="col2">(mm)</oasis:entry>
         <oasis:entry colname="col3">(mm)</oasis:entry>
         <oasis:entry colname="col4">(mm)</oasis:entry>
         <oasis:entry colname="col5">(mm)</oasis:entry>
         <oasis:entry colname="col6">(mm)</oasis:entry>
         <oasis:entry colname="col7">(mm)</oasis:entry>
         <oasis:entry colname="col8">(mm)</oasis:entry>
         <oasis:entry colname="col9">(mm)</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">ONS1 3<inline-formula><mml:math id="M123" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col2">0.70</oasis:entry>
         <oasis:entry colname="col3">0.56</oasis:entry>
         <oasis:entry colname="col4">0.67</oasis:entry>
         <oasis:entry colname="col5">0.52</oasis:entry>
         <oasis:entry colname="col6">0.67</oasis:entry>
         <oasis:entry colname="col7">0.52</oasis:entry>
         <oasis:entry colname="col8">0.66</oasis:entry>
         <oasis:entry colname="col9">0.50</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">ONS1 10<inline-formula><mml:math id="M124" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col2">0.74</oasis:entry>
         <oasis:entry colname="col3">0.63</oasis:entry>
         <oasis:entry colname="col4">0.69</oasis:entry>
         <oasis:entry colname="col5">0.59</oasis:entry>
         <oasis:entry colname="col6">0.70</oasis:entry>
         <oasis:entry colname="col7">0.58</oasis:entry>
         <oasis:entry colname="col8">0.67</oasis:entry>
         <oasis:entry colname="col9">0.56</oasis:entry>
       </oasis:row>
       <oasis:row rowsep="1">
         <oasis:entry colname="col1">ONS1 15<inline-formula><mml:math id="M125" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col2">0.80</oasis:entry>
         <oasis:entry colname="col3">0.79</oasis:entry>
         <oasis:entry colname="col4">0.75</oasis:entry>
         <oasis:entry colname="col5">0.78</oasis:entry>
         <oasis:entry colname="col6">0.75</oasis:entry>
         <oasis:entry colname="col7">0.76</oasis:entry>
         <oasis:entry colname="col8">0.72</oasis:entry>
         <oasis:entry colname="col9">0.76</oasis:entry>
       </oasis:row>
       <oasis:row rowsep="1">
         <oasis:entry namest="col1" nameend="col9">Correlation coefficients </oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">ONS1 3<inline-formula><mml:math id="M126" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col2">0.62</oasis:entry>
         <oasis:entry colname="col3">0.61</oasis:entry>
         <oasis:entry colname="col4">0.63</oasis:entry>
         <oasis:entry colname="col5">0.64</oasis:entry>
         <oasis:entry colname="col6">0.64</oasis:entry>
         <oasis:entry colname="col7">0.65</oasis:entry>
         <oasis:entry colname="col8">0.63</oasis:entry>
         <oasis:entry colname="col9">0.66</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">ONS1 10<inline-formula><mml:math id="M127" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col2">0.61</oasis:entry>
         <oasis:entry colname="col3">0.55</oasis:entry>
         <oasis:entry colname="col4">0.66</oasis:entry>
         <oasis:entry colname="col5">0.61</oasis:entry>
         <oasis:entry colname="col6">0.65</oasis:entry>
         <oasis:entry colname="col7">0.63</oasis:entry>
         <oasis:entry colname="col8">0.68</oasis:entry>
         <oasis:entry colname="col9">0.66</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">ONS1 15<inline-formula><mml:math id="M128" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col2">0.55</oasis:entry>
         <oasis:entry colname="col3">0.37</oasis:entry>
         <oasis:entry colname="col4">0.61</oasis:entry>
         <oasis:entry colname="col5">0.41</oasis:entry>
         <oasis:entry colname="col6">0.60</oasis:entry>
         <oasis:entry colname="col7">0.43</oasis:entry>
         <oasis:entry colname="col8">0.63</oasis:entry>
         <oasis:entry colname="col9">0.46</oasis:entry>
       </oasis:row>
     </oasis:tbody>
   </oasis:tgroup></oasis:table><table-wrap-foot><p id="d1e2175"><inline-formula><mml:math id="M112" display="inline"><mml:msup><mml:mi/><mml:mi mathvariant="normal">a</mml:mi></mml:msup></mml:math></inline-formula> GPS <inline-formula><mml:math id="M113" display="inline"><mml:mo>+</mml:mo></mml:math></inline-formula> Galileo, <inline-formula><mml:math id="M114" display="inline"><mml:msup><mml:mi/><mml:mi mathvariant="normal">b</mml:mi></mml:msup></mml:math></inline-formula> GPS <inline-formula><mml:math id="M115" display="inline"><mml:mo>+</mml:mo></mml:math></inline-formula> Glonass, <inline-formula><mml:math id="M116" display="inline"><mml:msup><mml:mi/><mml:mi mathvariant="normal">c</mml:mi></mml:msup></mml:math></inline-formula>GPS <inline-formula><mml:math id="M117" display="inline"><mml:mo>+</mml:mo></mml:math></inline-formula> Glonass <inline-formula><mml:math id="M118" display="inline"><mml:mo>+</mml:mo></mml:math></inline-formula> Galileo.</p></table-wrap-foot></table-wrap>

      <p id="d1e2615">Table <xref ref-type="table" rid="Ch1.T2"/> shows the WRMS differences and correlations of the east and the north gradients for the whole year of
2019. As expected, using the data from multi-GNSS, we note a significant improvement (an increase in the correlation of up to 20 % and a maximum
reduction in the WRMS difference of 11 %). Our interpretation is that the geometry of the observations is improved when observations from
additional systems are added, especially in the south–north direction (see Fig. <xref ref-type="fig" rid="Ch1.F1"/>).<?pagebreak page5598?> We also note that the GRE solution, in
general, gives the best agreement with the WVR gradients.</p>

      <?xmltex \floatpos{t}?><fig id="Ch1.F6" specific-use="star"><?xmltex \currentcnt{6}?><?xmltex \def\figurename{Figure}?><label>Figure 6</label><caption><p id="d1e2625">Correlations between estimated gradients from the GNSS and the WVR data calculated for each month.</p></caption>
          <?xmltex \igopts{width=384.112205pt}?><graphic xlink:href="https://amt.copernicus.org/articles/14/5593/2021/amt-14-5593-2021-f06.png"/>

        </fig>

      <p id="d1e2634">For the GPS-only solution, the highest correlation is obtained for the elevation cutoff angle of 3<inline-formula><mml:math id="M129" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula>, especially for the north gradients. This
is, however, not the case for the multi-GNSS solutions (GR, GE, and GRE), where the best correlation for the east gradient is obtained for the elevation
cutoff angle of 10<inline-formula><mml:math id="M130" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula>. For the GRE solutions, the correlations given by the 3 and 10<inline-formula><mml:math id="M131" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula> solutions are similar for the north
gradients. These results indicate that the choice of elevation cutoff angle is a compromise between having a good geometry and avoiding
elevation-angle-dependent systematic errors, e.g. multipath effects. Related to this is the error introduced by the assumption that the turbulent
atmosphere may be modelled with just a linear gradient when the elevation cutoff angle of 25<inline-formula><mml:math id="M132" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula> has to be used for the WVR observations in order
to avoid ground noise pickup. As is presented in Table <xref ref-type="table" rid="Ch1.T1"/>, the estimated size of the gradients was in general increasing when
using a decreasing sky coverage of the observations.</p>

