Distribution of landforms and buried sedimentary deposits during the growth of the Aceh River delta (Sumatra, Indonesia)

ABSTRACT

Fluvial and coastal landforms are indicative of landscape river delta evolution over time and provide clues for understanding coastal adjustments to sea-level and fluvial dynamics fluctuations, tectonic displacements, and extreme waves. We have mapped the surface and sub-surface footprints of fluvial and coastal geomorphological features in the Aceh River delta, northern Sumatra, using imagery dataset, vertical facies logging and helicopter electromagnetic surveys. The result is a geomorphological map at the scale of 1:75.000 which outlines the main features of the deltaic plain, including rivers, tidal and buried channels, fluvial levees, beach ridges, swales, tidal flats and lagoons. We compare their spatial distribution to the geometry of buried sediment bodies, revealed by boreholes and resistivity maps. Buried channel belts and floodplain deposits document the former locations of the distributary channels of the Aceh River. Coastal-parallel beach ridges evidence 7–8 km of asymmetric delta progradation since the mid-Holocene sea-level high stand.

KEYWORDS:

• Aceh River delta
• geomorphological map
• remote sensing
• stratigraphy
• Helicopter ElectroMagnetic (HEM) data
• Sumatra

1. Introduction

In the context of the current worldwide increase in population along coastlines and of rapid climate change, reconstructing the past evolution of deltaic landscapes helps predicting coastal adjustments to the current phase of sea-level rise (Ericson et al., 2006; Fanget et al., 2016; Stanley & Warne, 1994), to fluvial dynamic alterations (Syvitski et al., 2005), to tectonics (Higgins, 2016; Mathers & Zalasiewicz, 1999), and to extreme storm events (May et al., 2018, 2015; Aagaard et al., 2012). The spatio-temporal mapping of deltaic landforms provides insights into delta dynamics and can reveal their main controlling parameters (Aiello et al., 2013; Tarragoni et al., 2011). It also evidences the long-term evolutionary trajectories of fluvial-coastal systems (Ainsworth et al., 2011; Rossetti et al., 2015; Giacomelli et al., 2018; Mathers & Zalasiewicz, 1999; Rao et al., 2015; Rossetti et al., 2015). So far, the main targets have been the major deltaic systems (Amorosi & Milli, 2001; Anthony et al., 2014; Caldwell et al., 2019; Ericson et al., 2006; Korus & Fielding, 2015; Stanley & Warne, 1994), because they echo to broader research interest and have global environmental significance, thus determining funding allocations. Delta studies in Indonesia have focused on historical and short-term evolutions (Arjasakusuma et al., 2021; Beselly et al., 2021; Chapkanski, Brocard, Lavigne, Tricot, et al., 2022; Malawani et al., 2018; Septiangga and Mutaqin, 2021). Less attention has been paid to their Holocene evolution, except in the Mahakam delta (Storms et al., 2005). The need for background information regarding the long-term dynamics of these deltas is pressing, as they are now being rapidly affected by intensive land management, and can also rapidly evolve in response to volcanic eruptions (Beselly et al., 2021; Kuenzi et al., 1979), earthquakes (Kelsey et al., 2015; Meltzner et al., 2010; Pre et al., 2012) and permanent vertical displacement of the land surface, ground shaking, ground liquefaction, and tsunami landing (Latief et al., 2000; Monecke et al., 2008; Rubin et al., 2017; Sieh et al., 2015; Wassmer et al., 2015). Here, we conducted the first step study of a smaller delta in Indonesia, which long-term dynamics remains little studied.

The Aceh River delta, in Sumatra, was, by far, the delta most severely affected by the 2004 Indian Ocean tsunami, one of the largest tsunami recorded in human history (Doocy et al., 2007; Levy & Gopalakrishnan, 2005). In this specific context, scientific studies have been conducted to understand its recent dynamics (decennial timescales), in order to address societal (Scawthorn et al., 2006), sanitary (Prasetiyawan et al., 2006; Redwood-Campbell & Riddez, 2006) and environmental (Chapkanski, Brocard, Lavigne, Meilianda, et al., 2022; McLeod et al., 2010; Paris et al., 2009; Siemon et al., 2007) issues after the tsunami. Less attention has been paid to its long-term (centennial to millennial timescales) evolution, and how such evolution influences the current response of the delta to tsunami disturbances (Umitsu et al., 2007).

