The role of citizen science in environment management

Luigi Ceccaroni

The role of citizen science in environment management

The role of citizen science

in environment management

Forecasting based on

citizens’ observatories

Luigi Ceccaroni (1000001 Labs)

3rd International Marine Protected Areas Congress, Marseille (France), October 22, 2013 Università di Bologna – Scienze Ambientali, Ravenna (Italy), December 17, 2015, 2 pm

1

Luigi Ceccaroni


• Knowledge of the field of citizen science and current

trends that influence it

• Understand the principles and practical aspects of a

citizen science project

• Learn about data quality in citizen science

Learning outcomes 2

Luigi Ceccaroni


• Current activities in the area of citizen science online

and offline

• Typology of engagement in citizen science

• Citizen Science in a historical perspective and

underlying trends

• The case for low resolution in environmental

monitoring

• Forecasting based on citizens’ observatories

Content: an introduction to citizen science 3

Luigi Ceccaroni


Acquisition, processing, delivery: a new way 4

Luigi Ceccaroni


• Citizen science involving people on training cruises, on

sailing races, on holiday (e.g., scuba diving) in data

collection

Citizens’ participation via smart devices

Tall Ship Regatta

21st Sept. - 4th October 2013:

Barcelona - Toulon - La Spezia

5

Luigi Ceccaroni



Barcelona World

Race

2013-2015: around

the world sailboat

race



collection

6

Luigi Ceccaroni





collection

7

Luigi Ceccaroni


Information acquisition 8

Luigi Ceccaroni


Information delivery (Citclops) 9

Luigi Ceccaroni



Luigi Ceccaroni


Information delivery (WaterWatch) 11

Luigi Ceccaroni


Information delivery (WaterWatch) 12

Luigi Ceccaroni



Luigi Ceccaroni


• Water-color, oil-spill, river-status measurement

• Water transparency via phone pictures and

Secchi disc

• Retrieval of sensor measurements from low-

cost moorings

Applications 15

Luigi Ceccaroni


• Improvement of scuba-diving activities

• Best beaches’ ranking

• Early-warning systems for HABs and bio-

chemical hazards

Applications 16

Luigi Ceccaroni


• Andromeda

• Citclops

• CITI-SENSE

• Coastwatch

• COBWEB

• SeaWatchers

• Socientize

• WaterWatch

• WeSenseIt

European citizen-science projects: other examples 17

Luigi Ceccaroni


Coastwatch and Citclops 18

Luigi Ceccaroni


SeaWatchers 19

Luigi Ceccaroni


COBWEB, CITI-SENSE, WeSenseIt, Omniscientis 20

Luigi Ceccaroni


Andromeda project 21

Luigi Ceccaroni


Andromeda project 22

Luigi Ceccaroni


citizen science n. scientific work undertaken by

members of the general public, often in collaboration

with or under the direction of professional scientists and

scientific institutions.

citizen scientist n. (a) a scientist whose work is

characterized by a sense of responsibility to serve the

best interests of the wider community (now rare); (b) a

member of the general public who engages in scientific

work, often in collaboration with or under the direction

of professional scientists and scientific institutions; an

amateur scientist.

Citizen Science (Oxford English Dictionary 2014) 23

Luigi Ceccaroni


Indicators (of environment quality/perturbation) can be

measured by citizens in a more or less

voluntary,

active or

conscious way.

Typology of engagement in citizen science 24

Luigi Ceccaroni


• Involvement of citizens in: archaeology, astronomy,

ornithology, conservation, meteorology…

• No recognition, viewing volunteers as ‘untrustworthy’

contributors, that are better replaced by automated

instruments

Historical perspective: the era of professional

science (late 1800s – 1900s)

25

Luigi Ceccaroni


• Societal trends:

• Education and qualifications

• Leisure

• Sharing economies / peer production systems

• Technological trends:

• Internet access (broadband)

• Mobile devices

• Collaborative Web

• DIY electronics

Historical perspective: the era of opening science

(2000s)

26

Luigi Ceccaroni


• World (data by UNESCO): 95 M – 199 M

Increased level of education and qualifications 27

Luigi Ceccaroni


• Italy (data by UNESCO): 1.8 M - 2.0 M - 1.9 M

• Spain (data by UNESCO): 1.8 M - 1.8 M - 2.0 M

Increased level of education and qualifications 28

Luigi Ceccaroni


• Italy (data by OECD): 1850 - 1730

• Spain (data by OECD): 1750 - 1690

Increased level of leisure

1650

1700

1750

1800

1850

1900

1950

2000

2050

2000 2002 2004 2006 2008 2010 2012 2014 2016

Israel

Italy

Spain

OECD countries

Average Annual Hours Worked (per person)

29

Luigi Ceccaroni


• In many areas, especially in production and sharing of

information

Increased level of sharing economies 30

Luigi Ceccaroni


Increased (fixed) broadband affordability

Data by the International

Telecommunication Union

(ITU)

31

Luigi Ceccaroni


Increased (mobile) broadband affordability

Data by the International Telecommunication Union (ITU)

32

Luigi Ceccaroni


Increased mobile-subscription affordability

Data by the International Telecommunication Union (ITU)

33

Luigi Ceccaroni


Collaborative Web A map of the world, created by people like you and free to use under an open license

34

Luigi Ceccaroni


DIY electronics

Citclops’s KdUINO monitoring buoy based on Arduino

35

Luigi Ceccaroni


The case for low resolution 36

with Jaume Piera

Luigi Ceccaroni


Monitoring: resolution, accuracy μM = 10−6 mol/dm3 = 10−3 mol/m3

In situ ultraviolet spectrophotometer to

measure nutrients

Temporal and spatial variability?

