© 2

01

4,

Nete

ler

et

al. -

htt

p:/

/gis

.cri

.fm

ach

.it/

Tracking emerging diseases from space: Geoinformatics for human

health

Markus Neteler

Joint work withM. Metz, D. Rocchini, M. Marcantonio, A Rizzoli

IFGI symposium - 11 June 2014"Geoinformatics: Solving global challenges?"

© 2

01

4,

Nete

ler

et

al. -

htt

p:/

/gis

.cri

.fm

ach

.it/

Fondazione Edmund Mach, Trento, Italy

● Founded 1874 as IASMA - Istituto Agrario San Michele all'Adige (north of Trento, Italy)

● Research Centre + Tech. Transfer Center + highschool, ~ 800 staff

● … of those 300 staff in research (Environmental research, Agro-Genetic research, Food safety)

http://cri.fmach.eu/

S. Michele all'Adige

© 2

01

4,

Nete

ler

et

al. -

htt

p:/

/gis

.cri

.fm

ach

.it/

Did

You

Ever

See

This

Mosquito?

Photo (and host): M Neteler

© 2

01

4,

Nete

ler

et

al. -

htt

p:/

/gis

.cri

.fm

ach

.it/

Focus on zoonotic diseases

transmitted from animals to humans, usually by a vector (e.g., ticks, mosquitoes)reservoir hosts: wildlife and domestic animalszoonoses involve all types of agents (bacteria, parasites, viruses and others)

Zoonotic diseases cause major health problems in many countries.

They are driven by environmental andpathogen changes as well as political and cultural changes.

The problem: Emerging infectious diseases

http://healthmap.org/en/

© 2

01

4,

Nete

ler

et

al. -

htt

p:/

/gis

.cri

.fm

ach

.it/

http://www.thelancet.com/journals/laninf/article/PIIS1473-3099%2814%2970781-9/abstracthttp://www.ft.com/cms/s/2/dd4ac6c2-eb4b-11e3-bab6-00144feabdc0.html#axzz34GSWNYQU

© 2

01

4,

Nete

ler

et

al. -

htt

p:/

/gis

.cri

.fm

ach

.it/

Infections causing also “hidden” problems ...

© 2

01

4,

Nete

ler

et

al. -

htt

p:/

/gis

.cri

.fm

ach

.it/

JC Semenza and D Domanović, 2013

© 2

01

4,

Nete

ler

et

al. -

htt

p:/

/gis

.cri

.fm

ach

.it/

Yellow fever Dengue

West Nile

Saint Louis encephalitis

Chikungunya

Spread of the tiger mosquito (Aedesalbopictus): infectious disease vectorsand globalization

De Llamballerie et al., 2008: Chikungunya

● Tiger mosquito: Disease vector● Spreads in Europe and elsewhere● Small containers, used tires

and lucky bamboo plants are relevant breeding sites

● >250 cases of Chikungunya innorthern Italy in 2007 (CHIKv imported by India traveler andwas then spread by Ae. albopictus)

© 2

01

4,

Nete

ler

et

al. -

htt

p:/

/gis

.cri

.fm

ach

.it/

Medlock et al. 2013, Parasites & Vectors, 6:1

Distribution of the vector: Ixodes ricinus

Current known distribution of the tick species at ‘regional’ administrative level (NUTS3); based on published historical data and confirmed data provided by experts from the respective countries as part of the VBORNET project.

© 2

01

4,

Nete

ler

et

al. -

htt

p:/

/gis

.cri

.fm

ach

.it/

Medlock et al. 2013, Parasites & Vectors, 6:1

Distribution of the tick Ixodes ricinus: Remote Sensing and Geoinformatics

© 2

01

4,

Nete

ler

et

al. -

htt

p:/

/gis

.cri

.fm

ach

.it/

Environmental factors derived from Remote Sensing

Pathogens

Vectors

Hosts

Temperature

Precipitation

Topography

Land use

Vegetation cover

Moisture / Humidityproxies

(incomplete view)

e.g. SRTM, ASTER GDEM

e.g.MODIS based

e.g.MODIS based

… not covered here

© 2

01

4,

Nete

ler

et

al. -

htt

p:/

/gis

.cri

.fm

ach

.it/

What does remote sensing offer?Requirements: operational systems, regular observations, data access

Example: NASA Terra and Aqua satellites (MODIS sensor, 4 maps per day)

● Land surface temperature (LST)

● late frost periods

● hot summer temperatures

● autumnal temperature decrease

● annual/monthly minima/maxima

● Urban heat islands, …

● Normalized/Enhanced Difference Vegetation Index (NDVI/EVI)

● seasonal differences

● spring/autumn detection

● length of growing season

● Normalized Difference Water Index (NDWI)

● as humidity proxy (?)

