Evaluation of various wet tropospheric corrections Evaluation of various wet tropospheric corrections Evaluation of various wet tropospheric corrections Evaluation of various wet tropospheric corrections in the Indonesia region in the Indonesia region in the Indonesia region in the Indonesia region EkoYuliHandoko (1) (3) , M. Joana Fernandes (1) (2) , Clara Lázaro (1) (2) (1) Faculdadede Ciências, Universidadedo Porto, Portugal (2) Centro Interdisciplinarde InvestigaçãoMarinhae Ambiental, Porto, Portugal (3) Department of GeomaticsEngineering, Institut TeknologiSepuluhNopember, Surabaya, Indonesia Email : ekoyh@geodesy.its.ac.id Introduction Indonesiaisthebiggestarchipelagiccountry,with17,608islandsandacoastline81,000kmlong(Fig.1).Dueto its unique characteristics, understanding sea level changes is crucial for Indonesia. Since there are significant risks associated with sea level rise in the Indonesia region, the continuous monitoring of sea level variability becomes urgent. By observing sea level continuously with satellite altimetry, its variability can be therefore assessed. The aim of this research is to evaluate, in the Indonesia region, the present WTC applied to altimeter data (those available in RADS and that computed using the GPD methodology) in view to obtain improved coastal altimetry datasetsfor estimating sealevel variation in this region. For this purpose, various data types are required. These include satellite altimetry data from RADS and Zenith Total Delay solution s from IGS (International GNSS Service) stations and all available additional GNSS stations inIndonesia. Fig. 2 shows the mean SLA time series data (period 1992 – 2013). computed from RADS data using the DTU10 meanseasurface(MSS)andallrequiredcorrectionsincluding:ERAInterimdryandwettroposphere,smoothed dual-frequency ionosphere , MOG2D dynamic atmospheric correction, CLS sea state bias and Fes2004 ocean and load tides. The trend and seasonal signal effects were derived using a Seasonal-Trend Decomposition ProcedurebasedonLoess(STL).Itcanbeobservedthatthisseriesissignificantlydifferentfromtheglobalone. Abstract Thisstudyaddressestheevaluationofvariouswettroposphericcorrection(WTC)intheIndonesiaregion. For this purpose sea level anomaly (SLA) analysis being performed for various missions. Here we present the results for the three reference missions: TOPEX/Poseidon, Jason-1 and Jason-2. The WTC datasets used are: the correction derived from the onboard microwave radiometer (MWR) of each mission, from ECMWF operational model (ECM) , from the ECMWF reanalysis model ERA Interim and from the GNSS- derived path delay (GPD) algorithm. All altimeter data and WTC used in this study, except for GPD, are thosepresentinthe RadarAltimeterDatabaseSystem(RADS). In addition, GNSS-derived Zenith Total Delays (ZTD) were derived for a set of coastal stations in the Indonesiaregion,forfutureuseinaregionalimprovedWTC usingtheGPDalgorithm. Conclusions and future work • For all missions, ERA Interim reduces the SLA variance when compared to ECMWF operational model. This isparticularly noticeablefor T/P. • For T/P, GPD reveals a significant improvement wit respect to the MWR correction present in RADS, particularly near the coast. • For J1, GPD reduces the SLA variance with respect to the MWR WTC, except for very small distances. However, for the shortest distances, this statistical analysis not significant due to the small number of available measurements. • For J2, present GPD correction does not show improvement with respect to the GDR-D MWR correction present in RADS. It should be noted that, while for T/P the MWR correction present in RADS does not have any type of coastal improvement, for J1 and J2, the MWR correction is already a coastal improved one (Brown, 2010). • The comparison between MWR and ERA reveals that MWR is significantly better than ERA for J2, slightly betterforJ1andworseforT/P. • Results show that ITS ZTD derived using Gypsy have a precision