REACCTINGRESEARCH ON EMISSIONS AIR QUALITY, CLIMATE,

AND COOKING TECHNOLOGY IN NORTHERN GHANA

Dickinson et al., BMC Public Health (2015) 15:126

DOI 10.1186/s12889-015-1414-1

1

Investigators

National Center for Atmospheric Research• Christine Wiedinmyer, Katie Dickinson, Mary Hayden, Andrew

Monaghan, Dan Steinhoff, Scott Archer-Nicholls, Isaac Rivera

University of Colorado-Boulder• Mike Hannigan, Ricardo Piedrahita, Nick Masson, Evan Coffey, Didier

Muvandimwe, Conner Jacobsen

• Vanja Dukic, Yolanda Cecile-Hagar, Alexia Newgord, Sam Hsu

Navrongo Health Research Centre• Abraham Oduro, Ernest Kayomanse, Dominic Anaseba, James Adoctor,

Rex Alirigia, Achazanga Manies, Awine Rockson, Victor Adoctor

Relief International/Gyapa Enterprises• MacKenzie Dove, Atsu Titiati

2

NSF funded, GEO-1211668 EPA funded, RD-8354201

REACCTINGResearch of Emissions, Air Quality, Climate, and

Cooking Technologies in Northern Ghana

EMISSIONS

CO

PM

VOC

Ambient Air Quality

Regional Climate

Personal

Exposure

Health Outcomes

3

Why Ghana?• Strong ties with Navrongo Health Research Center

(NHRC) from prior research

• Poverty affects the Northeast region

• Mostly cook with biomass

• Safe and stable work environment

4

Study Area

Region: Upper East

District: K-N

Town: Navrongo

5

Traditional Cooking

• 3-stone stove

• Biomass for fuel

6

7

REACCTING

Intervention Study

Sources and

Exposure

Characterization

rural households (200)

urban and rural regions

(50 households)

8

Intervention Study Design

200

households

randomly

selected

from

Kassena-

Nankana

District in

Northern

Ghana

Baseline

household

survey

conducted

in all

households

Households

attend stove

distribution

meetings

and

randomly

draw into 4

intervention

groups

Follow-up surveys

conducted periodically

in all 200 households

Stove emissions,

personal exposure, and

microenvironment air

quality sampling

conducted on a rotating

basis in a subset of

households

November / December 2013 December 2013 – 2015

9

Intervention Groups

• 200 households, each with one woman

aged 18-55, one child aged 0-5

Group A: Group B:

Group C: Group D:

10

Kassena-Nankana District Snapshot

Variable K-N District

(N, E, S, W

Regions)

Intervention

Study

Sample

P-value

# households 25,458 200

Ethnicity: Kasem 51.5% 50.0% 0.68

Ethnicity: Nankam 44.5% 45.5% 0.79

Cooking fuel: Biomass 84.3% 100% 0.000***

Location: Rural 90.0% 100% 0.000***

Water source: Borehole 86.4% 100% 0.000***

Sanitation: No facility 95.5% 96.5% 0.32

Has Electricity 13.1% 6.0% 0.003***

Owns Motorcycle 14.9% 10.0% 0.049**

Owns Bike 77.5% 87.0% 0.001***

Owns Mobile phone 69.0% 81.0% 0.000***

# Livestock 7.07 10.9 0.000***

11

Measurement Approach• Surveys

• Cooking behavior

• Health burden

• Emission

• Field Portable Emission Monitor for cookstoves

• Other sources

• Ambient

• Regional low-cost monitors

• Reference monitors at NHRC

• Household

• Air (CO and PM2.5)

• Stove use

• Personal

• Air (CO and PM2.5)

• Location

12

http://mobilesensingtechnology.com/

Fourth survey pretest: Dominic (Nankai)

13

Behavioral Surveys

14

18

29

84 2

46

69

12

5

4

1

1

1

0

10

20

30

40

50

60

70

80

90

1 2 3 4 5

Nu

mb

er

of

Ho

us

eh

old

s

Total Number of Stoves in Household at Baseline

Coal + OtherCoal OnlyTraditional + Coal + OtherTraditional + CoalTraditional Only

Types of Stoves

19

households

(10% of

sample)

76

households

(38% of

sample)

81

households

(41% of

sample)

16

households

(8% of

sample)8

households

(4% of

sample)

15

What dishes were cooked yesterday on each

stove?

