Act of God: an extraordinary interruption by a natural cause (as a flood or earthquake) of the usual course of events that experience, prescience, or care cannot reasonably foresee or prevent.
—Merriam-Webster Online Dictionary
As shown in this forecast issued at 11 p.m. EDT on Friday, 28 August, the NOAA Tropical Prediction Center expected nearly 60 hours before landfall that Hurricane Katrina would strike just east of New Orleans. Twelve hours earlier, TPC had been calling for a landfall on the Florida Panhandle. Click here or on image to enlarge it. (Image courtesy TPC.)
More than two months after Hurricane Katrina tore into Louisiana and Mississippi, causing one of the worst "natural" disasters in U.S. history, it is a challenging task to say something that hasn't been said already about this storm. There are so many fascinating and horrifying strands to the story: short-term forecasting success, planning and response failure, population growth in a vulnerable area, development in sensitive protective wetlands, economic inequality, land subsidence and sea-level rise, and climate change.
There can be little doubt that the failure of society and government on all levels contributed so much to this disaster that it can hardly be called an act of God. Years of warning about the vulnerability of New Orleans and the Gulf Coast produced no focused, coordinated government response to guard against the well-known risks. Indeed, human activities over the years actually worsened the problem; for example, the development of fragile wetlands that had helped protect inland areas exacerbated the risk and vulnerability. Plans for responding to the inevitable emergency were not executed promptly when it occurred, and while a horrified nation watched the disaster unfold on television, government at all levels remained paralyzed for more than three days before acting. Thus, in spite of a nearly perfect 60-hour forecast of the hurricane track and landfall position (see figure right) and a good forecast of hurricane intensity, hundreds of people unnecessarily lost their lives, and property damage ran into the billions of dollars. Some estimates put the total costs at well over $100 billion.
|"New Orleans is a disaster waiting to happen."
—Scientific American, October 2001
"Hundreds of thousands would be left homeless, and it would take months to dry out the area and begin to make it livable . . . but officials say that right now, nothing can stop ‘the big one.' "
—New Orleans Times-Picayune, October 2001
(see "On the Web")
"Power outages will last for weeks . . . as most power poles will be down and transformers destroyed. Water shortages will make human suffering incredible by modern standards."
—Special weather statement, National Weather Service, Slidell, Louisiana, 10:11 a.m. CDT, Sunday, 28 August
This is not to say at all, however, that the slowness of government response and the societal issues associated with a huge underclass living in a bowl below sea level made the forecasts irrelevant. To the contrary, roughly 80% of the population of about a million people were given more than two days of warning and were able to evacuate to safety, many of them carrying irreplaceable personal property and records. Without this warning, many more lives would have been lost, and police, fire departments, and emergency response teams—stretched extremely thin as it was—would have had five times the number of people to evacuate from the wind, floods, and chaos. Thus the excellent forecasts proved their worth, and the wisdom of government investments in the weather forecast and warning enterprise over the years was well justified.
The Tropical Prediction Center (TPC) of the National Weather Service deserves enormous credit for its life-saving forecasts and warnings, and I join many colleagues in the community in congratulating them. It is worth noting, however, that the TPC stands at the most visible end of a complex scientific and technological system that has taken years to develop, and much of this system is largely invisible to the public and to most policy makers. It is worthwhile summarizing the reasons that the Katrina forecasts were as accurate as they were, where they could have been improved, and how to ensure that forecasts of future hurricanes are as good as possible.
The accuracy of forecasts of hurricane tracks has been increasing steadily over the years (below), and especially over the last decade. The reasons for this steady increase include five important and essential factors:
• the increasing diversity, accuracy, and coverage of atmospheric and ocean observations, especially satellite observations;
• better use of the billions of observations through the assimilation of the observations in numerical weather prediction (NWP) models;
• better NWP models, including higher resolutions and more accurate treatment of physical processes such as clouds, precipitation, radiation, and boundary layer effects; and
• more powerful computers that make it possible to assimilate the observations into the models and run the models fast enough in real time to be useful to forecasters.
Overarching all four of these factors is the fifth: scientific advances that allowed researchers and engineers to build the satellites and their instruments, to process the data, to assimilate them effectively in the models, and to improve the models. The excellent forecast shown on p. 2 and the improvements over the years did not just happen by chance; they happened as the result of public investment in science and technology in government laboratories and universities by NASA, NSF, and NOAA. This includes basic research in mathematics, Earth science, and computer science, as well as the education and training needed to bring the research findings to fruition and benefit society. These reasons for the excellent forecast must be understood and remembered, and they form the foundation for a strategy to move forward to improve forecasts of hurricanes and other natural disasters in the future.
