Academia.eduAcademia.edu

Figure 2 – uploaded by Anthony P Doulgeris

See full PDFdownloadDownload figure
The Noise Equivalent Sigma Zero (NESZ) is the background noise (noise floor) in the SAR system. The Due to the sea ice drift with time the HEM and the satellite data had to be corrected to overlap with one another (see section 4.3). In Figure 2 the two flights are shown after drift correction together with the relevant satellite images.

Figure 2 The Noise Equivalent Sigma Zero (NESZ) is the background noise (noise floor) in the SAR system. The Due to the sea ice drift with time the HEM and the satellite data had to be corrected to overlap with one another (see section 4.3). In Figure 2 the two flights are shown after drift correction together with the relevant satellite images.

Related Figures (20)

Figure 1. Helicopter-borne electromagnetic sounding campaign. The electromagnetic instrument is seer hanging underneath the helicopter. Photo: S. Gerland Figure 1. Helicopter-borne electromagnetic sounding campaign. The electromagnetic instrument is seen Previous studies, [e.g., Geldsetzer and Yackel, 2009; Dierking, 2010b; Fors et al.,2016a have studied the importance of different incidence angles for the sea ice classification. Geld. setzer and Yackel [2009] found that an incidence angle above 27° were favourable for sea ice discrimination and Dierking [2010b] further commented upon the importance of the noise floor. In this study, we have overlapping data with varying incidence angles ranging fror 18° to 37°, specifically nine TerraSAR-X (X-band), five Radarsat-2 (C-band) scenes and fou ALOS-2 (L-band) scenes. By using the ice thickness measurements and photographs we are able to provide explanations for the different responses of the different SAR wavelengths anc incidence angles to the types of sea ice within the overlapping satellite scenes.
Figure 1. Helicopter-borne electromagnetic sounding campaign. The electromagnetic instrument is seer hanging underneath the helicopter. Photo: S. Gerland Figure 1. Helicopter-borne electromagnetic sounding campaign. The electromagnetic instrument is seen Previous studies, [e.g., Geldsetzer and Yackel, 2009; Dierking, 2010b; Fors et al.,2016a have studied the importance of different incidence angles for the sea ice classification. Geld. setzer and Yackel [2009] found that an incidence angle above 27° were favourable for sea ice discrimination and Dierking [2010b] further commented upon the importance of the noise floor. In this study, we have overlapping data with varying incidence angles ranging fror 18° to 37°, specifically nine TerraSAR-X (X-band), five Radarsat-2 (C-band) scenes and fou ALOS-2 (L-band) scenes. By using the ice thickness measurements and photographs we are able to provide explanations for the different responses of the different SAR wavelengths anc incidence angles to the types of sea ice within the overlapping satellite scenes.
Figure 2. Study area in the Arctic Ocean north of Svalbard. The lower left figure show the sea ice con-
Figure 2. Study area in the Arctic Ocean north of Svalbard. The lower left figure show the sea ice con-
Figure 3. Sea ice plus snow thickness distribution (probability density functions (PDF)) along the HE flight line on 19 April (black) and 24 April (grey). Note that the flight tracks in their entirety have been included here and for both days only part of the tracks are overlapping the SAR scenes. ‘igure 3. Sea ice plus snow thickness distribution (probability density functions (PDF)) along the HEM form in the newly opened leads. Observations from the ship record that new ice was blown to the downwind side of the wider openings, leaving open water on the upwind side. By 24 April ice was forming on the remaining part of the leads and openings. As a consequence of this process some of the refrozen leads in the vicinity of R/V Lance had thicker ice on one side than on the other, e.g. young grey white ice (thickness 0.15-0.40 m) on one side and a mixture of young grey ice (thickness <0.15 m), nilas and open water on the other. The L-band scenes indicate a 14% sea ice cover of young ice on 23 April [Johansson et al., 2016].
Figure 3. Sea ice plus snow thickness distribution (probability density functions (PDF)) along the HE flight line on 19 April (black) and 24 April (grey). Note that the flight tracks in their entirety have been included here and for both days only part of the tracks are overlapping the SAR scenes. ‘igure 3. Sea ice plus snow thickness distribution (probability density functions (PDF)) along the HEM form in the newly opened leads. Observations from the ship record that new ice was blown to the downwind side of the wider openings, leaving open water on the upwind side. By 24 April ice was forming on the remaining part of the leads and openings. As a consequence of this process some of the refrozen leads in the vicinity of R/V Lance had thicker ice on one side than on the other, e.g. young grey white ice (thickness 0.15-0.40 m) on one side and a mixture of young grey ice (thickness <0.15 m), nilas and open water on the other. The L-band scenes indicate a 14% sea ice cover of young ice on 23 April [Johansson et al., 2016].
Table 1. Satellite data used
Table 1. Satellite data used
