Recently, interest in the impact of weather on human health has become an issue of much greater significance, especially because of the numerous predictions for warmer weather over the next century.  In fact, the WMO was a co-sponsor of a recent book published with the WHO and UNEP on climate change and human health, which I was proud to co-author.  In the book, a large variety of climate/health issues were addressed, including the impact of heat and cold on mortality and morbidity, effects of climate on infectious vector-borne diseases, and the impact of extreme weather events on food production, freshwater supplies, and environmental refugees.

The result of this book was the development of a large collaboration with a number of major partner organizations in the Climate Agenda, as well as national and municipal agencies, in the development of what we are calling Showcase Projects, which are designed to apply climate information and weather forecasts to the reduction of human deaths related to extreme heat waves.  The goal of this lecture is to describe the Showcase Project agenda to you, and to point out the unique and state-of-the-art climatological methodologies to be used in an attempt to reduce heat-related mortality in cities within the developed and developing world.  Thus, I will describe our work in the construction of an international set of heat/health warning systems. This project is not only part of a large UN agency collaboration, it is also part of a formal commission which I head as member of the executive board of the International Society of Biometeorology, the foremost organization of biometeorologists in the world.  Weather/health studies fall perfectly within the realm of the discipline of biometeorology, a field which has shown phenomenal research growth over the past few decades, and the International Society of Biometeorology has led the way in promoting our discipline as an important mainstream science.

The basic premise in understanding human response to the weather rests on the knowledge that we respond very well to the average or normal weather situation, but we do poorly when certain weather thresholds are exceeded.  These threshold exceedences can be short-term, and only several hours of unusually hot weather can lead to markedly increased death rates, upswings in hospital admissions, and increases in the number of individuals suffering from mental stress and depression.
 

Research that we have performed in China, Canada, Egypt, Italy, and the USA all support this range of tolerance concept.  In many of these locales, only the warmest 10-15 percent of all days in summer have an impact on human mortality.  Some of these increases in mortality are greater than 100 percent, and for a large number of cities in the developing and developed world, heat is the greatest weather-related killer.  It is interesting that people in certain areas of the world are much more susceptible to the impacts of heat stress than others, and the developing countries are not necessarily the most at risk.  With heat-related illnesses, those areas where heat occurs less frequently, such as the temperate zones, show a much more pronounced negative health response to excessive heat than areas which have consistently hot conditions.  It appears that weather variability is more responsible for increases in heat-related illness and death than the intensity of the heat itself.  Thus, cities such as Chicago, St. Louis, and Philadelphia, in the USA, have many more heat-related deaths than many tropical cities in the developing world.  This is true in spite of the presence of air conditioning in most homes and workplaces in these cities.  Of course, temperate cities elsewhere in the world are also vulnerable, and Shanghai is a good example of a city where the population is highly sensitive to heat-related problems.  People living in highly variable summer climates are ill-adapted to extreme heat, mainly because it occurs irregularly.  On the other hand, people living in tropical regions cope with excessive heat through adaptations in lifestyle, physiological acclimatization, and the adoption of a particular mental approach.  Cultural or social adjustments, including house design and general urban area structure may also explain why heat-related mortality is often less in hot climates with low variability.

Evidence also exists that long-term fluctuations in weather also affect human health.  In some cases, inter- and intra-seasonal climatic factors have a huge impact on heat-related mortality.  For example, during summer, early season heat waves are associated with much higher mortality than late season heat waves of similar magnitude.  This intra-seasonal variation is partially attributed to a behavioral acclimatization which occurs within the population as the hot season continues.  In addition, susceptible members of the population may die during the early heat waves, leaving less vulnerable people to die during later heat waves.  There might even be an inter-annual effect to note here.  After a number of cool summers, it is possible that a very hot summer will produce even greater numbers of deaths than one might expect.  Cool summers permit the number of susceptible individuals to accumulate, and the first severe heat wave could be catastrophic.  This may be a significant factor in why so many people died in Chicago in 1995 - the last very hot year before this was 1988.

