Continuous GPS Stations Research Articles

A high-resolution, time-variable afterslip model for the 2010 Maule Mw = 8.8, Chile megathrust earthquake

The excellent spatial coverage of continuous GPS stations in the region affected by the Maule Mw = 8.8 2010 earthquake, combined with the proximity of the coast to the seismogenic zone, allows us to model megathrust afterslip on the plate interface with unprecedented detail. We invert post-seismic observations from continuous GPS sites to derive a time-variable model of the first 420 d of afterslip. We also invert co-seismic GPS displacements to create a new co-seismic slip model. The afterslip pattern appears to be transient and non-stationary, with the cumulative afterslip pattern being formed from afterslip pulses. Changes in static stress on the plate interface from the co- and post-seismic slip cannot solely explain the aftershock patterns, suggesting that another process – perhaps fluid related – is controlling the lower magnitude aftershocks. We use aftershock data to quantify the seismic coupling distribution during the post-seismic phase. Comparison of the post-seismic behaviour to interseismic locking suggests that highly locked regions do not necessarily behave as rate-weakening in the post-seismic period. By comparing the inter-seismic locking, co-seismic slip, afterslip, and aftershocks we attempt to classify the heterogeneous frictional behaviour of the plate interface.

The Greenville Fault: Preliminary Estimates of Its Long-Term Creep Rate and Seismic Potential

Abstract Once assumed to be locked, we show that the northern third of the Greenville fault (GF) creeps at 2 mm/yr, based on 47 yr of trilateration net data. This northern GF creep rate equals its 11 ka slip rate, suggesting a low strain accumulation rate. In 1980, the GF, easternmost strand of the San Andreas fault system east of San Francisco Bay, produced an M w 5.8 earthquake with a 6 km surface rupture and dextral slip growing to ≥2 cm on cracks over a few weeks. Trilateration shows a 10 cm post‐1980 transient slip ending in 1984. Analysis of 2000–2012 crustal velocities on continuous Global Positioning System stations, allows creep rates of ∼2 mm/yr on the northern GF, 0–1 mm/yr on the central GF, and ∼0 mm/yr on its southern third. Modeled depth ranges of creep along the GF allow 5%–25% aseismic release. Greater locking in the southern two‐thirds of the GF is consistent with paleoseismic evidence there for large late Holocene ruptures. Because the GF lacks large (>1 km) discontinuities likely to arrest higher (∼1 m) slip ruptures, we expect full‐length (54 km) ruptures to occur that include the northern creeping zone. We estimate sufficient strain accumulation on the entire GF to produce M w 6.9 earthquakes with a mean recurrence of ∼575 yr. While the creeping 16 km northern part has the potential to produce an M w 6.2 event in 240 yr, it may rupture in both moderate (1980) and large events. These two‐dimensional‐model estimates of creep rate along the southern GF need verification with small aperture surveys. Online Material: Green net data file.

GPS and tectonic evidence for a diffuse plate boundary at the Azores Triple Junction

We use GPS, bathymetric/structural, and seismic data to define the pattern of present deformation along the northern half of the Azores plateau, where the Nubia–Eurasia plate boundary terminates at the axis of the Mid-Atlantic Ridge (MAR). New and existing campaign GPS velocities from the Azores islands reveal extension oblique to a series of en échelon volcanic ridges occupied by Terceira, S. Jorge, Pico, and Faial islands. In a frame of reference defined by 69 continuous GPS stations on the Eurasia plate, Terceira Island moves 2±1 mm/yr away from Eurasia, consistent with the islandʼs location within the Terceira Rift and plate boundary structure. The volcanic ridges south of the Terceira Rift move toward WSW at progressively faster rates, reaching a maximum of 3.5±0.5 mm/yr (2-σ) for the Pico/Faial volcanic ridge. The hypothesis that the Terceira Rift accommodates all Nubia–Eurasia plate motion is rejected at high confidence level based on the motions of sites on S. Jorge Island just west of Terceira Rift. All of the islands move relative to the Nubia plate, with Pico Island exhibiting the slowest motion, only 1±0.5 mm/yr (2-σ). Detailed bathymetry from the interior of the hypothesized Azores microplate reveals faults that crosscut young MAR seafloor fabric. These observations and the GPS evidence for distributed deformation described above argue against the existence of a rigid or semi-rigid Azores microplate, and instead suggest that Nubia–Eurasia plate motion is accommodated by extension across a ∼140-km-wide zone east of the MAR axis, most likely bounded to the north by the northern shoulder of the Terceira Rift. The MAR spreading rate along the western end of the Azores deformation zone (∼38.5°N–39.5°N) is intermediate between the Eurasia–North America rate measured at 39.5°N and the Nubia–North America rate measured at 38.5°N, consistent with the joint conclusions that the Nubia–Eurasia boundary is broad where it intersects the MAR, and the Azores Triple Junction is diffuse rather than discrete.

