Must New York City Have Its Own Katrina?

Relative sea level has been rising inexorably in New York City over the past 140 years at an average rate of $0.27\mathrm{m}$ $\left(10.7\mathrm{in.}\right)$ per century due to both geologic subsidence and the warming trend in the twentieth century. With global warming, it may rise by more than that in the next fifty years (NASA News Archive 2006). Soon forgotten is the fact that the city has already suffered coastal flooding in severe nor’easters in 1962, 1991, 1992, and 1993. In the December 11–12, 1992 storm, the water level at the Battery peaked at nearly $2.4\mathrm{m}$ $\left(8\mathrm{ft}\right)$ above mean sea level, flooding much of lower Manhattan and causing the almost complete shutdown of the New York metropolitan transportation system (Gornitz 2001).

The city is at risk from frequent nor’easters, like the 1992 storm, and infrequent—but inevitable—hurricanes. A 2002 evaluation by an insurance industry source ranked New York City second only to the Miami/Fort Lauderdale area in the potential total economic losses from a hurricane; New Orleans was rated fifth (“Top ten worst places” 2002). The Weather Channel has called New York one of the five most vulnerable cities in America, citing only Miami and New Orleans as more vulnerable to damage from a hurricane (Office of Rep. Anthony D. Weiner 2006).

The storm surge from a hurricane typically washes in to reach a peak within a few hours, then subsides, possibly to be followed by a second surge. In nor’easters, on the other hand, the surge may extend over several days with the risk of flooding at consecutive high tides. In the December 1992 storm that virtually closed down transportation in New York City, the hurricane barrier at nearby Stamford, Connecticut, operated for six consecutive high tides to protect the city (Wiegel 1993).

In a worst-case scenario postulated in a government study, the eye of a hurricane would strike the New Jersey coast, with the highest winds then striking New York City. The bridges surrounding Manhattan would quickly be closed. Windows would break and masonry would fall from the skyscrapers into the streets, forcing people to take shelter in the subways, which would then flood as the storm surge sweeps in, drowning perhaps thousands (U.S. Army Corps of Engineers et al. 1995).

New York City is presently planning to be flooded, not to prevent being flooded. An assessment of storm surge barriers is being deferred because they would not protect the entire coastline and would be costly (City of New York 2007). Attention is focused on evacuation and post-disaster recovery (New York City Office of Emergency Management 2007). The planned response to an oncoming hurricane is the evacuation of residents, with the number of evacuees variously estimated from one to 2.5 million people. Evacuees would be urged to use buses and trains running on their regular routes (New York State Assembly 2005). Unfortunately, a hurricane storm surge is likely to be preceded by torrential rains, and as we have seen more than once in the last few years, even a heavy rain can bring the subways to a halt (Neuman 2007). According to one evaluation, “people would stop evacuating simply because they were unable to evacuate” (Hurricane Evacuation Study Program 2005).

Looking at history—certainly the history of the past half-century—it is clear when it was that people have prepared for a flood: after the disaster. (Perhaps the last one to be prepared for a flood was Noah, and he was tipped off.) And how long do the vulnerable regions remain unprotected? Long after the disaster. The New England hurricane barriers protecting Providence, Rhode Island, New Bedford, Massachusetts, and Stamford, Connecticut, were completed in the early 1960s, more than twenty years after the 1938 hurricane (Morang 2007). The Thames River barrier below London, England, opened in 1984, thirty-one years after the deadly 1953 storm. The major barriers in Holland’s massive Delta Project that followed the 1953 storm were completed on the Eastern Scheldt in 1986, thirty-three years later, and most recently on the New Waterway to Rotterdam, in 1997, forty-four years later.

As coastal flooding becomes more frequent with global warming, a delay like these is intolerable. With the scientific attention that has been given to the implications of global warming, we now know what is coming. Must we again wait for the disaster to happen before we take action?

In a 2001 study of the effects of climate change on the New York metropolitan region, the NASA Goddard Institute of Space Studies projected a rise in sea level of 17.5 to $60\mathrm{cm}$ (6.9 to $23.7\mathrm{in.}$ ) by the 2050s, and 24 to $108\mathrm{cm}$ (9.5 to $42.5\mathrm{in.}$ ) by the 2080s. What is now the one-hundred-year flood [one chance in a hundred of rising $3\mathrm{m}$ $\left(9.7\mathrm{ft}\right)$ in any one year] might occur once in nineteen to sixty-eight years by the 2050s, and as often as every four years by the 2080s (Rosenzweig and Solecki 2001).

