Wildfires, power outages, and vulnerability in California
Transcript of Wildfires, power outages, and vulnerability in California
Mar2019
Apr May Jun Jul Aug Sep Oct Nov Dec Jan2020
1000
3000
10000
30000
1000
3000
10000
30000
Area burned (acres)
Wildfires, power outages, and vulnerability
in California
Mathew Kiang, ScDDepartment of Epidemiology and Population Health Stanford University School of Medicine
Shenyue Jia, PhD Center for Excellence in Earth Systems Modeling and Observations Chapman University
Mar2019
Apr May Jun Jul Aug Sep Oct Nov Dec Jan2020
1000
3000
10000
30000
Area burned (acres)
Area burned (acres)1000
3000
10000
30000
California Fires in 2019
(Source: CAL FIRE Red Book 2019)
Mar2019
Apr May Jun Jul Aug Sep Oct Nov Dec Jan2020
1000
3000
10000
30000
Area burned (acres)
Area burned (acres)1000
3000
10000
30000
Two Largest Fires in 2019
(Source: CAL FIRE Red Book 2019)
Kincade Fire(77,758)
Walker Fire(54,608)
Mar2019
Apr May Jun Jul Aug Sep Oct Nov Dec Jan2020
1000
3000
10000
30000
Area burned (acres)
Area burned (acres)1000
3000
10000
30000
Mar2019
Apr May Jun Jul Aug Sep Oct Nov Dec Jan2020
NearbyHouseholds
50000
100000
150000
200000
1000
3000
10000
30000
Acres burned (acres)
Fires that threatened households
(Source: CAL FIRE Red Book 2019)
Saddle Ridge Fire(204,503)
Getty Fire(190,804)
Mar2019
Apr May Jun Jul Aug Sep Oct Nov Dec Jan2020
NearbyHouseholds
50000
100000
150000
200000
1000
3000
10000
30000
Acres burned (acres)
Getty Fire
(Source: CAL FIRE Red Book 2019)
Mar2019
Apr May Jun Jul Aug Sep Oct Nov Dec Jan2020
NearbyHospitals
0
1
2
3
4
5
6
1000
3000
10000
30000
Acres burned (acres)
Fires that threatened hospitals
(Source: CAL FIRE Red Book 2019)
Hill Fire(6)
Power outages can be deadly
(Source: Casey, Fukurai, Hernandez, Balsari, & Kiang in Current Environmental Health Reports 2020)
• Carbon monoxide poisoning
• Gastrointestinal illness
• Heart attack• Injury• Mental health• Renal disease• Temperature-related
illness
Number of large power outages in the US, 2000–2016
0
1
2
3
Jan2019
Mar May Jul Sep Nov Jan2020
Custo
mer-h
ours
witho
ut po
wer (
in mi
llions
)
Alameda
Contra Costa
Los Angeles
San Diego
San Francisco
Santa Clara
Power outages for selected counties
(Data: PowerOutages.us)
0
1
2
3
Jan2019
Mar May Jul Sep Nov Jan2020
Custo
mer-h
ours
witho
ut po
wer (
in mi
llions
)Power outages are common
(Data: PowerOutages.us)
0
1
2
3
Jan2019
Mar May Jul Sep Nov Jan2020
Custo
mer-h
ours
witho
ut po
wer (
in mi
llions
)Rolling blackouts in LA
(Data: PowerOutages.us)
Los Angeles Rolling Blackouts
0
1
2
3
Jan2019
Mar May Jul Sep Nov Jan2020
Custo
mer-h
ours
witho
ut po
wer (
in mi
llions
)Public safety power shut-offs
(Data: PowerOutages.us)
Public Safety Power Shut-offs
Los Angeles Rolling Blackouts
0
1
2
3
Jan2019
Mar May Jul Sep Nov Jan2020
Custo
mer-h
ours
witho
ut po
wer (
in mi
llions
)
(Data: PowerOutages.us)
Public Safety Power Shut-offs• 115 PSPS “events” • 45 counties• Affected 358 county-days• 110 million household-days
Los Angeles Rolling Blackouts
Public safety power shut-offs
0
100
200
300
400
500
600
2h 4h 8h 12h 1d 3d 10dPower outage duration
Numb
er of
outag
e eve
nts
Alameda
Contra Costa
Los Angeles
San Diego
San Francisco
Santa Clara
(Data: PowerOutages.us)
Four Questions1. How frequent are medically-relevant power outages?