Kinematic interpretation of present-day crustal deformation in central Greece from continuous GPS measurements

We processed 30-s GPS data from continuous GPS stations in central Greece using the Kalman filtering approach and accounting for time-correlated noise content obtaining a velocity field in the ITRF2008 and the Eurasian-fixed reference frame. The station distribution allowed us to compute 1D strain through rates of baseline length changes as well as to construct the image of the 2D strain and rotation rate fields. The obtained baselines range in length from 11 to 132km and show rates from −1.95mm/yr up to 14.14mm/yr (estimated uncertainties from 0.3 to 0.8mm/yr), while the calculated 1D strain rate ranges from −27ns/yr up to 226ns/yr (average uncertainty ∼15ns/yr). Largest extension (192–226ns/yr) is observed in the western and central part of the Corinth rift while similar extension rates (80–120ns/yr) are obtained for the eastern part of the Corinth rift and its continuation in the south Viotia–south of Evia region and across the Sperchios–Kammena Vourla rift. The coherent picture of the velocity pattern for Attica and north-eastern Peloponnese (Corinth) stations indicates that these areas belong to the same crustal block, separating by the Viotia region by a nearly E–W crustal discontinuity along the Kaparelli–Asopos valley faults. However, some internal strain is present within Attica's crust as well as across the Saronic Gulf resulting in extension rates of the order of 25ns/yr. We also find extension (54–71ns/yr) across “rigid” Peloponnese taken by normal faults in the greater Kalavryta region.

Spatially variable fault friction derived from dynamic modeling of aseismic afterslip due to the 2004 Parkfield earthquake

AbstractWe investigate fault friction from dynamic modeling of fault slip prior to and following the Mw 6.0 earthquake which ruptured the Parkfield segment of the San Andreas Fault in 2004. The dynamic modeling assumes a purely rate‐strengthening friction law, with a logarithmic dependency on sliding rate: . The initial state of stress is explicitly taken into account, and afterslip is triggered by the stress change induced by the earthquake source model given a priori. We consider different initial stress states and two coseismic models, and invert for the other model parameters using a nonlinear inversion scheme. The model parameters include the reference friction μ*, the friction rate dependency characterized by the quantity a‐b, assumed to be either uniform or depth dependent. The model parameters are determined from fitting the transient postseismic geodetic signal measured at continuous GPS stations. Our study provides a view of frictional properties at the kilometers scale over the 0–15 km depth illuminated by the coseismic stress change induced by the Parkfield earthquake. The reference friction is estimated to be between 0.1 and 0.5. With independent a priori constraints on the amplitude of differential stress, the range of possible values narrows down to 0.1–0.17. The friction rate coefficient a‐b is estimated to be ∼ 10− 3 − 10− 2 with a hint that it increases upward from about 1–3 × 10–3 at 3–7 km depth to about 4–7 × 10–3 at 0–1 km depth. It is remarkable that our results are consistent with frictional properties measured on rock samples recovered from the fault zone thanks to the San Andreas Fault Observatory at Depth experiment.

