DYNAMICS TOPOGRAPHY MONITORING IN PEATLAND USING THE LATEST DIGITAL TERRAIN MODEL

Julzarika A., Aditya T., Subaryono S., Harintaka H. (2022) DYNAMICS TOPOGRAPHY MONITORING IN PEATLAND USING THE LATEST DIGITAL TERRAIN MODEL, Journal of Applied Engineering Science, 20(1), 246 253, DOI:10.5937/ jaes0-31522 Trias Aditya Department of Geodetic Engineering, Universitas Gadjah Mada (UGM), Yogyakarta, Indonesia Subaryono Subaryono Department of Geodetic Engineering, Universitas Gadjah Mada (UGM), Yogyakarta, Indonesia


INTRODUCTION
The Peatland is a landscape composed of the imperfect decomposition of vegetation from waterlogged trees so that the conditions are anaerobic [1], [2]. The organic material continues to accumulate for a long time to form layers with a thickness of more than 50 cm [3], [4]. Peatlands are a type of wetlands, which are among the most valuable ecosystems on Earth, and are generally located between two major rivers [5]. They are critical for preserving global biodiversity, providing safe drinking water, minimizing flood risk, and addressing climate change as the largest natural terrestrial carbon store [2]. Peatlands in Indonesia are found in the lowlands and highlands. In general, lowland peat swamps are found in tidal swamps and humidifying swamps, located between two large rivers in the physiography or landform of the back of the river (back swamp), swamp behind the beach (swalle), humidification plains (closed basin), and coastal plain [6]. Currently, Peatland presents one of the main ecological problems in Indonesia. Various incidents such as land fires, mining, floods, land subsidence, and water quality have occurred on Indonesia's Peatland [7], [8]. One of Indonesia's most extensive peatlands is located in Pulang Pisau Regency, which has the largest peat dome and a depth of > 25 m [9]. Pulang Pisau is located in Central Kalimantan Province, see Figure 1. This district has an area of 8,997 km² [9]. The geological map of the area shows that the geological formations in Pulang Pisau Regency are composed of alluvium formations [10]. This location were formed during the Holocene era, along with fiery rock formations [11]. Alluvium formations are composed of clay, kaolinite, dust, sand, peat, crust, and loose chunks, which are river and swamp deposits [10], [5]. The volcanic rock formation comprises greenish-gray volcanic breccia with components consisting of andesite, basalt, and chert. These materials are associated with basalt [10]. The Peatland in Pulang Pisau Regency is experiencing several ecological problems, such as drought and land fires during the dry season and flooding during the rainy season [12]. Therefore, Pulang Pisau Regency's peatland dynamics require regular monitoring using remote sensing satellite technology. This technology was chosen because it can be applied to a large area at a lower cost [13], [14]. The dynamics of Peatland are in the form of land subsidence [15]. Land subsidence in Peatland has been monitored using the latest Digital Terrain Model (DTM) every 6 to 7 months, starting in January 2018. The latest DTM is extracted from the integration of the DTM master with displacement. First, DTM master results from Interferometric Synthetic Aperture Radar (InSAR) extraction of ALOS PALSAR-2 images, and then it converted from Digital Surface Model (DSM) to DTM. DTM master began using data with the acquisition in December 2017. The displacement parameter is obtained from the Differential SAR (DInSAR) of Sentinel-1 imagery [16], [17]. The advantage of this latest DTM is that it can be updated consistently using the Sentinel-1 image. Monitoring land subsidence in Peatland usually only focuses on vertical displacement [18], [19], [20]. We use the latest DTM because it can visualize the current topographical conditions. Therefore, the latest DTM can be used to solve DTM problems in Indonesia. Currently,

