Table 1.

Information extracted from the pool of relevant studies.

Authors (Date)	Event Type	Country	Time-Period/Case Study	Pop	Study Objective & Method	Typology	Health Metrics	Health Impacts Monetization	Ecosystem Services Mentioned?	Distributional Analysis	Key Results
[20] (2008)	Extreme temperatures	U.S.	2005	All	Descriptive analysis of statistics on hospital stays resulting from excessive heat or cold exposure due to extreme weather conditions.	Burden evaluation (Group A)	Hospitalization admissions	Healthcare costs	No	Yes	Excessive temperatures costed U.S. hospitals $120 m in 2005. The average cost of a heat-related stay amounted to $6200 vs. $12,500 for cold-related stay. Patients admitted were older than the average hospital patient (6 to 7.6/100,000 people aged ≥65 years old). Hospitalizations were found to be about 2 to 2.5 times more common in the poorest communities than in the wealthiest ones and slightly more common in rural regions.
[21] (2007)	Flooding	Japan	2004 Toyooka flood	All	Estimation of willingness to pay (WTP) (contingent valuation) to avoid of mental damage caused by flood disaster (indirect approach using option value).	WTP study (Group C)	Mental damage	WTP estimated in the study	No	No	Individuals expressed a significant WTP to avoid mental damage: Mean WTP 44,769 yen.
[22] (2012)	Extreme temperatures	U.S.	2004–2005	U.S. Medicare pop. (≥65 years old or disabled)	Descriptive study to assess the health care burden of hypo- and hyper-thermia due to extreme weather conditions.	Burden evaluation (Group A)	Inpatient and outpatient visits	Healthcare costs	No	Yes	Hyperthermia-related visits were more frequent than hypothermia but less costly ($36 m vs. $98 m for hypothermia in 2004–2005). Black and Native Americans had a significantly higher relative risk of healthcare visits than their white counterparts.
[23] (2012)	Extreme temperatures	U.S.	1991–2004 for current burden; 2046–2065 and 2080–2099 for projections	All	Estimation of current and projected heat-related public health burden in New York State under a range of three IPCC climate scenarios.	Burden estimation (Group A)	Respiratory admissions	Healthcare costs and productivity loss due to days hospitalized. Adjustment for inflation and 3% discounting used. Costs normalised to $2004.	No	Yes	Hospital costs associated with heat-related respiratory admissions in NYS are currently estimated at $0.64 m p.a. and projected to increase to $5.5–7.5 m in 2046–2065 and to $26–76 m in 2080–2099. The public health burden is projected to be greater among females and in low-income groups.
[24] (2015)	Extreme temperatures	Australia	2000–2010 for current burden; years 2030 and 2060 for projections	All	Estimation of the current and projected burden of heat-related emergency department visits in Brisbane under a range of two IPCC climate scenarios.	Burden estimation (Group A)	Emergency department (ED) visits	Healthcare costs (normalised to AU$ 2013)	No	Yes	Higher relative risks of ED visits for adults aged 65+ than for their younger counterparts (RR for all ED visits = 1.09 vs. 1.06) on hot days (>35 degrees). ED visits are projected to increase considerably on hot days in the future under population growth and climate change scenarios. The excess number of visits by older patients is estimated to grow twice as much as the younger group. The excess demand is estimated to add an extra cost of around AU$78,000–260,000 in 2030 and AU$215,000–1,985,000 in 2060 (2013 prices).
[25] (2015)	Extreme temperatures	Spain	2002–2006	All	Estimation of: (i) the impact of excessive heat on mortality; (ii) the temperature threshold to mortality increase; (iii) the hospital cost of heat-attributable deaths.	Burden estimation (Group A)	Deaths	Healthcare costs	No	No	A statistically significant increase in mortality was observed when daily max temperature reached 38 ${}^{\circ }$C degrees. Over 2002–2006, excessive heat was found to be responsible for 107 (95% CI: 42–173) premature deaths, associated with a healthcare cost of €426,000 (€167,000–689,000).
[26] (2011)	Hurricanes	U.S.	Hurricanes Katrina and Rita in 2005	Single mothers	Estimation of the impact of exposure to hurricanes on the mental health resilience of single mothers versus the general population and computation of the related economic cost from lost productivity.	Burden estimation (Group A)	Days of poor mental health (reported in the last 30 days after event)	Direct private costs from absenteeism due to mental health disturbance	No	Yes	Following exposure to hurricanes Katrina and Rita, days of poor mental health was found to increase by 72% in single mothers vs. 18% in the total population. As a result, single mothers were expected to be absent from work 18.4 more days (vs. 3.6 more days of absence for the average person), leading to an income loss of $4.200/person (vs. $817 for the average person), thus exacerbating their economic vulnerability. Differential effects were found to persist one year after the events.