● Maximum snow extent (SNOW)

Time series are essential

© 2

01

4,

Nete

ler

et

al. -

htt

p:/

/gis

.cri

.fm

ach

.it/

Temperature in space and time

Temperaturetime series

Average

Minimum

Maximum

Seasonal temperature:Winter, spring, summer, autumn

Spring warming, Autumnal cooling

Anomalies, Cool Night Index

Growing Degree Days (GDD)

Land Surface Temperaturefrom satellite

Late frost periods

Selected references:

● Kilpatrick et al 2011 (WNV transmission)

● ECDC 2009 (Aedes albopictus risk maps)

● Roiz et al 2011(Aedes albopictus distribution map)

● Randolph 2004 (tick seasonality)

● Tersago et al 2009 (Hantavirus)● Rios et al 2000 (Tubercolosis)

● Kalluri et al 2007 (mosquito abundance)

● Epstein et al 2002 (infectious diseases)

● Morand et al 2013 (infectious diseases)

● Peréz-Rodriguez et al 2013 (VB parasites)

© 2

01

4,

Nete

ler

et

al. -

htt

p:/

/gis

.cri

.fm

ach

.it/

EuroLST: MODIS LST daily time seriesExample: Land surface temperature for Sep 26 2012, 1:30 pm

Metz, Rocchini, Neteler, 2014: Rem SensEuroLST: http://gis.cri.fmach.it/eurolst/

Belluno

Reconstructed,i.e. gap-filled data

© 2

01

4,

Nete

ler

et

al. -

htt

p:/

/gis

.cri

.fm

ach

.it/

New EuroLST dataset:Comparison to other datasets(and advantages of using remote sensing time series)

DegreeCelsius

(reconstructed)

Metz, M.; Rocchini, D.; Neteler, M. 2014: Surface temperatures at the continental scale: Tracking changes with remote sensing at unprecedented detail. Remote Sensing. 2014, 6(5): 3822-3840 (DOI | HTML | PDF)

© 2

01

4,

Nete

ler

et

al. -

htt

p:/

/gis

.cri

.fm

ach

.it/

MODIS LST daily time seriesExample of aggregated data

2003: Deviation from baseline average temperature(baseline: 2002-2012)

Aggregated by FEM PGIS

© 2

01

4,

Nete

ler

et

al. -

htt

p:/

/gis

.cri

.fm

ach

.it/

BIOCLIM from reconstructed MODIS LST at 250m pixel resolution

BIO1 BIO2 BIO3 BIO4

BIO5 BIO6 BIO7 BIO10

BIO11BIO1: Annual mean temperature (°C*10)BIO2: Mean diurnal range (Mean monthly (max - min tem))BIO3: Isothermality ((bio2/bio7)*100)BIO4: Temperature seasonality (standard deviation * 100)BIO5: Maximum temperature of the warmest month (°C*10)BIO6: Minimum temperature of the coldest month (°C*10)BIO7: Temperature annual range (bio5 - bio6) (°C*10)BIO10: Mean temperature of the warmest quarter (°C*10)BIO11: Mean temperature of the coldest quarter (°C*10)

Metz, Rocchini, Neteler, 2014: Rem SensEuroLST: http://gis.cri.fmach.it/eurolst/

© 2

01

4,

Nete

ler

et

al. -

htt

p:/

/gis

.cri

.fm

ach

.it/

Examples:

“Hot” year 2003and effects

MODIS Land Surface Temperature

January 2004: Lake Garda still “warm” after hot 2003 summer--> local heating effect = insect overwintering facilitated

Metz, Rocchini, Neteler 2014.

© 2

01

4,

Nete

ler

et

al. -

htt

p:/

/gis

.cri

.fm

ach

.it/

Ae. albopictus winter survival from MODIS LST

Neteler et al., 2011: Int J Health Geogr, 10:49

2009:Positive trapNegative trap

Lakes as “heating” (Jan 2004):

© 2

01

4,

Nete

ler

et

al. -

htt

p:/

/gis

.cri

.fm

ach

.it/

Growing Degree Days from gap-filled MODIS LST

2003

2006

Grey: threshold not reached

Grey: threshold not reached

Number of Day-Of-Year (DOY) to reach 440 accumulated growing degree days (GDD) in the years 2003 and 2006:

● proxy for life-stage survival analysis of insect

● satellite-derived GDD are delivered as map, each pixel is“measured”

Data: EuroLST

440 GDD threshold

© 2

01

4,

Nete

ler

et

al. -

htt

p:/

/gis

.cri

.fm

ach

.it/

Precipitation in space and time

Precipitation Long-term average

Annual sum

Seasonal precipitation:Winter, spring, summer, autumn

Length of dry season / drought

Length of raining season

Anomalies

Selected references:

● Semenza & Menne 2009 (precipitation)

● Kilpatrick et al 2011 (precipitation – WNV transmission)

● Estrada-Peña et al 2008 (seasonal precip.)