of 3-6 mm, being therefore suitable for useinthecomputationofaregionalGPDwettropospheric correction. • Future work includes the following two main topics: the develop an of an improved WTC for the IndonesiaregionbasedontheGPDapproach,usingGNSSlocalnetworksand,eventually,measurements from imaging MWR; the analysis of the space-time variations of sea level in Indonesia and comparison withtidegaugedata. Acknowledgements The authors would like to acknowledge the Radar Altimeter Database System (RADS) for providing the altimetricdata. This work was supported by a grant from Directorate General of Higher Education (DIKTI), Ministry of Education and Culture, Indonesia References • Brown, S. (2010) A Novel Near-Land Radiometer Wet Path-Delay Retrieval Algorithm: Application to the Jason-2/OSTM Advanced Microwave Radiometer. IEEE Trans. Geosci. Remote Sensing , 48, 1986- 1992. • Fernandes, M.J.; Lazaro, C.; Nunes, A.L.; Pires, N.; Bastos, L.; Mendes, V.B. (2010) GNSS-Derived Path Delay: An Approach to Compute the Wet Tropospheric Correction for Coastal Altimetry. Ieee GeoscienceandRemoteSensingLetters,7,596-600. • Fernandes, M.J.; Pires, N.; Lázaro, C.; Nunes, A.L. (2013) Tropospheric Delays from GNSS for Application inCoastal Altimetry.Advances in Space Research, 51(8). • Kouba, J. (2009) A guide to using International GNSS Service (IGS) Products. Geodetic Survey Division, NaturalResources Canada Zenith Total Delay (ZTD) IGS_vs_ITS IGS_vs_UPORTO No Station Mean Sigma Min Max Mean Sigma Min Max 1 COCO 2.6 3.1 -22.4 32.6 2.2 4.8 -40.0 47.5 2 CUSV 3.6 3.2 -47.0 46.6 - - - - 3 DARW 1.7 3.6 -28.1 48.0 - - - - 4 GUAM 1.5 4.9 -50.1 41.5 1.9 6.4 -55.1 68.2 5 KARR 0.2 3.1 -27.5 62.5 - - - - 6 NTUS 0.1 2.9 -52.3 37.5 - - - - 7 PIMO 2.2 6.1 -36.7 54.5 3.7 8.9 -47.0 53.0 8 XMIS -0.1 2.6 -22.6 36.1 - - - - • ZTD were derived from GNSS data for a set of 42 stations in Indonesia ( red points in Fig.1) , using the GIPSY software. These results, called ITS Solutions , were computed using the Precise Point Positioning (PPP) technique and the set of parameters summarised in Table 1. • ToanalysetheaccuracyoftheZTD ITSsolutions,thesehavebeencomparedwiththeZTDprovidedbyIGS (Kouba,2009)andthosecomputedatUPortousingtheGAMITsoftware(UPortoSolution,Fernandesetal. (2013)), for the common stations. IGS ITS UPORTO Software GIPSY GIPSY GAMIT Cutoffelevation angle 7 7 7 Mapping function used Niell and GMF VMF1 VMF1 Atmospheric parameter Interval sampling 5 minutes 5 minutes 15 minutes T/P J1 J2 TREND This analysis four SLA datasets have been computed using the fDTU10 MSS and the following set of range and geophysical corrections: • SLA wet ERA = dry ERA + wet ERA + iono GIM + ssb CLS + tides FES-2004 + invbar MOG2D • SLA wet MWR -from SLA ref replacing wet ERA by wet MWR • SLA wet GPD -from SLA ref replacing wet ERA by wet GPD SLA – function of distance from coast Table 1. Parameters used in ZTD processing Table 2. Statisticparameters for IGS, ITS, and UPorto solutions (milimetres) Fig. 5 Differences IGS-ITS ZTD solutions and IGS -UPorto ZTD solutions (millimetres) for stations: (a) COCO (b) PIMO (c) GUAM (a) COCO (b) PIMO (c) GUAM Fig. 2 Time series ofmeanSLA in the Indonesia Region Fig. 1 IndonesiaRegion. The reddotsrepresentGNSS stations. Fig. 3 Microwave Radiometric flag -Topex Poseidon Cycle 362 JASON-1 Fig. 4 Differences of SLA variance function of distance fromcoast for TOPEX/Poseidon(top), Jason-1 (middle) and Jason-2 (bottom) JASON-2 TOPEX POSEIDON The analysis of SLA variance , derived from different datasets, can be used to study the coastal effects of the various corrections applied to altimeter data and to assess their quality. The evaluation of the various WTC is performed using SLA variance analysis function of distance from the coast. In the present study, the following WTC are evaluated: 1) from the onboard MWR (MWR); 2) from GPD (GPD, Fernandeset al. (2010)); 3)from ECMWF operational model (ECM) ; 4) from ERA Interim model (ERA).