Considering the significance of heat-related illnesses in many of our most important urban areas around the world, a collaboration has been developed to construct effective heat-health watch/warning systems for vulnerable large cities worldwide.  This collaboration is fostered under the World Climate Applications and Services Program through the Interagency Committee on the Climate Agenda, and includes a number of partners, such as the World Meteorological Organization, The World Health Organization, the United Nations Environment Programme, the US Environmental Protection Agency, the Electric Power Research Institute, and the University of Delaware.  As a WMO Rapporteur on Climate and Human Health, my work is accomplished through the Commission for Climatology, and it is my honor to lead this very worthwhile project.
 

The heat-health warning systems for Rome and Shanghai are unique in two very important ways.  First, they are specifically tailored for each urban area, and take into account the individual climate, social structure, and urban landscape of each city.  Thus, the systems recognize the fact that individuals react very differently to weather from one city to the next - a situation leading to heat-related mortality is very different in London than in Cairo. In the few places where operational watch/warning systems exist, the criteria for issuing a heat advisory are often based on somewhat arbitrarily chosen individual weather elements which do not necessarily relate to any particular human response (such as an apparent temperature exceeding 41oC), and coordination between the local weather service, which issues the advisory, and the local health agency is often far from optimal.

Second, the systems under development are the first in operation which are based on actual weather/human health relationships.  These systems acknowledge that it is an entire “umbrella of air”, rather than particular weather elements, which potentially cause a negative human health response.  The systems designed here are based on identifying stressful weather situations, rather than just temperature, humidity, and other weather parameters which may be part of an overall oppressive situation.

The Showcase Project watch/warning system concept was based on recommendations of a Commission for Climatology meeting of experts attended by WMO, WHO, and UNEP representatives in Freiburg, Germany in 1997.  The results grew into a UNEP/University of Delaware Memorandum of Understanding, with the following goal: “To develop heat/health warning systems for cities in the developed and developing world using procedures that emphasize climate-health outcomes, and based on guidance from these systems, to assist in the implementation of intervention and mitigation methodologies.”  These guidelines are based on our development of two already-existing prototype systems in the United States, located in Philadelphia and Washington, DC.  The first city selected for system development under the Showcase Project umbrella is Rome, Italy.  I would like to use Rome as an example of how these systems represent the state-of-the-art approach for evaluating weather-health outcomes, and how we put system information into practice to save lives.  The Rome system has been funded entirely by the UNEP.

The following steps have been used in the development of the Rome system:
· determine historical relationships between heat-related mortality and weather using the most efficient climatic modeling possible.
· develop a means to identify the most significant weather conditions which lead to increases in heat-related mortality.
· develop a means to forecast weather conditions which lead to increases in heat-related mortality over the next 48-hour period using 6-hourly forecasts.
· develop a set of intervention plans so the city may mitigate the health damage when the system indicates that an advisory should be issued.
· once the system is operating, develop a method to check the effectiveness of the system in actually saving lives.
 

The Rome system, like the others in the Showcase Project, was developed at the University of Delaware, with cooperation from local agencies; in Rome, these are the Lazio Health Authority and the Italian Meteorological Service.  To achieve the first step, Lazio provided us with a daily set of Rome mortality data for a 12-year period from 1987 through 1998, subdivided by age (greater than or less than 65 years of age) and gender.  The Italian Meteorological Service provided us with an extended period set of meteorological data for Rome so we could construct the climatological model used in the project.

It appears that virtually all causes of death increase under stressful weather, so it was unnecessary to divide the mortality data by cause of death.   Once we received the mortality data, they were standardized in a fashion to remove any non-meteorological noise.  For example, in Rome it is clear that an exodus of citizens occurs during August, when Italians frequently leave the city to take vacation.  Thus, the standardization must account for such variations before any mortality/weather relationships can be developed.

We tested three different climatological modeling procedures to determine the best means of relating variations in summer weather to the standardized mortality data.  These included:
· an apparent temperature approach, using models derived in the US to evaluate the simultaneous impact of temperature and humidity,
· a perceived temperature approach, with data kindly donated by my colleague Dr. Gerd Jendritzky of the Deutscher Wetterdienst, which evaluates the energy budget of the body,
· a synoptic climatological approach, which places each day within an air mass type and determines the impact of the total weather situation on the individual.

