Science Space
The ÍCARO (ICARUS) PROJECT: 15 years of research and surveillance
The ÍCARO (Importância do Calor: Repercussão nos Óbitos – The Importance of Heat: Repercussion on Deaths) Project began as a simple model that registered the occurrence of heat waves and corresponding relationship to mortality in the district of Lisbon. It ended up inspiring a surveillance system of heat waves with the purpose of reducing the well-known substantial impact on mortality.
Fifteen years later, the ÍCARO Project, based at INSA (Dr. Ricardo Jorge National Health Institute), continues to have the same mission, which is to contribute to reducing the impact of heat on health. However, since its inception, many steps have been taken and its objectives are now multifold.
Objective: Anticipate and identify the occurrence of heat spells with potential adverse impacts on health; contribute to the measurement of those impacts; carry out research on measurement, forecasting, monitoring and surveillance models of extreme weather periods with a potential impact on health.
Scope: National Surveillance System.
Constituents of the ÍCARO Project:
• ÍCARO Regions;
• ÍCARO Models;
• ÍCARO- Indexes and ÍCARO Warning-Indexes;
• ÍCARO Surveillance System.
It must be stressed that the current goal of the ÍCARO Project is no longer heath waves but heat periods. In fact, the occurrence of extreme heat that informed the initial project and the models that were built are not the main focus anymore.
The fact that the ÍCARO Surveillance System now extends to the whole country is a major technical advance. Initially, the ÍCARO Surveillance System set up in 1999 (see chronology, Figure 1) was based on meteorological and mortality data in the district of Lisbon. It was a Portuguese version of what is internationally called Heat Health Watch/Warning Systems[1], which, basically, are surveillance systems centred in a single city.
Figure 1. Chronology of the ÍCARO Project
As the known data for Portugal with regard to mortality associated with major heat waves up to 2000 indicated a high correlation between what was observed in the district of Lisbon and what was seen in the rest of the country, when the surveillance system was built, the forecasting model for the district of Lisbon was taken as a good predictor for the rest of mainland Portugal[2].
The experience resulting from the surveillance carried out in the summers 2000-2002 demonstrated that the relationship between predictions for Lisbon and the country at large, although basically true with regard to particularly extreme heat occurrences, had major limitations with regard to local phenomena. The best solution would have been to build multiple surveillance systems/models for the several cities around the country, but that proved unfeasible because the other cities/districts did not possess the population density to provide the necessary evidence to the building of models. This constraint was overcome with the construction of ÍCARO Regions (Figure 2).
Figure 2. ÍCARO Regions
The first ÍCARO model, for the district of Lisbon, introduced the concept of Accumulated Thermal Overload (ATO), which is a rationale whereby, in the face of intense heat, the corresponding stress accumulates over the days while the air temperatures do not drop. The ATO variable presupposed the existence of a maximum temperature threshold and stems from the maximum temperature of a particular day above that threshold multiplied by the number of consecutive days the temperature was above the threshold. In 2005, the models for the district of Lisbon were improved and included two new ideas: dynamic temperature thresholds and the generalization of the ATO variable, integrating the idea of slow cooling (slow fading of cumulative stress) when temperatures fall below the threshold instead of total immediate cooling[3].
The definition of models around the ÍCARO Regions (sets of districts) was quite challenging, as there were multiple possible temperatures that had to be considered, not just one, as in the Lisbon model. The models obtained by regions included the ATO variables for two temperature thresholds, 32ºC and 35ºC, and all regions included temperatures from at least two or three districts.
The existence of good statistical models for the predictions of deaths led to the creation of the ÍCARO-indexes[4], which initially aimed to relativize the absolute numbers of expected deaths, according to the following formula:
ICARO Index =Nº of expected deaths due to the effect of heat
__________________________________________ - 1 Nº of expected deaths without to the effect of heat
When multiple ÍCARO models came into existence (after 2006), it was found that the interpretation of values between the distinct indexes was not immediate. Recent research has demonstrated that this index depends on the modelled average number of deaths, which has led to a standardization of the Index called ÍCARO-Warning Index, and this has made interpreting all the indexes for the different regions, and even on a national basis, easier (Figure 3).
