Quantifying Climate Change for Stormwater / Wastewater ... Approaches to Quantify Climate... · OHM-ADVISORS.COM Quantifying Climate Change for Stormwater / Wastewater Collection

OHM-ADVISORS.COM

Quantifying Climate Change for

Stormwater / Wastewater

Collection Systems

Charles Humphriss

OHM Advisors

MWEA Conference

December 3, 2014

ARCHITECTS. ENGINEERS. PLANNERS.

• Long design life of infrastructure

• Increasing precipitation

observed across the Midwest

• Models predict increases in both

average and extreme rainfall

• Statistics are changing

OHM-ADVISORS.COM ARCHITECTS. ENGINEERS. PLANNERS.

Motivation

• We state that the ‘design life’ is 50

years, BUT:

• Minneapolis, MN: Average pipe age is 80

years

• Minnesota: One third of state’s sewers are

older than 50 years

• Philadelphia, PA: Average age of sewers is

100 years, with some pipes as old as 190

years

• Philadelphia, PA: Wastewater sewer planned

replacement rate: 100-120 years

• New Jersey: Average age of sewers is 70

years


How Long will it Last?

• Rehabilitating sewers increases life

expectancy, BUT:

• Pipe lining typically results in a smaller effective

diameter

• Rehabilitated pipe can have a reduced flow

capacity

• We should assume that the pipe we

design will be in service 80-100 years

from now.


How Long will it Last?


Predicted Changes in

Average Precipitation

• 1986-2005 data compared to 2081-2100 multi-model mean

• 10-20% increase in Great Lakes Region

• IPCC Summary for Policy Makers


Predicted Changes in


• 1986-2005 data compared to 2081-2100 multi-model mean

• 10-20% increase in Great Lakes Region

• IPCC Summary for Policy Makers



Predicted to Increase

• National Climate

Assessment, US

Global Change

Research Program

• Total Annual

Average Rainfall:

1971-2000 vs 2041-

2070 with continued

emissions scenario


Heavy Precipitation Days



Assessment, US

Global Change

Research Program

• Top 2% of all rainfalls

each year

• 1971-2000 vs 2041-

2070 with continued

emissions scenario


Heavy Precipitation Days



Assessment, US

Global Change

Research Program

• Top 2% of all rainfalls

each year

• 1971-2000 vs 2041-

2070 with continued

emissions scenario


Observed Changes in Very

Heavy Precipitation


Assessment

• Heaviest 1% of

all observed daily

events from

1958-2012

• Midwest outside

range of natural

variability


Recent Rainfall Statistics

• NOAA Atlas 14

• More recent set of

rainfall statistics

(through 2012)

• Previous sources

(TP40 and Bulletin

71) are at least 20-50

years out of date

Source: NOAA Atlas 14, Volume 8 (2013)

Percent difference between NOAA Atlas 14 and TP40

for the 100-yr 24-hr rainfall

3.0

3.5

4.0

4.5

5.0

5.5

6.0

100-yr 24-hr

100-yr 3-hr 2.0

2.5

3.0

3.5

4.0

↑ 24%

↑ 38%

3.0

3.5

4.0

4.5

5.0

NO

AA

Atl

as

14

TP

40

Bu

lleti

n 7

1

LEGEND

10-yr 2-hr

100-yr 3-hr

↑ 26%

2.0

2.5

3.0

3.5

4.0

1.6

1.7

1.8

1.9

2.0

2.1 ↑ 14%

3.0

3.5

4.0

4.5

5.0

NO

AA

Atl

as

14

TP

40

Bu

lleti

n 7

1

LEGEND

2.0

2.5

3.0

3.5

4.0

1.6

1.7

1.8

1.9

2.0

10-yr 2-hr

100-yr 3-hr

↑ 40%

↑ 11%

3.0

3.5

4.0

4.5

5.0

NO

AA

Atl

as

14

TP

40

Bu

lleti

n 7

1

LEGEND

2.5

3.0

3.5

4.0

1.7

1.8

1.9

2.0

2.1

10-yr 2-hr

100-yr 3-hr

↑ 22%

↑ 11%

3.0

3.5

4.0

4.5

5.0

NO

AA

Atl

as

14

TP

40

Bu

lleti

n 7

1

LEGEND


Comparing Statistics

• August 11-12, 2014

• Catastrophic rainfall

event in Metro Detroit

• Local press coverage

categorized this as a

100-yr + storm

• However, using recent

statistics shows us it

wasn’t as rare as we

thought



City

Peak 3-hour

rainfall (in.) TP 40 Bulletin 71 NOAA Atlas 14Exceedance

Interval

Garden City 2.89 63 >100 28 25-yr to 50-yr

Detroit (west fringe) 3.21 >100 >100 48 50-yr to 75-yr

Romulus 2.84 56 >100 28 75-yr to 100-yr

Westland 2.91 66 >100 29 >100 yr

August 11-12, 2014 Rainfall Event

Exceedance Interval (years) Key



City

Peak 6-hour


Interval


Detroit (west fringe) 3.59 91 >100 43 50-yr to 75-yr

Romulus 3.65 98 >100 48 75-yr to 100-yr

Westland 3.49 79 >100 36 >100 yr





City

Peak 12-hour


Interval


Detroit (west fringe) 4.24 93 >100 36 50-yr to 75-yr

Romulus 4.03 77 >100 28 75-yr to 100-yr

Westland 3.94 70 >100 23 >100 yr



• Climate data based on past statistics only and are not forward-looking

• Flow rates and pipe sizes dependent on selection of climate

data

• Need new tools to address future climate patterns


Uncertainty in Design

?

