Profiling risk and sustainability in coastal deltas of the world

See allHide authors and affiliations

Science  07 Aug 2015:
Vol. 349, Issue 6248, pp. 638-643
DOI: 10.1126/science.aab3574

Deltas are growing centers of risk

Population growth, urbanization, and rising sea levels are placing populations living in delta regions under increased risk. The future resiliency and potential for adaptation by these populations depend on a number of socioeconomic and geophysical factors. Tessler et al. examined 48 deltas from around the globe to assess changes in regional vulnerability (see the Perspective by Temmerman). Some deltas in countries with a high gross domestic product will be initially more resilient to these changes, because they can perform expensive maintenance on infrastructure. However, short-term policies will become unsustainable if unaccompanied by long-term investments in all delta regions.

Science, this issue p. 638; see also p. 588


Deltas are highly sensitive to increasing risks arising from local human activities, land subsidence, regional water management, global sea-level rise, and climate extremes. We quantified changing flood risk due to extreme events using an integrated set of global environmental, geophysical, and social indicators. Although risks are distributed across all levels of economic development, wealthy countries effectively limit their present-day threat by gross domestic product–enabled infrastructure and coastal defense investments. In an energy-constrained future, such protections will probably prove to be unsustainable, raising relative risks by four to eight times in the Mississippi and Rhine deltas and by one-and-a-half to four times in the Chao Phraya and Yangtze deltas. The current emphasis on short-term solutions for the world’s deltas will greatly constrain options for designing sustainable solutions in the long term.

Deltas present a quintessential challenge for humans amid global environmental change. Home to some of the world’s largest urban areas, deltas are also highly dynamic landforms shaped by fluvial and coastal flooding (13). Human activities at the local and regional scales can perturb the water and sedimentary dynamics necessary to maintain a delta’s integrity, increasing the rate of relative sea-level rise (RSLR, the combination of land subsidence and offshore sea-level rise) and increasing flood risk (4, 5).

Delta sediments naturally compact over time, requiring new sediment fluxes from the upstream river network and deposition on the delta surface to maintain land elevation (4). Upstream dams and reservoirs trap sediment (6), and soil conservation practices can reduce the mobilization of sediment (7). River channelization on deltas inhibits depositional processes (8), whereas urban construction and groundwater extraction can accelerate sediment compaction (9, 10). Land subsidence is compounded by rising sea levels and the changing intensity and distribution of extreme events related to climate change (11, 12). Policies aimed at reducing the apparent levels of risk often employ costly engineering solutions that may be inherently unsustainable (1315). A framework to enable comparative risk assessment for deltas across the globe that specifically accounts for the dual natural and anthropogenic forces shaping these systems is a necessary precursor for strategies to improve their long-term resilience (16).

We present a systematic global-scale assessment of the changing risk profiles of coastal deltas. Most manifestations of risk are the immediate consequences of extreme events (3), such as fluvial and coastal flooding, that act over short time scales. Overall risk, however, is conditioned over longer periods by anthropogenically modified geomorphic processes such as changes in sediment supply, deposition, and compaction—which increase land subsidence and RSLR—and by the socioeconomic capability to prepare for and mitigate exposure to hazardous conditions (17). Focusing on interdelta differences in risk, we used sufficiently mature global data sets that depict factors with well-documented impacts on delta conditions, together with established methods for spatial integration of qualitatively distinct data types (1820). We focused on 48 major coastal deltas across a wide range of climate, biome, and socioeconomic contexts (Fig. 1A), with an estimated current combined population of over 340 million (21) (see the supplementary materials). At less direct risk are an additional 140 million people living within 25 km of these deltas, who, together with 3.5 billion people in upstream catchment basins, produce additional human impacts. We defined risk (R), or expected loss, to a delta population as a product of hazard (H), exposure (E), and vulnerability (V): R = HEV (17, 22, 23). Hazard is the probability of a damage-producing event, defined as fluvial or coastal flooding. Exposure is the expected number of people exposed to hazardous conditions for a given event, and vulnerability is the harm or loss caused by the exposure (eq. S1).

Fig. 1 Risk trends for deltas worldwide.

(A) Map showing the 48 deltas included in this study. (B) Phase diagram of contemporary risk assessment results, showing the three component proxy indices used to estimate per-capita R'. Color density represents a delta’s overall risk trend. Quadrant III deltas have predominantly low R', whereas quadrant II deltas have high R'. (C) Estimates of the relative rate of change in risk, or risk trend, for each delta due to increasing exposure associated with RSLR. The Krishna and Ganges-Brahmaputra deltas, despite being only moderately susceptible to short-term hazardous events, are increasingly at risk because of high rates of RSLR and high socioeconomic vulnerability. Ganges-Brahmaputra is abbreviated to “Ganges” in some panels for brevity.

