1. Introduction
Jakarta is one of the largest coastal megacities in the world, with a total population of over 10 million inhabitants. Although it has been facing many urban development issues [
1], land subsidence appears to currently pose one of the biggest challenges to its long-term ability to sustain human settlement. Subsidence rates along the coast of Jakarta ranged from 9.5 to 21.5 cm/year for the 2007 to 2009 period [
2]. The current pace of subsidence in Jakarta appears to be the fastest amongst large Asian cities, including cities such as Tokyo, Bangkok, Hanoi, and Can Tho [
3,
4,
5,
6]. The soil in Jakarta’s coast consists primarily of thick unconsolidated layers of clay and silt, partially filled with sand, up to a depth of about 300 m [
7]. Land subsidence has been mainly caused by the formally prohibited practice of extracting groundwater for industrial usage, and has led to severe damage to infrastructure and buildings, increasing the extent of flooded areas [
8,
9]. As a result, Jakarta is considered to be one of the most vulnerable cities in the world to coastal floods [
10]. At present, large portions of the coastal area are barely above sea water (see
Figure 1a, showing how Sunda Kelapa port, where the level of the water is already higher than the ground behind the narrow kerb/dyke that protects the port area from getting flooded), or are already below seawater (the Pluit area is already protected by a dyke which surpasses the average height of an adult, about 1.8 m high, see
Figure 1b). If groundwater extraction continues at the current rate, Jakarta is likely to sink a further 5–6 m by 2100 [
11], which would require coastal dykes to be built around the entire coastal perimeter of the city or people to be relocated away from these areas. Thus, the National Capital Integrated Coastal Development (NCICD) [
7] has urged Jakarta’s policy makers to prioritize solving the issue of land-subsidence as quickly as possible. If effective countermeasures to mitigate subsidence are not undertaken, future flood events will inevitably cause greater damage to Jakarta than what can be expected at present, especially given the increase in future exposure due to rapid economic and population growth. The potential areas at risk of being flooded in Jakarta are expected to drastically increase in the next decades, with subsidence being responsible for 88% of this increase [
12]. However, the flood projection by Takagi et al. [
12] was derived through a simulation that neglects the presence of coastal dykes, essentially investigating the potential flood extent without taking into account mitigation countermeasures against coastal floods. This assumption may cause an overestimation of the possible future impacts of flooding, at a time when the government of Jakarta has upgraded coastal protection in the area into a national project, as described later. In order to predict more realistic coastal floods until the year 2050, the authors developed a coastal hydrodynamic model that focused on the city of Jakarta. The aim of this model is to examine the effectiveness and limitation of coastal dykes, in order to address the fundamental question of how long the dykes will work under scenarios of continuing rapid land subsidence.
3. Results
This section shows the results of the projections of the flooded area and floodwater volume caused by coastal floods up until the year 2050, which demonstrate how land subsidence will be the main reason why the impact of flooding is likely to become more severe in the future. The results of the questionnaire survey regarding the current situation of coastal floods in the target coastal community are also described.
3.1. Projection of the Extent of Coastal Floods in Jakarta by 2050
Figure 6 presents the projection of the extent of the inundation areas between 2015 and 2050 for the range of scenarios shown in
Figure 2. The present study assumed the worst possible high tide condition, essentially reproducing the abnormal tidal event in November 2007. Since this event represents the case of a coastal flood that actually took place, the simulation correctly estimates that some parts of the coastline can already potentially be flooded. However, the maximum extent of the possible area that can be flooded is limited to within a few hundred meters from the coastline. If the dyke is finished within the next decade, as planned by the NCICD, it is expected that coastal floods could be completely prevented by 2030.