0
100
200
300
400
500
600
2h 4h 8h 12h 1d 3d 10dPower outage duration
Numb
er of
outag
e eve
nts
Alameda
Contra Costa
Los Angeles
San Diego
San Francisco
Santa Clara
(Data: PowerOutages.us)
Four Questions1. How frequent are medically-relevant power outages?2. Where are medically-relevant power outages occurring?
0
100
200
300
400
500
600
2h 4h 8h 12h 1d 3d 10dPower outage duration
Numb
er of
outag
e eve
nts
Alameda
Contra Costa
Los Angeles
San Diego
San Francisco
Santa Clara
(Data: PowerOutages.us)
Four Questions1. How frequent are medically-relevant power outages?2. Where are medically-relevant power outages occurring?3. Where are medically-vulnerable populations (in relation
to where outages are occurring)?
0
100
200
300
400
500
600
2h 4h 8h 12h 1d 3d 10dPower outage duration
Numb
er of
outag
e eve
nts
Alameda
Contra Costa
Los Angeles
San Diego
San Francisco
Santa Clara
(Data: PowerOutages.us)
Four Questions1. How frequent are medically-relevant power outages?2. Where are medically-relevant power outages occurring?3. Where are medically-vulnerable populations (in relation
to where outages are occurring)?4. What are the county-level characteristics associated
with medically-relevant power outages?
0
100
200
300
400
500
600
2h 4h 8h 12h 1d 3d 10dPower outage duration
Numb
er of
outag
e eve
nts
Alameda
Contra Costa
Los Angeles
San Diego
San Francisco
Santa Clara
(Data: PowerOutages.us)
Events affecting at least 500 households
0
100
200
300
400
500
600
2h 4h 8h 12h 1d 3d 10dPower outage duration
Numb
er of
outag
e eve
nts
Alameda
Contra Costa
Los Angeles
San Diego
San Francisco
Santa Clara
(Data: PowerOutages.us)
Lasting between 2h and 10d
0
100
200
300
400
500
600
2h 4h 8h 12h 1d 3d 10dPower outage duration
Numb
er of
outag
e eve
nts
Alameda
Contra Costa
Los Angeles
San Diego
San Francisco
Santa Clara
Large events by duration
(Data: PowerOutages.us)
0
100
200
300
400
500
600
2h 4h 8h 12h 1d 3d 10dPower outage duration
Numb
er of
outag
e eve
ntsLarge events by duration
(Data: PowerOutages.us)
0
100
200
300
400
500
600
2h 4h 8h 12h 1d 3d 10dPower outage duration
Numb
er of
outag
e eve
ntsMedically-relevant events: 8h and 1d
(Data: PowerOutages.us)
0
100
200
300
400
500
600
2h 4h 8h 12h 1d 3d 10dPower outage duration
Numb
er of
outag
e eve
ntsNearly 1,100 8h events
(Data: PowerOutages.us)
1,096 events across56 counties accounting for
111 million household-days
0
100
200
300
400
500
600
2h 4h 8h 12h 1d 3d 10dPower outage duration
Numb
er of
outag
e eve
nts149 events lasting at least 1 day
(Data: PowerOutages.us)
149 events across45 counties accounting for8.2 million household-days
1,096 events across56 counties accounting for
111 million household-days
0
100
200
300
400
500
600
2h 4h 8h 12h 1d 3d 10dPower outage duration
Numb
er of
outag
e eve
nts
(Data: PowerOutages.us)
8h
Spatial distribution of large events1d