This study aims at delineating of the spatial distribution of various fluvial and coastal landforms in the Aceh River delta in order to: (i) provide a geomorphological map of the Aceh River delta at the scale of 1:75.000 (refer to the Main Map), (ii) provide a framework to reconstruct its evolution since the mid-Holocene sea-level high stand and (iii) identify the prevailing processes that control its geomorphological evolution and sedimentary architecture. We first provide a detailed mapping of the delta surface and then compare the spatial distribution of landforms to the geometry of buried sediment bodies revealed by resistivity maps and boreholes. Borehole data and resistivity maps are used to assess the correlation between sediment grain size and resistivity values. We finally briefly outline the Holocene evolution of the delta.

2. General context

The Aceh River delta is located five degrees north of the equator in the northernmost part of Sumatra Island (Indonesia, Figure 1(A)). It lies between 0 and ∼+20 m above sea level (Figure 1(B)). Based on the Ulee Lhue tide station (https://srgi.big.go.id), tides along the delta coasts are semi-diurnal with minimum neap, mean and maximum Spring tide range of 0.4, 1.2 and 2.1 m, respectively. Waves along the shoreface track from north-west from October to March, and from north-east from April to September, inducing a long-shore drift toward the north-east and south-west, respectively (Diposaptono & Mano, 1998). The delta is fed mostly by the Aceh River in its central part, by small intermittent karst-sourced rivers along its southwest border, and by the minor Angan River and its tributaries along its northeastern border (Figure 1(C)). The Aceh River, which is partly diverted into the Alue Naga Floodway Canal a few kilometers from its mouth, is the major freshwater source of the delta, with a catchment of 1568 km² (Syvitski et al., 2014). It originates in the Barisan Mountains (Figure 1(B)), and then runs through a terraced valley for ∼130 km (Montagne, 1963) before reaching the delta plain ∼20 km from the sea. The hydrological regime of the Aceh River is influenced by equatorial convective and monsoonal rainfall, with mean and maximum flood discharge at the Banda Aceh gauging station (period of 9 years) of 176 m³/s and 1700 m³/s, respectively (Syvitski et al., 2014). Climate is tropical with annual temperatures ranging within 25–32°C, and high humidity (80–90%). Mean annual rainfall in the delta is ∼1600 mm/y but reaches up to 5000 mm/y in the headwaters (Barisan Mountains, Muis et al., 2016; Ploethner & Siemon, 2006). The driest season stretches from June to July, with less than 100 mm/month. The rainy season stretches from September to March, and peaks in November and December at 300 mm/month.

Figure 1. Location map of the Aceh fluvial-deltaic environment at different spatial scales. (A) The Sumatra Island, (B), the Aceh catchment and (C) the Aceh delta. Bathymetry, altitude, hydrological network and major toponyms are reported.

The Aceh River delta is situated at the northern tip of Sumatra (A). The delta elevation ranges from 0–15 m in the downstream of the Aceh River. The Seulawah Agam volcano is situated to the north. The Barisian mountains present the highest elevations and border the watershed to the west and south-west. Two later faults cross-cut the watershed (B). Beach ridges to the east show the higher elevations in the delta comparing to the western edges and along the coast. Banda Aceh is situated in the middle reaches where the Aceh River flows (C).

The Aceh River valley is cross-cut by two major left-lateral strike-slip faults, the Aceh fault in the southwest, and the Seulimeum fault in the northeast (Figure 1(B)). They represent the north-westernmost segments of the Great Sumatra fault (Fernández-Blanco et al., 2016; Genrich et al., 2000; Ito et al., 2012; Tabei et al., 2015). The Aceh River valley is bordered to the west by steep mountains made of massive Cretaceous limestones, and to the east by andesitic tuffs and flows, and to the south by Pliocene-Pleistocene fossiliferous tuffaceous sandstone and mixed lithic conglomerates (refer to the Main Map in Supplementary Material; Bennett et al., 1981; Culshaw et al., 1979). The Aceh River delta has filled the drowned lower reaches of this valley during the Holocene (Culshaw et al., 1979; Farr & Djaeni, 1975). The thickness of Quaternary sediment fill reaches 180 m in the center part of the Aceh delta (Culshaw et al., 1979).