 The procedure which produced the most robust results, by identifying the strongest relationships between summer weather variation and excess mortality, was the one to be selected for use in the watch/warning system.  This was tested by determining which kind of threshold: an apparent temperature value, a perceived temperature value, or an air mass type, best explained variations in mortality during oppressive situations.  The evaluation indicated that the synoptic climatological methodology was most capable of detecting high mortality days in Rome, and certain air masses are associated with elevated levels of mortality when they occur.  The system used here, called the Spatial Synoptic Classification (SSC), is an automated statistical procedure which classifies each day into one of eight air mass types:
· dry polar (DP)
· dry moderate (DM)
· dry tropical (DT)
· moist polar (MP)
· moist moderate (MM)
· moist tropical (MT)
· moist tropical plus (MT+), the most oppressive subset of MT
· transition (TR)

The system accounts for numerous meteorological parameters, as well as their variation throughout the day, and a synoptic approach allows for a determination of the entire umbrella of air over us at any given time, rather than just an evaluation of individual weather parameters.  It is suggested here that the evaluation of all these meteorological parameters simultaneously within such a synoptic approach is the reason for SSC’s superior results, as simpler models, such as an apparent temperature algorithm, are not capable of incorporating the entire suite of weather conditions present on any given day.  The results for Rome are consistent for those obtained for US cities, where a synoptic procedure is capable of isolating weather situations which lead to increased numbers of deaths.

When evaluating mean total mortality for Rome during the summer it is clear that two air masses in particular, MT+ and DT, have a mean mortality which is statistically significantly higher than the other 6 air masses, with an additional 5 to 7 deaths per day on average.  Considering that Rome averages about 40 deaths per day, this is a significant 12-17 percent daily mortality increase.  DT and MT+ comprise approximately 11 percent of all summer days in Rome.  DT contains the highest temperatures of any air mass, and the least cloud cover.  MT+, while lower in afternoon temperature, has the highest dew point of any air mass.  Both of these air masses typically have very high overnight temperatures.

Typically, the air masses which contain the highest mean daily mortality also contain the highest daily standard deviation of mortality as well.  This indicates that, although some DT and MT+ days have very high mortality totals, sometimes 15 or 20 deaths above the mean, other days within these air masses have very little excess mortality.  Hence, following the initial identification of these “offensive” air masses, tests are performed to discern which within-air mass parameters on offensive air mass days are most closely linked to mortality.  The parameters are listed on the slide, and include meteorological factors, such as maximum and minimum temperatures within the offensive air mass on that particular day, a sum of cooling degree hours, and mean daily cloud cover.  Several important non-meteorological factors are also evaluated, such as the number of consecutive days of the offensive air mass (three or four straight days of DT air are very detrimental to human health), and the time of season the offensive air mass day occurred (an MT+ day occurring in June kills more people than a similar MT+ day in August, and this must be accounted for).  The parameters which show a statistically significant relationship with mortality variability within the offensive air mass are identified using stepwise linear regression, and the developed algorithm can then be used to estimate excess mortality on days when weather emergencies are declared.

The regressions with the highest correlation were found for elderly people over 65 years of age; these were retained for system development.  On DT days, time of season is an important inverse factor, as mortality is higher earlier in the season within this air mass, before people are acclimatized to the summer weather.  Another reason for this is due to the large susceptible population early in the summer season, which is diminished as people die during subsequent heat waves.  Day in sequence is a direct factor, as persistent offensive air mass days exacerbate heat-related problems.  According to the regression analysis, every additional DT day adds an additional 1.61 deaths on average.   In terms of meteorological conditions, the morning’s minimum apparent temperature and yesterday’s mean apparent temperature are both contributing direct factors.

MT+ days have a different set of important criteria.  Perhaps because the air mass is always warm overnight, minimum temperatures are not important.   Rather, a total heat load factor of cooling degree hours does contribute to increased mortality.  This variable is less concerned with maximum temperature and more dependent upon consistent heat throughout the day.  Thus, a day with a high maximum temperature but with a cooling afternoon thunderstorm will have a lower cooling degree hour total than a consistently hot day, even if maximum temperatures were not as high as the thunderstorm day.  Like the DT relationship, time of season is also important, with more deaths earlier in the season when an MT+ air mass is present.

