Written by Wen Wen
Introduction
Marine and coastal environmental are very important location that need to be monitored and analyzed. Remote sensing technology can be used to handle this problem. Landsat TM 7 data has so many functions to assist analysis in the spatial analysis, especially it can be used to derive marine and coastal information such as sea surface temperature, total suspended sediment, chlorophyll -a concentration and coral reef.
However, in this discussion only focused on sea surface temperature analysis in North Java Sea that located in North Banten Coastal District. From sea surface temperature map that derived from Landsat 7 TM data only can explains about the temperature of the actual surface of the ocean. GRASS (Geographical Resources Analysis Support System) free open source software is a raster/vector GIS combined with integrated image processing and data visualization subsystems also will be used in this discussion, for processing Landsat 7 TM data that can help marine and coastal analysis. GRASS software was under GNU GPL (General Public License), therefore it can be used as the free software with open source philosophy. GRASS GIS software also will be compared with image processing software such as Er Mapper, ERDAS Imagine, etc.
Objectives
-
To estimate and monitor sea surface temperature with band 6 Landsat TM 7 data using GRASS.
Data & Methods
Band 6 of Landsat 7 TM data will be used to derive sea surface temperature, the spatial resolution for this band was 60 meter and the acquisition date was 07 August 2001. The coverage area was in the path row 123 064.The following flowchart will display the processing steps in this discussion that how to derive the sea surface temperature map from Landsat 7 TM data.
Figure 1 Flowchart of Deriving Sea Surface Temperature
Result and Discussion
First at all, the whole area from Landsat coverage 185 km x 185 km from path row 123 64 was cropped into small study area with specific area was in north Banten coastal area. After cropping study area region then the next step is to create masking area to separate between land and water area using band 5 or band 4 as the specific band that can be used to mask land area in the Band 6. Land and water areas will be separated with medium digital value for land and water is 20, however, this number was not fix value for water and land are it will depends o different observation. In GRASS this process was done using r.mapcalculator command that can be used to create masking area based on IF..Condition command. The following figure will be showed how to create the masking area.
r.mapcalculator emap=tm72001.5_atcor fmap=tm72001.6 formula= if (E < 20, F, null()) outfile=tm72001.6_mask
r.mapcalculator emap=tm72001.5_atcor fmap=tm72001.6′formula=if(E<=20, F, null())’outline=tm72001.6_mask(command line to create masking area for band 6(thermal) using band 5)
After creating masking area then the next step is to convert the digital number values into radiances. Conversion to radiance was needed because digital number values can not be used directly to derive sea surface temperature map. Radiance value will be more accurately to show the emissivity / heat of ocean surface. Thus, the metadata file that contain the information of the band 6 (thermal) was required from this data the information of gain and offset value can be known. The gain value for this band 6 (thermal) is 0.066823529412 or equal to 17.040 / 254, but for this band 61 the offset value is 0 so the calculation is only DN band 6 with gain value.
r.mapcalculator fmap=tm72001.6_mask ‘formula=F*(17.040/254)’outline=tm72001.6_mask_rad (conversion from DN values into radiances)
Figure 2 Sea Surface Temperature Map
Based on the result of sea surface temperature map, the minimum temperature for this area was 5.2° C and for maximum temperature was 24.4° C. Landsat 7 TM data was acquired at 10.00 am local time therefore also can be said that sea temperature described the temperature at that time. For the spatial distribution the high temperature was distributed near coastal area, the value was high because the area is located in the coastal settlement area and also river sediment that was fall in this area can affect the sea temperature. Moreover, several factory wastes were throwing away into river and after that the waste of the factory accumulated in the coastal area. However, there is one area that looks little bit strange that shows like line in the coastal area with sea surface temperature around 14.8° C, in this case the haze and slightly clouds also affected this area then the result of the sea surface temperature become slightly different. In remote sensing, there were many obstacles can affected the remote sensing data but in this case atmospheric is one of important things that influence the sea surface temperature value.
Conclusion
North Banten Coastal Area especially in the coastal area or shoreline has high temperature value 24° C compare with another place. Coastal area temperature was high because of river sediment, waste disposal that has throw from factory area and distributed into coastal area. On the other hand low temperature value occurred in the small islands area.
More over, Sea surface temperature map can be derived directly from Landsat 7 TM data with different constant values for k1 and k2 with Landsat 5 TM data. GRASS as free open source software also can be used to handle this spatial analysis as well as remote sensing commercial packages.
References
Neteler, M and Mitasova H. 2005. Open Source GIS :A GRASS GIS APPROACH, second edition. Kluwer Academic Publisher.
Mather, Paul M. 2004. Computer processing of remotely-sensed images: an introduction. John Wiley and Sons Ltd. 3rd ed. 324 pp.
Robinson, Ian S. 2004. Measuring the ocean from space: the principles and methods of satellite oceanography. 669 pp.
This is good Juragan, I wish t you can also make this study in Nusa Penida
By: Andreas Muljadi on July 30, 2010
at 8:09 am