3. Material and methods

Landforms and sub-surface sedimentary deposits of the Aceh River delta were mapped using a GIS-based approach (Figure 2). The dataset (Table 1) consists of (i) topographic and geological maps, (ii) aerial photographs, (iii) satellite images and (iv) digital elevation models (DEM), summarized, treated and georeferenced by Chapkanski, Brocard, Lavigne, Tricot, et al. (2022). In addition, this study integrates (v) bathymetric data, (vi) sub-surface vertical facies logging and (vii) helicopter electromagnetic data. To ensure the accurate overlay of each data layer, the entire dataset has been projected in WGS 84-UTM-Zone 46N. Uncertainties on georeferencing accuracy, digitizing and local hydrological context are provided in Chapkanski, Brocard, Lavigne, Tricot, et al. (2022).

Figure 2. Schematic demonstration of the methodological proceeding.

Historical maps, photographs, satellite images, digital elevation models and bathymetry are used for interpretation of land and sub-aquatic forms. These interpretations were used to provide the geomorphological map. Borehole and Resistivity data are used for the sub-surface interpretation. All the data is combined to provide evolution maps of the delta.

Table 1. Overview of the database used in the study.

Type of data		Year	Spatial resolution/scale	Data available from	Data reused from	Targeted features
Maps	Topographic map	1884	1:80 000	Online library (University of Leiden)	Chapkanski, Brocard, Lavigne, Tricot, et al., 2022	Rivers, tidal channels tidal flats, lagoons, beaches
	Topographic map	1896	1:40 000	National French Library (Gallica)
	Topographic map	1914	1:40 000	Online library (University of Leiden)
	Topographic map	1944	1:50 000	EOS database
	Topographic map	1978	1:50 000	TDMRC database
	Nautical chart map	1924	1: 100 000	Online library (University of Leiden)		Water depths
	Geological map	1981	1: 250 000	TDMRC database	Bennett et al., 1981	Geological units
Aerial photos	Panchromatic ortho-photographs	1996	pixel of 1,8 m	TDMRC database	Chapkanski, Brocard, Lavigne, Tricot, et al., 2022	Rivers
	Ortho-rectified photo mosaic	2005	pixel of 0,3 m	TDMRC database	Meilianda et al., 2010	Palaeochannels
Digital Elevation Models	High resolution Topographic data	2011	pixel of 2 m	EOS database	Chapkanski, Brocard, Lavigne, Tricot, et al., 2022	Beach ridges, swales, fluvial levees
	DEMNAS	2017	pixel of 7,5 m	INAGI
Satellite Imagery	KeyHole-7	1967	pixel of 2 m	USGS – Earthxplorer	Current study	Beach ridges, rivers, tidal channels, tidal flats, lagoons
	Landsat Legacy TM	1989	pixel of 28,5 m	USGS – Earthxplorer	Chapkanski, Brocard, Lavigne, Tricot, et al., 2022
	Landsat-7 ETM+	2001	pixel of 30 m	USGS – Earthxplorer
	SPOT-5	2004	pixel of 5 m	TDMRC database
	Sentinel-2	2019	pixel of 10 m	USGS – Earthxplorer
Bathymetry	Marine depth measurements	2007	0,5 m (vertical)	TDMRC database	Meilianda et al., 2010	Water depths
Stratigraphic data	Borehole data	1975, 1979, 1993	-	BGR database	Culshaw et al., 1979; Farr & Djaeni, 1975; Iwaco, 1993	Fluvial levees, palaeochannels, beach ridges, swells
	Borehole field missions	2018 and 2019	-	-	Current study
Geophysical data	Resistivity maps	2005	pixel of 100 m	BGR database	Siemon, Ploethner, et al., 2006; Siemon, Röttger, et al., 2006	Fluvial levees, beach ridges

EOS, Earth Observatory of Singapore; TDMRC, Tsunami Disaster Mitigation Research Center; INAGI, Indonesian Agency of Geospatial Information; USGS, U.S. Geological Survey; and BGR, Federal Institute for Geosciences and Natural Resources.