0% 20% 40% 60% 80% 100%

Three Stone

Charcoal

Percent of All Dishes Cooked on Each Type of Stove

Rice

Beans / Bombara beans

TZ

Tubani

Vegetables / Soup

Tubers

Corn / Maize

Porridge

Other

TZ is primarily

cooked on

Three Stone

stoves

Households use

charcoal stoves to

cook vegetables / soup

16

Health MeasuresFor Respondent and Children Under Five

• Self-Reported Illness

• Blood Spots

• Anthropometrics:

• Height

• Weight

• Arm circumference

17

Air Quality Measurement Timeline18

NHRC Reference Measurements

NO/NOx

CO

O3

CO2

PM2.5

Regional G-Pods

CO

NO

CO2

VOCs

Household Measurements

CO

CO2

PM2.5

SUMs

PATS/Beacon

Rural Personal Measurements

CO

PM2.5

Location

Urban Personal Measurements

CO

PM2.5

Location

2014

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

2015

Jan Feb Mar

2013

Sep Oct Nov Dec

NO/NOx

CO

O3

CO2

PM2.5

2014

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov

Dec

2015

Jan Feb

Mar

2013

Sep Oct Nov

Dec

NHRC Reference Measurements

• NO/NOx [2B Tech 401/410]

• CO [Thermo 48]

• O3 [2B Tech 202]

• CO2 [LiCOR 840]

• PM2.5 [6.5 LPM URG cyclone]

19

Ambient Air Quality at NHRC

O3

ppb

10

15

20

25

30

35

CO

ppm

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

CO2

ppm

375

380

385

390

395

400

405

410

415

420

425

20

Regional G-Pods

21

2014

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov

Dec

2015

Jan Feb

Mar

2013

Sep Oct Nov

Dec

Household Measurements

• Stove usage monitors (SUMs)

• 110 SUMs over 35 households

• CRECER/RESPIRE studies

• Proposed standard methods and algorithms for SUMs in Ruiz-Mercado

et al., 2012

• Microenvironment monitoring

• CO, PM2.5, CO2

22

CO

CO2

PM2.5

SUMs

PATS/Beacon

Household Measurements

23

Example SUMs DataoC

Delta T

(oC

/min

)

24

Ruiz-Mercado, I. et al., Biomass Bioenergy 2012.

25

Peak cooking time distribution by study arm

26

Stove Use Over Time

Assessing cooking event detection algorithm

• Develop a test data set to assess cooking event detection

• 8+ week of thermocouple in fire box + SUMs on 5-7 stoves

27

SUMs Stacking Results

28

G = Gyapa

T = 3-stone

P = Philips

29

G1

G2 3

G1,3

G2,3

SUMs Stacking Results

G1 G2

fraction

of

da

ys u

sed

G1,G

2

G1

P1 3

G1,3

P2,3

G1,P

1

G1 P1

fraction o

f days u

sed

P1

P2 3

P1,3

P2,3

P1,P

2

P1 P2

fraction o

f days u

sed

31

31,3

2

32

3fr

action o

f days u

sed

Personal Measurements• 48-hr personal exposure

monitoring periods rotating

through study households

• 4 families/households

monitored from each study

arm per deployment period

• 2 deployment periods per

week

• Mon-Wed, Wed-Fri

• Monitoring

• cooks

• school children

• children under 5

• available males

30

Personal MeasurementsCarbon monoxide (CO)

• Real-time CO loggers worn by

participants

– Around necks or in “child-proof” pockets

• Lascar EL-USB (electrolytic)

– 1-minute logging interval

– Range: 0-300 ppm

– 0.5 ppm resolution

• Battery operated

• Calibrated weekly at NHRC

31

Personal Measurements and Methods

Particulate Matter (PM2.5) Sampling• Monitors worn in small backpacks

or fanny packs

• SKC pump running at 2 LPM• Calibrated biweekly

• 25mm quartz filters with URG

impactor upstream

• Battery powered to last over 48

hours

• Filters returned to US for EC/OC

analysis

32

Measurements Stats

Personal PM Personal CO

Analyzed Sampling

IntervalNov 2013 – May 2014 Nov 2013 – July 2014

N Individuals

70 total

18 children

16 males

266 adults

112 children

from 118 households

(over 1.2 million

minutes)