Previous direct estimates of H, E, and V have been carried out in a number of local and regional studies (2325) using high-resolution data sets that are not currently available at the global scale. Exposure data at the necessary scales exist for select deltas; however, for global-scale analysis, the rate of change in exposure, E', is a more tractable measure for interdelta comparison. By reducing the relative elevation of a delta, RSLR results in increased population exposure to a given hazard and thus increased expected loss. The rate of change in expected loss (R') due to anthropogenic RSLR is termed the risk trend: R' = HE'V (see the supplementary materials).

We estimated H, E', and V for each delta, using empirical indices derived from global data sets. The Hazardous Event Index (HEI) is a proxy for hazard, based on empirical indicators of the probability and intensity of delta flood events. E' was estimated using the Anthropogenic Conditioning Index (ACI), which is built from measures of long-term anthropogenic drivers of RSLR (fig. S1). Vulnerability, which is strongly dependent on socioeconomic conditions in the delta (17, 22), was estimated as a function of per-capita gross domestic product (GDP), aggregate GDP, and government effectiveness. High GDP provides the financial capacity to make vulnerability-reducing investments from household to delta scales, when effective governments are present to leverage aggregate wealth to reduce vulnerability. An index representing the absence of this capacity, the Investment Deficit Index (IDI), was used as a proxy for V. The risk component estimates derived from the ACI, HEI, and IDI indices were used to estimate the risk trend resulting from RSLR (eq. S7).

We mapped the 48 deltas into a risk space defined by each delta’s specific anthropogenic, geophysical, and socioeconomic characteristics (Fig. 1). These estimates were made in an index space, comparing delta systems with each other on a relative per-capita basis. In quadrant I, which contains deltas with low ACI and high HEI scores, the Limpopo delta stands out because of its high vulnerability associated with a lack of infrastructure investment capacity. Most of the high-R' deltas fell in quadrant II, including the Krishna delta, which had the highest R'. In quadrant IV, characterized by high ACI and low to moderate HEI scores, the Ganges-Brahmaputra delta had a high R' due to high vulnerability. In contrast, quadrant III contains deltas with low R' due to both low ACI and low HEI scores, including the high-latitude Yukon, Lena, and Mackenzie deltas.

The Mississippi, Rhine, and Tone deltas had ACI and HEI scores similar to those of the Brahmani and Godavari deltas (between 0.4 and 0.7), but they had far lower IDI scores (<0.2). Their resulting risk levels were much more stable, with risk trends among the lowest of all the deltas in the study. Although the results presented here are for per-capita risk, the Ganges-Brahmaputra system had, by far, the greatest rate of change in risk when aggregating across delta populations. The Ganges-Brahmaputra had the second highest R' on a per-capita basis, and at more than 100 million people, it has more than twice the population of the Nile, the second most populous delta.

A low IDI, indicating a high capacity for investment in risk-reducing technologies, is the primary reason that several wealthy, developed deltas today have relatively low risk trends. For instance, after catastrophic flooding of the Rhine delta in 1953, the Dutch Delta Works were constructed to reduce future flood risk, using a network of storm-surge barriers, dams, levees, and other engineered structures. A long history of land subsidence, however, has left parts of this delta 6 m below sea level (26). Modernization and improvement costs across the Netherlands are projected to reach €1 billion to €2 billion per annum over the next century (27). The long-term sustainability of this and similar risk-reducing investments elsewhere has been called into question because of their heavy reliance on external financial and energy subsidies (14).

To examine the sensitivity of delta risk to reduced infrastructure investment benefits, we considered a future scenario in which infrastructure costs have increased. For our analysis, we used an increase in energy prices as a likely reason for a rise in the cost of infrastructure, but factors such as relative increases in the costs of labor and material or rising interest rates would have similar impacts on infrastructure cost. We modeled this scenario by adjusting IDI scores: We reduced GDP indicator weights based on 100-year projections of energy price growth in excess of GDP (28) (see the supplementary materials), reflecting expectations of higher costs for a given level of risk reduction. Estimates of future vulnerability increased for all deltas under this scenario (Fig. 2) but were greatest for systems with high GDP. The Mississippi, Rhine, Han, Chao Phraya, and Yangtze deltas had the greatest increases in vulnerability under this scenario, although others were also strongly affected, including the Parana, Rhone, and Pearl deltas. Deltas in low-GDP regions, such as the Irrawaddy, Tana, and Fly, were the least sensitive to these potential future changes. These and other less economically developed deltas were instead more sensitive to future risk increases stemming from increased exposure to hazardous events.

Fig. 2 Estimated future change in exposure and vulnerability for selected deltas.

The vertical axis, showing one scenario of change in vulnerability, is the difference between the 100-year future Investment Deficit Index, IDIF, and the current index, IDI. The origin represents current conditions. Both axes indicate a change over time for the two risk components: Increases in exposure are taken as proportional to the current rate of change, and increases in vulnerability are based on the difference between current and future estimates. Future change in risk is either associated predominately with increases in exposure (red) or increases in vulnerability (orange). Several delta systems (e.g., Mississippi, Rhine, and Yangtze) are at risk both from RSLR, leading to increased exposure to flooding, and from reduced effectiveness of risk reduction strategies that may not be sustainable on the century scale.