3.1. Imagery pretreatments and feature recognitions

Sentinel-2 Multi-Spectral images at 10 m resolution (Table 1) were processed to enhance the spatial contrasts in visible and near-infrared bands generated by lateral variations in soil moisture and hydric stress on vegetation (Huete, 2004; Lillesand et al., 2015). These variations help identifying buried paleochannels (Giacomelli et al., 2018). Combinations of red, green and blue (RGB normal color composite) and near-infrared, red and green (as RGB color-infrared composite) were chosen as they best reveal the landforms. Bathymetry was interpolated from sounding points (Table 1; Meilianda et al., 2010) by inverse distance weighted interpolation to generate bathymetric isobaths. Hillshading of the 2 m resolution DEM was used to improve the delineation of beach ridges and dunes across the delta plain. ArcGIS Hydrology tools and 8 m resolution DEM were used to generate the drainage network and delineate the watersheds of the rivers that feed the delta. Geomorphological features were manually drawn in GIS environment in order to avoid the misclassification of automatic procedures (Brandolini et al., 2020). The resulting shapefiles were assigned to the following geomorphological groups: fluvial, coastal, and marine (Figures 3 and 4), following an architectural classification approach (Ainsworth et al., 2011; Nanson et al., 2013; Vakarelov & Ainsworth, 2013). Sinuosity ratios of the palaeochannels were calculated following Malavoi and Bravard (2010) and compared to the sinuosity ratios of the sub-contemporary (1884–2019) Aceh River channels, as reported in Chapkanski, Brocard, Lavigne, Tricot, et al. (2022).

Figure 3. Simplified version of the geomorphological map of the Aceh River delta. Field observations and snapshot orientation are displayed (Figure 4). Borehole locations and associated topographic cross-sections (T1: SW – NE and T2: SE – NW) are shown (Figure 5).

Coastal-parallel beach ridge sets occur in the eastern delta while dense network of palaeochannels run WNW direction in the western delta. The coastal strip present lagoons, tidal flats and tidal channels. Multitude of topographic alignments evidences the cuspate shape of the delta since the early stages of the progradation. The topographic alignments evidence an asymmetric evolution of the delta. The transects from this figure run across the palaeochannels SW-NE and across the beach ridges SE-NW. The interpretation of the bathymetry shows lobate forms of the sub-aquatic sediment deposits seaward from the current Aceh river mouth.

Figure 4. Fieldwork observations. Locations of snapshots are shown in Figure 3. 1: White beach consisting of biogenic carbonate sands; 2: Paddy fields in the foreground and steep limestone hills in the background; 3: Dry paddy fields in former palaeochannel; 4: Tidal flats of the western delta; 5: Beach ridge and sand dunes in the eastern delta; 6: Lagoons and tidal flats downstream of the Angan River; 7: Pre-Holocene layered terraces; 8: Angan River.

Mosaic of photos. As an alternative solution to shorten the Text, the caption of the figure provides an explication for each photo.

3.2. Airborne geophysical data

A helicopter-borne electromagnetic (HEM) survey of the Aceh delta was flown during the summer (from the 23th of August to the 12th of September) of 2005 by the German Federal Institute for Geoscience and Natural Resources (BGR), as part of the German-Indonesian cooperation project HELP ACEH (HELicopter Project ACEH, Siemon, Röttger, et al., 2006). The project aimed finding suitable locations for drilling new freshwater wells after the Boxing Day tsunami, because electrical conductivity correlates positively with water mineralization, and therefore helps assessing the landward penetration of the coastal saline wedge (Ploethner & Siemon, 2006). Geophysical data processing is detailed in Siemon, Röttger, et al. (2006). Here, we used resistivity maps derived from 1D inversion models of these data at 3, 4, 5, 10, 15 and 20 m below ground level (BGL, Siemon, Ploethner, et al., 2006; Siemon et al., 2007; Siemon & Steuer, 2011). We use the resistivity maps to gain insight into the underground architecture of the delta, using the difference in resistivity between floodplains (relatively lower resistivity values of finer sediments) and sand bodies (relatively higher resistivity values of coarser sediments) such as river channels and beach ridges.