The recent NRC report Earth Science and Applications from Space: Urgent Needs and Opportunities to Serve the Nation (see "On the Web") emphasizes the importance of the Earth sciences and their applications to meeting societal needs. Hurricane Katrina reiterates this point, yet there are ominous signs that the nation's research and operational satellite programs are in danger. The NRC report sounded the alarm: "Today, this system of environmental satellites is at risk of collapse." Additional threats to the nation's satellite system have emerged since the report was released. Space News reports that NPOESS, the nation's future polar orbiting satellite system, is having significant developmental and budget problems, and the originally projected launch in 2010 may be delayed as much as four years. We can only hope that the government response to the NRC warning is more effective than the response to the warnings about the vulnerability of New Orleans to a major hurricane.
But you might ask, Aren't forecasts good enough already? How could we have improved on the forecast shown on p. 2, and what good would it have done, given the overwhelming societal issues in this case? There are several ways of answering this. First, although the forecasts 60 hours in advance of landfall were excellent, forecasts at longer lead times (three days and beyond) were not nearly as good. Second, the forecasts for Katrina within 60 hours of landfall were better than most; the average 60-hour error in forecast hurricane position is approximately 260 kilometers (160 miles). Hurricane Rita, a second Category 5 system in the Gulf of Mexico just two weeks after Katrina, proved more challenging to predict. Although Rita made landfall well within the 48-hour forecast zone, the landfall location (near the Texas-Louisiana border) was well east of the three-day projections. If three- to four-day forecasts had more skill and greater certainty, the massive evacuation of Houston might have been avoided. So there is significant room for continued improvements, and there is every reason to believe these improvements are possible following the formula above.
This graphic shows the steady reduction in error for hurricane track forecasts from NOAA's Tropical Prediction Center since 1970. TPC began issuing 4- and 5-day track forecasts (orange and purple) in 2001. Errors are shown in nautical miles; 1 nautical mile is about 1.15 miles or about 1.85 kilometers. (Image courtesy TPC.)
Looking to the future, it is essential to do more than address the Katrina disaster and then carry on with business as usual. Katrina was not an isolated event, and similar disasters (winter storms, floods, droughts, earthquakes) will happen again and again. It is essential to learn from this disaster and act in fundamentally different ways to prepare for the future, as outlined in the 2001 AMS and UCAR document A National Priority: Building Resilience to Natural Hazards (see "On the Web"). The university community, including UCAR and NCAR, stands ready to help in making the necessary changes through research, technology development and transfer, and education and training.
The unprecedented, but not unforeseen, disastrous impacts associated with Hurricane Katrina could have been largely mitigated, likely at a cost much less than what is now being spent on recovery. This lesson must be learned and remembered, and appropriate actions must be taken so that future hurricanes and other natural disasters do not result in a similar catastrophic loss of life and property. As Katrina showed, there are huge costs and consequences for not acting. Understanding, predicting, and preparing for disasters are all major issues for our nation's health, prosperity, and defense, and they must be treated as a long-term priority along with other issues such as the economy and traditional forms of national defense.
Hurricanes and global warming
The possible relationship of Hurricane Katrina and other recent powerful hurricanes to global climate warming has been a topic of much speculation in the media and scientific community. This is a complicated, important scientific topic that is currently under study, and it will be a number of years—perhaps many—before we know the relationships between climate change and the various characteristics of tropical cyclones (e.g., frequency, intensity, rainfall). However, a few well-accepted facts appear relevant.
(1) Climate is changing. The atmosphere and tropical oceans are warming, and the water vapor in the atmosphere
(2) Tropical cyclones owe their existence to warm, moist tropical air and high sea-surface temperature.
(3) The formation, intensity, size, and movement of tropical cyclones depend upon the large-scale atmospheric circulation and associated properties such as wind shear and static stability. Some aspects of large-scale circulation appear to be changing as the mean global climate warms.
Given these facts, there are several plausible ways that tropical cyclone characteristics could change in a warmer climate. It is tempting to hypothesize from (1) and (2) that both the frequency and the intensity of tropical cyclones will increase as Earth warms. However, hurricanes do not form everywhere in the tropics at once; at any given time, large-scale areas of ascent and descent tend to favor cyclone formation in some areas and disfavor it in others. Changes in these circulation patterns could have major implications, probably differing in different ocean basins. Two major studies this summer (see page 6) indicate a steady or even decreasing number of total tropical cyclones over the last several decades, coupled with greater intensities among the cyclones that do form. One possible explanation is that the climatological circulations and static stability distributions will become less favorable for tropical cyclogenesis, but the systems that do form will be more intense on average because of the higher ocean temperatures.
Whatever the relationship between hurricanes and global warming turns out to be, it is not likely to be simple, and we will never be able to attribute a single event like Katrina to a changed climate. We can only hope to understand changes in the statistics of many storms and their relationship to climate change.