<?xmltex \floatpos{t}?><table-wrap id="Ch1.T3" specific-use="star"><?xmltex \currentcnt{3}?><label>Table 3</label><caption><p id="d1e2678">The WRMS differences and correlations of the east and the north gradients, obtained from different satellite constellations, relative to the WVR data for June and July 2019.</p></caption><oasis:table frame="topbot"><oasis:tgroup cols="9">
     <oasis:colspec colnum="1" colname="col1" align="left"/>
     <oasis:colspec colnum="2" colname="col2" align="right"/>
     <oasis:colspec colnum="3" colname="col3" align="right" colsep="1"/>
     <oasis:colspec colnum="4" colname="col4" align="right"/>
     <oasis:colspec colnum="5" colname="col5" align="right" colsep="1"/>
     <oasis:colspec colnum="6" colname="col6" align="right"/>
     <oasis:colspec colnum="7" colname="col7" align="right" colsep="1"/>
     <oasis:colspec colnum="8" colname="col8" align="right"/>
     <oasis:colspec colnum="9" colname="col9" align="right"/>
     <oasis:thead>
       <oasis:row>
         <oasis:entry colname="col1">GNSS station</oasis:entry>
         <oasis:entry rowsep="1" namest="col2" nameend="col3" align="center" colsep="1">G(PS) </oasis:entry>
         <oasis:entry rowsep="1" namest="col4" nameend="col5" align="center" colsep="1">GE<inline-formula><mml:math id="M140" display="inline"><mml:msup><mml:mi/><mml:mi mathvariant="normal">a</mml:mi></mml:msup></mml:math></inline-formula></oasis:entry>
         <oasis:entry rowsep="1" namest="col6" nameend="col7" align="center" colsep="1">GR<inline-formula><mml:math id="M141" display="inline"><mml:msup><mml:mi/><mml:mi mathvariant="normal">b</mml:mi></mml:msup></mml:math></inline-formula></oasis:entry>
         <oasis:entry rowsep="1" namest="col8" nameend="col9" align="center">GRE<inline-formula><mml:math id="M142" display="inline"><mml:msup><mml:mi/><mml:mi mathvariant="normal">c</mml:mi></mml:msup></mml:math></inline-formula></oasis:entry>
       </oasis:row>
       <oasis:row rowsep="1">
         <oasis:entry colname="col1"><inline-formula><mml:math id="M143" display="inline"><mml:mo>+</mml:mo></mml:math></inline-formula> cutoff angle</oasis:entry>
         <oasis:entry colname="col2">East</oasis:entry>
         <oasis:entry colname="col3">North</oasis:entry>
         <oasis:entry colname="col4">East</oasis:entry>
         <oasis:entry colname="col5">North</oasis:entry>
         <oasis:entry colname="col6">East</oasis:entry>
         <oasis:entry colname="col7">North</oasis:entry>
         <oasis:entry colname="col8">East</oasis:entry>
         <oasis:entry colname="col9">North</oasis:entry>
       </oasis:row>
     </oasis:thead>
     <oasis:tbody>
       <oasis:row rowsep="1">
         <oasis:entry colname="col1">WRMS differences</oasis:entry>
         <oasis:entry colname="col2">(mm)</oasis:entry>
         <oasis:entry colname="col3">(mm)</oasis:entry>
         <oasis:entry colname="col4">(mm)</oasis:entry>
         <oasis:entry colname="col5">(mm)</oasis:entry>
         <oasis:entry colname="col6">(mm)</oasis:entry>
         <oasis:entry colname="col7">(mm)</oasis:entry>
         <oasis:entry colname="col8">(mm)</oasis:entry>
         <oasis:entry colname="col9">(mm)</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">ONS1 3<inline-formula><mml:math id="M144" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col2">0.82</oasis:entry>
         <oasis:entry colname="col3">0.67</oasis:entry>
         <oasis:entry colname="col4">0.79</oasis:entry>
         <oasis:entry colname="col5">0.62</oasis:entry>
         <oasis:entry colname="col6">0.78</oasis:entry>
         <oasis:entry colname="col7">0.62</oasis:entry>
         <oasis:entry colname="col8">0.78</oasis:entry>
         <oasis:entry colname="col9">0.61</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">ONSA 3<inline-formula><mml:math id="M145" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col2">0.82</oasis:entry>
         <oasis:entry colname="col3">0.65</oasis:entry>
         <oasis:entry colname="col4">0.79</oasis:entry>
         <oasis:entry colname="col5">0.62</oasis:entry>
         <oasis:entry colname="col6">0.78</oasis:entry>
         <oasis:entry colname="col7">0.61</oasis:entry>
         <oasis:entry colname="col8">0.77</oasis:entry>
         <oasis:entry colname="col9">0.60</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">ONS1 10<inline-formula><mml:math id="M146" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col2">0.83</oasis:entry>
         <oasis:entry colname="col3">0.71</oasis:entry>
         <oasis:entry colname="col4">0.77</oasis:entry>
         <oasis:entry colname="col5">0.66</oasis:entry>
         <oasis:entry colname="col6">0.77</oasis:entry>
         <oasis:entry colname="col7">0.65</oasis:entry>
         <oasis:entry colname="col8">0.73</oasis:entry>
         <oasis:entry colname="col9">0.63</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">ONSA 10<inline-formula><mml:math id="M147" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col2">0.82</oasis:entry>
         <oasis:entry colname="col3">0.70</oasis:entry>
         <oasis:entry colname="col4">0.75</oasis:entry>
         <oasis:entry colname="col5">0.65</oasis:entry>
         <oasis:entry colname="col6">0.75</oasis:entry>
         <oasis:entry colname="col7">0.64</oasis:entry>
         <oasis:entry colname="col8">0.72</oasis:entry>
         <oasis:entry colname="col9">0.63</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">ONS1 15<inline-formula><mml:math id="M148" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col2">0.91</oasis:entry>
         <oasis:entry colname="col3">0.88</oasis:entry>
         <oasis:entry colname="col4">0.83</oasis:entry>
         <oasis:entry colname="col5">0.85</oasis:entry>
         <oasis:entry colname="col6">0.83</oasis:entry>
         <oasis:entry colname="col7">0.83</oasis:entry>
         <oasis:entry colname="col8">0.79</oasis:entry>
         <oasis:entry colname="col9">0.83</oasis:entry>
       </oasis:row>
       <oasis:row rowsep="1">
         <oasis:entry colname="col1">ONSA 15<inline-formula><mml:math id="M149" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col2">0.92</oasis:entry>
         <oasis:entry colname="col3">0.85</oasis:entry>
         <oasis:entry colname="col4">0.84</oasis:entry>
         <oasis:entry colname="col5">0.80</oasis:entry>
         <oasis:entry colname="col6">0.83</oasis:entry>
         <oasis:entry colname="col7">0.80</oasis:entry>
         <oasis:entry colname="col8">0.79</oasis:entry>
         <oasis:entry colname="col9">0.78</oasis:entry>
       </oasis:row>
       <oasis:row rowsep="1">
         <oasis:entry namest="col1" nameend="col9">Correlation coefficients </oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">ONS1 3<inline-formula><mml:math id="M150" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col2">0.71</oasis:entry>
         <oasis:entry colname="col3">0.62</oasis:entry>
         <oasis:entry colname="col4">0.73</oasis:entry>
         <oasis:entry colname="col5">0.65</oasis:entry>
         <oasis:entry colname="col6">0.75</oasis:entry>
         <oasis:entry colname="col7">0.66</oasis:entry>
         <oasis:entry colname="col8">0.74</oasis:entry>
         <oasis:entry colname="col9">0.67</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">ONSA 3<inline-formula><mml:math id="M151" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col2">0.70</oasis:entry>
         <oasis:entry colname="col3">0.66</oasis:entry>
         <oasis:entry colname="col4">0.73</oasis:entry>
         <oasis:entry colname="col5">0.68</oasis:entry>
         <oasis:entry colname="col6">0.74</oasis:entry>
         <oasis:entry colname="col7">0.70</oasis:entry>
         <oasis:entry colname="col8">0.74</oasis:entry>
         <oasis:entry colname="col9">0.69</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">ONS1 10<inline-formula><mml:math id="M152" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col2">0.73</oasis:entry>
         <oasis:entry colname="col3">0.59</oasis:entry>
         <oasis:entry colname="col4">0.77</oasis:entry>
         <oasis:entry colname="col5">0.66</oasis:entry>
         <oasis:entry colname="col6">0.78</oasis:entry>
         <oasis:entry colname="col7">0.67</oasis:entry>
         <oasis:entry colname="col8">0.80</oasis:entry>
         <oasis:entry colname="col9">0.70</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">ONSA 10<inline-formula><mml:math id="M153" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col2">0.73</oasis:entry>
         <oasis:entry colname="col3">0.61</oasis:entry>
         <oasis:entry colname="col4">0.78</oasis:entry>
         <oasis:entry colname="col5">0.66</oasis:entry>
         <oasis:entry colname="col6">0.79</oasis:entry>
         <oasis:entry colname="col7">0.68</oasis:entry>
         <oasis:entry colname="col8">0.81</oasis:entry>
         <oasis:entry colname="col9">0.70</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">ONS1 15<inline-formula><mml:math id="M154" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col2">0.68</oasis:entry>
         <oasis:entry colname="col3">0.41</oasis:entry>
         <oasis:entry colname="col4">0.73</oasis:entry>
         <oasis:entry colname="col5">0.46</oasis:entry>
         <oasis:entry colname="col6">0.75</oasis:entry>
         <oasis:entry colname="col7">0.49</oasis:entry>
         <oasis:entry colname="col8">0.77</oasis:entry>
         <oasis:entry colname="col9">0.52</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">ONSA 15<inline-formula><mml:math id="M155" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col2">0.66</oasis:entry>
         <oasis:entry colname="col3">0.43</oasis:entry>
         <oasis:entry colname="col4">0.74</oasis:entry>
         <oasis:entry colname="col5">0.49</oasis:entry>
         <oasis:entry colname="col6">0.75</oasis:entry>
         <oasis:entry colname="col7">0.50</oasis:entry>
         <oasis:entry colname="col8">0.77</oasis:entry>
         <oasis:entry colname="col9">0.53</oasis:entry>
       </oasis:row>
     </oasis:tbody>
   </oasis:tgroup></oasis:table><table-wrap-foot><p id="d1e2681"><inline-formula><mml:math id="M133" display="inline"><mml:msup><mml:mi/><mml:mi mathvariant="normal">a</mml:mi></mml:msup></mml:math></inline-formula> GPS <inline-formula><mml:math id="M134" display="inline"><mml:mo>+</mml:mo></mml:math></inline-formula> Galileo, <inline-formula><mml:math id="M135" display="inline"><mml:msup><mml:mi/><mml:mi mathvariant="normal">b</mml:mi></mml:msup></mml:math></inline-formula> GPS <inline-formula><mml:math id="M136" display="inline"><mml:mo>+</mml:mo></mml:math></inline-formula> Glonass, <inline-formula><mml:math id="M137" display="inline"><mml:msup><mml:mi/><mml:mi mathvariant="normal">c</mml:mi></mml:msup></mml:math></inline-formula> GPS <inline-formula><mml:math id="M138" display="inline"><mml:mo>+</mml:mo></mml:math></inline-formula> Glonass <inline-formula><mml:math id="M139" display="inline"><mml:mo>+</mml:mo></mml:math></inline-formula> Galileo.</p></table-wrap-foot></table-wrap>

      <?pagebreak page5599?><p id="d1e3356">Even though the WVR and the GNSS sample different parts of the sky, it is noted that the agreement becomes worse for all GNSS solutions when the
15<inline-formula><mml:math id="M156" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula> elevation cutoff angle is used. Our interpretation is that it is because many important observations are removed, especially in the north
direction. We also note that even though there are slightly more observations from the Glonass system contributing to the south–north direction,
especially below the elevation angle of 20<inline-formula><mml:math id="M157" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula> (see Fig. <xref ref-type="fig" rid="Ch1.F1"/>), the GR solution does not give a significantly better
agreement with the WVR than the one for the GE solution.</p>

      <?xmltex \floatpos{t}?><fig id="Ch1.F7" specific-use="star"><?xmltex \currentcnt{7}?><?xmltex \def\figurename{Figure}?><label>Figure 7</label><caption><p id="d1e3381">The mean code multipath RMS calculated from ONS1 <bold>(a)</bold> and ONSA <bold>(b)</bold>.</p></caption>
          <?xmltex \igopts{width=412.564961pt}?><graphic xlink:href="https://amt.copernicus.org/articles/14/5593/2021/amt-14-5593-2021-f07.png"/>

        </fig>

      <?xmltex \floatpos{t}?><fig id="Ch1.F8" specific-use="star"><?xmltex \currentcnt{8}?><?xmltex \def\figurename{Figure}?><label>Figure 8</label><caption><p id="d1e3399">Correlations and WRMS differences between estimated gradients from the GNSS and the WVR data using different effective temporal resolutions, <inline-formula><mml:math id="M158" display="inline"><mml:mrow><mml:mi mathvariant="normal">Δ</mml:mi><mml:msub><mml:mi>t</mml:mi><mml:mtext>eff</mml:mtext></mml:msub></mml:mrow></mml:math></inline-formula>.</p></caption>
          <?xmltex \igopts{width=426.791339pt}?><graphic xlink:href="https://amt.copernicus.org/articles/14/5593/2021/amt-14-5593-2021-f08.png"/>