A given increase in vulnerability will not affect risk trends in all deltas equally; rather, its effect will be related to each delta’s ACI and HEI scores. We started with each system’s uncompensated R' based on anthropogenic and geophysical considerations alone (Fig. 3A). When considering contemporary vulnerability estimates (Fig. 3B), the wealthy but otherwise at-risk deltas such as the Rhine, Mississippi, and Han benefited substantially from their capacity to sustain engineering and infrastructure investments. Low-GDP deltas that are not able to make risk-reduction investments moved to higher risk rankings. The largest increases in R' ranks occurred for the Limpopo and Irrawaddy deltas. Improved economic development and associated vulnerability reductions would result in the largest decreases in risk in these deltas, akin to a transition from Fig. 3B to Fig. 3A. In the future vulnerability scenario (Fig. 3C), where investment capacity less effectively reduces risk trends, R' ranks for high-GDP deltas reverted back toward expectations based on geophysical hazards and anthropogenic change alone. Although contemporary estimates of risk trends were highest in South Asia (Fig. 4A), future increases in R', relative to current estimates, were greatest in the Rhine, Mississippi, Han, Tone, Chao Phraya, and Yangtze deltas (Fig. 4B). These systems are highly stressed by anthropogenic activities and regularly contend with hazardous events, so future increases in vulnerability will have a disproportionately large risk impact relative to other deltas. Management strategies that address the drivers of RSLR, particularly sediment supply and deposition, will be a core determinant of long-term sustainability over the next century.

Fig. 3 Current and future investment impacts on risk-trend rankings.

(A) Contemporary risk trend when considering only the anthropogenic and geophysical setting of each delta. (B) When also considering relative vulnerability, which is low for deltas that can make risk-reducing investments, the overall contemporary risk trend changes, in many cases dramatically. The Mississippi and Rhine deltas show substantially reduced risk. (C) Current risk-reduction strategies become more expensive and less sustainable in a more energy-constrained future scenario (14). In the long term, deltas that today are protected by substantial compensating infrastructure are likely to see their risk profiles approach those in (A).

Fig. 4 Contemporary risk trend and future risk-trend growth.

(A) High contemporary values of R' are distributed globally among the highly populated deltas of Southeast Asia and deltas in developing parts of Africa and the Middle East. (B) Estimated 50-year growth in the future risk trend, R'F, relative to the current R'. The highest relative increases in R'F are in the Rhine, Mississippi, Han, Tone, and Chao Phraya deltas, all systems where current risk is reduced through investments enabled by high GDP and energy costs that are affordable relative to the future scenario.

Future changes in the intensity and distribution of hazardous events, which are highly uncertain at the local scale, are also an important driver of future risk trends. Broad evidence suggests that climate change is affecting tropical cyclone intensity and river flooding (11, 29), global sea-level rise is accelerating (12, 30), and local sea-level rise may be substantially different than the global mean in some coastal areas because of regional patterns of ocean heat uptake and glacial isostatic adjustment (30). Land subsidence, taken as constant over time in our study, is also likely to change as future global population growth occurs predominantly in urban areas (31), driving further anthropogenic change in already stressed deltas. Growth in population, urbanization, and economic development is leading to increased interest in expanding hydropower infrastructure, which is already proliferating across many river systems (32, 33). Expanding hydropower infrastructure will reduce sediment transport and the discharge capacity of river systems that are essential for nourishing deltas (6). Delta shorelines, for instance, are highly sensitive to the balance between sediment supply and absolute wave energy (4).

Future environmental, geophysical, and societal changes will reposition, in many cases considerably, most of the world’s deltas into a future space of elevated risk. Although potential geophysical changes require additional research at the regional and local scales, our study demonstrates that economic ability and decisions to deploy engineering solutions will be key factors in determining how sustainable deltas become in the long term. Investments that manage the drivers of RSLR, rather than its symptoms, will be necessary to sustain deltas. Although the time horizons are long, acting now is essential, given that rehabilitation will be difficult (if not impossible) to achieve once ground is lost to rising seas.

Supplementary Materials

Materials and Methods

Figs. S1 to S4

Tables S1 to S3

References (3452)


  1. Acknowledgments: Data are available as supplementary materials on Science Online. This work was supported by NASA (Land Cover/Land Use Change Program grant NNX12AD28G) and NSF (Belmont Forum Coastal Vulnerability awards 1343458 and 1342944, and Dynamics of Coupled Natural and Human Systems award 1115025). The authors report no conflicts of interest. The authors thank B. Fekete and P. Green for helpful comments on the manuscript.
View Abstract


Navigate This Article