All these tools have enabled the surveillance of heat periods for 13 consecutive years. When it was set up in 1999, the system relied on the INSA and on the Weather Forecast Institute from a technical stance, and on the Directorate General for Health and on Civil Protection for a more active role. After all these years, the structure is still basically the same, although now guided by the heat contingency plans that were introduced throughout Europe from 2004 onwards, following the major heat wave of 20035.
Despite the long way the ÍCARO Project and corresponding Heat Wave Surveillance System (the first in Europe) have gone, there are still many research challenges, such as finding more evidence of impact on morbidity, building the corresponding models, studying the effectiveness of contingency plans, and contributing to improve them.
Paulo Jorge Nogueira
Institute of Preventive Medicine of the Faculty of Medicine of the University of Lisbon
pnogueira@fm.ul.pt
__________________
References
1. Kalkstein LS, Jamason PF, Greene JS, Libby J, Robinson L. (1996). "The Philadelfia Hot Weather - Health Watch/Warning System: Development and Application, Summer 1995." Bulletin of the American Meteorological Society 77(7): 1519-1528.
2. Nogueira P (2005). Examples of Heat Health Warning Systems: Lisbon’s ICARO’s surveillance system, summer of 2003. Extreme weather events and Public Health Responses. European Public Health Association.
3. Nogueira P e Paixão E (2008). "Models for mortality associated with heatwaves: update of the Portuguese heat health warning system." International Journal of Climatology 28(4): 545-562.
4. Nogueira P, Nunes B, Dias CM, Falcão JM. (1999). "Um sistema de vigilância e alerta de ondas de calor com efeitos na mortalidade: o índice Ícaro." Revista Portuguesa de Saúde Pública. I: 79-84.
5. Robalo J, Dieges P, Batalha L, (2011). Plano de contingência para temperaturas extremas adversas 2001 - módulo Calor. www.dgs.pt, Direcção Geral da Saúde.
Fifteen years later, the ÍCARO Project, based at INSA (Dr. Ricardo Jorge National Health Institute), continues to have the same mission, which is to contribute to reducing the impact of heat on health. However, since its inception, many steps have been taken and its objectives are now multifold.
Objective: Anticipate and identify the occurrence of heat spells with potential adverse impacts on health; contribute to the measurement of those impacts; carry out research on measurement, forecasting, monitoring and surveillance models of extreme weather periods with a potential impact on health.
Scope: National Surveillance System.
Constituents of the ÍCARO Project:
• ÍCARO Regions;
• ÍCARO Models;
• ÍCARO- Indexes and ÍCARO Warning-Indexes;
• ÍCARO Surveillance System.
It must be stressed that the current goal of the ÍCARO Project is no longer heath waves but heat periods. In fact, the occurrence of extreme heat that informed the initial project and the models that were built are not the main focus anymore.
The fact that the ÍCARO Surveillance System now extends to the whole country is a major technical advance. Initially, the ÍCARO Surveillance System set up in 1999 (see chronology, Figure 1) was based on meteorological and mortality data in the district of Lisbon. It was a Portuguese version of what is internationally called Heat Health Watch/Warning Systems[1], which, basically, are surveillance systems centred in a single city.
Figure 1. Chronology of the ÍCARO Project
As the known data for Portugal with regard to mortality associated with major heat waves up to 2000 indicated a high correlation between what was observed in the district of Lisbon and what was seen in the rest of the country, when the surveillance system was built, the forecasting model for the district of Lisbon was taken as a good predictor for the rest of mainland Portugal[2].