        </fig>

      <p id="d1e3421">In order to study any seasonal variability we compare the estimated gradients from the GNSS and the WVR for each month. The GNSS gradients are
obtained using different constellations and a cutoff angle of 3<inline-formula><mml:math id="M159" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula> (see Fig. <xref ref-type="fig" rid="Ch1.F6"/>). The change in the correlation is large
from month to month, and these changes seem to be related to the amplitude of the ZWD (see Fig. <xref ref-type="fig" rid="Ch1.F5"/>). In general, a large ZWD
variability results in a larger dynamic range for the gradients and consequently also a larger correlation. Figure <xref ref-type="fig" rid="Ch1.F6"/> shows higher
gradient correlations for June and July, which is consistent with the results given by <xref ref-type="bibr" rid="bib1.bibx6" id="text.26"/>. To investigate how well the GNSS data
capture large gradients, we carried out the same comparison using data from June and July only. Another reason for focusing on these 2 months is to
include the ONSA GNSS station for validation purposes. Multi-GNSS data from ONSA started to be acquired only in April 2019.</p>
      <?pagebreak page5600?><p id="d1e3442">The results are summarized in Table <xref ref-type="table" rid="Ch1.T3"/>. Because of the larger gradients the WRMS differences increase in all
cases. However, at the same time we see higher correlations between GNSS and WVR gradients compared to the whole data set (see
Table <xref ref-type="table" rid="Ch1.T2"/>). The two GNSS stations (ONS1 and ONSA) show very similar agreements with the WVR gradients. This is partly expected
since they are located close to each other, and therefore the gradients from the two stations are estimated based on the same observational directions
and are affected by common error sources, such as orbit errors. However, it is of interest to note that the different antenna mountings (see
Fig. <xref ref-type="fig" rid="Ch1.F3"/>) do not have a significantly different impact on the estimated gradients. This can also be seen in Fig. <xref ref-type="fig" rid="Ch1.F7"/>, which shows the mean code multipath RMS values for both ONS1 and ONSA calculated using Anubis software <xref ref-type="bibr" rid="bib1.bibx22" id="paren.27"/> and an elevation cutoff
angle of 5<inline-formula><mml:math id="M160" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula>. Therefore, we continue to use only ONS1 GNSS data in the following.</p>
</sec>
<sec id="Ch1.S3.SS3">
  <label>3.3</label><?xmltex \opttitle{Effective temporal resolutions from 5\,{$\unit{{min}}$} to 24\,{$\unit{{h}}$}}?><title>Effective temporal resolutions from 5 <inline-formula><mml:math id="M161" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">min</mml:mi></mml:mrow></mml:math></inline-formula> to 24 <inline-formula><mml:math id="M162" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">h</mml:mi></mml:mrow></mml:math></inline-formula></title>
      <?pagebreak page5601?><p id="d1e3490">As mentioned in the introduction, GNSS gradients have been compared to other independent estimates over different timescales and temporal
resolutions. We therefore averaged the GNSS and the WVR gradients by applying a Gaussian window with different FWHM from <inline-formula><mml:math id="M163" display="inline"><mml:mo>±</mml:mo></mml:math></inline-formula> 2.5 to
<inline-formula><mml:math id="M164" display="inline"><mml:mo>±</mml:mo></mml:math></inline-formula> 720 <inline-formula><mml:math id="M165" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">min</mml:mi></mml:mrow></mml:math></inline-formula> for further comparisons. As described in Sect. <xref ref-type="sec" rid="Ch1.S3.SS2"/>, when an FWHM of <inline-formula><mml:math id="M166" display="inline"><mml:mo>±</mml:mo></mml:math></inline-formula> 2.5 <inline-formula><mml:math id="M167" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">min</mml:mi></mml:mrow></mml:math></inline-formula> is used, only the
WVR gradients were interpolated to the epochs in the GNSS time series. For all FWHMs larger than <inline-formula><mml:math id="M168" display="inline"><mml:mo>±</mml:mo></mml:math></inline-formula> 2.5 <inline-formula><mml:math id="M169" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">min</mml:mi></mml:mrow></mml:math></inline-formula>, both the GNSS and the WVR
gradients are interpolated to the epochs at 0, 5, 10, …, and 55 <inline-formula><mml:math id="M170" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">min</mml:mi></mml:mrow></mml:math></inline-formula> after the hour. The requirement to calculate a
value at a specific epoch is that at least half of the original data points (with a 5 <inline-formula><mml:math id="M171" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">min</mml:mi></mml:mrow></mml:math></inline-formula> resolution) exist within the FWHM. In the following
we refer to and use the FWHM as an effective temporal resolution, <inline-formula><mml:math id="M172" display="inline"><mml:mrow><mml:mi mathvariant="normal">Δ</mml:mi><mml:msub><mml:mi>t</mml:mi><mml:mtext>eff</mml:mtext></mml:msub></mml:mrow></mml:math></inline-formula>, from 5 to 1440 <inline-formula><mml:math id="M173" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">min</mml:mi></mml:mrow></mml:math></inline-formula> (1 d), although the
time series will still have a value every 5 <inline-formula><mml:math id="M174" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">min</mml:mi></mml:mrow></mml:math></inline-formula>.</p>

      <?xmltex \floatpos{t}?><fig id="Ch1.F9"><?xmltex \currentcnt{9}?><?xmltex \def\figurename{Figure}?><label>Figure 9</label><caption><p id="d1e3596">Correlations between estimated east gradients from the GNSS, given by the GRE solution with a 3<inline-formula><mml:math id="M175" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula> elevation cutoff angle, and the WVR data. Four different values of <inline-formula><mml:math id="M176" display="inline"><mml:mrow><mml:mi mathvariant="normal">Δ</mml:mi><mml:msub><mml:mi>t</mml:mi><mml:mtext>eff</mml:mtext></mml:msub></mml:mrow></mml:math></inline-formula> are used: <bold>(a)</bold> 5 <inline-formula><mml:math id="M177" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">min</mml:mi></mml:mrow></mml:math></inline-formula>, <bold>(b)</bold> 2 <inline-formula><mml:math id="M178" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">h</mml:mi></mml:mrow></mml:math></inline-formula>, <bold>(c)</bold> 12 <inline-formula><mml:math id="M179" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">h</mml:mi></mml:mrow></mml:math></inline-formula>, and <bold>(d)</bold> 24 <inline-formula><mml:math id="M180" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">h</mml:mi></mml:mrow></mml:math></inline-formula>.</p></caption>
          <?xmltex \igopts{width=241.848425pt}?><graphic xlink:href="https://amt.copernicus.org/articles/14/5593/2021/amt-14-5593-2021-f09.png"/>

        </fig>

      <?xmltex \floatpos{t}?><fig id="Ch1.F10"><?xmltex \currentcnt{10}?><?xmltex \def\figurename{Figure}?><label>Figure 10</label><caption><p id="d1e3675">Correlations between estimated north gradients from the GNSS, given by the GRE solution with a 3<inline-formula><mml:math id="M181" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula> elevation cutoff angle, and the WVR data. Four different values of <inline-formula><mml:math id="M182" display="inline"><mml:mrow><mml:mi mathvariant="normal">Δ</mml:mi><mml:msub><mml:mi>t</mml:mi><mml:mtext>eff</mml:mtext></mml:msub></mml:mrow></mml:math></inline-formula> are used: <bold>(a)</bold> 5 <inline-formula><mml:math id="M183" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">min</mml:mi></mml:mrow></mml:math></inline-formula>, <bold>(b)</bold> 2 <inline-formula><mml:math id="M184" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">h</mml:mi></mml:mrow></mml:math></inline-formula>, <bold>(c)</bold> 12 <inline-formula><mml:math id="M185" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">h</mml:mi></mml:mrow></mml:math></inline-formula>, and <bold>(d)</bold> 24 <inline-formula><mml:math id="M186" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">h</mml:mi></mml:mrow></mml:math></inline-formula>.</p></caption>
          <?xmltex \igopts{width=241.848425pt}?><graphic xlink:href="https://amt.copernicus.org/articles/14/5593/2021/amt-14-5593-2021-f10.png"/>

        </fig>

      <p id="d1e3751"><?xmltex \hack{\newpage}?>The resulting WRMS differences and correlations are shown in Fig. <xref ref-type="fig" rid="Ch1.F8"/>. It is clear that the WRMS differences decrease when the
<inline-formula><mml:math id="M187" display="inline"><mml:mrow><mml:mi mathvariant="normal">Δ</mml:mi><mml:msub><mml:mi>t</mml:mi><mml:mtext>eff</mml:mtext></mml:msub></mml:mrow></mml:math></inline-formula> increases, and more variations in the gradients are averaged out. Over all GNSS solutions, the highest correlation for the east
gradients is obtained when a <inline-formula><mml:math id="M188" display="inline"><mml:mrow><mml:mi mathvariant="normal">Δ</mml:mi><mml:msub><mml:mi>t</mml:mi><mml:mtext>eff</mml:mtext></mml:msub></mml:mrow></mml:math></inline-formula> of 2 <inline-formula><mml:math id="M189" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">h</mml:mi></mml:mrow></mml:math></inline-formula> is used, while for the north gradients, the best agreement is seen for a <inline-formula><mml:math id="M190" display="inline"><mml:mrow><mml:mi mathvariant="normal">Δ</mml:mi><mml:msub><mml:mi>t</mml:mi><mml:mtext>eff</mml:mtext></mml:msub></mml:mrow></mml:math></inline-formula> of 6 <inline-formula><mml:math id="M191" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">h</mml:mi></mml:mrow></mml:math></inline-formula>. Figures <xref ref-type="fig" rid="Ch1.F9"/> and <xref ref-type="fig" rid="Ch1.F10"/> depict the gradients estimated from the GNSS data and the
GRE solution for a 3<inline-formula><mml:math id="M192" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula> elevation cutoff angle against the gradients obtained from the WVR data for four different values for <inline-formula><mml:math id="M193" display="inline"><mml:mrow><mml:mi mathvariant="normal">Δ</mml:mi><mml:msub><mml:mi>t</mml:mi><mml:mtext>eff</mml:mtext></mml:msub></mml:mrow></mml:math></inline-formula> (5 <inline-formula><mml:math id="M194" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">min</mml:mi></mml:mrow></mml:math></inline-formula>, 2 <inline-formula><mml:math id="M195" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">h</mml:mi></mml:mrow></mml:math></inline-formula>, 12 <inline-formula><mml:math id="M196" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">h</mml:mi></mml:mrow></mml:math></inline-formula>, and 24 <inline-formula><mml:math id="M197" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">h</mml:mi></mml:mrow></mml:math></inline-formula>). It illustrates that even when the gradients are averaged over 1 d, i.e. applying
<inline-formula><mml:math id="M198" display="inline"><mml:mrow><mml:mi mathvariant="normal">Δ</mml:mi><mml:msub><mml:mi>t</mml:mi><mml:mtext>eff</mml:mtext></mml:msub></mml:mrow></mml:math></inline-formula> of 24 <inline-formula><mml:math id="M199" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">h</mml:mi></mml:mrow></mml:math></inline-formula>, there are substantial variations left which still give a clear correlation between the GNSS and the WVR data.</p>

