Nontidal ocean loading: amplitudes and potential effects in GPS height time series

in Journal of Geodesy (2012), 86(11), 1043-1057

Ocean bottom pressure (OBP) changes are caused by a redistribution of the ocean’s internal mass that are driven by atmospheric circulation, a change in the mass entering or leaving the ocean, and/or a ... [more ▼]

Ocean bottom pressure (OBP) changes are caused by a redistribution of the ocean’s internal mass that are driven by atmospheric circulation, a change in the mass entering or leaving the ocean, and/or a change in the integrated atmospheric mass over the ocean areas. The only previous global analysis investigating the magnitude of OBP surface displacements used older OBP data sets (van Dam et al. in J Geophys Res 129:507–517, 1997). Since then significant improvements in meteorological forcing models used to predict OBP have been made, augmented by observations from satellite altimetry and expendable bathythermograph profiles. Using more recent OBP estimates from the Estimating the Circulation and Climate of the Ocean (ECCO) project, we reassess the amplitude of the predicted effect of OBP on the height coordinate time series from a global distribution of GPS stations. OBP-predicted loading effects display an RMS scatter in the height of between 0.2 and 3.7 mm, larger than previously reported but still much smaller (by a factor of 2) than the scatter observed due to atmospheric pressure loading. Given the improvement in GPS hardware and data analysis techniques, the OBP signal is similar to the precision of weekly GPS height coordinates. We estimate the effect of OBP on GPS height coordinate time series using the MIT reprocessed solution, mi1. When we compare the predicted OBP height time series with mi1, we find that the scatter is reduced over all stations by 0.1 mm on average with reductions as high as 0.7 mm at some stations. More importantly we are able to reduce the scatter on 65 % of the stations investigated. The annual component of the OBP signal is responsible for 80 % of the reduction in scatter on average.We find that stations located close to semi-enclosed bays or seas are affected by OBP loading to a greater extent than other stations. [less ▲]

Impact of loading effects on determination of the International Terrestrial Reference Frame

in Advances in Space Research (2010), 45(1), 144-154

The International Terrestrial Reference Frame (ITRF), as a realization of the International Terrestrial Reference System (ITRS), is represented by a set of station positions and linear velocities. They ... [more ▼]

The International Terrestrial Reference Frame (ITRF), as a realization of the International Terrestrial Reference System (ITRS), is represented by a set of station positions and linear velocities. They are intended to be used as regularized coordinates to which some corrections should be added to access instantaneous coordinates. The latest ITRS realization is the ITRF2005, which has integrated time series of station positions to form long-term solutions for the four space geodetic techniques. Currently, a purely linear model is used to parameterize station displacements in the estimation process, plus occasional discontinuities in case of earthquakes or equipment changes. However the input data have been derived without applying surface loading models and so surface loading effects are supposed to be embedded in the coordinates as measured quantities. We evaluate the effect of applying a posteriori loading corrections, which include the effect of atmospheric, non-tidal ocean, and continental water loading, to time series of positions estimated from Satellite Laser Ranging (SLR), Very Long Baseline Interferometry (VLBI), and Global Positioning System (GPS) data. We notice that they reduce about 50% or more of the annual signals in the translation and scale parameter time series of the SLR and VLBI techniques, except in SLR Z translation. In general, the estimated secular frame definition is negligibly affected and estimated positions and velocities are not significantly modified for stations that have accumulated a large number of observations. A multi-technique combination of such derived frames allows concluding that, for some cases, loading model corrections might degrade co-located station coordinates almost as much as they benefit them. However, most significant improvement of the estimated secular coordinates is observed for stations with less than 100 estimated positions as demonstrated with a multi-technique combination. [less ▲]

Effect of the satellite laser ranging network distribution on geocenter motion estimation

in Journal of Geophysical Research (2009), 114

SLR network translations estimated between a quasi-instantaneous station position set, theoretically expressed with respect to the center of mass of the Earth (CM), and a secular reference frame are the ... [more ▼]

SLR network translations estimated between a quasi-instantaneous station position set, theoretically expressed with respect to the center of mass of the Earth (CM), and a secular reference frame are the signature of the motion of the CM with respect to the Earth crust. Geocenter motion is defined here to be the motion of the CM with respect to the geometric center of the solid Earth surface (CF). SLR translational variations cannot be rigorously interpreted as identical to geocenter motion due to the sparse and nonuniform distribution of the SLR network. Their difference is called the network effect, which should be dominated at subdecadal timescales by loading signals.We have computed translation time series of the SLR network using two independent geophysically based loading models. One is a displacement model estimated from surface fluid data (Green’s function approach), called forward model, and the other is a displacement model estimated from GPS and ocean bottom pressure (OBP) data, called inverse model. The translation models have been subtracted from their respective geocenter motion models computed from degree-1 mass load coefficients in order to evaluate their network effect biases. Scatter due to the SLR network effect is at the level of 1.5 mm RMS. It could slightly shift the phase of the annual SLR geocenter motion estimate by less than 1 month and could affect X and Z annual geocenter motion amplitudes at the 1-mm level, which is about one third of the expected signal. Two distinct methods are suggested to account for network effect when comparing SLR translations to geocenter motion models. The first is to add the network effect term predicted by a displacement model to the geocenter motion loading model. The second relies on an adequate combination of SLR and GPS products to estimate SLR translation that could be better compared with geocenter motion. [less ▲]