Near-field surface deformation during the April 20, 2013, Ms7. 0 Lushan earthquake measured by 1-Hz GNSS

The April 20, 2013, M s7. 0 Lushan earthquake was successfully recorded by closely spaced Continuous Global Positioning System (CGPS) stations owned by the Crustal Movement Observation Network of China (CMONC). The 1-Hz GNSS data from eight CGPS stations, which are located between 30 km and 200 km from the hypocenter, were processed within quasi-real-time. The near-field surface deformation indicated the following characteristics: the near-field movements were limited to several centimeters; the peak of the deformation wave was significantly larger than the static permanent offset; at the beginning of the event, the north wall of the fault moved to the southeast as the south wall moved to the southwest; station SCTQ, which was the closest station to the hypocenter at 30 km, had the largest static permanent displacement of 2 cm; the peaks of the deformation waves were 1. 5 cm, 5 cm and 3 cm, to the east, the south and vertically upward, respectively ; and the peaks of velocity and acceleration, derived from the deformation, were 3. 4 cm/s and 5. 3 cm/s 2 , respectively.

The coseismic displacements of the 2013 Lushan Mw6.6 earthquake determined using continuous global positioning system measurements

Abstract: Based on Continuous GPS (CGPS) observation data of the Crustal Movement Observation Network of China (CMONOC) and the Sichuan Continuous Operational Reference System (SCCORS), we calculated the horizontal coseismic displacements of CGPS sites caused by the 2013 Lushan Mw 6. 6 earthquake. The results indicate that the horizontal coseismic deformations of CGPS stations are consistent with thrust-compression rupture. Furthermore, the sites closest to the epicenter underwent significant coseismic displacements. Three network stations exhibited displacements greater than 9 mm (the largest is 20. 9 mm at SCTQ), while the others were displaced approximately 1–4 mm.

Seismotectonics of the 2008 and 2009 Qaidam Earthquakes and its Implication for Regional Tectonics

Abstract:Three magnitude >6 earthquakes struck Qaidam, Qinghai province, China, in November 10th 2008, August 28th and 31st 2009 respectively. The Zongwulongshan fault has often been designated as the active seismogenic structure, although it is at odd with the data. Our continuous GPS station (CGPS), the Xiao Qaidam station, located in the north of the Qaidam basin, is less than 30 km to the southwest of the 2008 earthquake. This CGPS station recorded the near field co‐seismic deformation. Here we analyzed the co‐seismic dislocation based on the GPS time series and the rupture processes from focal mechanism for the three earthquakes. The aftershocks were relocated to constrain the spatial characteristics of the 2008 and 2009 Qaidam earthquakes. Field geological and geomorphological investigation and interpretation of satellite images show that the Xitieshan fault and Zongwulongshan fault were activated as left lateral thrust during the late Quaternary. Evidence of folding can also be identified. Integrated analyses based on our data and the regional tectonic environment show that the Xitieshan fault is the fault responsible for the 2008 Qaidam earthquake, which is a low dip angle thrust with left lateral strike slip. The Zongwulongshan fault is the seismogenic fault of the 2009 earthquakes, which is a south dipping back thrust of the northern marginal thrust system of the Qaidam basin. Folding takes a significant part of the deformation in the northern marginal thrust system of the Qaidam basin, dominating the contemporary structure style of the northern margin of the Qaidam basin and Qilianshan tectonic system. In this region, this fault and fold system dominates the earthquake activities with frequent small magnitude earthquakes.

Hybrid conventional and Persistent Scatterer SAR interferometry for land subsidence monitoring in the Tehran Basin, Iran

Excessive groundwater extraction has caused land subsidence in a large rural area located southwest of Tehran, Iran. We used radar images to estimate the temporal and spatial variation in the magnitude of the subsidence. Due to the large perpendicular baselines and rapid temporal decorrelation of the available data, the application of conventional synthetic aperture radar interferometry (InSAR) to monitor the deformation was not possible. Instead, we applied a recently developed Persistent Scatterer InSAR (PSI) method but found that displacements were underestimated in some areas due to high deformation rates that cause temporal aliasing of the signal. We therefore developed a method that combines conventional InSAR and PSI to estimate the high deformation rates in the southwestern Tehran Basin. We used rates estimated from conventional small temporal baseline interferograms to reduce the likelihood of aliasing and then applied PSI to the residual phase. The method was applied to descending and ascending ENVISAT ASAR images spanning from 2003 to 2009. Mean line-of-sight velocities obtained from both orientations that were further decomposed into horizontal and vertical deformation components were highly compatible with each other, indicating the high performance of the applied method. The mean precision of the estimated vertical component is 2.5mm/yr. We validated our results using measurements from a continuous GPS station located in one of the subsiding areas. The results also compare favourably with levelling data acquired over a different time interval. Finally, we compared the estimated displacements to hydraulic head variations and geologic profiles at several piezometric wells. We found that the geology is the most important factor controlling the subsidence rate in the southwestern Tehran Basin, regardless of the water level decline.