After 2030, however, the potential extent of the flooding depends on the scenario considered. In particular, continuing land subsidence greatly influences the outcome of the simulation. If land subsidence is stopped by 2030, the risk of coastal floods will be virtually eliminated, as SLR alone may not cause a drastic increase in the coastal flood risk, at least not until 2050. However, if land subsidence continues after 2030, the effectiveness of a 1-m high dyke will be nullified within 10 years of the dyke’s completion. Particularly, the western coast (which experiences the fastest rate of subsidence in the city) will start to get inundated as early as 2040, with inundation areas gradually extending towards the east coast and central districts. By 2050, the extent of the flooded area with a 1-m dyke would become almost identical to the case where no dyke was present, because if the dyke is overflowed, water will rapidly propagate over land, almost as if there was no coastal protection.
A 3-m high dyke will obviously preserve its functionality for a longer period, though even in this case its effectiveness will probably be limited to just 10 or 15 years after its completion in 2030. However, it is worth noting that its performance would suddenly drop just before 2050, resulting in a rapid expansion of the potential flooded areas within a very short transition period. In spite of the fact that a 3-m high dyke seems to be sufficiently high for present needs, the dyke will inevitably sink if the ground underneath it continues to subside, and will likely end up being totally submerged before 2050.
Since a 5-m high dyke appears to be too high to be overflowed, as proven by
Figure 7, the flooded area will substantially increase until at least 2050, irrespective of the scenarios assumed, and, thus,
Figure 6 does not show the results for the 5-m dyke.
3.2. Floodwater Volume and Required Pump Facilities
Pluit Pump Station is the largest pump in Jakarta (situated adjacent to the sea, with a discharge capacity of about 50 m
3/s), followed by Cideng–Siantar Pump Station (40 m
3/s, [
29]). The total discharge capacity of the 17 pump facilities in Jakarta is estimated to be around 200 m
3/s. Among them, five stations are located within the potential coastal areas that can be flooded, as shown in
Figure 6. The present study, thus, assumes that these five pumps could work to discharge floodwater up to their total capacity of 77 m
3/s (see
Table 1).
Figure 7 shows the projections of floodwater volumes, which significantly vary depending on the year, scenario, and tidal conditions considered. The scenario that considers no dyke, continuing land subsidence, and SLR (top left in
Figure 7) indicates that the rate of increase in the potential flooded area would be accelerated in later years, demonstrating an exponential increase in the impact of flooding as time progresses.
Figure 7 also indicates the time required to pump out all the floodwater using the five pumping stations, which was simply estimated by dividing the total water volume by the discharge capacity of 77 m
3/s. Although tidal ranges in Jakarta are relatively small (less than 1.2 m), the total water volume that could enter the land as a consequence of a high tide would be considerably larger than those during mean or low tide. Therefore, floodwater volumes and water levels need to take into account the maximum possible tides, as the worst flood damage would obviously take place during such events.
Nevertheless,
Figure 7 also shows that the influence of the tidal fluctuations on the extent of the flooding will diminish as the coastal dykes become higher, as higher walls will trap more floodwater than a smaller wall when the water returns to the sea during the ebb tidal phase. This implies that a higher dyke is not necessarily better than a smaller dyke, and could sometimes cause a more persistent flood. For example, a 3-m dyke may work well to prevent coastal floods until 2045. However, afterwards the volume of floodwater that would have to be removed from behind the dykes would suddenly jump to about 60 million m
3, irrespective of the tidal phase considered.
Compared with the case where no dyke is present, the 3-m dyke appears to be more effective in reducing the maximum volume of floodwater. However, the results show that unless the pumping facilities are drastically expanded or upgraded, it may take as long as 9 days by 2050 to remove all the potential floodwater once a flooding event occurs. However, if land subsidence can be stopped by 2030, a 1-m dyke would effectively protect Jakarta from coastal floods, at least until 2050. Furthermore, in this case, it is probably not necessary to upgrade the present pumping systems, which appear adequate to handle the occasional coastal flood.