MATERIALS AND METHODS
During this research, peatland dynamics detection used a remote sensing approach to monitor the dynamic topography in Peatland using the latest DTM every 6 to 7 months. Then, a comparative analysis of changes in dynamics was carried out with profiling on peatlands.
The latest DTM was built using the ALOS PALSAR-2 and Sentinel-1 integration. ALOS PALSAR-2 is used to generate a DTM master. The vertical displacement is extracted using Sentinel-1. The DTM Master and the displacement used a similar height reference field. The height reference field of DTM Master and the displacement both used EGM 2008. The latest DTM was obtained from the integration of the DTM master with the displacement. The latest DTM is extracted by using three combination methods. The methods are InSAR, DSM to DSM master conversion, and displacement using DIn-SAR. DTM master is the result produced by InSAR ALOS PALSAR-2. The steps taken included data preparation, interferogram generation and adaptive filters, phase unwrapping, refinement, re-flattening, DSM extraction with phase to height, height error correction, and converting DSM to DTM [22], [23]. The ALOS PALSAR-2 data used was the raw level data (1.0). There were at least two data sets used. Both data sets functioned as master data and slave data [24]. This data preparation included focusing and Single Looks Complex (SLC) [25]. The SLC file is the primary data used for the InSAR stages [26]. The next step was interferogram generation and adaptive filtering. This stage aimed to determine the coherence value of master data and slave data used [27]. Goldstein filter was selected to display smoother processing results [28], [29]. The third stage was the phase of unwrapping. This stage aimed to reduce the noise that occurred in the interferogram and facilitate the image's interpretation. The method used in this phase unwrapping was minimum cost flow (MCF) [16]. The fourth stage was refinement and re-flattening. Around 100 tie points were made on both images simultaneously, with a minimum error of 3σ [30]. The fifth stage was DSM extraction with phase to height. However, the resulting DSM still included height errors, and it was necessary to carry out height error correction [31]. This correction aimed to eliminate anomalies in the DSM and the eight neighboring pixels with a tolerance of 2σ [32]. The next stage was DSM to DTM master conversion. The parameter used at this stage was a radius of 20 m with a slope angle of 200. The final result on the ALOS PALSAR-2 data was the DTM master. This DTM master was used as the reference DTM for the latest DTM extraction. Displacement extraction used DInSAR of Sentinel-1 data. Displacement was extracted every 6 to 7 months. In this study, the displacement periods were January 2018, August 2018, January 2019, July 2019, January 2020, and June 2020. Displacement extraction during one of these periods used a minimum of two Sentinel-1 images [27], [33], [34]. Both Sentinel-1 images used SLC level data [34]. The stages taken were DInSAR. These stages of this method were mostly similar to those of InSAR but included one different stage. It involved replacing the creation of DSM displacement with "phase to displacement" [17], [19], [27], [35], [36].
The displacement results during this process also need to be corrected for height errors [37], [38]. The next stage was the latest DTM extraction with the integration of the DTM master and displacement. The latest DTM produced was in January 2018, August 2018, January 2019, July 2019, January 2020, and June 2020. This condition is a dynamic condition that occurs in Peatland. The six combinations of the latest DTM will describe the condition of surface change in peatlands. Six of the latest DTM with a different period (rain and dry season). Dynamic topographic conditions can be determined by making a cross-section profile. The cross-section is made to visualize the profile of rivers and swamps in the lowland. This condition will visualize the dynamic topographic conditions based on six periods of the latest DTM. The subsidence and uplift conditions in rivers and swamps on peatlands are monitored based on the cross-section profile. The latest DTM that has been produced will require a vertical accuracy evaluation. The latest DTM was obtained from the combined DTM master and the latest vertical displacement. Vertical shifts occur in tectonic areas with dip-slip faults, while horizontal shifts occur in tectonic areas with strike-slip faults. Therefore, the latest DTM represents the latest topographic conditions.