3.3. Sub-surface stratigraphy database

We dug 50 hand-auger boreholes as deep as 4 m into the Aceh plain in 2018 and 2019, targeting geomorphological elements such as palaeochannels, beach ridges, flood plains and swales, in order to provide a sub-surface validation of surface landforms. To log the boreholes, sediments were characterized in the field, based on sediment color, hand texture, hardness, the presence of oxidation, shells, diffuse organic matter, and plant remains (Figure 5). Stratigraphic information published in former surveys and reports (Culshaw et al., 1979; Farr & Djaeni, 1975; Iwaco, 1993) was collected, and standardized to be compared and combined with the 2018–2019 boreholes. The entire dataset comprises 890 stratigraphic units from 206 core logs, up to 20 m long (see locations on Figure 1(C)).

Figure 5. Topographic cross-sections and borehole stratigraphy. For better readability, boreholes are presented with independent scale frames from the topographic cross-sections. Resistivity values at 2, 4, 6 and 8 m below the ground level were extracted at specific core locations and reported along the boreholes.

Two topographic transects showing variations in the elevation between fluvial levees (sandy deposits) and inter-levee floodplain (silty-clayey deposits) as well as between ridge sets and swells, both sandy environments. The general elevation for the sand ridges decreases seaward. Resistivity values are higher for sands and lower for clayey-silts.

3.4. Correspondence between resistivity maps and borehole stratigraphic data

The resistivity at 3, 4, 5, 8, 10 and 20 m was extracted in GIS from the resistivity map at each borehole location. Borehole logs were then compared to the extracted resistivity values. Stratigraphic data relative to grain size were grouped into three classes: clay–silt, sand and clay-to-sand. Resistivity values are expected to vary with grain size (Siemon et al., 2007). Therefore, the two data matrices (resistivity values and grain size) were combined and subjected to descriptive statistics using the software IBM SPSS statistics 20.0 (Armonk, NY, USA). Minimum, first 1st and 3rd quartiles, median and mean values were calculated to quantify the spread of resistivity values in clay–silt, sand and clay to sand units at different depths below the ground surface (Figure 6). The results were used to evaluate down-depth resistivity attenuation and to select values used for the discretization of resistivity values on corresponding resistivity maps. Despite the low resolution of the maps and saltwater intrusions in the lower delta, sands in logs were found to consistently exhibit higher resistivity than silty-clays (Figure 6). We therefore assumed that resistant bodies are composed of coarser sediment and we used this property to map buried sandy fluvial levees and beach ridges hosting freshwater (Figure 5). Similar observations were made by Siemon et al. (2007) demonstrating that the current Aceh River channel spatially correlates with relatively high values in resistivity maps. That also coincides with observations in other survey areas in Germany where low resistivity values correlate with clay–silt deposits in former lakes and higher resistivity rates with sandy to gravel deposits (Siemon et al., 2020).

Figure 6. Boxplots of resistivity value distributions for specific textures and at different depths below the ground level.

Seven group of boxplots plotting the distributions of resistivity values by type of texture, extracted at the location of each borehole for each resistivity map depths. The mean values are progressively decrease with depth.

4. Results and discussion

4.1. General architecture of delta landforms

The lower delta is a ∼14-km wide coastal strip. It is made of tidal flats, lagoons, aquaculture ponds, and tidal channels (Figure 3; see the Geomorphological map in Supplementary Material). Upstream, the delta plain surfaces rise up to 20 m above sea level. West and south-west of the Aceh River, we found palaeochannel belts and associated natural levee deposits that run parallel to the west delta margin, and are separated from the margin by eroded terraces rising ∼ +30 m above the delta plain, and by small alluvial fans. Irregularly spaced beach and chenier ridges, sub-parallel to the current coastline, were identified, across the delta plain, as far as 8 km inland (Figure 3). East of the Aceh River in the upper delta, we identified pre-Holocene strandplain standing ∼+20 m above the Holocene delta plain and incised by a dense network of small valleys (Figure 3).