● Epstein et al 2001 (seasonalprecip.)

● Reusken & Heyman 2013 (snow - hantavirus)

● Morand et al 2013 (precipitation – Typhoid fever)

● Greer et al. 2008 (precipitation – Trichinosis)

● Tourre et al. 2008 (precipitation – Epidemics)

● Srivatsava et al. 2001 (precipitation – Malaria vectors)

● Reisen et al 2008 (precipitation– mosquito abundance)

Presence of snow

© 2

01

4,

Nete

ler

et

al. -

htt

p:/

/gis

.cri

.fm

ach

.it/

ECA&D: gridded meteo data

Average annual precipitation

mm

ECA&D 25 km (gridded from meteo stations, daily)average annual precipitation (1981-2010)

http://www.ecad.eu/

© 2

01

4,

Nete

ler

et

al. -

htt

p:/

/gis

.cri

.fm

ach

.it/

GPCP V1.2 data: remote sensing based

Average annual precipitation

mm

NASA GPCP 1 degree (satellite based, daily)average annual precipitation (1997-2012)

http://precip.gsfc.nasa.gov/

Globallyavailable

© 2

01

4,

Nete

ler

et

al. -

htt

p:/

/gis

.cri

.fm

ach

.it/

PostGISomics

© 2

01

4,

Nete

ler

et

al. -

htt

p:/

/gis

.cri

.fm

ach

.it/

Moisture/Humidity in time and space

Moisture / humidity proxies

Long-term average

Saturation deficit

Water stress

Tasseled Cap

DWSI (Disease Water Stress Index)

Selected references:

● DWSI: Brown et al. 2008● NDWI: Estallo et al. 2012

● Saturation deficit: Perret et al. 2000● Tasseled cap: Rodgers & Mather 2006

● Hashizume et al. 2008 (low humidity – Gastroenteritis)

● Baylis et al. 1998 (soil moisture– mosquito vectors)

● Kalluri et al 2007 (relative humidity– VB diseases)

NDWI (Normalized Differences Water Index)

© 2

01

4,

Nete

ler

et

al. -

htt

p:/

/gis

.cri

.fm

ach

.it/

NDWI in time and spaceNDWI (Normalized Differences Water Index): June 2011

Venice

Israel

Madrid

NDWI time series

Derived fromMODIS VIS

Humidity proxy:Usability yetto be verified!

Aggregated by FEM PGIS

© 2

01

4,

Nete

ler

et

al. -

htt

p:/

/gis

.cri

.fm

ach

.it/

Emerging concern: West Nile Virus spread in Europe

To better address the ongoing spread of WNVin Europe, there is

● a need for an early warning system of potential outbreaks which is crucial in order to timely raise the awareness of the clinicians and speed up diagnosis to implement the blood safety regulation

Specifically, we need● to identify predictors of WNV circulation and outbreaks● Modelling: to consider the continental scale in order to

apply these predictors across Europe and neighbouring countries

© 2

01

4,

Nete

ler

et

al. -

htt

p:/

/gis

.cri

.fm

ach

.it/

WNV in Europe: complicated pattern

Marcantonio et al. (in prep.)

© 2

01

4,

Nete

ler

et

al. -

htt

p:/

/gis

.cri

.fm

ach

.it/

MethodologyWe are testing the association between West Nile Fever incidence and a wide range of potential predictors:

● including temperature data, land use, human presence, urbanization, water body density, landscape fragmentation and heterogeneity, protected areas

● avoiding weak interpolation methods from sparse point data by use of spatially continuous input data

● Use of multi-model inference to gain a consensus from multiple linear mixed models predicting WNV incidence at a scale of NUTS3/GAUL1 administrative units

Anomalies from LST PGIS HPC Precipitation NDWI Biomes: Anthromes

Globcover: land useNDVIMarcantonio et al. (in prep.)

© 2

01

4,

Nete

ler

et

al. -

htt

p:/

/gis

.cri

.fm

ach

.it/

Markus NetelerFondazione E. Mach (FEM)Centro Ricerca e InnovazioneGIS and Remote Sensing Unit38010 S. Michele all'Adige (Trento), Italyhttp://gis.cri.fmach.ithttp://www.osgeo.org

Conclusions

● Emerging diseases need to be considered among the “emerging themes” to be covered by integrated research strategies because of their dramatic impact on well being and economy

● Current and potential distribution of disease vectors (like Ae. albopictus) can be modelled at high resolution, relevant to many health projects

● New reconstructed high temporal resolution datasets allow for real modelling

● ... bring it all together in Geoinformatics!