Figure 4. Meteorological data measured at 22 ma.s.l. aboard R/V Lance for 17 April until 25 April 2015. Downward-looking photographs were taken with a camera (GoPro model YHDC5170) every two seconds. The camera was mounted on the helicopter, which flew at 35-40 m above the ice surface while HEM measurements were collected. The photographs were used to identify the sea ice surface, e.g. snow cover, presence of frost flowers (FF), rafting or ridges; during the ice thickness processing and to support interpretation of our results.
Figure 4. Meteorological data measured at 22 ma.s.l. aboard R/V Lance for 17 April until 25 April 2015. Downward-looking photographs were taken with a camera (GoPro model YHDC5170) every two seconds. The camera was mounted on the helicopter, which flew at 35-40 m above the ice surface while HEM measurements were collected. The photographs were used to identify the sea ice surface, e.g. snow cover, presence of frost flowers (FF), rafting or ridges; during the ice thickness processing and to support interpretation of our results.
‘igure 5. Radar backscatter oy-1, scenes from 2015-04-23 for a) TerraSAR-X, b) Radarsat-2 and c) ALOS- Only two of the four flight lines for 24 April overlap with all three of the different satellite scenes, L3 and L4. The L3 flight line seen in Figure 5 is a north-north-east (-16°) bound flight line and the L4 flight line is a south west (222 °) bound flight line. 2. Figures 5d-5f are close-ups in of the scenes Figure 5a-5c. The yellow lines indicate 3 of the overlapping
‘igure 5. Radar backscatter oy-1, scenes from 2015-04-23 for a) TerraSAR-X, b) Radarsat-2 and c) ALOS- Only two of the four flight lines for 24 April overlap with all three of the different satellite scenes, L3 and L4. The L3 flight line seen in Figure 5 is a north-north-east (-16°) bound flight line and the L4 flight line is a south west (222 °) bound flight line. 2. Figures 5d-5f are close-ups in of the scenes Figure 5a-5c. The yellow lines indicate 3 of the overlapping
Figure 6. Radar backscatter intensity oj, values on 23 April 2015 along two of the HEM flight lines a) 5.1 SAR backscatter signal, HEM ice thickness and Surface Roughness L3 and b) L4 as shown in Figure 5. The ovr q for X-band (green), C-band (red) and L-band (blue) are shown
Figure 6. Radar backscatter intensity oj, values on 23 April 2015 along two of the HEM flight lines a) 5.1 SAR backscatter signal, HEM ice thickness and Surface Roughness L3 and b) L4 as shown in Figure 5. The ovr q for X-band (green), C-band (red) and L-band (blue) are shown
Figure 7. 9,1, values for the X-band scenes at 19 April 13:16 are plotted against the 19 April 13:18 C- band scenes o?,,, values. The two straight blue lines indicate the mean incidence angle noise floor for the respective scenes. The orange line represents a linear fit to the data. Figure 7. 07,1 values for the X-band scenes at 19 April 13:16 are plotted against the 19 April 13:18 C- When comparing the C- and X-band o9,,, values for the near coinciding 13:18 and 13:16 scenes from 19 April we observe a similar trend in the data (Figure 7). The orange line in Figure 7 show a linear fit for the data. The C-band data is at least 4 dB above the noise floor and hence can be said to contain an actual signal. It should be noted that the X-band scenes from 13:16 are outside nominal performance range of TerraSAR-X and that the noise floor calculations [Airbus, 2014] were optimised for dual-polarization data.
Figure 7. 9,1, values for the X-band scenes at 19 April 13:16 are plotted against the 19 April 13:18 C- band scenes o?,,, values. The two straight blue lines indicate the mean incidence angle noise floor for the respective scenes. The orange line represents a linear fit to the data. Figure 7. 07,1 values for the X-band scenes at 19 April 13:16 are plotted against the 19 April 13:18 C- When comparing the C- and X-band o9,,, values for the near coinciding 13:18 and 13:16 scenes from 19 April we observe a similar trend in the data (Figure 7). The orange line in Figure 7 show a linear fit for the data. The C-band data is at least 4 dB above the noise floor and hence can be said to contain an actual signal. It should be noted that the X-band scenes from 13:16 are outside nominal performance range of TerraSAR-X and that the noise floor calculations [Airbus, 2014] were optimised for dual-polarization data.
Figure 8. Sea ice roughness versus snow and ice thickness measurements for a) 19 April 2015 and b) 24 April 2015. The blue dots represent individual measurements whilst the background colour represents the measurement density. Figure 8. Sea ice roughness versus snow and ice thickness measurements for a) 19 April 2015 and b) 24
Figure 8. Sea ice roughness versus snow and ice thickness measurements for a) 19 April 2015 and b) 24 April 2015. The blue dots represent individual measurements whilst the background colour represents the measurement density. Figure 8. Sea ice roughness versus snow and ice thickness measurements for a) 19 April 2015 and b) 24
Figure 9. Details from lead A underneath the L3 HEM flight line. (a) shows the o°,;, and of,1, values for the X-band (green), the C-band (red) and the L-band (blue) scenes as well as the HEM thickness (black). On the x-axis the combination of o9,;,; and oy is denoted as o°,. (b) yoo values for the X-band (green), C-band (red) and L-band (blue) as well as the roughness measure (black). (c) photographs from the part of the scenes for L3. The vertical lines denoted with Roman numerals in Figure 9a indicate the location of the corresponding photographs. The red and white line seen in the photographs is the HEM instrument. Figure 9. Details from lead A underneath the L3 HEM flight line. (a) shows the o;;, and oy-y values
