GNSS RTK Application to Determine Coastline Case Study at Northen Area of Sulawesi and Gorontalo

Article 12 of Indonesia Government Regulation No.4-year 2011 about Geospatial Information [5], coastline is an element of base map. Article 13 subsection 1 states that the boundary form between the land and the ocean were affected by the tides. In subsection 2, coastline is divided into three types, namely lowest, highest, and mean sea level tides. Studies about coastline is important to support many maritime activities such as to determine administrative area (such as province, etc.), determine the length of coastline, coastal and marine management, also inland and marine spatial planning. Due to the importance of coastline, data with good accuracy is needed. Coastline mapping technology has developed rapidly. Previous method of coastline mapping is to trace marine debris carried by wave to the coast. At present time, because of needs of coastline data with higher accuracy and in the bigger scales, the GNSS RTK technology is starting to be used. GNSS technology generated coordinates and elevation and use tides values as reference datum. After the above, Digital Elevation Model generated from those points are combined with bathymetry and land elevation data resulting three types of coastline in accordance with the Indonesia Government Regulation about Geospatial Information Article 13 Subsection 2.


Introduction
As a maritime nation, coastline determination in Indonesia is necessary to support many maritime activities such as to determine administrative area (such as province, etc.), determine the length of coastline, coastal and marine management, also inland and marine spatial planning.
Previously, coastline mapping is to trace marine debris carried by wave to the coast, using GPS handheld. As states on Indonesia Government Regulation No. 4-year 2011 about Geospatial Information [5], article 12 states that coastline is one of the elements of base map. Article 13 subsection 1 states that the boundary form between the land and the ocean were affected by the tides. In subsection 2, coastline is divided into three types, namely lowest, highest, and mean sea level coastline. After the regulation has been legalized, the GNSS handheld is no longer relevant to be used because the three types of coastline cannot be generated. Coastline determination has been described in many studies with different methods. Khomsin [6] used rectified SPOT 7 Imagery to determine coastline in high water level by interpreting the visual appearance of objects natural color composite as the firm boundaries between sea and land. in term of the advantage of using the GNSS technology, Gonclaves [3] described that GNSS-based method offers a quicker, all-weather, highly accurate and continuously updatable shoreline positional time series relevant for monitoring and management task. Mendonça [8] used kinematic GNSS for shoreline monitoring as it has advantages of being able to produce consistent data for temporal analysis. the result shows that for temporal analysis of five years, the shoreline constantly changes in different periods.
We organize the study as follows: section 2 explains about method used in the study, area, and data collection. Section 3 explains about the results from the study and analysis. The study is concluded in section 4.

Method
Data needed in the study are:  Tidal measurement data,  Bathymetry data,  Benchmark (BM) measurement data,  Levelling measurement from BM to tide station,  Coastline data tracking using GNSS RTK. Flowchart of the study can be seen on Fig. 1.

Study Area
The study area for this paper is located in the Northern area of Sulawesi Island, that included to North Sulawesi and Gorontalo Province of Indonesia, see Fig. 2. Area of study is along the coastline of the study area with approximately 750 kilometres in length. Data acquisition for this study retrieved in 2017.

Data Collection
Data collection for this study contains several activities. Those are tidal observation, levelling from BM to tide station, bathymetry survey, and coastline tracking. Fig.3 represented the process to determine coastline using GNSS RTK.

Tides and Levelling
Tidal observation needs to be done at least 30 days of observation. This observation can be done using AWLR (Automatic Water Level Recorder) with an observation interval every 5 minutes. The observation can be assisted with manual tide observation, which has observation interval every hour. Along with this observation, levelling process needs to be done to determine the value of highest, lowest, and mean sea level according to BM. For tracking measurement using RTK, the final result of tides values is used as a reference. In the survey area, 19 tides station are built to find out the characteristics of tides. Because of the device and environmental factor, the data which can be used only from 15 tides stations. The tides station installation will be depending on the characteristics of the area, therefore there are some terms and conditions refers to the national standard of Indonesia about tides station installation (SNI 7924:2013). Those are: must not be placed in estuary, must be placed in stable position which is not affected by direct wave but still have a direct access to the sea in both conditions (high and low), can be accessed in every weather and conditions, the depth value around that area is homogeny. The area of this study is using two types of tides station, port and temporary tide building, see Fig. 4.