[27] (2009)	Hurricanes	U.S.	Hurricane Katrina in 2005	Patients with diabetes	Observational before/after study of the impact of Katrina on healthcare management of patients with diabetes in 3 different healthcare systems and projected health and healthcare costs consequences of treatment disruption over patients remaining lifetime.	Burden estimation (Group A)	Life expectancy (LE); quality-adjusted life expectancy (QALE)	Health care costs from treatment disruption over patients’ remaining lifetime	No	Yes	Treatment disruption in patients with diabetes following Katrina was projected to result in substantial health care costs ($504 m for the affected pop.) due to co-morbidities/disease complications in the long run. The impact reflects the high prevalence of the disease (about 9% of U.S. pop.) and the large size of the population affected. The disaster exacerbated inequalities in access to healthcare and resulting health disparities between socio-economic subgroups.
[28] (2011)	Multiple events	U.S.	2000–2009	All	Estimation and comparison of the health costs associated with 6 climate change related-events : (i) California heatwave 2006; (ii) ozone air pollution (for daily levels above national standards; impacts computed for the years 2000–2002); (iii) Florida hurricane season 2004 (4 hurricanes in one month); (iv) West Nile virus outbreak (vector-borne disease) in Louisiana; (v) red river flooding in North Dakota in 2009; (vi) Southern California Wildfires in 2003.	Burden study (Group A)	Deaths, hospitalizations, emergency department visits and outpatient healthcare use	VSL for mortality ($7.8 m in $2008); healthcare costs and loss-work productivity for morbid endpoints	No	No	Events associated with the greatest number of premature deaths were associated with the highest costs. The costliest weather-related extreme event in terms of health impacts was California’s 2006 heat wave ($5.4 bn), followed by Florida hurricane season ($1.4 bn), California wildfires ($600 m), West Nile infectious disease outbreak ($207 m) and red river flooding ($20 m). When normalised to 1000 people, the cost of river flooding was, however, nearly as high as the cost of heatwave ($150 k/1000 person).
[29] (2014)	Pollution peak	Malaysia	2004–2009	All	Estimation of the change in hospital admissions for a change in pollution concentrations (dose-response function) and evaluation of the associated economic burden.	Burden estimation (Group A)	Respiratory and cardiovascular hospital admissions	Healthcare costs and productivity loss	No	No	On average, over 2004–2008 for Kuala Lumpur and some areas in Selangor state (equivalent to 25% of the Malaysian population), smoke haze occurrences were found to be associated with an increase in inpatients visits by 2.4/10,000 people per year, representing a 31% increase from normal days. The associated economic loss amounted to $91,000 per year. Under no change in haze recurrence, over 20 years, this would represent a cost of $1.7 m, discounting at 5% p.a.
[30] (2015)	Pollution peak	China	Severe haze event in January 2013	All	Modelling of $P{M}_{2.5}$ concentrations during the haze episode and estimation of the associated acute mortality and morbidity impacts and associated health care costs.	Burden estimation (Group A)	Deaths, cases of acute bronchitis and asthma, hospital admissions	VSL for mortality ($274 k); WTP or healthcare costs and productivity loss for morbid endpoints	No	No	The total economic cost of the haze-related health impacts was estimated, under conservative assumptions, at $253 m, i.e., about 0.8% of the annual GDP of Beijing.
[31] (2005)	Flooding	Turkey	1970–1996 (624 floods recorded)	All	Descriptive analysis of seasonal and regional trends in the mortality and economic impacts of flooding based on registered flood reports.	Descriptive analysis (Group B)	Deaths	N.A. (see Section 2.2 and Section 3.5)	No	No	Seasonal and regional trends in terms of human deaths and economic impacts were determined. Most floods and deaths happened in the summer season. Most of the floods and deaths occurred in the Black Sea region.
[32] (2013)	Flooding	Turkey	Large floods between 1955–2005 in the eastern Black Sea basin (EBSB)	All	Descriptive analysis of flood occurrence and meteorological conditions and identification of trends in damages and human lives loss.	Descriptive analysis (Group B)	Deaths	N.A. (see Section 2.2 and Section 3.5)	Yes	No	Between 1995–2005, 51 floods occurred in EBSB causing 258 deaths and $500 m of damages to assets. Most floods occurred during summer months when snow melt is combined with heavy rainfall in mountainous valleys. Despite the absence of an increasing trend in extreme rain values and flood frequency, an upward trend in terms of both death and damages was found. The latter was attributed to human factors, such as illegal land use, urbanization in flood-prone areas, road construction in stream beds, deforestation and insufficient drainage structures. Alongside structural measures, watershed management and reduced deforestation were suggested to reduce vulnerability to flood.