      <?xmltex \floatpos{p}?><fig id="Ch1.F11" specific-use="star"><?xmltex \currentcnt{11}?><?xmltex \def\figurename{Figure}?><label>Figure 11</label><caption><p id="d1e3895">ZWD (top) and gradient time series from the WVR and the GNSS solution using an elevation cutoff angle of 3<inline-formula><mml:math id="M200" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula>. The east and north gradients are shown in the second, third, and fourth row of graphs for <inline-formula><mml:math id="M201" display="inline"><mml:mrow><mml:mi mathvariant="normal">Δ</mml:mi><mml:msub><mml:mi>t</mml:mi><mml:mtext>eff</mml:mtext></mml:msub></mml:mrow></mml:math></inline-formula> of 5 <inline-formula><mml:math id="M202" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">min</mml:mi></mml:mrow></mml:math></inline-formula>, 2 <inline-formula><mml:math id="M203" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">h</mml:mi></mml:mrow></mml:math></inline-formula>, and 24 <inline-formula><mml:math id="M204" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">h</mml:mi></mml:mrow></mml:math></inline-formula>, respectively. Note the different scales for the east and the north gradient graphs.</p></caption>
          <?xmltex \igopts{width=384.112205pt}?><graphic xlink:href="https://amt.copernicus.org/articles/14/5593/2021/amt-14-5593-2021-f11.png"/>

        </fig>

      <p id="d1e3950">We also study a specific event of short-lived gradients, associated with rapid changes in the ZWD, starting from 00:00 UTC on 23 July (see
Fig. <xref ref-type="fig" rid="Ch1.F11"/>). There is a passage of a warm front, indicated by a sudden increase in the ZWD, during the late evening of 24 July. As a
result, we see a large gradient towards the west direction. They are detected by both the GNSS and WVR, but the amplitudes of the gradients from the GNSS are
much smaller. For the north direction, there are also some large gradients, i.e. around 08:00 and 17:00 UTC on 23 July, which are detected by both
the GNSS and WVR. Also in this case the amplitudes from the GNSS are smaller. In addition, the multi-GNSS solutions have a slightly higher possibility of
capturing sudden short-lived gradients.</p>

<?xmltex \floatpos{t}?><table-wrap id="Ch1.T4" specific-use="star"><?xmltex \currentcnt{4}?><label>Table 4</label><caption><p id="d1e3958">The changes in WRMS differences and correlations of east and north gradients when a weaker constraint is applied in the GNSS data processing. The changes are the results from the constraint of 1.0 <inline-formula><mml:math id="M205" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">mm</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:msqrt><mml:mi mathvariant="normal">h</mml:mi></mml:msqrt><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> relative to 0.3 <inline-formula><mml:math id="M206" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">mm</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:msqrt><mml:mi mathvariant="normal">h</mml:mi></mml:msqrt><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> (in Table <xref ref-type="table" rid="Ch1.T3"/>).</p></caption><oasis:table frame="topbot"><oasis:tgroup cols="9">
     <oasis:colspec colnum="1" colname="col1" align="left"/>
     <oasis:colspec colnum="2" colname="col2" align="right"/>
     <oasis:colspec colnum="3" colname="col3" align="right" colsep="1"/>
     <oasis:colspec colnum="4" colname="col4" align="right"/>
     <oasis:colspec colnum="5" colname="col5" align="right" colsep="1"/>
     <oasis:colspec colnum="6" colname="col6" align="right"/>
     <oasis:colspec colnum="7" colname="col7" align="right" colsep="1"/>
     <oasis:colspec colnum="8" colname="col8" align="right"/>
     <oasis:colspec colnum="9" colname="col9" align="right"/>
     <oasis:thead>
       <oasis:row>
         <oasis:entry colname="col1"/>
         <oasis:entry rowsep="1" namest="col2" nameend="col3" align="center" colsep="1">G(PS) </oasis:entry>
         <oasis:entry rowsep="1" namest="col4" nameend="col5" align="center" colsep="1">GE<inline-formula><mml:math id="M214" display="inline"><mml:msup><mml:mi/><mml:mi mathvariant="normal">a</mml:mi></mml:msup></mml:math></inline-formula></oasis:entry>
         <oasis:entry rowsep="1" namest="col6" nameend="col7" align="center" colsep="1">GR<inline-formula><mml:math id="M215" display="inline"><mml:msup><mml:mi/><mml:mi mathvariant="normal">b</mml:mi></mml:msup></mml:math></inline-formula></oasis:entry>
         <oasis:entry rowsep="1" namest="col8" nameend="col9" align="center">GRE<inline-formula><mml:math id="M216" display="inline"><mml:msup><mml:mi/><mml:mi mathvariant="normal">c</mml:mi></mml:msup></mml:math></inline-formula></oasis:entry>
       </oasis:row>
       <oasis:row rowsep="1">
         <oasis:entry colname="col1"/>
         <oasis:entry colname="col2">East</oasis:entry>
         <oasis:entry colname="col3">North</oasis:entry>
         <oasis:entry colname="col4">East</oasis:entry>
         <oasis:entry colname="col5">North</oasis:entry>
         <oasis:entry colname="col6">East</oasis:entry>
         <oasis:entry colname="col7">North</oasis:entry>
         <oasis:entry colname="col8">East</oasis:entry>
         <oasis:entry colname="col9">North</oasis:entry>
       </oasis:row>
     </oasis:thead>
     <oasis:tbody>
       <oasis:row rowsep="1">
         <oasis:entry colname="col1">WRMS differences</oasis:entry>
         <oasis:entry colname="col2">(mm)</oasis:entry>
         <oasis:entry colname="col3">(mm)</oasis:entry>
         <oasis:entry colname="col4">(mm)</oasis:entry>
         <oasis:entry colname="col5">(mm)</oasis:entry>
         <oasis:entry colname="col6">(mm)</oasis:entry>
         <oasis:entry colname="col7">(mm)</oasis:entry>
         <oasis:entry colname="col8">(mm)</oasis:entry>
         <oasis:entry colname="col9">(mm)</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">3<inline-formula><mml:math id="M217" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col2">0.01</oasis:entry>
         <oasis:entry colname="col3">0.03</oasis:entry>
         <oasis:entry colname="col4">0.03</oasis:entry>
         <oasis:entry colname="col5">0.04</oasis:entry>
         <oasis:entry colname="col6">0.04</oasis:entry>
         <oasis:entry colname="col7">0.04</oasis:entry>
         <oasis:entry colname="col8">0.03</oasis:entry>
         <oasis:entry colname="col9">0.04</oasis:entry>
       </oasis:row>
       <oasis:row rowsep="1">
         <oasis:entry colname="col1">10 <inline-formula><mml:math id="M218" display="inline"><mml:mrow class="unit"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:mrow></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col2"><inline-formula><mml:math id="M219" display="inline"><mml:mo>-</mml:mo></mml:math></inline-formula>0.04</oasis:entry>
         <oasis:entry colname="col3">0.02</oasis:entry>
         <oasis:entry colname="col4"><inline-formula><mml:math id="M220" display="inline"><mml:mo>-</mml:mo></mml:math></inline-formula>0.03</oasis:entry>
         <oasis:entry colname="col5">0.01</oasis:entry>
         <oasis:entry colname="col6"><inline-formula><mml:math id="M221" display="inline"><mml:mo>-</mml:mo></mml:math></inline-formula>0.03</oasis:entry>
         <oasis:entry colname="col7">0.00</oasis:entry>
         <oasis:entry colname="col8"><inline-formula><mml:math id="M222" display="inline"><mml:mo>-</mml:mo></mml:math></inline-formula>0.02</oasis:entry>
         <oasis:entry colname="col9">0.01</oasis:entry>
       </oasis:row>
       <oasis:row rowsep="1">
         <oasis:entry namest="col1" nameend="col9">Correlation coefficients </oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">3<inline-formula><mml:math id="M223" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col2">0.02</oasis:entry>
         <oasis:entry colname="col3">0.01</oasis:entry>
         <oasis:entry colname="col4">0.00</oasis:entry>
         <oasis:entry colname="col5">0.01</oasis:entry>
         <oasis:entry colname="col6"><inline-formula><mml:math id="M224" display="inline"><mml:mo>-</mml:mo></mml:math></inline-formula>0.01</oasis:entry>
         <oasis:entry colname="col7">0.00</oasis:entry>
         <oasis:entry colname="col8">0.00</oasis:entry>
         <oasis:entry colname="col9"><inline-formula><mml:math id="M225" display="inline"><mml:mo>-</mml:mo></mml:math></inline-formula>0.01</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">10<inline-formula><mml:math id="M226" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula></oasis:entry>
         <oasis:entry colname="col2">0.03</oasis:entry>
         <oasis:entry colname="col3">0.04</oasis:entry>
         <oasis:entry colname="col4">0.03</oasis:entry>
         <oasis:entry colname="col5">0.03</oasis:entry>
         <oasis:entry colname="col6">0.02</oasis:entry>
         <oasis:entry colname="col7">0.04</oasis:entry>
         <oasis:entry colname="col8">0.01</oasis:entry>
         <oasis:entry colname="col9">0.03</oasis:entry>
       </oasis:row>
     </oasis:tbody>
   </oasis:tgroup></oasis:table><table-wrap-foot><p id="d1e3999"><inline-formula><mml:math id="M207" display="inline"><mml:msup><mml:mi/><mml:mi mathvariant="normal">a</mml:mi></mml:msup></mml:math></inline-formula> GPS <inline-formula><mml:math id="M208" display="inline"><mml:mo>+</mml:mo></mml:math></inline-formula> Galileo, <inline-formula><mml:math id="M209" display="inline"><mml:msup><mml:mi/><mml:mi mathvariant="normal">b</mml:mi></mml:msup></mml:math></inline-formula> GPS <inline-formula><mml:math id="M210" display="inline"><mml:mo>+</mml:mo></mml:math></inline-formula> Glonass, <inline-formula><mml:math id="M211" display="inline"><mml:msup><mml:mi/><mml:mi mathvariant="normal">c</mml:mi></mml:msup></mml:math></inline-formula> GPS <inline-formula><mml:math id="M212" display="inline"><mml:mo>+</mml:mo></mml:math></inline-formula> Glonass <inline-formula><mml:math id="M213" display="inline"><mml:mo>+</mml:mo></mml:math></inline-formula> Galileo.</p></table-wrap-foot></table-wrap>

