The big picture of marsh loss

Article metrics

A landscape-scale experiment shows that excessive nutrient levels can cause the loss of salt marshes — a result that was not seen in smaller studies. This illustrates the value of large-scale, long-term studies in ecology. See Letter p.388

The widespread use of fertilizers has greatly altered global nitrogen and phosphorus cycles, which are now dominated by anthropogenic rather than natural processes1,2. A substantial portion of fertilizers are ultimately transported into rivers and to the sea. One consequence of this increased nutrient supply is the proliferation of 'dead zones' at the mouths of the world's rivers — areas where the decomposition of fertilizer-driven algal blooms has used up most of the oxygen, leading to the death of fish and other large organisms3. Wetlands could provide a solution to this problem, because they chemically transform nutrients as they flow from land to sea. In particular, coastal salt marshes transform nitrate, the most abundant form of nitrogen in fresh waters, into nitrogen gas, thereby limiting the amount of nitrogen reaching the ocean4. But does providing this 'ecosystem service' to humanity have consequences for wetlands? On page 388 of this issue, Deegan et al.5 show that the price may be a high rate of global wetland loss.

Coastal wetlands are valuable. In addition to transforming nutrients, they support fisheries, protect coasts from storms and provide habitat for wildlife. To determine how wetlands may be affected by increased nutrient supplies, scientists have experimentally applied nutrients to wetlands in dozens of studies. Almost all of these studies used small plots, rarely larger than 1 square metre, and applied nutrient levels far higher than those that occur in even the most enriched habitats. These experiments have shown that nutrient addition causes a decrease in the mass ratio of plant roots to shoots, an increase in above-ground biomass, a change in plant species composition, and an increased vulnerability of plants to herbivores6. Whether these results can be scaled up to entire estuaries, however, has remained an open question.

Deegan et al. took a new approach to studying nutrient effects in estuaries. Working at the scale of entire tidal creeks (approximately 30,000 square metres), they added nutrients to the water flowing into a marsh on rising tides, replicating conditions typical of nutrient-enriched estuaries. They then compared the responses of the studied region with those of control creeks over nine years. Because Deegan and colleagues had the resources and time to conduct a large, long-term experiment, they were able to study all the microhabitats that occur in marshes. This broad scope paid off in an unexpected way: the authors observed striking changes to the stability and morphology of creek banks, a microhabitat that is usually not included in plot-based studies. Moreover, the geomorphological changes were not apparent until several years had passed, and so would probably have been overlooked by a standard research project.

Some of the authors' findings were expected on the basis of previous work: plants grew fewer roots because nutrients were easier to find, and the decomposition of organic matter in the soil increased because the added nutrients enabled bacteria to metabolize organic matter more easily. But other results transform our understanding of how nutrients affect coastal wetlands. In particular, Deegan et al. observed that the loss of roots and organic matter reduced the stability of creek-bank soils, leading to the collapse of creek banks and the consequent conversion of salt marsh into mudflat (Fig. 1). Although only the area within a few metres of creeks was affected, most salt marshes are highly dissected by a network of tidal creeks; the potentially affected habitat could thus represent a large proportion of overall marsh area.

Figure 1: Nutrients and salt-marsh loss.

Over the course of nine years, Deegan et al.5 added nutrients to the water at the mouth of tidal creeks at Plum Island Estuary, Massachusetts, when the tide was rising, so that the nutrients entered the marshes around the creeks. The marshes in this region are dominated by two grasses: Spartina patens on the marsh platform and Spartina alterniflora along creek banks. a, When nutrient supplies to estuaries are normal, marshes are dissected by narrow tidal creeks, the grasses produce substantial root growth and the soil contains lots of organic matter. b, By contrast, when nutrients were added, the authors observed that plants produced fewer roots, organic matter decomposed more rapidly and creek banks collapsed, leading to wider creeks and less vegetated marsh, with narrower bands of S. alterniflora.

Could increased nutrient availability be contributing to the loss of coastal wetlands worldwide? Deegan et al. worked at a relatively high-latitude salt marsh in the northeastern United States. To answer this question, their work needs to be replicated at other sites, particularly at lower-latitude salt marshes, which typically have soils with lower levels of organic matter7. Replicating such a large experiment would be a daunting task, but observational studies comparing marshes in high- and low-nutrient environments might also identify effects on creek morphology. Indeed, Deegan et al. have taken a first step along these lines, by showing that the history of rapid marsh loss in Long Island Sound and Jamaica Bay (also both in the northeastern United States) is consistent with the timing of nutrient enrichment in these areas.

Because salt marshes are threatened by many factors worldwide8, including sea-level rise, human development of marshes and adjacent terrestrial habitats, and increased livestock grazing, creek-bank collapse is unlikely to be the only cause of marsh loss. It is likely, however, that nutrient-driven creek-bank collapse will interact with other drivers of marsh loss in unexpected ways. Future studies need to investigate these potential interactions, which may be key to predicting the future health of salt marshes. Moreover, further study is needed to determine whether the process of creek-bank collapse will continue indefinitely, or whether it is self-limiting and will stabilize once creeks equilibrate into a new morphology.

Perhaps the broadest lesson from Deegan and colleagues' work is that ecologists must use caution when extrapolating the results of experiments at modest scales to entire landscapes, because findings from small-scale experiments may be incomplete or even misleading. For example, a typical outcome of plot-based studies is that the composition of plant communities changes in response to increased nutrient levels9, but Deegan et al. found no evidence for such changes. In addition, as Deegan and colleagues discovered, large-scale experiments may uncover processes that have been totally overlooked by small-scale studies. Landscape-level studies are expensive and time-consuming, but they may be indispensable if we are to understand how environmental challenges affect the ecosystems that we depend on.


  1. 1

    Galloway, J. et al. Science 320, 889–892 (2008).

  2. 2

    Elser, J. & Bennett, E. Nature 478, 29–31 (2011).

  3. 3

    Diaz, R. J. & Rosenberg, R. Science 321, 926–929 (2008).

  4. 4

    Valiela, I. & Cole, M. L. Ecosystems 5, 92–102 (2002).

  5. 5

    Deegan, L. A. et al. Nature 490, 388–392 (2012).

  6. 6

    Bertness, M. D., Silliman, B. R. & Holdredge, C. in Human Impacts on Salt Marshes: A Global Perspective (eds Silliman, B. R., Bertness, M. D. & Grosholz, E. D.) Ch. 8 (Univ. California Press, 2009).

  7. 7

    McCall, B. D., & Pennings, S. C. Oecologia (2012).

  8. 8

    Silliman, B. R., Bertness, M. D. & Grosholz, E. D. (eds) Human Impacts on Salt Marshes: A Global Perspective (Univ. California Press, 2009).

  9. 9

    Pennings, S. C. et al. Oikos 110, 547–555 (2005).

Download references

Author information

Correspondence to Steven C. Pennings.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Pennings, S. The big picture of marsh loss. Nature 490, 352–353 (2012) doi:10.1038/490352a

Download citation

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.