Bathymetry Data
Depth values of survey area generated from bathymetry survey using single beam echosounder. Bathymetry survey focused on shallow area, which depth is 5 metres or less. The survey was done near to the coast, as near as it can be, and also on the high tide during the survey time.
The collected depth values were later referenced to defined datum, which in this study, lowest tide is used. This data is used as a complement for tracking data inshore to intertidal zone. Bathymetry data acquisition will start after lowest tide zone. Both of the data will be combined to form elevation profile for coastal area up to shallow area. See

GNSS RTK Tracking Data
For coastline tracking data acquisition using differential method of radio RTK. Every point must be bounded to Benchmark which already attached to tides station. Illustration of data tracking acquisition can be seen in     Tidal data processing has a purpose to get the value of highest, lowest, and mean sea level elevation. Tidal data processed by least square method with the help of UTide MATLAB program [2]. UTide processed tidal data to produce tidal harmonic constituents. Tidal harmonic constituents are needed to generate tidal prediction data in one year after time of tides measurement. Value of highest, lowest, and mean sea level elevation are derived from this tidal prediction data. The results of tidal processing data in 15 from 19 tides station at survey area can be seen in Table.1 The value of highest, lowest, and mean sea level elevation still referenced to each local tides station, and location tides station can be seen in Fig.9.
We also processed tides data from BIG tides station in Anggrek and Manado to compare with several tidal processing data results from tides station survey area. The results can be seen in Fig.8 We chose one value of highest, lowest, and mean sea level elevation from 15 results of tidal processed data as a reference to generate coastline vector from DEM interpolation. We did this because the coastline vector elevation must be in one value and seamless line. Lowest sea level elevation from tide station was chosen as a reference elevation for all data.

Bathymetry and RTK Tracking Data Processing
First, bathymetric data that obtained from bathymetric survey using Singlebeam Echosounder needs to be cleaned from spike or outlier data. Then, cleaned data needs barcheck correction and tidal correction to get final bathymetric data with good accuracy. To make sure that the data has a good accuracy is to compare bathymetric data from main sounding survey line and cross sounding survey line. In this research, we use Orde 1b IHO SP 44 standard [4] to check the data. The result is we got 95,02% confidence level that indicated the data are good. The last step is to reference bathymetric data to lowest sea level.
GNSS RTK tracking data consists of horizontal coordinates and height value that referenced to benchmark point. GNSS RTK tracking data precision level is under 1 meter in order to fulfil level of accuracy in large-scale mapping. GNSS RTK tracking data processing only reference height value of RTK tracking data to lowest sea level. The method is reference height value of benchmark point to lowest sea level using height difference from benchmark point to tidal station.

Generating DEM and Coastline
Bathymetric data and RTK tracking data point that referenced to lowest sea level are used to generating DEM that used to extract coastline data. In this paper, we use Natural Neighbor method in ArcGIS to create the DEM after considered distribution of the data that spread not evenly and with the least consume time to create. The Natural Neighbors interpolation method was introduced by Sibson [9]. The model works by finding the nearest subset of samples for a given cell without a measured value, and then applies weights to the samples based on the proportional area they occupy [9]. This method is well known for its ability to interpolate scattered and unevenly distributed data [7,10]. The result can be seen in   We extract coastline data from DEM with the value of highest, lowest, and mean sea level that we choose from tidal processing data that referenced to lowest sea level value. The result can be seen in

Conclusions
GNSS Real Time Kinematic (RTK) method can be used as a method to determine highest, lowest, and mean sea level coastline. In order to extract three coastline vector data needs DEM that created from RTK tracking point and bathymetric data that referenced to lowest sea level. We use Natural Neighbor method to creates DEM then to extract coastline data. This method create DEM with less time than Kriging method. But, in the next research, we will try to use Kriging method to create DEM and compared the result with Natural Neighbor method.
Advantage of this method is accuracy level of coastline data that can be used in large-scale mapping especially in 1:10.000. This is because the coastline data that generated from DEM that formed from RTK tracking data and bathymetric data were good and met the required criteria for large-scale mapping. To make sure, validation of coastline data still needed in the future. Another advantage is if the area measured not to large has a low budget operation compared to another method such as LiDAR or photogrammetry.
Disadvantage of this method is the time taken to complete the measurement of RTK tracking data will be long. In this paper, RTK tracking measurement needed three until four months to completed. Another shortcoming is we can measure RTK tracking data only in the sandy coast area. Coast which has cliff or mangrove characteristic cannot be measured. In this area, we need referenced topographic data that obtained from LiDAR or photogrammetry method.
However, GNSS RTK method to determine coastline is still applicable in most area in Indonesia and with the need of large-scale mapping of coastline data.