Figure 9. Details from lead A underneath the L3 HEM flight line. (a) shows the o°,;, and of,1, values for the X-band (green), the C-band (red) and the L-band (blue) scenes as well as the HEM thickness (black). On the x-axis the combination of o9,;,; and oy is denoted as o°,. (b) yoo values for the X-band (green), C-band (red) and L-band (blue) as well as the roughness measure (black). (c) photographs from the part of the scenes for L3. The vertical lines denoted with Roman numerals in Figure 9a indicate the location of the corresponding photographs. The red and white line seen in the photographs is the HEM instrument. Figure 9. Details from lead A underneath the L3 HEM flight line. (a) shows the o;;, and oy-y values
Figure 11. co versus sea ice thickness for lead A underneath (a) L3 and (b) L4.
Figure 11. co versus sea ice thickness for lead A underneath (a) L3 and (b) L4.
Figure 12. Details from lead B underneath the L3 flight line. (a) shows the fy, values as well as the HEM thickness and (b) the yco in dB as well as the roughness standard deviation. South of the wider lead B is a narrower lead, denoted as C in Figure 5. This lead is connected to lead B. Photographs of this lead show newly formed sea ice with some rafting but without frost flowers. The roughness values indicate areas of small scale deformation similar to those seen at lead A. The yoo values for this lead are approximately | dB for both the C- and L-band scenes.
Figure 12. Details from lead B underneath the L3 flight line. (a) shows the fy, values as well as the HEM thickness and (b) the yco in dB as well as the roughness standard deviation. South of the wider lead B is a narrower lead, denoted as C in Figure 5. This lead is connected to lead B. Photographs of this lead show newly formed sea ice with some rafting but without frost flowers. The roughness values indicate areas of small scale deformation similar to those seen at lead A. The yoo values for this lead are approximately | dB for both the C- and L-band scenes.
Figure 13. yco in GB for (a) the 14:51 X-band scene and (b) the 20:18 L-band scene on 19 April. The higt Yco areas are lead D. The red horizontal lines indicate the values used for the yco calculations. Figure 13. co in dB for (a) the 14:51 X-band scene and (b) the 20:18 L-band scene on 19 April. The high
Figure 13. yco in GB for (a) the 14:51 X-band scene and (b) the 20:18 L-band scene on 19 April. The higt Yco areas are lead D. The red horizontal lines indicate the values used for the yco calculations. Figure 13. co in dB for (a) the 14:51 X-band scene and (b) the 20:18 L-band scene on 19 April. The high
Figure 14. Roughness versus thickness for a subsection of the HEM flight line for 19 April 2015. The subsection features smooth FYI/MYI, deformed FYI/MYI as well as ridges. Figure 14. Roughness versus thickness for a subsection of the HEM flight line for 19 April 2015. The
Figure 14. Roughness versus thickness for a subsection of the HEM flight line for 19 April 2015. The subsection features smooth FYI/MYI, deformed FYI/MYI as well as ridges. Figure 14. Roughness versus thickness for a subsection of the HEM flight line for 19 April 2015. The
‘igure 15. Histograms showing the number of data points (N) within each of the following sea ice classes,
‘igure 15. Histograms showing the number of data points (N) within each of the following sea ice classes,
Figure 16. Classification results for (a) X-band, (b) C-band and (c) L-band on 23 April 2015. The black line indicates the flight tracks. Zoomed in classification results for (d) X-band, (e) C-band and (f) L-band on 23 April 2015. Figure 16. Classification results for (a) X-band, (b) C-band and (c) L-band on 23 April 2015. The black
Figure 16. Classification results for (a) X-band, (b) C-band and (c) L-band on 23 April 2015. The black line indicates the flight tracks. Zoomed in classification results for (d) X-band, (e) C-band and (f) L-band on 23 April 2015. Figure 16. Classification results for (a) X-band, (b) C-band and (c) L-band on 23 April 2015. The black
Figure 17. The of, vs. o%;, values for each of the different sea ice classes seen underneath the L3 HEM flight line 23 April 2015. The incidence angle (IA) is denoted for the different frequencies. (a) is the X-band data, (b) the C-band data and (c) the L-band data.
Figure 17. The of, vs. o%;, values for each of the different sea ice classes seen underneath the L3 HEM flight line 23 April 2015. The incidence angle (IA) is denoted for the different frequencies. (a) is the X-band data, (b) the C-band data and (c) the L-band data.
* The type of sea ice is given as it is denoted in the respective articles. For the study by Onstott [1992] the ice type denoted by a * corresponds to the o%,, values, and the unmarked ice t Table 4. Newly formed sea ice observations from previous studies®.
* The type of sea ice is given as it is denoted in the respective articles. For the study by Onstott [1992] the ice type denoted by a * corresponds to the o%,, values, and the unmarked ice t Table 4. Newly formed sea ice observations from previous studies®.
Related topics:
GeologyRemote Sensing
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved