Sensitivity of large dengue epidemics in Ecuador to long-lead predictions of El Niño

Desislava Petrova; Rachel Lowe; Anna Stewart-Ibarra; Joan Ballester; Siem Jan Koopman; Xavier Rodó

doi:10.1016/j.cliser.2019.02.003

Sensitivity of large dengue epidemics in Ecuador to long-lead predictions of El Niño

Desislava Petrova^*, Rachel Lowe, Anna Stewart-Ibarra, Joan Ballester, Siem Jan Koopman, Xavier Rodó

^*Corresponding author for this work

Research output: Contribution to Journal › Article › Academic › peer-review

Abstract

Long-lead forecasts of El Niño events are lacking despite their enormous societal and economic impacts. These climatic events lead to floods and droughts in many tropical regions, and damage agriculture and the economy in poor countries. Due to their impact on local climate, they can also affect human health by increasing the risk of vector-borne diseases, such as dengue fever. Physical processes at the origin of this complex coupled ocean-atmosphere phenomenon are just beginning to be better understood, with subsurface processes and stored heat as two of the main driving forces leading to the development of El Niño in a quasi-periodic manner. Taking advantage of this new knowledge, a statistical dynamic components model, using a state space approach and predictors relevant to the El Niño evolution, was specifically tailored to forecast warm events at lead times of about 2 years (well beyond the traditional spring barrier limit in El Niño predictability). This forecasting scheme provides skilful information on the amplitude of El Niño events, their duration, and the peak time of the sea surface temperature anomalies at a sufficient lead time as to efficiently serve preventive public health actions. The long-lead El Niño predictions were coupled to a statistical dengue model to estimate dengue cases during the 1998 and the 2010 epidemics in El Oro Province in Ecuador, where dengue is hyper-endemic. The dengue model correctly estimated these two largest dengue epidemics even at a 2-year simulation lead time. Thus, information is successfully passed from the El Niño forecast domain to the dengue estimation domain, and the long-lead El Niño predictions are shown to potentially anticipate the magnitude of dengue epidemics in the peak season. The results validate the sensitivity of large dengue epidemics in the region to the El Niño forecasts within the proposed model coupling set-up and imply a potential for increasing lead-time in dengue prediction. This coupled model framework and exploratory analysis, based on El Niño predictions, could be easily extended to other similarly transmitted diseases in tropical and subtropical countries, which are directly and severely affected by the large-scale temperature and precipitation teleconnections occurring before, during and after El Niño events.

Original language	English
Article number	100096
Pages (from-to)	1-9
Number of pages	9
Journal	Climate Services
Volume	15
DOIs	https://doi.org/10.1016/j.cliser.2019.02.003
Publication status	Published - Aug 2019

Funding

The research leading to these results was supported by the DENFREE project (FP7-HEALTH.2011.2.3.3-2; 282378) and EUPORIAS project (FP7-ENV.2012.6.1-1; 308291), funded by the European Commission’s Seventh Framework Research Programme . Many thanks to colleagues in Ecuador for providing the original data used in the analysis: Raul Mejia at the National Institute of Meteorology and Hydrology, Jhonny Real from the Ministry of Health of Ecuador, and Efrain Beltran and other colleagues from the National Vector Control Service (SNEM). RL was supported by a Royal Society Dorothy Hodgkin Fellowship . AMSI was additionally supported by NSF DEB EEID 1518681 and NSF DEB RAPID 1641145. JB gratefully acknowledges funding from the European Union’s Horizon 2020 research and innovation programme under grant agreements No 727852 (project Blue Action), 730004 (project PUCS), and 737480 (Marie Sklodowska-Curie fellowship ACCLIM). SJK acknowledges the support from CREATES (DNRF78) at Aarhus University, Denmark, funded by the Danish National Research Foundation. XR was in receipt of funds by the New Indigo project PARA-CLIM-CHANDIGARGH (Mineco, 2013). The funding sources had no direct involvement in the conduct of the research and the preparation of the article. Appendix A

Funders	Funder number
CREATES	DNRF78
DENFREE	282378, FP7-ENV.2012.6.1-1, 308291, FP7-HEALTH.2011.2.3.3-2
Efrain Beltran
European Commission’s Seventh Framework Research Programme
European Union’s Horizon 2020	Action
Ministry of Health of Ecuador
NSF DEB EEID 1518681
NSF DEB RAPID 1641145	DEB RAPID 1641145
National Vector Control Service
SNEM
National Science Foundation	DEB EEID 1518681, 1641145
Aarhus Universitet
Horizon 2020 Framework Programme
H2020 Marie Skłodowska-Curie Actions
Royal Society
Danmarks Grundforskningsfond
Seventh Framework Programme
Horizon 2020	737480, 727852, 730004

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access to Document

10.1016/j.cliser.2019.02.003

Persistent URL (handle)

Persistent link

Cite this

@article{6d6b36e08cbf41b5987a6c41b1d13ab3,

title = "Sensitivity of large dengue epidemics in Ecuador to long-lead predictions of El Ni{\~n}o",

abstract = "Long-lead forecasts of El Ni{\~n}o events are lacking despite their enormous societal and economic impacts. These climatic events lead to floods and droughts in many tropical regions, and damage agriculture and the economy in poor countries. Due to their impact on local climate, they can also affect human health by increasing the risk of vector-borne diseases, such as dengue fever. Physical processes at the origin of this complex coupled ocean-atmosphere phenomenon are just beginning to be better understood, with subsurface processes and stored heat as two of the main driving forces leading to the development of El Ni{\~n}o in a quasi-periodic manner. Taking advantage of this new knowledge, a statistical dynamic components model, using a state space approach and predictors relevant to the El Ni{\~n}o evolution, was specifically tailored to forecast warm events at lead times of about 2 years (well beyond the traditional spring barrier limit in El Ni{\~n}o predictability). This forecasting scheme provides skilful information on the amplitude of El Ni{\~n}o events, their duration, and the peak time of the sea surface temperature anomalies at a sufficient lead time as to efficiently serve preventive public health actions. The long-lead El Ni{\~n}o predictions were coupled to a statistical dengue model to estimate dengue cases during the 1998 and the 2010 epidemics in El Oro Province in Ecuador, where dengue is hyper-endemic. The dengue model correctly estimated these two largest dengue epidemics even at a 2-year simulation lead time. Thus, information is successfully passed from the El Ni{\~n}o forecast domain to the dengue estimation domain, and the long-lead El Ni{\~n}o predictions are shown to potentially anticipate the magnitude of dengue epidemics in the peak season. The results validate the sensitivity of large dengue epidemics in the region to the El Ni{\~n}o forecasts within the proposed model coupling set-up and imply a potential for increasing lead-time in dengue prediction. This coupled model framework and exploratory analysis, based on El Ni{\~n}o predictions, could be easily extended to other similarly transmitted diseases in tropical and subtropical countries, which are directly and severely affected by the large-scale temperature and precipitation teleconnections occurring before, during and after El Ni{\~n}o events.",

author = "Desislava Petrova and Rachel Lowe and Anna Stewart-Ibarra and Joan Ballester and Koopman, {Siem Jan} and Xavier Rod{\'o}",

year = "2019",

month = aug,

doi = "10.1016/j.cliser.2019.02.003",

language = "English",

volume = "15",

pages = "1--9",

journal = "Climate Services",

issn = "2405-8807",

publisher = "Elsevier B.V.",

}

TY - JOUR

T1 - Sensitivity of large dengue epidemics in Ecuador to long-lead predictions of El Niño

AU - Petrova, Desislava

AU - Lowe, Rachel

AU - Stewart-Ibarra, Anna

AU - Ballester, Joan

AU - Koopman, Siem Jan

AU - Rodó, Xavier

PY - 2019/8

Y1 - 2019/8

N2 - Long-lead forecasts of El Niño events are lacking despite their enormous societal and economic impacts. These climatic events lead to floods and droughts in many tropical regions, and damage agriculture and the economy in poor countries. Due to their impact on local climate, they can also affect human health by increasing the risk of vector-borne diseases, such as dengue fever. Physical processes at the origin of this complex coupled ocean-atmosphere phenomenon are just beginning to be better understood, with subsurface processes and stored heat as two of the main driving forces leading to the development of El Niño in a quasi-periodic manner. Taking advantage of this new knowledge, a statistical dynamic components model, using a state space approach and predictors relevant to the El Niño evolution, was specifically tailored to forecast warm events at lead times of about 2 years (well beyond the traditional spring barrier limit in El Niño predictability). This forecasting scheme provides skilful information on the amplitude of El Niño events, their duration, and the peak time of the sea surface temperature anomalies at a sufficient lead time as to efficiently serve preventive public health actions. The long-lead El Niño predictions were coupled to a statistical dengue model to estimate dengue cases during the 1998 and the 2010 epidemics in El Oro Province in Ecuador, where dengue is hyper-endemic. The dengue model correctly estimated these two largest dengue epidemics even at a 2-year simulation lead time. Thus, information is successfully passed from the El Niño forecast domain to the dengue estimation domain, and the long-lead El Niño predictions are shown to potentially anticipate the magnitude of dengue epidemics in the peak season. The results validate the sensitivity of large dengue epidemics in the region to the El Niño forecasts within the proposed model coupling set-up and imply a potential for increasing lead-time in dengue prediction. This coupled model framework and exploratory analysis, based on El Niño predictions, could be easily extended to other similarly transmitted diseases in tropical and subtropical countries, which are directly and severely affected by the large-scale temperature and precipitation teleconnections occurring before, during and after El Niño events.

AB - Long-lead forecasts of El Niño events are lacking despite their enormous societal and economic impacts. These climatic events lead to floods and droughts in many tropical regions, and damage agriculture and the economy in poor countries. Due to their impact on local climate, they can also affect human health by increasing the risk of vector-borne diseases, such as dengue fever. Physical processes at the origin of this complex coupled ocean-atmosphere phenomenon are just beginning to be better understood, with subsurface processes and stored heat as two of the main driving forces leading to the development of El Niño in a quasi-periodic manner. Taking advantage of this new knowledge, a statistical dynamic components model, using a state space approach and predictors relevant to the El Niño evolution, was specifically tailored to forecast warm events at lead times of about 2 years (well beyond the traditional spring barrier limit in El Niño predictability). This forecasting scheme provides skilful information on the amplitude of El Niño events, their duration, and the peak time of the sea surface temperature anomalies at a sufficient lead time as to efficiently serve preventive public health actions. The long-lead El Niño predictions were coupled to a statistical dengue model to estimate dengue cases during the 1998 and the 2010 epidemics in El Oro Province in Ecuador, where dengue is hyper-endemic. The dengue model correctly estimated these two largest dengue epidemics even at a 2-year simulation lead time. Thus, information is successfully passed from the El Niño forecast domain to the dengue estimation domain, and the long-lead El Niño predictions are shown to potentially anticipate the magnitude of dengue epidemics in the peak season. The results validate the sensitivity of large dengue epidemics in the region to the El Niño forecasts within the proposed model coupling set-up and imply a potential for increasing lead-time in dengue prediction. This coupled model framework and exploratory analysis, based on El Niño predictions, could be easily extended to other similarly transmitted diseases in tropical and subtropical countries, which are directly and severely affected by the large-scale temperature and precipitation teleconnections occurring before, during and after El Niño events.

UR - http://www.scopus.com/inward/record.url?scp=85062157868&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85062157868&partnerID=8YFLogxK

U2 - 10.1016/j.cliser.2019.02.003

DO - 10.1016/j.cliser.2019.02.003

M3 - Article

AN - SCOPUS:85062157868

SN - 2405-8807

VL - 15

SP - 1

EP - 9

JO - Climate Services

JF - Climate Services

M1 - 100096

ER -

Sensitivity of large dengue epidemics in Ecuador to long-lead predictions of El Niño

Abstract

Funding

UN SDGs

Access to Document

Persistent URL (handle)

Other files and links

Fingerprint

Cite this