0 1 5 10 20 40 80
Number of outages
Continuous 8 hours Continuous 24 hours
0 10000 20000 30000 0 10000 20000 300000
5
10
15
20
25
0
20
40
60
80
Number of DME users
Numb
er of
powe
r outa
ges (
>500
hous
ehold
s)Highly vulnerable populations
(Data: PowerOutages.us and HHS emPOWER)
Number of MedicareDME users
a
aa
10000
20000
30000
10000
20000
30000
Number of MedicareDME users
Continuous 8 hours Continuous 24 hours
0 10000 20000 30000 0 10000 20000 300000
5
10
15
20
25
0
20
40
60
80
Number of DME users
Numb
er of
powe
r outa
ges (
>500
hous
ehold
s)Highly vulnerable populations
(Data: PowerOutages.us and HHS emPOWER)
Number of MedicareDME users
a
aa
10000
20000
30000
10000
20000
30000
Number of MedicareDME users
Los Angeles (33,839)
Continuous 8 hours Continuous 24 hours
0 10000 20000 30000 0 10000 20000 300000
5
10
15
20
25
0
20
40
60
80
Number of DME users
Numb
er of
powe
r outa
ges (
>500
hous
ehold
s)Highly vulnerable populations
(Data: PowerOutages.us and HHS emPOWER)
Number of MedicareDME users
a
aa
10000
20000
30000
10000
20000
30000
Number of MedicareDME users
Riverside (13,818)
Orange (11,047)
San Bernadino(11,722)
Los Angeles (33,839)
R = 0.29, p = 0.028 R = 0.24, p = 0.07
Continuous 8 hours Continuous 24 hours
1 2 3 4 1 2 3 40
5
10
15
20
25
0
20
40
60
80
US Forest Service Average Wildfire Hazard Potential Score
Num
ber o
f pow
er o
utag
es (>
500
hous
ehold
s)Correlated with wildfire potential
R = −0.074, p = 0.58 R = −0.18, p = 0.18
Continuous 8 hours Continuous 24 hours
40000 60000 80000 100000 120000 40000 60000 80000 100000 1200000
5
10
15
20
25
0
20
40
60
80
Median household income
Numb
er of
powe
r outa
ges (
>500
hous
ehold
s)Not correlated with income
(Data: PowerOutages.us and HHS emPOWER)
R = 0.03, p = 0.83 R = −0.1, p = 0.46
Continuous 8 hours Continuous 24 hours
20 40 60 80 20 40 60 800
5
10
15
20
25
0
20
40
60
80
Average HPI Score (higher is better)
Numb
er of
powe
r outa
ges (
>500
hous
ehold
s)Not correlated with social vulnerability
(Data: PowerOutages.us and HHS emPOWER)
R = −0.33, p = 0.011 R = −0.36, p = 0.0061
Continuous 8 hours Continuous 24 hours
0.3 0.5 0.7 0.9 0.3 0.5 0.7 0.90
5
10
15
20
25
0
20
40
60
80
Fraction BIPOC
Num
ber o
f pow
er o
utag
es (>
500
hous
ehold
s)Correlated with less diverse counties
(Data: PowerOutages.us and HHS emPOWER)
Next Steps
• Sub-county level analyses• Hospitalization and mobility data• Creating a dashboard with
historic and near-real time data• Shelter and hospital locations• Long-term care facilities• Evacuation notices and mobility
0 1 5 10 20 40 80
Number of outages
8h 1d
Additional Slides
0
100
200
300
400
500
600
2h 4h 8h 12h 1d 3d 10dPower outage duration
Numb
er of
outag
e eve
nts
(Data: PowerOutages.us)
8h
Distribution of households impacted1d
0.2 1.7 16.7 166.7
Number of customer-hours (in millions)