      <p id="d1e4391">The results from the comparisons using different <inline-formula><mml:math id="M227" display="inline"><mml:mrow><mml:mi mathvariant="normal">Δ</mml:mi><mml:msub><mml:mi>t</mml:mi><mml:mtext>eff</mml:mtext></mml:msub></mml:mrow></mml:math></inline-formula> indicate that due to the poor geometry of the GNSS observations (especially
in the south–north direction), if the gradient happens suddenly, the GNSS data do not capture the full picture of the gradients. When a small <inline-formula><mml:math id="M228" display="inline"><mml:mrow><mml:mi mathvariant="normal">Δ</mml:mi><mml:msub><mml:mi>t</mml:mi><mml:mtext>eff</mml:mtext></mml:msub></mml:mrow></mml:math></inline-formula> is used, all gradients are kept, including the ones which are not correctly detected by the GNSS data. As a result, the correlation
between the GNSS and the WVR data is deteriorated, and this is the case when a <inline-formula><mml:math id="M229" display="inline"><mml:mrow><mml:mi mathvariant="normal">Δ</mml:mi><mml:msub><mml:mi>t</mml:mi><mml:mtext>eff</mml:mtext></mml:msub></mml:mrow></mml:math></inline-formula> of 5 <inline-formula><mml:math id="M230" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">min</mml:mi></mml:mrow></mml:math></inline-formula> is applied. However, when
<inline-formula><mml:math id="M231" display="inline"><mml:mrow><mml:mi mathvariant="normal">Δ</mml:mi><mml:msub><mml:mi>t</mml:mi><mml:mtext>eff</mml:mtext></mml:msub></mml:mrow></mml:math></inline-formula> is too large (i.e. 24 <inline-formula><mml:math id="M232" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">h</mml:mi></mml:mrow></mml:math></inline-formula>), the larger gradients which are captured correctly by both the GNSS and the WVR data will
also be averaged out. The range of variations and the correlation decrease.</p>

      <?xmltex \floatpos{t}?><fig id="Ch1.F12" specific-use="star"><?xmltex \currentcnt{12}?><?xmltex \def\figurename{Figure}?><label>Figure 12</label><caption><p id="d1e4465">Gradient time series from the WVR and from the GNSS (GRE) applying two different constraint values (0.3 and 1.0 <inline-formula><mml:math id="M233" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">mm</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:msqrt><mml:mi mathvariant="normal">h</mml:mi></mml:msqrt><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>) and using elevation cutoff angles of 3<inline-formula><mml:math id="M234" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula> <bold>(a, b)</bold> and 10<inline-formula><mml:math id="M235" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula> <bold>(c, d)</bold>.</p></caption>
          <?xmltex \igopts{width=384.112205pt}?><graphic xlink:href="https://amt.copernicus.org/articles/14/5593/2021/amt-14-5593-2021-f12.png"/>

        </fig>

</sec>
<sec id="Ch1.S3.SS4">
  <label>3.4</label><title>Different constraints for gradient variability</title>
      <?pagebreak page5603?><p id="d1e4524">All GNSS-derived gradients were so far estimated using a random walk model with a constraint value of 0.3 <inline-formula><mml:math id="M236" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">mm</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:msqrt><mml:mi mathvariant="normal">h</mml:mi></mml:msqrt><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> (see
Sect. <xref ref-type="sec" rid="Ch1.S2.SS1"/>). This value may be too small, especially when we have more observations from multi-GNSS constellations, to allow the GNSS
data to detect sudden large gradients. In order to investigate this issue, we have processed the GNSS data from June and July again with the
5 <inline-formula><mml:math id="M237" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">min</mml:mi></mml:mrow></mml:math></inline-formula> temporal resolution, applying a weaker constraint of 1.0 <inline-formula><mml:math id="M238" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">mm</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:msqrt><mml:mi mathvariant="normal">h</mml:mi></mml:msqrt><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>. The changes in WRMS differences and correlations,
relative to the solution using the constraint value of 0.3 <inline-formula><mml:math id="M239" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">mm</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:msqrt><mml:mi mathvariant="normal">h</mml:mi></mml:msqrt><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>, are shown in Table <xref ref-type="table" rid="Ch1.T4"/>, where
the GNSS gradients are estimated applying the elevation cutoff angles of 3<inline-formula><mml:math id="M240" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula> and 10<inline-formula><mml:math id="M241" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula>. The formal errors obtained when using the weak
constraint of 1.0 <inline-formula><mml:math id="M242" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">mm</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:msqrt><mml:mi mathvariant="normal">h</mml:mi></mml:msqrt><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> are about twice as large compared to the ones obtained when using the constraint of
0.3 <inline-formula><mml:math id="M243" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">mm</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:msqrt><mml:mi mathvariant="normal">h</mml:mi></mml:msqrt><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>.</p>
      <p id="d1e4648">When a weak constraint is applied with an elevation cutoff angle of 3<inline-formula><mml:math id="M244" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula>, the WRMS differences increase for both the east and the north
gradients, while the correlations are almost the same. We note that for the 10<inline-formula><mml:math id="M245" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula> solution, the east gradients obtained from using the weak
constraint values result in smaller WRMS differences compared to the WVR gradients. A slight improvement is also seen in the correlations for both the
east and the north gradients. We interpret the result as the compromise between capturing large gradients and including more noise. When a low
elevation cutoff angle is used, the GNSS measurements will be more sensitive to the noise from the environment, i.e. multipath effects. In addition,
the GNSS signals will have a lower signal-to-noise ratio due to the longer path through the atmosphere. When a weak constraint is applied, individual
observations will have a larger influence on the estimated gradients.</p>
      <p id="d1e4669">More details are seen in Fig. <xref ref-type="fig" rid="Ch1.F12"/>, depicting the time series of the gradients for the 2.5 d, starting at
00:00 UTC on 23 July, from the WVR and the GNSS data based on the GRE solutions. There are several peaks, i.e. large gradients, shown for these 2.5 d, i.e. east gradients at 21:00 UTC on 24 July and north gradients at 08:00 and 17:00 UTC on 23 July. When a weak constraint is
applied, there is a clear improvement in tracking those larger gradients when the elevation cutoff angle of 10<inline-formula><mml:math id="M246" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula> is used. This is not the case
when the lower elevation cutoff angle of 3<inline-formula><mml:math id="M247" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula> is used, possibly because the sampled volume of the atmosphere is more different compared to that
observed by the WVR.</p>
</sec>
</sec>
<?pagebreak page5604?><sec id="Ch1.S4" sec-type="conclusions">
  <label>4</label><title>Conclusions</title>
      <p id="d1e4701">We have estimated linear horizontal gradients using 1 year of data acquired from the GNSS station ONS1 located on the Swedish west coast. The
GNSS-derived gradients were compared to the ones obtained from a collocated WVR. Overall the multi-GNSS solutions, i.e. combinations of GPS, Glonass,
and Galileo, show small but significant improvements with the WVR gradients compared to the GPS-only solution (Tables <xref ref-type="table" rid="Ch1.T2"/>
and <xref ref-type="table" rid="Ch1.T3"/>).</p>
      <p id="d1e4708">For the GPS-only solution, the best agreement, in terms of the correlation coefficient with the WVR gradients, is obtained when using an elevation
cutoff angle of 3<inline-formula><mml:math id="M248" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula>. For the multi-GNSS solution using all three constellations, the best agreement with the WVR data is obtained for the
solution with an elevation cutoff angle of 10<inline-formula><mml:math id="M249" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula>. The difference is largest for the east component, which has the better sky coverage
(Fig. <xref ref-type="fig" rid="Ch1.F1"/> and Tables <xref ref-type="table" rid="Ch1.T2"/> and <xref ref-type="table" rid="Ch1.T3"/>). This indicates that if there are a
sufficient number of observations, the low-elevation observations are not that important. This is especially true when the comparison is made to a WVR
using observations evenly spread over the sky above an elevation angle of 25<inline-formula><mml:math id="M250" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula>. It is also an indication that a linear model for horizontal
variations in the wet refractivity does not describe the turbulent atmosphere well during all conditions.</p>
      <p id="d1e4745">We investigated different effective temporal resolutions, <inline-formula><mml:math id="M251" display="inline"><mml:mrow><mml:mi mathvariant="normal">Δ</mml:mi><mml:msub><mml:mi>t</mml:mi><mml:mtext>eff</mml:mtext></mml:msub></mml:mrow></mml:math></inline-formula>, of the compared time series. For all GNSS solutions, the highest
correlations obtained for the east and the north gradients are for a <inline-formula><mml:math id="M252" display="inline"><mml:mrow><mml:mi mathvariant="normal">Δ</mml:mi><mml:msub><mml:mi>t</mml:mi><mml:mtext>eff</mml:mtext></mml:msub></mml:mrow></mml:math></inline-formula> of 2 and 6 <inline-formula><mml:math id="M253" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">h</mml:mi></mml:mrow></mml:math></inline-formula>, respectively
(Fig. <xref ref-type="fig" rid="Ch1.F8"/>). When these <inline-formula><mml:math id="M254" display="inline"><mml:mrow><mml:mi mathvariant="normal">Δ</mml:mi><mml:msub><mml:mi>t</mml:mi><mml:mtext>eff</mml:mtext></mml:msub></mml:mrow></mml:math></inline-formula> values are applied, strong gradients of short duration detected by the WVR but not by
the GNSS are averaged out, and as a result the correlation increases. When estimating GNSS gradients the choice of <inline-formula><mml:math id="M255" display="inline"><mml:mrow><mml:mi mathvariant="normal">Δ</mml:mi><mml:msub><mml:mi>t</mml:mi><mml:mtext>eff</mml:mtext></mml:msub></mml:mrow></mml:math></inline-formula> is a compromise
between getting a high correlation and losing track of rapid gradient variations. However, when <inline-formula><mml:math id="M256" display="inline"><mml:mrow><mml:mi mathvariant="normal">Δ</mml:mi><mml:msub><mml:mi>t</mml:mi><mml:mtext>eff</mml:mtext></mml:msub></mml:mrow></mml:math></inline-formula> is even larger, e.g.
24 <inline-formula><mml:math id="M257" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">h</mml:mi></mml:mrow></mml:math></inline-formula>, all gradients are further averaged, and the dynamic range of gradient size and the correlation decreases (Figs. <xref ref-type="fig" rid="Ch1.F9"/>,
<xref ref-type="fig" rid="Ch1.F10"/>, and <xref ref-type="fig" rid="Ch1.F11"/>).</p>
      <p id="d1e4838">Furthermore, weakening the constraint used when estimating the GNSS gradients from 0.3 to 1.0 <inline-formula><mml:math id="M258" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">mm</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:msqrt><mml:mi mathvariant="normal">h</mml:mi></mml:msqrt><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> helps the GNSS data to track
short-lived gradients, approaching a timescale of 5 <inline-formula><mml:math id="M259" display="inline"><mml:mrow class="unit"><mml:mi mathvariant="normal">min</mml:mi></mml:mrow></mml:math></inline-formula>, however at the cost of increased formal errors
(Table <xref ref-type="table" rid="Ch1.T4"/> and Fig. <xref ref-type="fig" rid="Ch1.F12"/>).</p>
      <p id="d1e4872">Possible improvements to study in similar future work would be to include BeiDou observations and use a WVR with better stability. It would also be
of interest to carry out a similar study at low-latitude GNSS stations, where the sky coverage is better, and perhaps also the atmosphere is more
variable. In addition, the role of the geometry of GNSS observations (see Fig. <xref ref-type="fig" rid="Ch1.F1"/>) can be further studied. For example, one can
remove observations in a certain direction and investigate the change in the estimated gradients and their formal uncertainties for different
observation geometries.</p><?xmltex \hack{\newpage}?>
</sec>

