Page 194 - NGTU_paper_withoutVideo
P. 194

Modern Geomatics Technologies and Applications















                                          (b)                                                                                  (a)
          Fig.3. a) Instantaneous water level time series for pass154 and cycle7 with fitted linear model (red line) b) The fitted model to
                                 the SST time series for pass154 and the identified outlier (red point)

          5.  Results and Discussion
                      Considering that the aim of re-tracking is to overcome the corrupted waveform issues, the number and percentage
                 of multi-peak (distorted) waveforms analyzed over the study area, mentioned in table-2. As a result, the pass numbers
                 154, 595 and 709 showing the most distorted waveforms which they were selected for further process in this study
                 (Table(2)). The results and SST time series figures of  Level-2 data and  waveform re-tracking evaluation (process
                 mentioned in section 3) are presenting in this section.

                   Table 2 Multi-pick waveform analysis                Table 3 RMS (cm) of water level from
                                   Multi-  Percentage                   different re-trackers in Level-2 data
                   Pass#    Total
                                    pick     (%)                         Retracker    595   709   154

                 154         1263   580      45.92
                 382         1465   270      18.43                        Tracker     116   70    97
                 268         989    143      14.46                        Ocean       62    71    77
                 481         1494   171      11.44                        OCOG        66    67    62
                 595         1446   447      30.97                        Sea-ice     63    103   322
                 709        10571   4081     38.60                       Ice-sheet    63    116   64


          Table-3  shows  the  results  that  the  Ocean  and  OCOG  re-trackers  with  RMSE  62  and  67  cm  having  the  best  performance
          comparing with the other available re-trackers in level-2 data. Also mathematical of this re-trackers having maximum proportion
          with waveforms of study passes.

           Table 4 RMSE (cm) of water level from the threshold     Table 5 RMSE (cm) of water level from the threshold
           algorithm on the original waveform/first sub-waveform   algorithm on the mean waveform/Max Corr.
                           compared to TG                                waveform compared to TG
           Threshold     595        709          154               Threshold      595     709     154

           10%           48/47      44/40        77/76             10%            47/63   93/67   100/66
           20%           43/48      46/41        69/75             20%            40/34   50/44   103/67
           30%           44/48      43/40        70/74
           40%           43/49      47/39        68/73             30%            36/33   43/39   117/67
           50%           44/48      57/40        69/72             40%            34/33   44/40   115/67
           60%           45/48      66/48        69/73             50%            32/34   44/39   82/69
           70%                      5/48   80/48   70/74           60%            30/32   44/38   90/72
                         4
           80%           46/47      90/47        72/73             70%            28/29   44/38   90/73
           90%           48/47      90/47        76/71             80%            28/28   43/38   91/83
                                                                   90%            28/28   61/39   91/84




                                                                                                               7
   189   190   191   192   193   194   195   196   197   198   199