3.3. Current Situation of Coastal Floods Revealed through the Interview
Figure 8a indicates that 85.5% of respondents (
n = 200) have experienced some sort of damage caused by coastal floods in the course of their lives, with about 68% of them answering that their properties/houses suffered major damage or were completely destroyed (
Figure 8b). This is not surprising considering that the last major floods took place in 2007, as reported in Takagi et al. [
12,
28]. Essentially, this was an important benchmark event that could be used to calibrate the responses of those interviewed, giving the authors confidence in the results of the questionnaire. As a result of this 2007 event, few respondents appeared to have been spared the consequences of this event, with 14.5% of the respondents replying that they have not experienced any sort of damage, and 11.1% replying that there was
no damage due to coastal flooding. As these two answers should theoretically coincide, the slight difference points out that there is some confusion regarding flood type among local people, who may not necessarily distinguish the mechanism behind the floods.
From the results, it is clear that events where the seawater overtops the dykes and floods vulnerable communities adjacent to Jakarta bay are already taking place.
Figure 8c shows that, according to respondents, the frequency of coastal floods is between
once a week to
once every few years, indicating that the chance of inundation is not uniformly distributed even in this small community.
Regarding the most severe flood they had experienced during their lives, 95% of respondents replied that water was 1 m deep or more (
Figure 8d). As the authors had not originally foreseen that such deep inundation was taking place, the questionnaire could not accurately capture the ranges for higher inundations. Nevertheless, some respondents indicated that the flood waters reached up to the first floor of their houses, which would represent a maximum inundation of nearly 2 m.
Finally, respondents were also asked about what measures the government should implement to address the issue of flooding in the Pluit district (
Figure 9), with most of them indicating that the dyke should be strengthened, and about half of respondents also favoring an even bigger “seawall” in the Jakarta bay (as will be discussed later).
4. Discussion
As land subsidence continues, the extent of the area at risk of flooding is likely to quickly increase, and these events will start to affect not only coastal areas but also the city center, where the busiest financial and commercial districts in the country are concentrated. Land subsidence can eventually result in an adverse slope in the sewage systems that are connected to the sea, which will further help seawater come into urban areas. If water is not quickly removed after a flooding event, it can eventually cause many problems, including traffic jams, degradation of water quality and sanitary conditions, disease outbreaks (e.g., dengue fever, cholera), and electric blackouts, which can lead to significant economic and human risks. If heavy rain or fluvial floods coincide with the timing of an extensive coastal flood, the combined event could have catastrophic consequences to the Jakarta metropolitan area. To respond to the possibility of unprecedented floods taking place in the future, large numbers of pumping stations would need to be installed to remove the huge quantities of seawater that could accumulate behind the dykes.
The NCICD has urged policy makers and municipal officials to prioritize attempting to solve the current land-subsidence problem. Nevertheless, effective and tangible countermeasures against land subsidence have yet to be implemented, and would obviously require alternative provisions of clean water, in itself not easy given that this is a megacity with a population of over 10 million. Otherwise, the NCICD program has also proposed a comprehensive roadmap to eliminate coastal floods, which is composed of multiple phases over the next four decades. The first stage of the project, referred to as “phase A”, aims to strengthen current sea walls and build new dykes during the course of the next 10 years, costing about
$1.9 billion [
7].
As corroborated by the present study, a similar project in Jakarta would be expected to contribute to the reduction of flood damage, at least over the course of the next 10 to 20 years. The authors’ questionnaire survey confirmed that local people consider that the strengthening of the present dyke should be the first countermeasure to be implemented (
Figure 9).
However, the possibility of extensive floods affecting the region would resurface by the middle of the 21st century if land subsidence continues, potentially allowing seawater to reach several kilometers inland, effectively affecting downtown Jakarta. For instance, due to land subsidence, a 3-m dyke would completely lose its flood-protecting function by the year 2040.