Slow slip event in the Mexican subduction zone: Evidence of shallower slip in the Guerrero seismic gap for the 2006 event revealed by the joint inversion of InSAR and GPS data

Slow slip events (SSEs) in subduction zones have been observed in the last decade with continuous GPS stations. Some of them could be related to the lateral segmentation of subduction interface that seems to be a critical parameter for the propagation of large subduction earthquakes. In 2006, one of the largest SSEs recorded so far was captured by a dozen continuous GPS stations, in the Guerrero area (Mexico) along the Mexican subduction zone. Previous studies based on these data suggested a lateral variation of the updip depth of the SEE at the Guerrero seismic gap, but suffered from a lack of resolution east of the gap. Here, we show the ability of InSAR technique to capture a part of the 2006 SSE cumulative displacement east of the Guerrero gap by a stacking approach. We processed long strip Envisat interferograms corrected for orbital errors and interseismic signal using GPS data. We first use a forward modelling approach to test InSAR sensitivity to the amount of slip, depth and width of the slipping area on the subduction interface. Due to its high spatial resolution, InSAR allows one to comprehensively sample the North–South spatial wavelength of the SSE deformation, complementing the sparse GPS network. InSAR locates the maximum of uplift and subsidence caused by the SSE more precisely than the GPS data, giving better constraints on the updip slip limit of the SSE. We then inverted the InSAR and GPS data separately to understand how each inversion resolves the slip at depth. Finally, we performed a joint inversion of InSAR and GPS data, which constrained the SSE slip and its location on the plate interface over the entire Guerrero area. The joint inversion shows significant lateral variation of the SSE slip distribution along the trench with a shallower updip edge in the Guerrero seismic gap, west of Acapulco, and a deeper slip edge further east.

SqueeSAR™ and GPS ground deformation monitoring of Santorini Volcano (1992–2012): Tectonic implications

The Santorini Volcanic Complex (SVC) has been in a dormant state for the last 60years until January 2011 when upward influx of magma reawakened the volcano with intense radial ground deformation and inter-caldera seismicity that lasted until January 2012 but declined afterwards. This paper aims to study the ground deformation and the inferred tectonic implications of the SVC for the period 1992–2012 mainly based on the SqueeSAR™ technique and DGPS campaign results of our local network which incorporates available data on Internet from several continuous GPS stations established on the island. The spatial deformation of the SVC during the quiet period 1992–2010 was deduced by joint analysis of ERS1 and 2 and ENVISAT. It was found that the intra caldera Palaea Kammeni shield volcano was being uplifted (2–3mm/yr) with increasing rate, whilst the adjacent Nea Kammeni shield volcano was being subsided (up to 6mm/yr) with increasing rate. The rest of the SVC showed a velocity field varying from −1 to +2mm/yr, indicating a rather linear deformation during that period. The results from the GPS network are in full agreement with the SqueeSAR results. Based on the results of SqueeSAR analysis of 12 ENVISAT images, and DGPS/CGPS data to end 2012, the deformation for the unrest period 2011–2012 was non-linear being characterized by strong radial deformation in the northern part of the caldera (50–120mm/yr), and accelerating values (>130mm/yr2). Combined GPS/SqueeSAR Mogi modeling indicated a source located north of Nea Kammeni at a shallow depth. However, a progressively decreasing rate in deformation was noted at most GPS/CGPS station components after January 2012, indicating magma settlement consistent with the constantly decreasing rate of the inter-caldera seismicity. The faulting features seem to have a key role in the evolution of the deformation, which continues up the end 2012, but at a very low level.