4.1.1. Fluvial channels and levees.

The Aceh River currently flows along the median axis of the delta (Figure 3). Its course is highly sinuous and the meanders have migrated rapidly during the twentieth Century (refer to the Main Map in Supplementary Material). The sinuosity ratio increased during twentieth century from 1.75 to 1.91 before dropping to 1.36 in 1996 (Chapkanski, Brocard, Lavigne, Tricot, et al., 2022) when extensive containment dykes were constructed, cutting several meanders (Diposaptono & Mano, 1998). Wide palaeochannels, already abandoned by the twentieth century, are found to the north-east of the Aceh River, between the Aceh River and the Alue Naga Canal (Figure 3). They were partly excavated and reused during the construction of the derivation canal. East of the Alue Naga Canal, coast-parallel paleochannels are found between the sand ridges. Their layout is reminiscent of present-day tidal channels. They connect to the Agan River, which collects water along eastern delta plain, at the toe of the volcanoclastic hills, and represents the only fluvial system east of the Aceh River that manages to cross the strand plain and reach the sea. We identified four belts of buried paleochannels and associated levees to the west of the Aceh River (Figure 3). They form a series of ribbons, parallel to the modern river, that stretch over the entire delta plain. The westernmost belt lies close to the delta west margin and is highly sinuous (sinuosity index of 1.68). The other three belts show similar sinuosity but tend to take a more northerly course in the downstream direction. The boreholes dug in these paleochannels expose similar stratigraphic superpositions (Figure 5-T1). Weathered silty-clay deposits fill the upper 2 m of the channels. They overlay sands showing a decreasing grain-size upward (Figure 5-T1). The levees are made of weathered silty-sand (Figure 5-T1). The surrounding floodplains consist of silty-clay layers that extend ∼3–5 m below the fluvial levees.

4.1.2. Beach ridges and swales

Prominent (ridgelines 10–15 m amsl) sandy beach ridges form a belt that extends from the present-day coastline to 3–6 km inland. They are separated by swales filled by silty sands (Figures 3 and 5-T2). Imbrication of beach ridges occur ∼3–4 km inland. The height of the beach ridges and swales decreases seaward (Figure 5-T2). A more localized, pronounced drop in height occurs ∼ 3 km from the shoreline, were the ridge height decreases to 3–5 m amsl. Seaward, swales then tend to be filled by shelly and clayey sandy tidal flats 0–5 m amsl (Figure 5-T2).

Long-shore drift along the Aceh delta coast is predominantly to the north-east, owing to the predominance of the waves tracking from the north-west (Diposaptono & Mano, 1998). Sediment drift generates coastal ridges that are concave seaward in the western delta and convex seaward in the eastern delta. Beach ridges between the Aceh River and Alue Naga mouths are truncated as a result of the erosion of a formerly more prominent river mouth which was surrounded by more cuspate shorelines, a trait typical of quickly prograding, wave-dominated deltas (Anthony, 2015). The delta lobe has since been smoothed by the coastal retrogradation (Chapkanski, Brocard, Lavigne, Tricot, et al., 2022).

4.2. Submarine delta

The most significant morphological feature of the Aceh River prodelta is an offlap break quasi parallel to the coastline at the transition between a relatively flat topset (∼0.3°) and a steeper foreset (∼2°). North of the contemporary river mouth, the offlap break lies ∼3 km from the coastline, at a depth of ∼26 m. In the eastern delta the slope break is located farther offshore, ∼6 km from the coastline. Concave lobate forms are observed offshore and may represent the remnants of former mouths of the Aceh River.

4.3. HEM results and interpretations

The distribution of resistivity at depth is reported on Figures 6, 7 and 8. Low resistivity values (blue) correspond to seawater-saturated sediments of the lower delta. Sand and coarser-grained sediment filled by freshwater are represented by orange to brown pixels. Low to medium resistivity (green pixels) corresponds to silty-clay or/and coarser sediment filled by brackish water (Siemon et al., 2007; Siemon & Steuer, 2011). The general decrease of resistivity values with depths is probably induced by higher groundwater mineralization at depth, or/and by biases introduced by the comparison of point information in the boreholes with volume information in the electromagnetic data.

Figure 7. Resistivity maps at 20, 15 and 10 m below ground level and associated interpretations of resistivity bodies. Texture in boreholes at each specific depth are shown.

At 20 m bgl, resistive bodies are rare and scattered along the deltaic plain. At 15 m bgl, resistive bodies become clustered and run to the north direction. At 10 m bgm, resistive bodies correlate with palaeochannels.

Figure 8. Resistivity maps at 5, 4 and 3 m below ground level and associated interpretations of resistivity bodies. Texture in boreholes at each specific depth is shown.

At 5 m bgl, resistive bodies are very well correlated with palaeochannels. Also, wide belt of resistive bodies runs parallel to the coastline and correlates with beach ridges. At 4 m bgl, the resistive bodies progress further in the seaward direction, probably because of the delta progradation. At 3 m bgl the configuration of resistive bodies remains similar excepting slight advancement around the current Aceh River mouth in the center of the delta.