It is obvious that the higher the dyke, the greater the flood mitigation functionality. Indeed, the results for a 5-m dyke demonstrate that it would be effective to almost completely protect coastal areas until 2050 (
Figure 7). However, as
Figure 10a illustrates, SLR and land subsidence will both increase the relative difference between the sea surface and the ground, which would create the risk of a
dyke-break induced tsunami taking place [
28]. Although analyzing the structural stability of the dykes is out of the scope of this study, the thin coastal dykes found in many communities of Jakarta could suddenly collapse due to material aging, the occurrence of high waves, SLR, land subsidence, landslides, a collision with ships, earthquakes, or other adverse mechanisms. Takagi et al. [
28] in the course of their field visits investigated the thin dyke (with the height of 1.8 m) that protects the Pluit area, and already found that there is substantial leakage of seawater through a concrete joint. As even such a relatively low dyke is not capable of completely holding back seawater, there are serious questions about what could happen during the lifetime of a 5-m high dyke that made use of similar construction techniques, which could potentially be structurally unstable and even less impermeable.
The NCICD [
7] has also proposed a second stage of the project, Phase B, which is collectively called the Giant Sea Wall and Great Garuda project. This project involves the construction of an outer seawall over 35 km long and land reclamation of 1.250 ha, which will attempt to protect Jakarta against coastal floods by creating a large water storage basin between the sea and the land. Many concerns have been voiced by local residents, activists, and scientists about the impacts of such a project [
30,
31], though, according to our questionnaire survey, over half of respondents support these massive national projects, rather than more reasonable solutions such as constructing higher buildings, planting mangroves, or raising the lands.
It is worth noting, however, that there are a number of other grave concerns relating to such an ambitious project. In the event that the great seawalls fail to shut out seawater or the project is suspended or postponed due to economic turmoil, engineering difficulties, environmental impact, or political decisions, and assuming that efforts to reduce land subsidence are not carried out, downtown Jakarta would eventually become submerged. This could result in environmental displacement [
32,
33], eventually forcing the relocation of hundreds of thousands, if not millions, of Jakarta’s citizens, unless alternative adaptation strategies could be attempted.
To make matters worse, contemplating such grandiose national plans could lower the awareness amongst people of the most pressing problem, which is clearly land subsidence. The construction of a dyke along the perimeter of the city could dilute these concerns and the necessity to implement measures to stop ground water extraction.
If the Phase A project is implemented, this will mean that Jakarta will follow the path of Tokyo, Japan, which experienced similar land subsidence over the course of the 20th century, resulting in the construction of large dykes and many pumping stations to protect areas such as Koto and Edogawa wards (currently 2–4 m under sea level, [
34]). However, the robustness of the dykes in Tokyo appears to be much more reliable than those found in Jakarta. If Jakarta was to prioritize hard structures, far more robust dykes should be built, similar, for example, to those protecting low-lying coasts in Tokyo (
Figure 10b).
5. Conclusions
The present study projected coastal floods around Jakarta until the year 2050 to address concerns over how long the dykes will remain effective under the rapid land subsidence being experienced in Jakarta. To do so, the authors developed a coastal hydrodynamic model that considers land subsidence, sea-level rise, and tidal range. The tidal constituents were adjusted to reproduce the observed abnormal tides that took place in November 2007 and which lead to extensive coastal flooding. The changing effectiveness of the dykes over the coming decades was investigated by assuming multiple coastal protection scenarios, including a long dyke being built along the coastline of Jakarta, with heights ranging from 1 m to 5 m. Two further scenarios were considered, one where land subsidence was stopped by 2030, and the other where it continued past this date. The simulations showed that the dyke can effectively prevent coastal floods in the early stage of its lifetime, though it would soon lose its effectiveness unless land subsidence is stopped. Once seawater overflows the dyke, extensive flooding would immediately occur, potentially reaching a few kilometers inland into downtown Jakarta. For example, a 3-m dyke is expected to be sufficiently high for present day conditions, but it would have entirely lost its flood-protecting function by the year 2040. On the other hand, even a short structure such as a 1-m dyke is expected to stop coastal floods if land subsidence is stopped. Hence, the present study demonstrates that any actions that attempts to stop land subsidence would be the most effective countermeasure to mitigate future coastal floods, and should clearly be prioritized by the national government.