      
      </body>
    <back><notes notes-type="dataavailability"><title>Data availability</title>

      <p id="d1e4883">The input GNSS data, in RINEX format, are available from EUREF (<uri>https://igs.bkg.bund.de/</uri>, <xref ref-type="bibr" rid="bib1.bibx8" id="altparen.28"/>). The ERA5 data are accessible from <ext-link xlink:href="https://doi.org/10.5065/BH6N-5N20" ext-link-type="DOI">10.5065/BH6N-5N20</ext-link> <xref ref-type="bibr" rid="bib1.bibx7" id="paren.29"/>. The estimated gradients from GNSS and WVR data have been registered and archived by the Swedish National Data Service (SND): <ext-link xlink:href="https://doi.org/10.5878/fyt8-bs80" ext-link-type="DOI">10.5878/fyt8-bs80</ext-link> <xref ref-type="bibr" rid="bib1.bibx5" id="paren.30"/>.</p>
  </notes><notes notes-type="authorcontribution"><title>Author contributions</title>

      <p id="d1e4908">The two authors (TN and GE) planned the work and the structure of the paper together. TN performed the GNSS data analyses and GE performed the WVR data analyses. Both contributed to the writing of the manuscript and approved it before the submission.</p>
  </notes><notes notes-type="competinginterests"><title>Competing interests</title>

      <p id="d1e4914">The authors declare that they have no conflict of interest.</p>
  </notes><notes notes-type="disclaimer"><title>Disclaimer</title>

      <p id="d1e4920">Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.</p>
  </notes><ack><title>Acknowledgement</title><p id="d1e4926">We thank Tobias Nilsson for providing the hydrostatic gradients from the ERA5 data.</p></ack><notes notes-type="reviewstatement"><title>Review statement</title>

      <p id="d1e4931">This paper was edited by Roeland Van Malderen and reviewed by two anonymous referees.</p>
  </notes><ref-list>
    <title>References</title>

      <ref id="bib1.bibx1"><?xmltex \def\ref@label{{Baldysz et~al.}(2018)}?><label>Baldysz et al.(2018)</label><?label baldysz_etal2018?><mixed-citation>Baldysz, Z., Nykiel, G., Figurski, M., and Araszkiewicz, A.:
Assessment of the impact of GNSS processing strategies on the long-term parameters of 20 years IWV time series, Remote Sens.-Basel,
10, 496, <ext-link xlink:href="https://doi.org/10.3390/rs10040496" ext-link-type="DOI">10.3390/rs10040496</ext-link>, 2018.</mixed-citation></ref>
      <ref id="bib1.bibx2"><?xmltex \def\ref@label{{Bar-Sever et~al.}(1998)}?><label>Bar-Sever et al.(1998)</label><?label barsever_etal1998?><mixed-citation>Bar-Sever, Y.-E., Kroger, P. M., and Börjesson, J. A.:
Estimating horizontal gradients of tropospheric path delay with a single GPS receiver,
J. Geophys. Res.,
103, 5019–5035, <ext-link xlink:href="https://doi.org/10.1029/97jb03534" ext-link-type="DOI">10.1029/97jb03534</ext-link>, 1998.</mixed-citation></ref>
      <ref id="bib1.bibx3"><?xmltex \def\ref@label{{Boehm et~al.}(2006)}?><label>Boehm et al.(2006)</label><?label boehm_etal2006?><mixed-citation>Boehm, J., Werl, B. and Schuh, H.:
Troposphere mapping functions for GPS and very long baseline interferometry from European Centre for Medium-Range Weather Forecasts operational analysis data,
J. Geophys. Res.,
111, B02406, <ext-link xlink:href="https://doi.org/10.1029/2005JB003629" ext-link-type="DOI">10.1029/2005JB003629</ext-link>, 2006.</mixed-citation></ref>
      <ref id="bib1.bibx4"><?xmltex \def\ref@label{{Davis et~al.}(1993)}?><label>Davis et al.(1993)</label><?label davis_etal1993?><mixed-citation>Davis, J. L., Elgered, G., Niell, A. E., and Kuehn, C. E.:
Ground-based measurement of gradients in the “wet” radio refractivity of air,
Radio Sci.,
28, 1003–1018, <ext-link xlink:href="https://doi.org/10.1029/93RS01917" ext-link-type="DOI">10.1029/93RS01917</ext-link>, 1993.</mixed-citation></ref>
      <ref id="bib1.bibx5"><?xmltex \def\ref@label{{Elgered and Ning}(2021)}?><label>Elgered and Ning(2021)</label><?label elgered_ning2021?><mixed-citation>Elgered, G. and Ning, T.: High temporal resolution wet delay gradients estimated from multi-GNSS and microwave radiometer observations, Chalmers University of Technology, Space, Earth and Environmental Science, Swedish National Data Service [data set], Version 1, <ext-link xlink:href="https://doi.org/10.5878/fyt8-bs80" ext-link-type="DOI">10.5878/fyt8-bs80</ext-link>, 2021.</mixed-citation></ref>
      <?pagebreak page5605?><ref id="bib1.bibx6"><?xmltex \def\ref@label{{Elgered et~al.}(2019)}?><label>Elgered et al.(2019)</label><?label elgered_etal2019?><mixed-citation>Elgered, G., Ning, T., Forkman, P., and Haas, R.: On the information content in linear horizontal delay gradients estimated from space geodesy observations, Atmos. Meas. Tech., 12, 3805–3823, <ext-link xlink:href="https://doi.org/10.5194/amt-12-3805-2019" ext-link-type="DOI">10.5194/amt-12-3805-2019</ext-link>, 2019.</mixed-citation></ref>
      <ref id="bib1.bibx7"><?xmltex \def\ref@label{ECMWF(2019)}?><label>ECMWF(2019)</label><?label ECMWF2019?><mixed-citation>European Centre for Medium-Range Weather Forecasts (ECMWF): ERA5 Reanalysis (0.25 Degree Latitude-Longitude Grid), updated monthly, Research Data Archive at the National Center for Atmospheric Research, Computational and Information Systems Laboratory [data set], <ext-link xlink:href="https://doi.org/10.5065/BH6N-5N20" ext-link-type="DOI">10.5065/BH6N-5N20</ext-link>, 2019.</mixed-citation></ref>
      <ref id="bib1.bibx8"><?xmltex \def\ref@label{{Federal Agency for Cartography and Geodesy}(2021)}?><label>Federal Agency for Cartography and Geodesy(2021)</label><?label FederalAgen2021?><mixed-citation>Federal Agency for Cartography and Geodesy, Germany: <uri>https://igs.bkg.bund.de/</uri>, last access: 12 August 2021.</mixed-citation></ref>
      <ref id="bib1.bibx9"><?xmltex \def\ref@label{{Graffigna et~al.}(2019)}?><label>Graffigna et al.(2019)</label><?label graffigna_etal2019?><mixed-citation>Graffigna, V., Hernández-Pajares, M., Gende, M. A., Azpilicueta, F. J., and Antico, P. L.: Interpretation ofthe tropospheric gradients estimated with GPS during Hurricane Harvey, Earth Space Science,
6, 1348–1365, <ext-link xlink:href="https://doi.org/10.1029/2018EA000527" ext-link-type="DOI">10.1029/2018EA000527</ext-link>, 2019.</mixed-citation></ref>
      <ref id="bib1.bibx10"><?xmltex \def\ref@label{{Jarlemark et~al.}(1998)}?><label>Jarlemark et al.(1998)</label><?label jarlemark1998?><mixed-citation>Jarlemark, P. O. J., Emardson, T. R., and Johansson, J. M.:
Wet delay variability calculated from radiometric measurements and its role in space geodetic parameter estimation,
Radio Sci.,
33, 719–730. <ext-link xlink:href="https://doi.org/10.1029/98RS00551" ext-link-type="DOI">10.1029/98RS00551</ext-link>, 1998.</mixed-citation></ref>
      <ref id="bib1.bibx11"><?xmltex \def\ref@label{{Ka\v{c}ma\v{r}\'{i}k et~al.}(2019)}?><label>Kačmařík et al.(2019)</label><?label kacmarik12019?><mixed-citation>Kačmařík, M., Douša, J., Zus, F., Václavovic, P., Balidakis, K., Dick, G., and Wickert, J.: Sensitivity of GNSS tropospheric gradients to processing options, Ann. Geophys., 37, 429–446, <ext-link xlink:href="https://doi.org/10.5194/angeo-37-429-2019" ext-link-type="DOI">10.5194/angeo-37-429-2019</ext-link>, 2019.</mixed-citation></ref>
      <ref id="bib1.bibx12"><?xmltex \def\ref@label{{Li et~al.}(2015)}?><label>Li et al.(2015)</label><?label li_etal2015?><mixed-citation>Li, X., Zus, F., Lu, C., Ning, T., Dick, G., Ge, M., Wickert, J., and Schuh, H.:
Retrieving high-resolution tropospheric gradients from multiconstellation GNSS observations,
Geophys. Res. Lett.,
42, 4173–4181, <ext-link xlink:href="https://doi.org/10.1002/2015GL063856" ext-link-type="DOI">10.1002/2015GL063856</ext-link>, 2015.</mixed-citation></ref>
      <ref id="bib1.bibx13"><?xmltex \def\ref@label{{Lu et~al.}(2016)}?><label>Lu et al.(2016)</label><?label lu_etal2016?><mixed-citation>Lu, C., Li, X., Li, Z., Heinkelmann, R., Nilsson, T., Dick, G., Ge, M., and Schuh, H.:
GNSS tropospheric gradients with high temporal resolution and their effect on precise positioning,
J. Geophys. Res.-Atmos.,
121, 912–930, <ext-link xlink:href="https://doi.org/10.1002/2015JD024255" ext-link-type="DOI">10.1002/2015JD024255</ext-link>, 2016.</mixed-citation></ref>
      <ref id="bib1.bibx14"><?xmltex \def\ref@label{{Lyard et~al.}(2006)}?><label>Lyard et al.(2006)</label><?label Lyard2006?><mixed-citation>Lyard, F., Lefevre, F., Letellier, T., and Francis, O.:
Modelling the global ocean tides: Modern insights from FES2004,
Ocean Dynam., 56, 394, <ext-link xlink:href="https://doi.org/10.1007/s10236-006-0086-x" ext-link-type="DOI">10.1007/s10236-006-0086-x</ext-link>, 2006.</mixed-citation></ref>
      <ref id="bib1.bibx15"><?xmltex \def\ref@label{{Meindl et~al.}(2004)}?><label>Meindl et al.(2004)</label><?label meindl_etal2004?><mixed-citation>Meindl, M., Schaer, S., Hugentobler, U., and Beutler, G.:
Tropospheric gradient estimation at CODE: results from global solutions,
J. Meteorol. Soc. Jpn.,
82, 331–338, <ext-link xlink:href="https://doi.org/10.2151/jmsj.2004.331" ext-link-type="DOI">10.2151/jmsj.2004.331</ext-link>, 2004.
</mixed-citation></ref><?xmltex \hack{\newpage}?>
      <ref id="bib1.bibx16"><?xmltex \def\ref@label{{Morel et~al.}(2015)}?><label>Morel et al.(2015)</label><?label morel_etal2015?><mixed-citation>Morel, L., Pottiaux, E., Durand, F., Fund, F., Boniface, K., de Oliveira, P. S., and Van Baelen, J.:
Validity and behaviour of tropospheric gradients estimated by GPS in Corsica,
Adv. Space Res.,
55, 135–149, <ext-link xlink:href="https://doi.org/10.1016/j.asr.2014.10.004" ext-link-type="DOI">10.1016/j.asr.2014.10.004</ext-link>, 2015.</mixed-citation></ref>
      <ref id="bib1.bibx17"><?xmltex \def\ref@label{{Ning et~al.}(2012)}?><label>Ning et al.(2012)</label><?label ning2012?><mixed-citation>Ning, T., Haas, R., Elgered, G., and Willén, U.:
Multi-technique comparisons of ten years of wet delay estimates on the west coast of Sweden,
J. Geodesy,
86, 565–575, <ext-link xlink:href="https://doi.org/10.1007/s00190-011-0527-2" ext-link-type="DOI">10.1007/s00190-011-0527-2</ext-link>, 2012.</mixed-citation></ref>
      <ref id="bib1.bibx18"><?xmltex \def\ref@label{{Prange et~al.}(2020)}?><label>Prange et al.(2020)</label><?label prange2020?><mixed-citation>Prange, L., Villiger, A., Sidorov, D., Schare, S., Beutler, G., Dach, R., and Jäggi, A.: Overview of CODE's MGEX solution with the focus on Galileo,
Adv. Space Res.,
66, 2786–2798, <ext-link xlink:href="https://doi.org/10.1016/j.asr.2020.04.038" ext-link-type="DOI">10.1016/j.asr.2020.04.038</ext-link>, 2020.</mixed-citation></ref>
      <ref id="bib1.bibx19"><?xmltex \def\ref@label{{Saastamoinen}(1973)}?><label>Saastamoinen(1973)</label><?label saastamoinen1973?><mixed-citation>Saastamoinen, J.:
Contributions to the theory of atmospheric refraction,
B. Geod.,
107, 13–34, <ext-link xlink:href="https://doi.org/10.1007/BF02521844" ext-link-type="DOI">10.1007/BF02521844</ext-link>, 1973.</mixed-citation></ref>
      <ref id="bib1.bibx20"><?xmltex \def\ref@label{{Schmid et~al.}(2007)}?><label>Schmid et al.(2007)</label><?label schmid2007?><mixed-citation>Schmid, R., Steigenberger, P., Gendt, G., Ge, M., and Rothacher, M.:
Generation of a consistent absolute phase center correction model for GPS receiver and satellite antennas,
J. Geodesy,
81, 781–798, <ext-link xlink:href="https://doi.org/10.1007/s00190-007-0148-y" ext-link-type="DOI">10.1007/s00190-007-0148-y</ext-link>, 2007.</mixed-citation></ref>
      <ref id="bib1.bibx21"><?xmltex \def\ref@label{{Steigenberger et~al.}(2007)}?><label>Steigenberger et al.(2007)</label><?label steigenberger_etal2007?><mixed-citation>Steigenberger, P., Tesmer, V., Krügel, M., Thaller, D., Schmid, R., Vey, S., and Rothacher, M.:
Comparison of homogeneously reprocessed GPS and VLBI long time-series of tropospheric zenith delays and gradients,
J. Geodesy,
81, 503–514, <ext-link xlink:href="https://doi.org/10.1007/s00190-006-0124-y" ext-link-type="DOI">10.1007/s00190-006-0124-y</ext-link>, 2007.</mixed-citation></ref>
      <ref id="bib1.bibx22"><?xmltex \def\ref@label{{V\'{a}clavovic et~al.}(2016)}?><label>Václavovic et al.(2016)</label><?label vaclavovic2016?><mixed-citation>Václavovic, P. and Douša, J.:
G-Nut/Anubis: open-source tool for multi-GNSS data monitoring,
in: IAG Symposia Series, Springer,
143, 775–782, <ext-link xlink:href="https://doi.org/10.1007/1345_2015_97" ext-link-type="DOI">10.1007/1345_2015_97</ext-link>, 2016.</mixed-citation></ref>
      <ref id="bib1.bibx23"><?xmltex \def\ref@label{{Zhou et~al.}(2017)}?><label>Zhou et al.(2017)</label><?label zhou_etal2017?><mixed-citation>Zhou F., Li, X., Li, W., Chen, W., Dong, D., Wickert, J., and Schuh, H.:
The Impact of Estimating High-Resolution Tropospheric Gradients on Multi-GNSS Precise Positioning,
Sensors,
17, 756, <ext-link xlink:href="https://doi.org/10.3390/s17040756" ext-link-type="DOI">10.3390/s17040756</ext-link>, 2017.</mixed-citation></ref>
      <ref id="bib1.bibx24"><?xmltex \def\ref@label{{Zumberge et~al.}(1997)}?><label>Zumberge et al.(1997)</label><?label zumberge1997?><mixed-citation>Zumberge, J. F., Heflin, M. B., Jefferson, D. C., Watkins, M. M., and Webb, F. H.: Precise Point Positioning for the Efficient and Robust Analysis of GPS Data from Large Networks, J. Geophys. Res.,
102, 5005–5017, <ext-link xlink:href="https://doi.org/10.1029/96JB03860" ext-link-type="DOI">10.1029/96JB03860</ext-link>, 1997.</mixed-citation></ref>
      <ref id="bib1.bibx25"><?xmltex \def\ref@label{{Zus et~al.}(2019)}?><label>Zus et al.(2019)</label><?label zus_etal2019?><mixed-citation>Zus, F., Douša, J., Kačmařík, M., Václavovic, P., Dick, G. and Wickert, J.:
Estimating the Impact of Global Navigation Satellite System Horizontal Delay Gradients in Variational Data Assimilation,
Remote Sens.-Basel,
11, 41, <ext-link xlink:href="https://doi.org/10.3390/rs11010041" ext-link-type="DOI">10.3390/rs11010041</ext-link>, 2019.</mixed-citation></ref>

