Water vapour estimates over Antarctica from 12 years of globally reprocessed GPS solutions Ian Thoma

1
Water vapour estimates over Antarctica from 12
years of globally reprocessed GPS solutions
Ian Thomas, Matt King, Peter Clarke Newcastle
University, UK email ian.thomas_at_ncl.ac.uk

Introduction Atmospheric water vapour - a natural
greenhouse gas - is often poorly monitored in
spatial and temporal resolution, particularly in
remote and inhospitable regions such as
Antarctica. GPS is known to be able to provide
accurate estimates of precipitable water (PW),
and is well suited to contribute useful
information in such inhospitable regions.
Continual advances in GPS modelling and
processing strategies over the last decade have
meant that GPS time series of geodetic parameters
(e.g. those derived using IGS final orbits) are
most often inhomogeneous. Such inconsistencies
limit the usefulness of GPS for the reliable
detection of small secular trends (e.g. climatic
signals, vertical surface velocities).
Here, we present the results of a homogenous
global reprocessing of 12 years of GPS data,
1995-2006, from which we derive a 2 hourly
precipitable water (PW) data set at 12 Antarctic
locations.
Stability of reprocessed time series
To confirm stability of the reprocessed GPS,
daily solutions were mapped to ITRF2005 by a
7-parameter Helmert Transformation.
Helmert parameters show GPS solutions to have
good long term stability with respect to ITRF2005
particularly scale parameter bias of -0.4 ppb
and scale drift 0.002 ppb.yr-1 (left plot).
Drift in X, Y and Z translation parameters of
0.25 0.40, 0.02 mm.yr-1 respectively (right
plot). A strong near annual signal is evident in
the scale of GPS relative to ITRF2005.


Plots, left to right and top to bottom
1) Radiosonde versus GPS PW scatter plots for 6
Antarctic sites. 2) Monthly mean of GPS PW (blue)
, radiosonde PW (green) and temperature (red,
right scale). 3) 1995-2004 GPS radiosonde PW, a
nd temperature at site DAV1. 4) As 3) for site AM
UN. 5) Spectral analysis for site MAW1. (black a
ll, red summer, blue winter) Signals are
typical of E. Antarctica.
Antarctic water vapour results
We compare our two hourly GPS PW estimates with
measurements derived from radiosonde and from
instruments aboard NASAs AQUA satellite - AIRS,
MODIS, AMSR-E. Plot below shows measurements for
2004, for site CAS1. Results and inter-technique
agreement at CAS1 are typical of East Antarctic
coastal locations. Strong annual signal in PW, 1
mm winter to 9 mm summer. The Antarctic Peninsu
la is more humid (12 mm PW in summer), The South
Pole (AMUN) is drier (2mm PW in summer).

.
Global reprocessing strategy Daily 60 station glo
bal GPS networks (e.g. 1st Jan 2004, above)
processed using GAMIT 10.3 software. Most sites
are in the IGS. Parameters estimated station coo
rdinates, satellite orbital parameters, Earth
orientation parameters, tropospheric zenith
delays and horizontal gradients.
Observation models Absolute satellite and receiv
er antenna phase centre variations
(igs05_1421.atx). VMF1 wet and dry mapping functi
ons. Elevation cut-off angle of 7.5 (10 for Ant
arctic sites). OTL displacements computed from FE
S2004 model. Atmospheric loading displacements co
mputed from 6-hourly NCEP reanalysis data.
Ambiguities resolved - GAMIT LC_HELP
LC_AUTCLN. 3-hourly surface temperature press
ure data used to compute a priori ZHD and for
ZWD?PW conversion (Bevis et al.,1994).
Conclusions Reprocessed GPS provides a temporally
stable PW measurement technique that compare
favourably with radiosonde and AIRS satellite
measurements for the dry Antarctic climate.
GPS PW agrees best with radiosonde
(sub-millimetre), followed by AIRS, MODIS and
AMSR-E. GPS captures the dry climate of Antarctic
a, and the dominant annual signal, plus a
semi-annual signal at East Antarctic sites.
A summer-time diurnal (S1) signal is observed at
many sites. Possible secular increases in PW in E
ast Antarctica and at the South Pole, although -
even with reprocessing - longer time series are
required for GPS to be useful as climate tool.
Acknowledgement We thank Merrit N. Deeter of
NCAR, Boulder, for computing and providing the
AMSR-E PW estimates. Reference Bevis, M., S. Bu
singer, and Chiswell, S.R, Herring, T.A., Anthes,
A., Rocken, C., Ware, R., 1994 GPS Meteorology
Mapping zenith wet delays onto precipitable
water, J. Appl. Met., 33, 379-386