From 20 to 15 m bgl, a high resistivity belt along the valley axis seems to be composed of sand and gravel deposited by the Aceh River in its upper delta. By contrast, in the middle delta, resistive patches are aligned crosswise to the river course and may correspond to the first beach ridges to have formed after the maximum flooding of the Aceh River valley.

From 10 to 5 m bgl, some resistive patches are aligned parallel to the valley margins. They are elongate and display some sinuosity consistent with paleochannels and levees of the Aceh River. Four major SE-NW-striking belts can be clearly seen. In the middle delta, a ∼3 km wide belt of high resistivity lies parallel to the coastline and coincides with the beach ridges of the eastern delta. This belt is composed of sands that evidence the sub-surface continuation of the eastern delta ridges towards the southwest, below urbanized areas of the delta. Compared to eastern delta ridges, western delta ridges appear more discontinuous, and interrupted by the former courses of the Aceh River across the western delta. Sporadically, coast-parallel unidentified ridges are preserved, allowing us to track the beach or chenier ridges across the western delta. Intervening conductive layers are made of former inter-ridge lagoons or tidal flats and are nowadays covered by flat silty-clay floodplains.

Starting at 5–4 m bgl and above, the paleo-channel belts become fragmented, indicating that they were no longer fed by the Aceh River. In the lower delta, lobate, resistive bodies may correspond to former delta lobes beyond the former river mouths.

4.4. Processes and evolution of the Aceh River delta

At the beginning of the Holocene, the sea level of the Andaman Sea (Scheffers et al., 2012; Tjia, 1996) and Malacca strait (Geyh et al., 1979) was about – 70 m below the current sea level. A marine high stand as high as + ∼ 3 m has been documented during the mid-Holocene ∼5500 years ago; it was followed by a gradual lowering to the current sea level. Based on the landward extent of the delta plain, it appears that this sea-level high stand led to a Holocene marine transgression that flooded the Aceh valley as far as 6 km inland (Culshaw et al., 1979). The resistive sandy bodies buried 20–15 m below ground may belong to a transgressive tract deposited before sea level stabilized, sheltering the flooded valley farther inland from swale, thus allowing the upper valley to be filled by lagoons, swamps, or tidal flats. The Aceh River then covered this area with channel belts, fluvial levees and silty-clay to peaty floodplain deposits, as evidenced by resistivity maps above 10 m bgl. As delta progradation proceeded, the Aceh River expanded over the central part of the upper delta, feeding a series of now-buried channels, which may represent successive courses of the Aceh River or synchronous distributary channels. The courses of the most deeply imaged sand belts match the course of the modern Aceh River and of the abandoned channels still visible in the topography.

The presence of abandoned meander belts in the western part of the delta (Figure 3) implies that the Aceh River has flown for some time across the western delta where ridges are interrupted. The downstream terminations of the meander belts most likely correspond to former mouths of the Aceh River, which occupied the estuary, east from Ulee Lheue where beach ridges show curved shape (Figures 1 and 3). Layouts of the Aceh and Angan River at different time-spans (refer to the Main Map in Supplementary Material), show that, since the end of nineteenth century, both rivers have maintained the same courses across the delta (Chapkanski, Brocard, Lavigne, Tricot, et al., 2022). Until the end of twentieth century, the mouth of the Aceh River progressively retreated before migrating westward at the beginning of 21th century, after the destruction of the up-drift ridge.