These formulae can be used to estimate the number of heat-related deaths if either of the two offensive air masses are forecast to arrive in Rome.  For example, if a DT air mass is forecast for tomorrow with the following characteristics:
· occurs on 1 July
· third consecutive DT day
· 22 degree C minimum apparent temperature
· 25 degree C minimum apparent temperature for tomorrow

the DT formula estimates that there will be 11 extra deaths attributed to heat.

Certain 2-day combinations of air masses can lead to greatly increased daily mortality.  For example, in Rome, a day of DT air followed by a transition situation leads to almost 16 extra deaths per day.  A MT+ air mass day, followed by a DT air mass day, leads to over 12 extra deaths.  These particularly dangerous combinations need to be accounted for in watch/warning system development
The identification of the oppressive air masses and the subsequent formulae comprise the central part of the heat/health warning system.  During a recent meeting with the Lazio Health Authority and the Italian Meteorological Service, it was decided that the forecast of an offensive air mass within the next 48 hours will trigger a warning of ATTENTION, indicating that conditions will soon be conducive to health problems if the forecasted offensive air mass arrives.  If an offensive air mass is forecast within the next 24 hours, and if the formulae predict two or more excess deaths from the heat, an ALARM will be issued by the Lazio Health Authority to warn of these dangerous conditions.  Particular care was taken to assure that the terminology used was designed to prompt action from the local citizens, and may vary from city to city.

Construction of the system represents only one part of an effective heat-health alert plan.  It also requires close coordination with the local meteorological service, who issues the weather forecasts necessary to operate the system.  In the case of Rome, the Italian Meteorological Service will provide 48 hour forecasts daily with all the meteorological variables necessary to determine the day’s air mass type and to calculate the formulae if the air mass is offensive.  The forecast data will be fed into a password-protected website that will be available to Lazio health officials, and each morning, the website will automatically determine whether an offensive air mass is predicted over the next two days, and if so, calculate the number of excess deaths.  The system will be run in experimental mode this summer in Rome to make certain that operation is smooth, and to determine if the system is actually predicting excess deaths with some accuracy.  Full system operation is scheduled for summer, 2001.

Another key to the success of this system will be the development of an effective intervention plan which is put into place once a warning is called.  Accompanying me to Rome was Dr. Lawrence Robinson, Deputy Health Commissioner for the city of Philadelphia, who has developed a very efficient intervention plan whenever the Philadelphia Health Commissioner calls a heat emergency. Philadelphia uses a system which we designed at the University of Delaware which is very similar to the Rome system and tailored specifically for Philadelphia.  The illustration on the screen shows the steps taken by the Philadelphia Department of Public Health every time a heat emergency is declared.  First, a media contact person at the Health Department calls the television and radio stations to inform the public that the Health Commissioner has declared a heat emergency and that the weather is conducive to heat illnesses.  The media then broadcasts to the population steps that should be taken to lessen the likelihood of a heat problem, such as staying in air conditioning if possible, drinking plenty of fluids, and taking a cool bath.

However, the intervention system is much more complex than this.  Among other things, the city of Philadelphia has established a “buddy system”, where an assigned individual for each street goes door-to-door to check on the elderly and infirm.  If someone is ill, the “buddy” provides support and advice, and can call an ambulance if necessary.  Buddies are trained by the Philadelphia Department of Public Health.  In addition, the Philadelphia Corporation for Aging, an agency which deals with nursing homes and other elderly problems, assigns extra staff to make certain that these nursing facilities deal with heat-related illnesses in the proper manner.  The city also staffs a “heatline”, which is a special telephone number broadcast by the media to be used by individuals who are becoming ill by the heat.  Specially trained individuals can then tell citizens over the telephone what to do if they are feeling badly.  The Department of Public Health also contacts local utility companies, specifically the water department and the electric company, warning that they should not terminate service to anyone during the heat emergency, even if individuals have not paid their bills.  The city opens special air conditioned shelters for individuals who need cooling relief.  A special team of volunteers travels to areas in the city where the homeless congregate, to make certain that they are handling the situation satisfactorily.  There are other steps taken as well, and it is clear that the city must expend dollars and considerable effort to make certain that its citizens are dealing with the heat as well as possible.  Therefore, the system must be accurate to make certain that we have a minimum of “false positives”, when a heat warning is issued when none was necessary, or even worse, “false negatives”, when a heat warning is not called but should have been, leading to excess unexpected deaths.  During the severe heat wave of 1995, when the system was in its first year of operation, the Philadelphia Department of Public Health declared that they estimated a saving of 300 lives because of the use of this system. They are continuing to use the system successfully to this day, with close cooperation from the National Weather Service.  At present, the Lazio Health Authority in Rome is designing its own intervention plan, which will go into effect whenever the system declares an ALARM in Rome.