[33] (1999)	Multiple Events	U.S.	1968–1995	All	Descriptive analysis of trends in the frequency of extreme events and their associated fatalities and economic losses.	Descriptive analysis (Group B)	Deaths	N.A. (see Section 2.2 and Section 3.5)	No	No	The upward trends in human fatalities and economic losses from extreme events was found to be essentially related to an increased vulnerability stemming from a growing population in coastal areas and lifestyle and demographic (population ageing) changes.
[34] (2012)	Flooding	Vietnam	2007 floods	All	Estimation via contingent valuation of the welfare loss from flood-related illnesses and well-being reduction following flood disaster in a developing country. Willingness to contribute in kind was used to estimate WTP to avoid this welfare loss.	WTP study (Group C)	Flood-related illnesses; well-being reduction	WTC in-kind estimated in the study multiplied by an estimate of the opportunity cost of labour time	No	Yes	Flood damage was estimated on average to represent about 20% of households’ annual income. However, it was not possible to disentangle the welfare loss from morbidity and well-being reduction from the welfare loss due to damages to assets. Poor households were found to be more vulnerable to floods as the associated damage made up a significantly larger portion of their annual income. Households heavily dependent on agricultural activities were also found to be more vulnerable.
[35] (2010)	Extreme temperatures	Taiwan	1971–2006	All	Estimation of (i) the impact of climatic conditions on cardiovascular deaths in Taiwan over 1971–2006 and (ii) WTP (contingent valuation method) to avoid the increase in cardiovascular deaths projected under climate change.	WTP study (Group C)	Cardiovascular deaths	WTP estimated in the study.	No	No	Cardiovascular deaths are projected to increase by 1.2% to 4.1% in Taiwan under alternative IPCC climate scenarios, and each individual would be willing to pay annually $51 to $97 to avoid such an increase in mortality risk.
[36] (2015)	Flooding	Thailand	2011 floods	All	Survey of flood victims in three-severely affected provinces of Thailand to capture health-related and non-health related costs of damage.	Population-based survey (Group D)	Flood-related diseases	Healthcare costs from flood-related diseases	No	Yes	Health-related costs were negligible in contrast to losses to tangible assets (property, valuable etc). Few households experienced health-related losses (11% of sample). Evacuation rates varied between poor and non-poor households: 65% of poor households had some members evacuate vs. 77% for non-poor households.
[37] (2014)	Flooding	Pakistan	Pakistan floods in 2010	All	Comparison of the economic impacts and time-to-recovery after floods in Pakistan versus after an earthquake in Haiti using cross-sectional cluster surveys.	Population-based survey (Group D)	Death and injuries	N.A. (see Section 2.2 and Section 3.5)	No	No	Injuries and deaths were much greater in Haiti. Whilst a decline in income was widespread in both countries, relative household income loss was greater in Pakistan because of damages to the agricultural economy. Housing recovery was however quicker in Pakistan, and food insecurity was smaller than in Haiti, due to greater receipt of food aid.
[38] (2004)	Extreme temp.	U.S.	1995–1998	≥65 years old	Retrospective statistical analysis of the effectiveness of Philadelphia’s heat warning system (PWWS) in terms of reduced excess mortality.	Economic appraisal (Group E)	Deaths	VSL ($4 m)	No	No	117 lives are expected to have been “saved” (with substantial uncertainty around this estimate) over the 3-year period thanks to PWWS. This is equivalent to a gross benefit of $468 m that is much higher than the cost of running the system ($210 k).
[39] (2013)	Hurricanes & tornadoes	U.S.	1989–2005	All	Testing for a potential relationship between hurricane and tornadoes-related casualties and work routine.	“Other” (Group F)	Deaths and injuries	N.A. (see Section 2.2 and Section 3.5)	No	Yes	Daily variation in casualties from hurricanes and tornadoes is affected by the work routine. All things being equal, hurricanes, which provide the at-risk population with some lead time, lead to greater casualties during weekdays since the opportunity cost (namely income loss) of adopting protection measures (e.g., evacuating) is much larger than during weekends. On the opposite, tornadoes, which provide little lead-time, lead to larger casualties during weekends as the acquisition of risk information is harder on week-ends and workplaces and schools are safer than private homes. Casualty risk from tornadoes was found to reduce by 6%–8% for every $1000/per capita income added at the county level.