RESULTS
The DTM Master obtained is the result of interferometric processing from ALOS PALSAR-2; see Figure 2. This area has an elevation between -3 and 50 m. Land subsidence generally occurs in areas with an elevation of <25 m. The spatial resolution of this DTM master is 5 m, with a vertical accuracy of 2-3 m. River basins and swamp areas have relatively flat landforms with elevation values of around -3 to 5 m. is the weakness of those methods. Therefore, there is significant subsidence; meanwhile, in June 2020, there was a dry season. In August 2018, there was an increase in Peatland. One of the causes for this was the supply of surface water in the study area to peatland and oil palm plantations. In January 2019, there was an increase in the surface of Peatland. This condition was due to an increase in surface water flow as the product of rainwater supply. During this month, there was high rainfall in the studied area. As a result, the displacement value that occurred was around -0.360 to 0.520 m. Therefore, the dominance of displacement in this period is positive. In July 2019, the displacement that occurred was -0.120 to 0.428 m. In general, during this period, the displacement that occurred was -0.120 to 0.2 m. In January 2020, the land subsidence that occurred was -0.72 to 1.20 m. During this period, land subsidence was dominated by a decline of -0.25 m. Areas that experience land subsidence is located around swamps, rivers, and oil palm plantations.

DISCUSSION
Based on the results obtained, it can be seen that the location of this study is predominantly experiencing land subsidence. The latest DTM with DEMNAS (oldest DTM) can be visualized the difference in vertical accuracy and the difference in topographic elevation values at the same point. In Figure 3, we can see a comparison between the latest DTM with DEMNAS (old DTM). The red color is the topographical height value in 2010, while the blue color was the topographical height value in June 2020. The difference between the two topographical heights is about 0.5 to 1.5 m. This difference indicates the occurrence of subsidence from 2010 to 2020. In some areas, there has been an uplift, but it is not of great value. In general, the peatland area is experiencing subsidence. This condition indicates that DEMNAS (old DTM) cannot be used for mapping applications in areas with high dynamics, such as peatlands and areas of high vertical deformation. The latest DTM can be used as a solution to the problem of DTM scarcity in Indonesia.

Comparison of Land Subsidence from the Latest DTM 2018-2020
Land subsidence extraction from the latest DTM can be monitored by checking the cross-section profile in the study area. A random cross-section profile check is car- In general, the study area experienced a decrease in land subsidence. Pulang Pisau Regency also experienced a decline, as did oil palm plantations, swamps, and peatlands. Areas that have experienced an increase in land subsidence are located near rivers and dense forests on peatlands. In the river area, there is sedimentation due to the dry season. From January 2018 to January 2020, information on land sub-sidence distribution on the peatlands was also collected. In this period, there was a decrease in vertical deformation of 0 to -0.72 m. The decline occurred in Pulang Pisau Regency, plantation land, and peatlands. The area that experienced an increase was the northeastern part of the Pulang Pisau Regency. One possibility for this area to experience an increase could be the area clearing oil palm land. When clearing oil palm land, flow is often carried out in the opening canals. This condition requires enormous amounts of water so that peatlands in plantation lands experience an increase in elevation due to water seeping into the peatlands. In the period from January 2018 to July 2019, this region experienced an increase in land subsidence. In general, the increase in land subsidence lies in the value range of 0 to 0.428 m. The area that is experiencing decline is located in the swampy part of the eastern city of Pulang Pisau. The range of values for the decline is from 0 to -0.12 m. From January 2018 to January 2019, in general, this region experienced an increase in land subsidence. However, only a small proportion of them experienced a decrease in land subsidence. The increase in land subsidence lies in the range of 0 to 0.52 m. It increased due to an increase in surface water flow during the period from February to January 2019. This increase can be seen in the information on land subsidence in the early January to August 2018 period. It is located in the range of 0 to 1.75 m. Although the land subsidence comparison can be made based on the latest DTM, comparisons can be made according to the period. Figure 4 compares land subsidence based on the latest DTM on Peatland from January 2018 to June 2020.