The Stony Brook Storm Surge (SBSS) model was developed to understand regional storm surge dynamics and as a first step to alert the city for increasing coastal flooding threats. It combines the MM5 meteorological model and the ADCIRC hydrodynamic model and applies them to the eastern seaboard from Maine to North Carolina. The MM5 model is routinely used at Stony Brook for daily sixty-hour forecasts of weather conditions, which are posted on the Internet (Colle et al. 2003; Jones et al. 2007). Applied retrospectively, it can be used to study past storms using archived weather data to calculate local winds and barometric pressure. These drive ADCIRC (Luettich et al. 1992), a two-dimensional hydrodynamic model, which then calculates tidal and local wind-driven water currents and sea levels. The geographic distribution of points at which the calculations are made consist of a grid of triangular elements, the size of which depends upon the local variation to be expected, with $5–8\mathrm{m}$ $(16–26\mathrm{ft})$ grid point spacing across narrow waterways up to a spacing of more than $70\mathrm{km}$ $\left(43\mathrm{mi}\right)$ in the open ocean. A seamless bathymetric-topographic database up to the 8-m (26-ft) contour line was prepared to allow SBSS to flood low-lying coastal regions (Bowman et al. 2004).

The model was validated by comparing sea level elevations calculated at a number of coastal stations with the historical NOAA sea level coastal records of two storms, the 1992 nor’easter and the 1999 Hurricane Floyd. The calculated rise and fall of sea level closely tracked the observed data to within $\pm 10$ percent over a period of days through the initial storm surge.

To study coastal flooding, the wind speeds in Floyd, which had been downgraded to an extratropical storm by the time it reached the New York region, were arbitrarily increased by 60 percent in the model, effectively doubling the wind force, to produce a higher surge able to flood coastal areas (Van Lenten 2005). The effectiveness of the barriers in protecting the inner city from Super Floyd is seen in Fig. 2, which shows the water levels inside and outside the three barriers and at the Battery, located at the foot of Manhattan.

But would floodwaters rise even higher where they are blocked just outside the barriers? If the East River barrier were placed in the constricted waterway downstream from its mouth on Long Island Sound, the surge would rise another 0.3 $\left(1\mathrm{ft}\right)$ in Super Floyd. At the Narrows and the Arthur Kill, which face the open ocean, the rise would be negligible.

Would the torrential rains that accompany these storms cause freshwater flooding inside the barriers while they are closed? To investigate this possibility, the runoff from Floyd was examined. Floyd set records for rainfall up and down the East Coast, including locations in New Jersey to the west of New York City and in the Hudson River watershed. The deviations from the normal rise and fall of the tide at the Battery are shown in Fig. 3. The storm surge peak accompanying the storm, followed by a deep trough, is clearly evident. A secondary peak, closely following the storm surge, can be attributed to the storm runoff from nearby rivers in northern New Jersey that empty into New York Harbor. This peak, about $0.3\mathrm{m}$ $\left(1\mathrm{ft}\right)$ high, can be compared to the normal tidal range of about $1.2–1.5\mathrm{m}$ $(4–5\mathrm{ft})$ at the Battery. Thus, if the gates were closed at or below midtide, the flash flood from New Jersey rivers would not flood New York City.

Runoff from the much larger Hudson River watershed followed Floyd’s storm surge by about 2.5 days, and lasted for another 2.5 days. The peak runoff from the Hudson River tributaries during this period was on the order of 3,000 $(\mathrm{100,000}{\mathrm{ft}}^3∕\mathrm{s})$ , much higher than the typical monthly Hudson discharge of $\mathrm{1,200}{\mathrm{m}}^3∕\mathrm{s}$ $(\mathrm{43,000}{\mathrm{ft}}^3∕\mathrm{s})$ in the peak month of April (Abood 1978). However, this is a small fraction of average tidal transport at the Narrows of $\mathrm{42,500}{\mathrm{m}}^3∕\mathrm{s}$ $(1.5\mathrm{million}{\mathrm{ft}}^3∕\mathrm{s})$ (Bowman 1994). Again, if the barriers were closed at mean tide or below, comparatively little freshwater would accumulate inside the barriers from the Hudson River runoff.