The experience resulting from the surveillance carried out in the summers 2000-2002 demonstrated that the relationship between predictions for Lisbon and the country at large, although basically true with regard to particularly extreme heat occurrences, had major limitations with regard to local phenomena. The best solution would have been to build multiple surveillance systems/models for the several cities around the country, but that proved unfeasible because the other cities/districts did not possess the population density to provide the necessary evidence to the building of models. This constraint was overcome with the construction of ÍCARO Regions (Figure 2).
Figure 2. ÍCARO Regions
The first ÍCARO model, for the district of Lisbon, introduced the concept of Accumulated Thermal Overload (ATO), which is a rationale whereby, in the face of intense heat, the corresponding stress accumulates over the days while the air temperatures do not drop. The ATO variable presupposed the existence of a maximum temperature threshold and stems from the maximum temperature of a particular day above that threshold multiplied by the number of consecutive days the temperature was above the threshold. In 2005, the models for the district of Lisbon were improved and included two new ideas: dynamic temperature thresholds and the generalization of the ATO variable, integrating the idea of slow cooling (slow fading of cumulative stress) when temperatures fall below the threshold instead of total immediate cooling[3].
The definition of models around the ÍCARO Regions (sets of districts) was quite challenging, as there were multiple possible temperatures that had to be considered, not just one, as in the Lisbon model. The models obtained by regions included the ATO variables for two temperature thresholds, 32ºC and 35ºC, and all regions included temperatures from at least two or three districts.
The existence of good statistical models for the predictions of deaths led to the creation of the ÍCARO-indexes[4], which initially aimed to relativize the absolute numbers of expected deaths, according to the following formula:
ICARO Index =Nº of expected deaths due to the effect of heat
__________________________________________ - 1 Nº of expected deaths without to the effect of heat
When multiple ÍCARO models came into existence (after 2006), it was found that the interpretation of values between the distinct indexes was not immediate. Recent research has demonstrated that this index depends on the modelled average number of deaths, which has led to a standardization of the Index called ÍCARO-Warning Index, and this has made interpreting all the indexes for the different regions, and even on a national basis, easier (Figure 3).
Figure 3. ÍCARO Warning Levels (operationalisation of the ÍCARO-Warning Indexes)
All these tools have enabled the surveillance of heat periods for 13 consecutive years. When it was set up in 1999, the system relied on the INSA and on the Weather Forecast Institute from a technical stance, and on the Directorate General for Health and on Civil Protection for a more active role. After all these years, the structure is still basically the same, although now guided by the heat contingency plans that were introduced throughout Europe from 2004 onwards, following the major heat wave of 20035.
Despite the long way the ÍCARO Project and corresponding Heat Wave Surveillance System (the first in Europe) have gone, there are still many research challenges, such as finding more evidence of impact on morbidity, building the corresponding models, studying the effectiveness of contingency plans, and contributing to improve them.
Paulo Jorge Nogueira
Institute of Preventive Medicine of the Faculty of Medicine of the University of Lisbon
pnogueira@fm.ul.pt
__________________
References
1. Kalkstein LS, Jamason PF, Greene JS, Libby J, Robinson L. (1996). "The Philadelfia Hot Weather - Health Watch/Warning System: Development and Application, Summer 1995." Bulletin of the American Meteorological Society 77(7): 1519-1528.
2. Nogueira P (2005). Examples of Heat Health Warning Systems: Lisbon’s ICARO’s surveillance system, summer of 2003. Extreme weather events and Public Health Responses. European Public Health Association.
3. Nogueira P e Paixão E (2008). "Models for mortality associated with heatwaves: update of the Portuguese heat health warning system." International Journal of Climatology 28(4): 545-562.
4. Nogueira P, Nunes B, Dias CM, Falcão JM. (1999). "Um sistema de vigilância e alerta de ondas de calor com efeitos na mortalidade: o índice Ícaro." Revista Portuguesa de Saúde Pública. I: 79-84.
5. Robalo J, Dieges P, Batalha L, (2011). Plano de contingência para temperaturas extremas adversas 2001 - módulo Calor. www.dgs.pt, Direcção Geral da Saúde.