The interseismic velocity field of the central Apennines from a dense GPS network

<p>Since 1999, we have repeatedly surveyed the central Apennines through a dense survey-style geodetic network, the Central Apennines Geodetic Network (CAGeoNet). CAGeoNet consists of 123 benchmarks distributed over an area of ca. 180 km × 130 km, from the Tyrrhenian coast to the Adriatic coast, with an average inter-site distance of 3 km to 5 km. The network is positioned across the main seismogenic structures of the region that are capable of generating destructive earthquakes. Here, we show the horizontal GPS velocity field of both CAGeoNet and continuous GPS stations in this region, as estimated from the position–time series in the time span from 1999 to 2007. We analyzed the data using both the Bernese and GAMIT software, rigorously combining the two solutions to obtain a validated result. Then, we analyzed the strain-rate field, which shows a region of extension along the axis of the Apennine chain, with values from 2 × 10<span><sup>–9</sup></span> yr<span><sup>–1</sup></span> to 66·× 10<span><sup>–9</sup></span> yr<span><sup>–1</sup></span>, and a relative minimum of ca. 20 × 10<span><sup>–9</sup></span> yr<span><sup>–1</sup></span> located in the L'Aquila basin area. Our velocity field represents an improved estimation of the ongoing elastic interseismic deformation of the central Apennines, and in particular relating to the area of the L'Aquila earthquake of April 6, 2009.</p>

Deformación cortical de las Béticas Orientales observada mediante GPS y su relación con el terremoto de Lorca

On May 11 th of 2011, a seismic series occurred near the city of Lorca (Murcia). The main earthquake of magnitude Mw 5.2 has been attributed to the Alhama de Murcia Fault, one of the most active faults in the SE Iberian Peninsula. We analyzed data from 5 GPS campaigns of the CuaTeNeo network conducted between 1997 and 2011. The velocities of the stations closest to the Alhama de Murcia Fault show the reverse and strike-slip direction of motion. Stations located on the southeastern side of the fault have the maximum velocities in the area (between 1.4 and 1.8 mm/yr), oriented towards NNW direction, obliquely to the trace of the fault. The kinematics of the fault and the strain rate directions obtained from the CuaTeNeo network GPS measurements matches the calculated focal mechanism of Lorca earthquake. Detailed analysis of the time-series from the continuous GPS station at the Lorca city allows the detection of co-seismic offset of ~6 mm to the North.

Assessment of Pseudorange Multipath at Continuous GPS Stations in Mexico

We conducted a study to quantify the amount of pseudorange multipath at continuous Global Positioning System (CGPS) stations in the Mexican territory. These CGPS stations serve as reference stations enabling rapid high-precision three-dimensional positioning capabilities, supporting a number of commercial and public safety applications. We studied CGPS data from a large number of publicly available networks spanning Mexico. These include the RGNA (National Active Geodetic Network) administered by INEGI (National Institute of Statistics and Geography), the PBO network (Plate Boundary Observatory) funded by the National Science Foundation (NSF) and operated by UNAVCO (University NAVstar Consortium), the Southern California Integrated GPS Network (SCIGN), which is a collaboration effort of the United States Geological Survey (USGS), Scripps Institution of Oceanography and the Jet Propulsion Laboratory (JPL), the UNAM network, operated by the National Seismological System (SSN) and the Institute of Geophysics of the National Autonomous University of Mexico (UNAM), the Suominet Geodetic Network (SNG) and the CORS (Continuously Operating Reference Station) network, operated by the Federal Aviation Administration (FAA). We evaluated a total of 53 CGPS stations, where dual-frequency geodetic-grade receivers collected GPS data continuously during the period from 1994 to 2012. Despite carefully selected locations, all GPS stations are, to some extent, affected by the presence of signal multipath. For GPS network users that rely on pseudorange observables, the existence of pseudorange multipath could be a critical source of error depending on the time scale of the application. Thus, to identify the most and the least affected GPS stations, we analyzed the averaged daily root mean square pseudorange multipath variations (MP1-RMS and MP2-RMS) for all feasible satellites tracked by the CGPS networks. We investigated the sources of multipath, including changes associated with hardware replacement (i.e., receiver and antenna type) and receiver firmware upgrades.