High accuracy

37

Luigi Ceccaroni


Monitoring: frequency

120 100 90 80 130 140 110 150

20

15

5

10

25

0

Nitrate (μM)

Julian days

1 sample every month

(at 10 am on the first day of the month)

1 sample every week

(at 10 am on the first day of the week)

Monterey Bay, California

38

Luigi Ceccaroni


The sunlight cycle

30 35 40

0

50

100

150

200

250

Julian days

Sunlight intensity (W/m2h)

Day

Night

39

Luigi Ceccaroni



30 35 40

0

50

100

150

200

250

Julian days


1 sample every week (at 4 pm on the first day)

This is a typical frequency in

biological domains

40

Luigi Ceccaroni



30 35 40

0

50

100

150

200

250

Julian days


1 sample every day (at 4 pm)

41

Luigi Ceccaroni



30 35 40

0

50

100

150

200

250

Julian days


2 samples every day (at 4 pm and 4 am)

42

Luigi Ceccaroni


1.very high

2.high

3.low

4.very low

5.no sunlight

Monitoring: resolution

vs.

Questionnaire to citizens

43

Luigi Ceccaroni


Monitoring: resolution

0

50

100

150

200

250

Julian days


1 sample every day (at 4 pm) Sunlight intensity

25 27 29 31 33 35 37 39 41

very high

high

low

very low

no sunlight

44

Luigi Ceccaroni


Forecasting based on citizens’ observatories 45

with Filip Velickovski

Luigi Ceccaroni


• Extract data from source

• Convert to data array

Parsing

• Convert data into feature vectors + target value

Feature preparation • Select feature

variables, amount of past time points

Feature Selection

• Train and test a set of models (decision trees, SVM, Bayesian)

Model selection • Evaluate best

model on unseen test set

Final evaluation

Machine-learning pipeline 47

Luigi Ceccaroni


Converting to three class classification problem 48

Luigi Ceccaroni


Using mean daily values (example: TSM)

6th May 30th Apr 24th May 12th May 18th May 2014

Luigi Ceccaroni


Using mean daily values, sometimes few data points

(example: TSM)

50

18th Dec 14th Dec 22nd Dec 26th Dec

2013

Luigi Ceccaroni


Short-term correlation to tide (Dominant semi-diurnal

lunar tidal component (M2) with a period of 12:25 h)

51

28th Sep 1st Oct 4th Oct 2013

Luigi Ceccaroni


Knowledge coming from pre-processing 52

• Clear relation among color, tides, and other drivers:

wind (which steers both waves and currents),

river/sluice discharge and biology

• Short-term forecasting (within a tidal period): not

possible without de-tiding the signals

• Longer-term forecasting: possibly improved by de-

tiding the signals

Luigi Ceccaroni


Feature and target-variable preparation

Luigi Ceccaroni


FU

Chl-a

TSM

Feature configuration

Number of examples for training

ML technique

Support Vector Machine • C (soft-margin) : 1.0 • Kernel : radial basis

Distribution of target FU classes

415

stable

decrease increase

27% 48%

Blind predictor benchmark

48%

25%

Accuracy (10-fold cross- validation)

53%

Wave height

Forecasting 54

Luigi Ceccaroni


FU

Chl-a

TSM



ML technique

Random Forest • n_trees: 10

359

33 % 32 %

35%

35 % 52 %

Wave height

55



stable

decrease increase Accuracy (10-fold cross- validation)

Forecasting

Luigi Ceccaroni


FU

Chl-a

TSM



ML technique

Decision Tree • max depth: 10

345

34 % 31 %

35%

35 % 45 %

Wave height

56



stable

decrease increase Accuracy (10-fold cross- validation)

Forecasting

Luigi Ceccaroni


http://www.1000001labs.org/

http://www.citclops.eu/

http://eyeonwater.org/

http://ecsa.citizen-science.net/

http://www.coastwatch.org/

http://creekwatch.researchlabs.ibm.com/

http://www.ecy.wa.gov/programs/eap/fw_riv/index.html

http://projectbaseline.org

http://crowd.cs.umass.edu/

https://it.wikipedia.org/wiki/Citizen_science

http://senseable.mit.edu/

http://research.cens.ucla.edu/aquatic/

http://www.secchidipin.org/

http://www.earthobservations.org/geo_me_201211_geo9_ec.shtml

http://www.nurp.noaa.gov/Spotlight/Observatory.htm

Resources 57

Luigi Ceccaroni


58

The role of citizen science in environment management

Environment

Transcript of The role of citizen science in environment management