In sum, the meteorological and hydrodynamic studies of storm surge barriers to protect the heart of New York City from coastal flooding reached the following results:

To proceed with the development of storm surge barriers to protect New York City, preliminary engineering studies are needed now to establish the technical feasibility of the barriers. These tasks should include the following:Will the engineering community now provide the leadership that is needed to advance this development?

The Katrina disaster is seen as largely a systemic failure. Much of the destruction was the result of engineering and engineering-related policy failures, compounded by organizational and political failures. What failed is the process. What is to be learned is not to depend upon that process.

The engineering failures were complex and involved numerous decisions by many people within many organizations over a long period of time. The U.S. Army Corps of Engineers designed for a “standard project hurricane” with maximum wind speeds of $101–111\mathrm{mph}$ , much less than the “probable maximum hurricane” with speeds of $151–169\mathrm{mph}$ , defined by the National Weather Service. Strangely, the Corps was not allowed to consider water levels above “authorized levels” by federal Office of Management and the Budget and Congress, hardly hydrologic experts (Anonymous 2007).

Fifty levee breeches were due to overtopping or weakening of levees by the storm surge. Why was the city not protected from a storm surge? Initially, the Corps wanted to build a barrier to prevent water from the Gulf of Mexico from reaching Lake Pontchartrain. According to The New York Times (Drew and Revkin 2005), by the late 1970s the Corps had abandoned that approach after the project was delayed by lawsuits from environmental groups—a charge denied by an environmental spokesman. (“That’s a bunch of baloney.” Harold Schoeffler, Sierra Club and Audubon Society, in The New York Times, September 30, 2005).

Clearly, however, the present process for protecting cities from flooding by storm surges has proven to be inadequate.

Would three barriers protecting inner New York City be prohibitively expensive? The engineering feasibility study and cost estimates are needed to answer this question with assurance. However, an idea of the order of magnitude of the cost can be obtained by comparing the structures with those previously built in Europe. The Eastern Scheldt barrier in The Netherlands, 80 percent wider than the Narrows and in places 15 percent deeper, but without provision for passing vessels, cost 3 billion Dutch guilders on its completion in 1986. Converting that to U.S. dollars at the exchange rate at the time and inflating the cost by the Bureau of Reclamation Construction Cost Composite Trend, it cost about $1.8 billion in present dollars. The Thames Barrier, spanning a river about the width of the Arthur Kill and the upper East River, cost 600 million British pounds on its completion in 1984, equal to $1.6 billion today. This suggests a ballpark for the three barriers of about $5 billion.

By comparison, by latest estimates the Corps will spend $14.7 billion to upgrade the levees surrounding New Orleans (Schwartz 2007).

There is little doubt that New York City will be exposed to major coastal flooding within the next several decades as sea level rises and storms may become more frequent and severe. In the present process, no measures will be taken to protect the city until Congress authorizes the Corps to do its own studies. Congress is likely to take this step only when there is sufficient support from the public, which is now essentially oblivious to the problem. If history is any guide, nothing will be done until New York City has its own Katrina. Then, decades more will pass before storm surge protection is in place.

To provide timely protection to the New York metropolitan region, the burden for action therefore lies with the professional community—engineers and scientists inside and outside government who are aware of the problem and have the capability to do something about it. We can take a step forward by establishing the engineering feasibility of the storm surge barriers now.

With the scientific foresight now available to anticipate coastal flooding, will engineering leadership materialize to bring about the protection needed for New York City? Or are we doomed to wait for the process that failed New Orleans?

The research reported here was performed by the Stony Brook Storm Surge Research Group, School of Marine and Atmospheric Sciences, State University of New York at Stony Brook The group is led by Malcolm J. Bowman and consisted at the time of Brian Colle, Roger Flood, Douglas Hill, Robert E. Wilson, Frank Buonaiuto, Peng Cheng, and Yi Zheng. The work was supported by New York Sea Grant, New York City Department of Environmental Protection, Eppley Foundation, and the URECA program of Stony Brook University. Helpful advice for this article was provided by two anonymous reviewers, Christine Van Lenten, and Francis J. Lombardi of the Port Authority of New York and New Jersey.