Surface waves of the 2011 Tohoku earthquake: Observations of Taiwan's dense high‐rate GPS network

AbstractSurface waves generated by the 2011 Mw 9.0 Tohoku, Japan earthquake were recorded by both the high‐rate GPS and broadband seismic stations in Taiwan. In this study, we investigate the precision of high‐rate GPS displacements and estimate the feasibility of using GPS for seismology study in Taiwan. One hertz observations of 210 continuous GPS stations in Taiwan and neighboring islands were processed using a precise point positioning technique to estimate the absolute epoch‐by‐epoch positions generated by the Tohoku earthquake. Modified sidereal filtering was used to correct for near‐daily periodical variations of high‐rate position time series that may be influenced by multipath effects, and a band‐pass filter was used to remove noises from marginal frequencies. For all the 210 stations, the GPS precision after modified sidereal filtering improved from 8.2 to 7.4 mm in the horizontal and from 19.5 to 12.7 mm in the vertical components except for about 20% of the data with large position errors. After applying the band‐pass filter in the teleseismic frequency band (0.008–0.08 Hz), excluding 54 stations with noisy data, 156 continuous GPS stations were selected. Surface wave displacements derived from both high‐rate GPS and broadband seismometers are highly consistent, and the correlation coefficients are enhanced by a band‐pass filter (0.008–0.08 Hz) from 0.85 to 0.95 in the horizontal and from 0.58 to 0.85 in the vertical components. We consider continuous GPS can be used as an alternative approach to study characteristics of surface wave propagation in Taiwan.

Volcano deformation at active plate boundaries: Deep magma accumulation at Hekla volcano and plate boundary deformation in south Iceland

Most magmatic systems on Earth are located at actively deforming plate boundaries. In these systems, the magmatic and plate boundary deformation signals are intertwined and must be deconvolved to properly estimate magma flux and source characteristics of the magma plumbing system. We investigate the inter‐rifting and inter‐seismic deformation signals at the Eastern Volcanic Zone (EVZ) – South Iceland Seismic Zone (SISZ) ridge ‐ transform intersection and estimate the location, depth, and volume rate for magmatic sources at Hekla and Torfajökull volcanoes, which are located at the intersection. We solve simultaneously for the source parameters of the tectonic and volcanic deformation signals using a new ten‐year velocity field derived from a dense network of episodic and continuous GPS stations in south Iceland. We find the intersection of the axes of the EVZ and the SISZ is located within the Torfajökull caldera, which itself is subsiding. Deformation at Hekla is statistically best described in terms of a horizontal ellipsoidal magma chamber at 24−2+4 km depth aligned with the volcanic system and increasing in volume by 0.017−0.002+0.007 km3 per year. A spherical magma chamber centered at 24−2+5 km depth with a volume rate of 0.019−0.002+0.011 km3per year, or a vertical pipe‐shaped magma chamber between 10−1+3 km and 21−4+7 km with a volume rate of 0.008−0.001+0.003 km3 per year are also plausible models explaining the deformation at Hekla. All three models indicate magma accumulation in the lower crust or near the Moho under Hekla.

The coseismic and postseismic deformation of the L’Aquila, 2009 earthquake from repeated GPS measurements

We analyze more than 100 GPS time series of continuous and discontinuous GPS stations located in the Abruzzi region (Italy) surrounding the epicentres of the L'Aquila 2009 seismic sequence. The purpose of this work is to reconstruct the coseismic displacement field caused by the 6th April (Mw 6.3) main shock from a dense network of survey-mode stations surrounding the epicentral area and to characterize the early postseismic deformation field. In the months following the main shock, an extensive GPS survey was carried out on the existing Central Apennines Geodetic Network (CAGeoNet), with the intention of collecting a robust data set and to study the co- and postseismicdeformation field of this Apenninic normal faulting earthquake. The analysis is carried out with two independent procedures and software (Bernese and Gamit) in order to provide reliable and validated geodetic solutions. The analysis of the postseismic transients and the knowledge of long-term inter-seismic velocities at all GPS stations, issued from permanent and CAGeoNet sites, allow us to derive a dense co- and postseismic displacement field for the L'Aquila Mw 6.3 main shock in a wide area around the epicentre. The highest deformation rate occurs during the first 4-5 months after the main shock and persists in the following at slightly slower rate throughout the whole monitoring period. Fast deformation rates imply that most of the observed deformation is due to a process different from a pure viscoelastic relaxation of the stress perturbation. Since the observed rates would imply a too low effective viscosity value (below 1017 Pa s), we rather suggest that most of the observed deformation in the first months after the earthquake is due to different processes, most likely frictional afterslip possibly modulated by the presence of fluids. The new coseismic displacement field is used to invert for the main shock fault geometry, analysing the consistency among the different geodetic solutions and the combined one, with the goal of validating the two data sets.