• Several Independent Methods:

EPA Stormwater Calculator

Updating Rainfall Statistics

Comparing Biases in Flow Data

Confidence Intervals


Tools for Addressing Climate Trends


Frequency Analysis

• Annual maxima determined with historical data

available from NOAA

• Plot Annual Max vs Probability

• Use Log-Pearson Type III Statistical Distribution to fit

data – Commonly accepted method in hydrologic

community

• Method also used in USGS Streamflow calculations


Frequency Analysis Example

Plot Data



Probability

Axis is Log

Scale



Calculate Log

Pearson Type

III Distribution



Add 95%

Confidence

Intervals



Determine Storm

Statistics:

10-yr storm = 1.64 in

• Historical data for Detroit City Airport available for 1949-2012

• Comparison of Annual Maxima between 1949-2012 and 1990-

2012

• 1 hour and 24 hour storms

• Increase in rainfall depths

• 7.4% increase for 10 year, 24 hour storm

• 10.2% increase for 10 year, 1 hour storm









7.4% increase

for 10-yr storm





10.2%

increase


EPA Stormwater Calculator

http://www2.epa.gov/water-research/national-stormwater-calculator

Estimates annual

amount of

rainwater &

frequency of

runoff from a

specific site based

on local soil

conditions, land

cover, and historic

rainfall records.


EPA Stormwater Calculator:

Climate Change Options


• Scenarios:

Hot/Dry

Median Change

Warm/Wet

• Time Period:

Near Term

(2020-2049)

Far Term

(2045-2074)





• Scenarios:

Hot/Dry

Median Change

Warm/Wet

• Time Period:

Near Term

(2020-2049)

Far Term

(2045-2074)

Warm/Wet

Scenario =

Higher Average

Rainfall





• Scenarios:

Hot/Dry

Median Change

Warm/Wet

• Time Period:

Near Term

(2020-2049)

Far Term

(2045-2074)

Hot/Dry

Scenario =

Larger Extreme

Storms



Ann Arbor Example


Applying

Rainfall

generated

to a

Sanitary

Sewer

Analysis



Ann Arbor Example


Far Term

(2045-2074)

Warm/Wet

Scenario



Ann Arbor Example



Ann Arbor Example

Perform

Frequency

Analysis

using

Hydrologic

Model

Output and

60 years of

historic rain

data



Ann Arbor Example

Perform

Frequency

Analysis

using

Hydrologic

Model

Output and

60 years of

hourly

historic rain

data



Ann Arbor Example

Perform a

second

Frequency

Analysis using

Hydrologic

Model Output

and 60 years

of hourly rain

data scaled

monthly using

output from

EPA’s SW

Calculator



Ann Arbor Example

Perform a

second

Frequency

Analysis using

Hydrologic

Model Output

and 60 years

of hourly rain

data scaled

monthly using

output from

EPA’s SW

Calculator



Ann Arbor Example • In Ann Arbor, there were 20 districts each with different changes to frequency

flows due to climate change.

• Use average of downstream meters to quantify system wide peak flow changes.

• Consider a 10% increase in peak flows to account for climate change.

14.3% 8.6%

7.0% 11.4%

11.1%

Overall

Increase

= 10.4%

25 year Frequency Analysis Comparison


NOAA Atlas 14


http://www.nws.noaa.gov/oh/hdsc/index.html

• Frequency estimates for Detroit, MI

• 90% confidence

intervals given • Upper bounds

are 20% higher than average

• Lower bounds are 13.3% lower than average


NOAA Atlas 14


http://www.nws.noaa.gov/oh/hdsc/index.html

In many cases, you can increase the duration while staying within the confidence interval



• 90% probability of observing a value within the

range

• Justifies use of a value closer to the upper

limit

• All methods show an increase in precipitation

values so use a value towards the upper

bound of the confidence interval


Novi Study

• Results of a 2014 sanitary sewer capacity study for the City of

Novi indicated increased heavy precipitation in recent years.

• Hydrologic model was calibrated and validated using 15 years

(2000-2014) local rain data and outlet meter flow data.

• Long-term flow simulation was performed using 64 years of

NOAA rain data from nearby Detroit City Airport.

• 10-year frequency flow based on most recent NOAA rain data

was significantly larger than that based on all 64 years of

NOAA rain data.


Novi Study


Novi Study


Novi Study

10-year flow = 22.5 cfs

based on running all 64

years (1949-2012) of

rainfall through model


Novi Study


Novi Study

22.5 cfs = 10-year flow

based on running 1949-

2012 rain data through

model

24.1 cfs = 10-year flow

based on running 2000-

2012 rain data through

model

7% increase

• Recent statistics indicate rainfall depths for design storms have been increasing in Michigan

• Climate models project further increases in rainfall depth for future storms

• Tools are available to account for higher rainfall for design purposes

• Tools provide a basis for increasing future design flows ~10% above existing


Conclusions

Quantifying Climate Change for Stormwater / Wastewater ... Approaches to Quantify Climate... · OHM-ADVISORS.COM Quantifying Climate Change for Stormwater / Wastewater Collection

Documents