Analysis TypeEC/OC Sunset

AnalyzerMatlab code

FutureOver 100 more to

analyze

Over 500,000 more

minutes to analyze

33

Personal EC/OC of PM2.5

µ µ

Under

5 y

r

Over

5 y

r

Fem

ale

Male

Harm

attan

Oth

er

Under

5 y

r

Over

5 y

r

Fem

ale

Male

Harm

attan

Oth

er

Under

5 y

r

Over

5 y

r

Fem

ale

Male

Harm

attan

Oth

er

EC/OCOC EC

µg/m

3

µg/m

3

µg/m

3

34

Example CO personal exposure

35

Personal CO Doses

G/PP/PG/G 3-stone Female Male Over 5 yr Under 5 yr

Dose (

ppm

-min

per

48 h

r)36

Personal CO Doses

G/PP/PG/G 3-stone Female Male Over 5 yr Under 5 yr

Dose (

ppm

-min

per

48 h

r)37

Why aren’t we seeing differences in CO exposure?

Personal CO Doses

G/PP/PG/G 3-stone Female Male Over 5 yr Under 5 yr

Dose (

ppm

-min

per

48 h

r)38

Why aren’t we seeing differences in CO exposure?

No difference

Which stove is actually being used

Other CO sources driving dose

Improve link between CO exposure and stove use

• Additional microenvironmental monitoring

• UCB PATS

• Personal monitoring with iBeacons

39

Apportioning CO exposureR

SS

sig

nal str

ength

Cook proximity to Gyapa

Cook proximity to 3 stone

Cook CO exposure

40

Where do we want to end up …

41

COdose

= Σ functioni(stovei T(t), proximity to stovei(t)) + COother sources

derive via regression

i

measure using SUMs

measure using beacon

Estimate …

PM2.5 too

Cooking Emissions Measurements

42

CCanopy

Stove

Sensor

Box

Fan

Evaluation of Stove Performance

• Using Controlled Cooking Test

• Efficiencies

• Heat Transfer Efficiencies ( HTE) and

Combustion Efficiency (FCE)

• Overall Thermal Efficiency (OTE)

• Emission Factors

• CO

• NO and NO2

• total VOCs

• PM2.5 (cumulative)

• speciation of PM2.5

Jetter et al. (2012)

43

Sampling Procedures/Techniques

44

• Be at household before evening cooking starts

• Let cook follow her normal practices

• Set up PEMs and measure emissions during

the entire cooking period (2-4 hours)

• PM collection on 90 mm quartz-filters for

integrated sample

• Gas phase emissions (CO, CO2, NO, NO2,

TVOCs)

• Measure the weight of the fuel used and food

cooked

Total samples collectedGyapa Phillips Three-Stone Total

North 4 6 5 15

South 4 3 3 10

East 4 4 4 12

West 4 4 4 12

Residential/Total 14 15 14 49

As of 02/22/2015

Typical Emission Factor (mass of pollutant per mass of fuel used)• Emissions= Function (stove type, fuel type, fuel moisture content, ambient T, food type)

Secondary analysis• Task based emission factors (mass of pollutant per mass of food cooked)

• Heat transfer efficiency in the field

45

Initial PM2.5 Results

46

January

2014 Data

Collected

OC

EC

Em

issio

n F

acto

r (g

/kg o

f fu

el)

EC/OC ratio

Other source sampling

24 events to date

Trash burning

Traffic

Pito brewing

Commercial cooking

47

A Norwegian an American and a Rwandan are sitting at a campfire in the middle

of a Boulder, CO parking lot.

The Rwandan has two pairs of pants, one fleece, a winter jacket and a beanie on…

whileThe Norwegian has shorts and a light

sweater on and is asking for sunscreen…while

The American is wondering what fast food restaurant to swing by for dinner!

The Pleasures of Cookstove Assessment....

Research translates to engaging class.

Questions?

49