  </ref-list></back>
    <!--<article-title-html>High-temporal-resolution wet delay gradients estimated from multi-GNSS and microwave radiometer observations</article-title-html>
<abstract-html><p>We have used 1 year of multi-GNSS observations at the Onsala Space Observatory on the Swedish west coast to estimate the linear horizontal
gradients in the wet propagation delay. The estimated gradients are compared to the corresponding ones from a microwave radiometer. We have
investigated different temporal resolutions from 5&thinsp;min to 1&thinsp;d. Relative to the GPS-only solution and using an elevation cutoff angle of
10° and a temporal resolution of 5&thinsp;min, the improvement obtained for the solution using GPS, Glonass, and Galileo data is an
increase in the correlation coefficient of 11&thinsp;% for the east gradient and 20&thinsp;% for the north gradient. Out of all the different GNSS
solutions, the highest correlation is obtained for the east gradients and a resolution of 2&thinsp;h, while the best agreement for the north
gradients is obtained for 6&thinsp;h. The choice of temporal resolution is a compromise between getting a high correlation and the possibility of
detecting rapid changes in the gradient. Due to the differences in geometry of the observations, gradients which happen suddenly are either not
captured at all or captured but with much less amplitude by the GNSS data. When a weak constraint is applied in the estimation of process, the GNSS
data have an improved ability to track large gradients, however, at the cost of increased formal errors.</p></abstract-html>
<ref-html id="bib1.bib1"><label>Baldysz et al.(2018)</label><mixed-citation>
Baldysz, Z., Nykiel, G., Figurski, M., and Araszkiewicz, A.:
Assessment of the impact of GNSS processing strategies on the long-term parameters of 20 years IWV time series, Remote Sens.-Basel,
10, 496, <a href="https://doi.org/10.3390/rs10040496" target="_blank">https://doi.org/10.3390/rs10040496</a>, 2018.
</mixed-citation></ref-html>
<ref-html id="bib1.bib2"><label>Bar-Sever et al.(1998)</label><mixed-citation>
Bar-Sever, Y.-E., Kroger, P. M., and Börjesson, J. A.:
Estimating horizontal gradients of tropospheric path delay with a single GPS receiver,
J. Geophys. Res.,
103, 5019–5035, <a href="https://doi.org/10.1029/97jb03534" target="_blank">https://doi.org/10.1029/97jb03534</a>, 1998.
</mixed-citation></ref-html>
<ref-html id="bib1.bib3"><label>Boehm et al.(2006)</label><mixed-citation>
Boehm, J., Werl, B. and Schuh, H.:
Troposphere mapping functions for GPS and very long baseline interferometry from European Centre for Medium-Range Weather Forecasts operational analysis data,
J. Geophys. Res.,
111, B02406, <a href="https://doi.org/10.1029/2005JB003629" target="_blank">https://doi.org/10.1029/2005JB003629</a>, 2006.
</mixed-citation></ref-html>
<ref-html id="bib1.bib4"><label>Davis et al.(1993)</label><mixed-citation>
Davis, J. L., Elgered, G., Niell, A. E., and Kuehn, C. E.:
Ground-based measurement of gradients in the “wet” radio refractivity of air,
Radio Sci.,
28, 1003–1018, <a href="https://doi.org/10.1029/93RS01917" target="_blank">https://doi.org/10.1029/93RS01917</a>, 1993.
</mixed-citation></ref-html>
<ref-html id="bib1.bib5"><label>Elgered and Ning(2021)</label><mixed-citation>
Elgered, G. and Ning, T.: High temporal resolution wet delay gradients estimated from multi-GNSS and microwave radiometer observations, Chalmers University of Technology, Space, Earth and Environmental Science, Swedish National Data Service [data set], Version 1, <a href="https://doi.org/10.5878/fyt8-bs80" target="_blank">https://doi.org/10.5878/fyt8-bs80</a>, 2021.
</mixed-citation></ref-html>
<ref-html id="bib1.bib6"><label>Elgered et al.(2019)</label><mixed-citation>
Elgered, G., Ning, T., Forkman, P., and Haas, R.: On the information content in linear horizontal delay gradients estimated from space geodesy observations, Atmos. Meas. Tech., 12, 3805–3823, <a href="https://doi.org/10.5194/amt-12-3805-2019" target="_blank">https://doi.org/10.5194/amt-12-3805-2019</a>, 2019.
</mixed-citation></ref-html>
<ref-html id="bib1.bib7"><label>ECMWF(2019)</label><mixed-citation>
European Centre for Medium-Range Weather Forecasts (ECMWF): ERA5 Reanalysis (0.25 Degree Latitude-Longitude Grid), updated monthly, Research Data Archive at the National Center for Atmospheric Research, Computational and Information Systems Laboratory [data set], <a href="https://doi.org/10.5065/BH6N-5N20" target="_blank">https://doi.org/10.5065/BH6N-5N20</a>, 2019.
</mixed-citation></ref-html>
<ref-html id="bib1.bib8"><label>Federal Agency for Cartography and Geodesy(2021)</label><mixed-citation>
Federal Agency for Cartography and Geodesy, Germany: <a href="https://igs.bkg.bund.de/" target="_blank"/>, last access: 12 August 2021.
</mixed-citation></ref-html>
<ref-html id="bib1.bib9"><label>Graffigna et al.(2019)</label><mixed-citation>
Graffigna, V., Hernández-Pajares, M., Gende, M. A., Azpilicueta, F. J., and Antico, P. L.: Interpretation ofthe tropospheric gradients estimated with GPS during Hurricane Harvey, Earth Space Science,
6, 1348–1365, <a href="https://doi.org/10.1029/2018EA000527" target="_blank">https://doi.org/10.1029/2018EA000527</a>, 2019.
</mixed-citation></ref-html>
<ref-html id="bib1.bib10"><label>Jarlemark et al.(1998)</label><mixed-citation>
Jarlemark, P. O. J., Emardson, T. R., and Johansson, J. M.:
Wet delay variability calculated from radiometric measurements and its role in space geodetic parameter estimation,
Radio Sci.,
33, 719–730. <a href="https://doi.org/10.1029/98RS00551" target="_blank">https://doi.org/10.1029/98RS00551</a>, 1998.
</mixed-citation></ref-html>
<ref-html id="bib1.bib11"><label>Kačmařík et al.(2019)</label><mixed-citation>
Kačmařík, M., Douša, J., Zus, F., Václavovic, P., Balidakis, K., Dick, G., and Wickert, J.: Sensitivity of GNSS tropospheric gradients to processing options, Ann. Geophys., 37, 429–446, <a href="https://doi.org/10.5194/angeo-37-429-2019" target="_blank">https://doi.org/10.5194/angeo-37-429-2019</a>, 2019.
</mixed-citation></ref-html>
<ref-html id="bib1.bib12"><label>Li et al.(2015)</label><mixed-citation>
Li, X., Zus, F., Lu, C., Ning, T., Dick, G., Ge, M., Wickert, J., and Schuh, H.:
Retrieving high-resolution tropospheric gradients from multiconstellation GNSS observations,
Geophys. Res. Lett.,
42, 4173–4181, <a href="https://doi.org/10.1002/2015GL063856" target="_blank">https://doi.org/10.1002/2015GL063856</a>, 2015.
</mixed-citation></ref-html>
<ref-html id="bib1.bib13"><label>Lu et al.(2016)</label><mixed-citation>
Lu, C., Li, X., Li, Z., Heinkelmann, R., Nilsson, T., Dick, G., Ge, M., and Schuh, H.:
GNSS tropospheric gradients with high temporal resolution and their effect on precise positioning,
J. Geophys. Res.-Atmos.,
121, 912–930, <a href="https://doi.org/10.1002/2015JD024255" target="_blank">https://doi.org/10.1002/2015JD024255</a>, 2016.
</mixed-citation></ref-html>
<ref-html id="bib1.bib14"><label>Lyard et al.(2006)</label><mixed-citation>
Lyard, F., Lefevre, F., Letellier, T., and Francis, O.:
Modelling the global ocean tides: Modern insights from FES2004,
Ocean Dynam., 56, 394, <a href="https://doi.org/10.1007/s10236-006-0086-x" target="_blank">https://doi.org/10.1007/s10236-006-0086-x</a>, 2006.
</mixed-citation></ref-html>
<ref-html id="bib1.bib15"><label>Meindl et al.(2004)</label><mixed-citation>
Meindl, M., Schaer, S., Hugentobler, U., and Beutler, G.:
Tropospheric gradient estimation at CODE: results from global solutions,
J. Meteorol. Soc. Jpn.,
82, 331–338, <a href="https://doi.org/10.2151/jmsj.2004.331" target="_blank">https://doi.org/10.2151/jmsj.2004.331</a>, 2004.