Coast-parallel beach ridges across a strandplain as large as 7–8 km require a substantial amount of delta progradation since the mid-Holocene. As beach ridges are formed from available sediments along the coast, dispersed by the swash of waves, and that inter-ridge swales lie within the intertidal zone, the height and morphology of these features may reflect sea-level change, as well as the succession of depositional and erosive phases (Otvos, 2000). Low, widely spaced beach ridges form under relatively high sediment influx and rapid vertical aggradation, by contrast to tall, closely spaced ridges, which tend to form under reduced sedimentation rates (Taylor & Stone, 1996). The observed decrease in ridge and swale height seaward may indicate a lowering of sea level (Taylor & Stone, 1996). Therefore, the first cluster (∼5–6 km from the coastline; Figure 5-T2) of high, wide and imbricate beach ridges may reflect a relatively slow rate of delta progradation. The progressive seaward decrease in the elevation of the intervening swales may track the post-high stands lowering of sea-level since the mid-Holocene. The lower-standing and more widely spaced character of the more recent beach ridge may result from high rate of sedimentation and relatively faster delta progradation. Accelerating delta progradation may also be responsible for the arrangement of the beach ridges located around the mouth of the Aceh River (Figure 3). Based on the truncated aspect of ridges and their curvature increasing seaward, the cuspate shape and asymmetric character of delta would have been more pronounced in the last visible stage of the delta progradation (Figure 3). Coastal retrogradation has prevailed at least since the late of nineteenth century (∼100 m of coastal retreat during the century prior to 2004 Indian Ocean Tsunami, and ∼120 m of land loss following the tsunami events; Chapkanski, Brocard, Lavigne, Tricot, et al., 2022).

5. Conclusion

Mapping of landforms and buried sediment deposits in the Holocene tract of the Aceh River delta was performed using historical maps, aerial photographs, satellite images, digital elevation models and airborne electromagnetic data. Particular attention was paid to the depiction of fluvial levees, palaeochannels, tidal channels, lagoons, tidal flat, swales and beach ridges such as to provide keys for deciphering the evolution of the fluvial-deltaic system. The morphology of fluvial levees and beach ridges were documented through topographic transects. The airborne electromagnetic data was correlated with extensive in-situ stratigraphic data in order to evaluate the down-depth attenuation of silty-clay and sand layers. The results were used to provide relevant discretization of resistivity maps at specific depth layers. The landform layouts revealed some degree of consistency with the underlying buried sediments structures. The results were compiled to produce a geomorphological map of the Holocene delta of the Aceh River. The results were interpreted, and we proposed a brief discussion concerning the lines of the delta evolution in relation to the sea-level trends. Research involving radiocarbon dates, sedimentological analyses and provenance tracing is currently in progress to establish robust chronological and morpho-sedimentary frame of the Holocene evolution of the Aceh River delta.

Software

Pretreatments, feature recognition and digitizing were conducted using ArcGis Software 10.3 (ESRI, California, USA). Figures and final maps were improved using Illustrator 23.0.3 (Adobe Systems Incorporated, California, USA). Descriptive statistics were conducted using the software IBM SPSS statistics 20.0 (Armonk, NY, USA).

Acknowledgements

The authors thank Cédric Bataille, Camille Tricot and Axel Heitz-Retamal for the help with the digitizing. We warmly thank Tomy Afrizal, Tarmizi, Fajarul Aulia, Rifqi Irvansyah, Junaidi Saman, Hestia Melani, Nila Kanti and Andreas Surya Adiwinata Suherman for field work logistical and technical assistance as well as Diah Novitasari from the French School for Far East Studies (EFEO) in Jakarta for the collaboration and all the help with the administrative procedure and research permit authorizations. The authors are grateful to the three reviewers (Mike Shand, Bachtiar Wahyu Mutaquin and Colin Woodroffe) for their conscientious review and feedback that helped to improve earlier versions of the manuscript. We express our gratitude to the journal editorial membership for the kind communication and careful editorial handling.

Disclosure statement

No potential conflict of interest was reported by the authors.

Data availability statement

All the data that support the findings of this study are available from the corresponding author upon reasonable request except the HEM data provided by the Federal Institute for Geosciences and Natural Resources (BGR).

Additional information

Funding

Permit to conduct fieldwork was granted by the Ministry of Research and Technology of the Republic of Indonesia in cooperation with the Universitas Syiah Kuala. This study was funded (i) by the French Ministry of Foreign Affairs, through a Partenariat Hubert Curien – Nusantara grants (attributed to F. Lavigne in 2017 and to J-Ph. Goiran in 2019), (ii) by the University of Paris 1 Pantheon-Sorbonne, through an International Mobility Grant awarded to S. Chapkanski and STRATI project, (iii) by the French Ministry of Higher Education, Research and Innovation, through the Institut Universitaire de France (IUF) attributed to F. Lavigne and by the Laboratory of Physical Geography (LGP). Additional financial support was provided by LabEx DynamiTe (ANR-11-LABX-0046), as a part of the ‘Investissements d’Avenir’ program.