Performance of Network-Based RTK GPS in Low-Latitude Region: A Case Study in Thailand

Thailand has established a Network-based Real Time Kinematic (NRTK) GPS system using the Virtual Reference Station (VRS) concept since 2008. Currently, the Thai NRTK consists of 11 reference stations located in the central part of Thailand with averaged reference station spacing at 60 km. A Previous test suggested some problems with the low rate of ambiguity-fixing and the large number of position jumps even when the ambiguity-fixed solutions could be obtained. In this paper, the position performance of NRTK is tested with a large number of GPS observations (31 consecutive days) and different reference receiver spacing, 10-20, 30-50, 50-60 and 60-80 km, with the use of all available Continuous GPS (CGPS) stations in the central part of Thailand. Test results indicate that the NRTK positioning performance is degraded when the reference station spacing is increased. It is also found that the ionospheric bias is the main error source that affects the performance of NRTK in Thailand. Even with the 10-20km reference station spacing, reliable ambiguity-fixed solutions could hardly be obtained during the period of high ionospheric variation. Thus, it is recommended that the reference station spacing should be kept less than 30km. In order to achieve the relaible NRTK solutions especially in low-latitude region like in Thailand, the ionospheric bias should be properly handled.

Seasonal hydrological loading in southern Alaska observed by GPS and GRACE

We compare vertical seasonal loading deformation observed by continuous GPS stations in southern Alaska and modeled vertical displacements due to seasonal hydrological loading inferred from GRACE. Seasonal displacements are significant, and GPS‐observed and GRACE‐modeled seasonal displacements are highly correlated. We define a measure called the WRMS Reduction Ratio to measure the fraction of the position variations at seasonal periods removed by correcting the GPS time series using a seasonal model based on GRACE. The median WRMS Reduction Ratio is 0.82 and the mean is 0.73 ± 0.26, with a value of 1.0 indicating perfect agreement of GPS and GRACE. The effects of atmosphere and non‐tidal ocean loading are important; we add the AOD1B de‐aliasing model to the GRACE solutions because the displacements due to these loads are present in the GPS data, and this improves the correlations between these two geodetic measurements. We find weak correlations for some stations located in areas where the magnitude of the load changes over a short distance, due to GRACE's limited spatial resolution. GRACE models can correct seasonal displacements for campaign GPS measurements as well.

Modeling aquifer-system compaction and predicting land subsidence in central Taiwan

Changhua is a county in central Taiwan that will house several major economic development projects. Extracting groundwater has caused large-scale land subsidence in Changhua, with the largest cumulative subsidence being 210cm over 1992–2010. A multi-sensor monitoring system consisting of continuous GPS stations, a leveling network, multi-layer compaction monitoring wells and groundwater wells is deployed to monitor land subsidence and its mechanism in Changhua. Data from the monitoring well CGSG in Dacheng Township of Changhua show a cumulative compaction of 110.6cm over 1997–2010, occurring mainly at two aquifers and one aquitard. A novel combination of GPS and monitoring well data was used to determine the stress–strain relations. The stratum compaction turns from plasticity to elastoplasticity after a long-term compaction at the second aquifer below the surface. Four hydrogeological parameters of the three sediment layers in Dacheng, vertical hydraulic conductivity, elastic skeletal specific storage, inelastic skeletal specific storage, and the initial maximum preconsolidation stress, in the one‐dimensional compaction model COMPAC, are estimated using the genetic algorithm. With the parameters, COMPAC predicts compactions to an accuracy consistent with in situ measurements, and the mean absolute percentage errors of prediction are below 10%. The result provides a key reference for water management in central Taiwan.