</mixed-citation></ref-html>
<ref-html id="bib1.bib16"><label>Morel et al.(2015)</label><mixed-citation>
Morel, L., Pottiaux, E., Durand, F., Fund, F., Boniface, K., de Oliveira, P. S., and Van Baelen, J.:
Validity and behaviour of tropospheric gradients estimated by GPS in Corsica,
Adv. Space Res.,
55, 135–149, <a href="https://doi.org/10.1016/j.asr.2014.10.004" target="_blank">https://doi.org/10.1016/j.asr.2014.10.004</a>, 2015.
</mixed-citation></ref-html>
<ref-html id="bib1.bib17"><label>Ning et al.(2012)</label><mixed-citation>
Ning, T., Haas, R., Elgered, G., and Willén, U.:
Multi-technique comparisons of ten years of wet delay estimates on the west coast of Sweden,
J. Geodesy,
86, 565–575, <a href="https://doi.org/10.1007/s00190-011-0527-2" target="_blank">https://doi.org/10.1007/s00190-011-0527-2</a>, 2012.
</mixed-citation></ref-html>
<ref-html id="bib1.bib18"><label>Prange et al.(2020)</label><mixed-citation>
Prange, L., Villiger, A., Sidorov, D., Schare, S., Beutler, G., Dach, R., and Jäggi, A.: Overview of CODE's MGEX solution with the focus on Galileo,
Adv. Space Res.,
66, 2786–2798, <a href="https://doi.org/10.1016/j.asr.2020.04.038" target="_blank">https://doi.org/10.1016/j.asr.2020.04.038</a>, 2020.
</mixed-citation></ref-html>
<ref-html id="bib1.bib19"><label>Saastamoinen(1973)</label><mixed-citation>
Saastamoinen, J.:
Contributions to the theory of atmospheric refraction,
B. Geod.,
107, 13–34, <a href="https://doi.org/10.1007/BF02521844" target="_blank">https://doi.org/10.1007/BF02521844</a>, 1973.
</mixed-citation></ref-html>
<ref-html id="bib1.bib20"><label>Schmid et al.(2007)</label><mixed-citation>
Schmid, R., Steigenberger, P., Gendt, G., Ge, M., and Rothacher, M.:
Generation of a consistent absolute phase center correction model for GPS receiver and satellite antennas,
J. Geodesy,
81, 781–798, <a href="https://doi.org/10.1007/s00190-007-0148-y" target="_blank">https://doi.org/10.1007/s00190-007-0148-y</a>, 2007.
</mixed-citation></ref-html>
<ref-html id="bib1.bib21"><label>Steigenberger et al.(2007)</label><mixed-citation>
Steigenberger, P., Tesmer, V., Krügel, M., Thaller, D., Schmid, R., Vey, S., and Rothacher, M.:
Comparison of homogeneously reprocessed GPS and VLBI long time-series of tropospheric zenith delays and gradients,
J. Geodesy,
81, 503–514, <a href="https://doi.org/10.1007/s00190-006-0124-y" target="_blank">https://doi.org/10.1007/s00190-006-0124-y</a>, 2007.
</mixed-citation></ref-html>
<ref-html id="bib1.bib22"><label>Václavovic et al.(2016)</label><mixed-citation>
Václavovic, P. and Douša, J.:
G-Nut/Anubis: open-source tool for multi-GNSS data monitoring,
in: IAG Symposia Series, Springer,
143, 775–782, <a href="https://doi.org/10.1007/1345_2015_97" target="_blank">https://doi.org/10.1007/1345_2015_97</a>, 2016.
</mixed-citation></ref-html>
<ref-html id="bib1.bib23"><label>Zhou et al.(2017)</label><mixed-citation>
Zhou F., Li, X., Li, W., Chen, W., Dong, D., Wickert, J., and Schuh, H.:
The Impact of Estimating High-Resolution Tropospheric Gradients on Multi-GNSS Precise Positioning,
Sensors,
17, 756, <a href="https://doi.org/10.3390/s17040756" target="_blank">https://doi.org/10.3390/s17040756</a>, 2017.
</mixed-citation></ref-html>
<ref-html id="bib1.bib24"><label>Zumberge et al.(1997)</label><mixed-citation>
Zumberge, J. F., Heflin, M. B., Jefferson, D. C., Watkins, M. M., and Webb, F. H.: Precise Point Positioning for the Efficient and Robust Analysis of GPS Data from Large Networks, J. Geophys. Res.,
102, 5005–5017, <a href="https://doi.org/10.1029/96JB03860" target="_blank">https://doi.org/10.1029/96JB03860</a>, 1997.
</mixed-citation></ref-html>
<ref-html id="bib1.bib25"><label>Zus et al.(2019)</label><mixed-citation>
Zus, F., Douša, J., Kačmařík, M., Václavovic, P., Dick, G. and Wickert, J.:
Estimating the Impact of Global Navigation Satellite System Horizontal Delay Gradients in Variational Data Assimilation,
Remote Sens.-Basel,
11, 41, <a href="https://doi.org/10.3390/rs11010041" target="_blank">https://doi.org/10.3390/rs11010041</a>, 2019.
</mixed-citation></ref-html>--></article>