A meeting was held in Shanghai last fall to initiate the second heat/health system under the Showcase Project umbrella for Shanghai.  The same team which developed the Rome system was augmented with representatives from the WHO and its Regional Office in Manila, as well as members of the Chinese Meteorological Administration, the Shanghai Meteorological Bureau, the Shanghai Health Administration and other Chinese experts.  The Shanghai heat/health system is being funded jointly by WMO (through US support of China’s Voluntary Cooperation Program request) and WHO, with contributions from other agencies.  At least five individuals as part of a Chinese team will travel to the University of Delaware this summer to begin construction of the Shanghai heat/health system, and to provide the necessary mortality and meteorological data for system development.  My students and I at the University of Delaware will transfer technology to the Chinese team by teaching them our synoptic air mass methodology and providing them with software so they can operate the system on site.

Further systems are being developed or are in the planning stages.  Three large cities in Ohio, USA are funding a large-scale heat-health system that we are developing for operation commencement this summer.  This system will affect over 4,000,000 people in a large section of southwestern Ohio.  The Toronto Department of Health has submitted a grant proposal with us to implement a system in that city; we are optimistic that the province of Ontario will fund this project.  We are presently collaborating with the National Institute of health in Lisbon, Portugal to update a system there.  Finally, I recently traveled to Phoenix, USA, in the desert, to plan the development of a system in that city.  It is clear that there is a great demand to deal with the problem of heat and its impact on human health, and with the possibility of a climate change and an associated increasing frequency of stressfully hot conditions, it is even more imperative that vulnerable cities be prepared to deal with the heat.

We are now deluged with a growing number of requests to install systems in other major vulnerable cities around the world.  Plans are under development for the best approach to assist in building the capacity for local systems worldwide.  WMO will continue to participate in multiagency projects of this kind, and as a rapporteur for WMO on climate/health issues, I am proud to be playing a part in this important initiative.  The primary goal of Phase II of Showcase Project initiative, which is being led by Dr. Gerd Jendritzky, will be the development and distribution of guidance materials and programs to transfer techniques worldwide.  Part of the goals of Phase II is to make a critical review of available methods for assessing thermal comfort with equal attention to all climatological and physiological models.  In addition, workshops will be planned to permit health officials and meteorological experts from developing countries to coordinate their activities and to learn how to operate such systems.  Outreach to local responsible agencies will increase as the Showcase Project matures, and it is hoped that systems such as the ones under development in Rome, Shanghai, and Ohio, will some day be commonplace in large urban areas around the world.

Let me conclude by thanking the WMO, and especially the World Climate Programme Department for insuring the essential participation of the WMO Permanent Representatives for the cities with which we are working.  I would also like to thank audience members for their support. Dr. Carlos Corvalan of the WHO has done an outstanding job working with the public health community within our cities.  Dr. Kris Ebi from the Electric Power Research Institute has given financial support to this worthwhile endeavor.  Professor Gerd Jendritzky has provided guidance, data, support, and some very useful advice.  And to my family in the audience, I offer special thanks for putting up with my repeated absences because of travel.  One family member has even participated significantly in my research, and I offer special thanks to my son, Adam, a student at the University of Virginia, who spends his summer as part of our research team at the University of Delaware.

Thank you for your attention.