Vertical Accuracy Test of the Latest DTM in Peatland
Checking the vertical accuracy test on the latest DTM was carried out in two ways. The methods used were checking the river profile and checking the swamp in the lowland profile. This profile check was meant to determine the dynamics profile changes at each of the latest DTM periods.Profile checking was carried out on the shape of the river. The latest DTM displays the curvature of the river so that it resembles the real conditions. The shape of the river's curvature is checked on each river, water channel, and stream. Overall, the latest DTM can display rivers, waterways, and swamps in lowlands; see Figure 5. However, the latest DTM is biased for surface runoff and contouring in areas around rivers and lowlands. Checking this curvature still requires checking with Global Navigation Satellite System (GNSS)-leveling. This step was used to check the cross-section profile of the swamps in the lowlands area. It is located at an elevation of -3 to 5 m. These lowland areas are generally swamps located adjacent to rivers. This area is checked by making an extended profile. The elevation value was extracted at each latest DTM from January 2018, August 2018, January 2019, July 2019, January 2020, and June 2020. All the latest DTM were compared to their elevation values. It was done in order to obtain a dynamics visualization of the peatlands. The elevation in August 2018 had a higher elevation than other DTMs. The latest DTM in January had the lowest elevation. The rainy season and dry season affected the dynamics of these peatlands. Land subsidence is influenced by the amount of water contained in the peatlands. Weather anomalies that occur in Kalimantan also affect the dynamics of Peatland. As in August 2018, there should have been a dry season, but there was heavy rain around the image acquisition date, so that the land dynamics would be high elevation. Under average weather conditions, June, July, and August will experience land subsidence. As in the latest DTM in July 2019, this land experienced a dry season, and land fires occurred. As a result, almost all elevations on Peatland experienced land subsidence of 10 to 80 cm. Another land subsidence anomaly also occurred in January 2020. Since it rarely rains from December to January 2020, this causes drought on the peatlands. The drought caused land subsidence of 10 to 100 m. Checks for lowland areas were also carried out elsewhere. The location is far from the river. This region has an elevation of -0.5 to 4 m. The rainy season and dry season affect the land subsidence condition in the peatlands. Weather anomaly factors also affect the wetness of peatlands.
The highest land subsidence in this area occurred in January 2019.Another check was carried out in peatland areas, which already had a slight variation in topography between -0.5 and 5 m. The cross-section profile created was located between 2 low hills and had a low elevation of less than 20 m. In this region, land subsidence conditions are more diverse. At an elevation of -0.5 to 3 m, there are visible land dynamics and higher land subsidence conditions. In another case, at an elevation of 4 to 5 m, the land subsidence decreased. Changes in land subsidence every 6 to 7 months in the latest DTM are low value, and peatland dynamics are starting below. A vertical accuracy test is done by comparing the latest DTM with field data (GNSS-levelling). Based on

CONCLUSION
This research concludes that monitoring peatland dynamics would benefit from the use of the latest DTM. In the study area in Pulang Pisau Regency, peatland monitoring used six periods with the latest DTM. The periods monitored were January 2018, August 2018, January 2019, July 2019, January 2020, and June 2020. The dynamics of peatlands varied from -1.5 to 1.5 m. It was due to the stability of the soil and the water levels in the peatlands. The elevation in the latest DTM varies with land subsidence and uplift. Different conditions in each period were also influenced by the rainy and dry seasons. The latest DTM is extracted from the integration of InSAR ALOS PALSAR-2 with DInSAR Sentinel-1. The latest DTM has an RMSE(z) of 0.705 m on the field measurement. This vertical accuracy test uses 15 measurement points from GNSS-leveling. Based on the RMSE (z) obtained, the latest DTM vertical accuracy is 1.3818 at the 95% confidence level. The latest DTM has a spatial resolution of 5 m and can be used at a mapping scale of 1: 10,000-1: 20,000.

AUTHOR STATEMENT
We would like to submit this manuscript named "Dynamic topography monitoring in peatland using the latest Digital Terrain Model" for possible evaluation for publishing. We certify that the paper is the original work and has not been published before.