Document Type : Original Manuscript
Authors
Department of Natural Geography, Faculty of Literature and Human Sciences, University of Mohaghegh Ardabili, Ardabil, Iran.
Abstract
Abstract
Remote sensing data has played an important role in natural resource management studies in recent years. These data, especially in water resources studies and research, have many uses. Among water-related studies, the use of water indexes in recent years has been widely considered. These indexes have grown and developed with the advancement and production of satellite images and their precision increased dramatically. In this research, Landsat 8 and Sentinel A2 satellite images were used on the coast of Bushehr on the Persian Gulf. 8 water indexes were selected and executed on images. Despite the fact to exist two classes of water and land unsupervised classification were applied to images Finally, the overall accuracy and kappa coefficient values range from 77.0% to 99.6% and 0.55 to 0.99 respectively. For Landsat images, the Modified Normalized Difference Water Index (MNDWI) and the Normalized Difference Pond Index (NDPI) were the best indexes. Water Ratio Index (WRI) and Normalized Difference Turbidity Index (NDTI) were recognized as the worst index. For Sentinel 2A images, the Modified Normalized Difference Water Index (MNDWI) and the Normalized Water Difference Index (NDWI), respectively, were the best. and the Automatic Water Extraction Index (AWEI_NSH) had the worst result. In general, the performance of the water indexes, and the accuracy level of the sentinel 2A images was significantly higher than the Landsat 8 images This factor can be due to the higher spatial resolution of Sentinel images. For both Landsat 8 and Sentinel A2 images the Modified Normalized Difference Water Index (MNDWI) has the best results.
Keywords: Sentinel 2A, landsat8, NDWI Index, MNDWI Index, NDPI Index
INTRODUCTION
More than 70% of the earth's surface is covered by water, then the use of remote sensing data to extract information from oceans, seas, and closed waters is very important (Alavipanah, 2004). Remote sensing data with different spatial, spectral, and temporal resolutions have provided a valuable resource for evaluating the water level and its changes in recent decades (Jawak et al., 2015). Extracting water from satellite images is more than two decades old. The use of satellite images for a general overview of phenomena and terrestrial resources recording the characteristics of phenomena by sensors and finally analyzing them in this field can help us a lot (Zarghami, 2011). activities such as checking water quality including salinity studies, checking suspended substances and sediments, checking watercolor, and chlorophyll level, and also quantitative studies of water sources are among the actions that can be done using remote sensing (Hashemi et al., 2018). Using such a technique to better control and manage the environment in advanced countries is considered a strategic technology (Mobasheri, 2014).
MATERIALS AND METHODS
The studied area is a 130-kilometer stretch from Bushehr Beach in Bushehr province. The area of the study area is about 182,650 hectares, which is located at 50° 45ʹ to 51° 6ʹ E, 28° 42ʹ to 29° 10ʹ N. The Landsat 8 satellite image of March 15, 2018, and July 17, 2001, as well as the SentinelA2 image of March 26, 2018, were used. Both images were taken with a short time interval. Landsat 8 images had a spatial resolution of 30 meters. In the Sentinel images, bands with a resolution of 10 meters were used, and for the use of other bands, the bands were unified using the Fusion operation. In this research, the 1:50000 topographic map of the Geographical Organization of the Armed Forces was also used. Also, soil maps from the Jihad Agriculture and Geology Organization with a scale of 1: 100,000 were obtained from the Mapping Organization and used. 8 indices were used: Normalized Difference Water Index (NDWI), Modified Normalized Difference Water Index (MNDWI), Normalized Difference pond index (NDPI), Normalized Difference Turbidity index (NDTI), Water Ratio Index (WRI), Normalized Difference Water Index (NDWI_A)), AWEI, AWEI_SH. To determine the correctness of the classified map, the control points prepared from the Google Heritage images were used along with 35 points taken using GPS. The Kappa coefficient was evaluated
RESULTS
By applying blue indices on both Landsat 8 and Sentinel A2 images, the blue areas were separated from the dry environment. In Landsat 8 images, the MNDWI index with an overall accuracy of 98% and a kappa coefficient of 0.97 had the best results, and the WRI index with an overall accuracy of 78% and a kappa coefficient of 0.55 presented the worst results. In Sentinel A2 images, the MNDWI index with an overall accuracy of 99% and a kappa coefficient of 0.99 had the best result, and the AWEI_NSH index with an accuracy of 77% and a kappa coefficient of 0.55 presented the worst result. For Landsat images, among the implemented indices, the modified normalized water difference index (MNDWI) and the normalized lake, wetland index (NDPI) were the best indices. Water ratio indices (WRI) and normalized turbidity difference index (NDTI) were recognized as the worst indices. For Sentinel A2 images, the modified normalized water difference index (MNDWI) and normalized water difference index (NDWI) had the best results, and the automatic water extraction index (AWEI_NSH) had the worst results. In general, in the implementation of blue indices, the level of accuracy and precision in the Cetinel A2 images was significantly higher than that of the Landsat 8 images.
CONCLUSION
In the present study, in general, the accuracy and precision of Sentinel A2 images were better than Landsat 8 in most of the indicators, and the higher spatial resolution of Sentinel images can be a reason for presenting better results. For both images, the modified normalized water difference index (MNDWI) was recognized as the best index that could distinguish water phenomena from other phenomena well; Although in the mentioned index, different results can be presented due to the combination of different bands. In both Landsat 8 and Sentinel A2 images, the automatic water extraction index (AWEI_NSH) presented the worst result in this index due to the high reflection of phenomena such as rocks, sand, the presence of shadows, and clouds with negative effects on the mentioned index. It is confused with water phenomena. On the other hand, this index showed that they provide poor results for revealing shallow waters (such as estuaries). In general, it can be said that the overall accuracy obtained in most of the applied indices shows reliable values, which is proof of the optimal choice of thresholds in these indices. Finally, the monitoring of the beaches of Bushehr from 2001 to 2018 shows that the beaches of Bushehr have undergone many changes during this period, and these changes were in the form of a regression of the coastline. The imbalance between the processes of erosion and sedimentation on this coast has caused the channel and estuary of most estuaries to shift and change in width.
References
Alavipanah, S.K., 2004. Application of remote sensing in the earth sciences (soil). University of Tehran Press. Tehran, Iran. (In Persian)
Hashemi, S.N., Akbarinasab, M. and Safarrad, T., 2018. The Detection of the Plume of the Arvand River Using Satellite Images. Hydrogeomorphology, 4(13), pp.147-164.
Jawak, S.D., Kulkarni, K. and Luis, A.J., 2015. A review on extraction of lakes from remotely sensed optical satellite data with a special focus on cryospheric lakes. Advances in Remote Sensing, 4(3), pp.196-213. DOI: 10.4236/ars.2015.43016
Mobasheri, M., 2014. Principles of physics in remote sensing and satellate technology. Khaje nasir university of technology publishers. Tehran, Iran. (In Persian).
Zarghami, M., 2011. Effective watershed management; case study of Urmia Lake, Iran. Lake and Reservoir Management, 27(1), pp.87-94. https://doi.org/10.1080/07438141.2010.541327.
Keywords
Main Subjects
)