PRESENTATION NAME

nwri.usa.org

PRESENTATION NAME

Use of Physicochemical Changes

in NOM to Evaluate THM

The Relationship of Size, Polarity and THMFP

of DOC in Northern Colorado Watersheds

Formation During Snowmelt,

After Wildfires, and in Pre-

Ozonation Water Treatment

July 27, 2011

Costa Mesa

Alex D. Revchuk, D.Env.

Issam Najm, Ph.D. (2)

I.H. (Mel) Suffet, Ph.D. (1)

Judy Billica, Ph.D. (3)

(1) (2)

(1)

Environmental Science & Engineering, UCLA

(2)

Water Quality & Treatment Solutions (WQTS) Inc.

(3)

City of Fort Collins Water Utilities, Ft. Collins, CO


Objectives

To demonstrate the utility of various NOM

characterization techniques to evaluate the

relationship between physicochemical behavior of

NOM and THM formation in the study of diverse

watershed and water treatment applications.

Characterize NOM:

• Size Distribution by Ultrafiltration (UF)

• THM Formation by SDS Test

• Polarity by Polarity Rapid Assessment Method

(PRAM)

• Origin and Make Up by Excitation Emission

Matrix (EEM) Fluorescence

2


Environmental Challenges

Changes in NOM characteristics and THM

formation studied in three settings:

(1)before, during, and after the annual snowmelt

runoff at Fort Collins, CO water supplies

(2) before and up to a year after Santa Barbara,

CA wildfires

(3) throughout conventional treatment pilot

system with:

• pre-chlorination

• pre-ozonation

• PAC, polyaluminium- and ferric-based

coagulants

3


a

Objective 1:

Snowmelt

4


DOM Seasonal Variability

TOC: Raw Poudre @ FCWTF 2004 - 2007

12

11

10

9

TOC (mg/L)

8

7

6

5

4

3

2

1

0

1/14/2004

7/17/2004

1/18/2005

7/22/2005

1/23/2006

7/27/2006

1/28/2007

8/1/2007

2/2/2008

5


Ft. Collins Water Supplies

6


Ft. Collins Water Supplies

• February: snow cover – low DOM, low flow

• April: start of runoff – medium DOM, low flow

• May: max runoff – peak DOM, high flow caused by

snow melt

Month

Flows During Sampling [cfs]

E

(Controlled)

B

(Free flow)

P

(Free flow)

N

(Controlled)

Feb 250 No data No data 10

April 500 40 40 10

May 550 350 170 10

7


Bulk DOC Trends

DOC

[mg/L]

E

SUVA

[L/mg*M]

DOC

[mg/L]

B P N

SUVA

[L/mg*M]

DOC

[mg/L]

SUVA

[L/mg*M]

DOC

[mg/L]

SUVA

[L/mg*M]

Feb 3.5 2.3 1.7 2.5 1.7 2.8 4.3 2.2

Apr 3.3 2.7 4.3 4.0 1.7 2.8 3.8 2.3

May 3.5 2.5 9.4 3.7 10.7 1.4 3.7 1.9

Variable

Stable

• Flow controls above sampling locations appear to be responsible for

stabilization of both carbon concentrations and aromaticity

• Variable sources: B = SUVA increase (aromatic)

P = SUVA decrease (aliphatic)

• Stable sources: Stable SUVA (intermediate)

8


Size Fraction Trends

mg/L

DOC

Apr Feb DOC

Apr DOC

May DOC

3

3

3

2

2

2

1

1

1

0

E

L/mg*m

6

4

B P N

Feb SUVA

test

Feb SUVA

EPAT BTR PANF NFBS

0

6

4

E

B

P

N

EPAT BTR PANF NFBS

Apr SUVA

test

Apr SUVA

UV: February

0

6

4

E

B

P

N

EPAT BTR PANF NFBS

May SUVA

test

May SUVA

2

2

0.070

2

0

0

EPAT BTR PANF NFBS

test

UV Abs [cm -1]

0.060

EPAT BTR PANF NFBS

0.050

• Shift from unimodal to 0.040 bimodal size distributions in May

0.030

• 1-5 kDa and >10 kDa most dominant size fractions

0.020

• No general aromaticity trend

0.010

• The Poudre River system 0.000(P and N) is dominated 10 kDa

EPAT BTR PANF NFBS

9

10 - 5 kDa

5 - 1 kDa

< 1 kDa


Correlations and Source Selection

UV- and DOC- THM CORRELATIONS

• Bulk UV and DOC, and 10 kDa size fractions are

key to THM production, and serve as good predictors of

THM (R 2 0.75 – 0.99)

•Surprisingly, the ratio of UV/DOC, SUVA does not correlate

with TTHMFP

SOURCE SELECTION

•Low Flow: B and P should be used due to their low DOC

content - uncontrolled sources

•High Flow: E and N should be used during the May peak to

reduce THM formations - controlled sources

10


a

Objective 2:

Wildfires

11


Fires and Sampling Locations

12


Zaca Fire, Precipitation, and DOC

DOC, mg/L

7

6

5

4

3

2

1

0

Oct

06

2.8 mg/L

Jan

07

Apr

07

Zaca

Fire

Jul

07

4.8 mg/L

Oct

07

Jan

08

Apr

08

Jul

08

Sep

08

WTP Source

Water DOC

Dec

08

Mar

09

4.1 mg/L

Precipitation

Jesusita

Fire

Jun

09

• Raw water DOC increased 70% after Zaca Fire

• Increase followed seasonal precipitation

• 2 nd season precipitation led to dilution of DOC

Sep

09

Dec

09

Mar

10

100

90

80

70

60

50

40

30

20

10

0

13

Precipitation, cm


Source Water THM Before & After Fire

300

250

Before Fire

After Fire

THM, µg/L

200

99% confidence

interval

150

THM = 34.8 DOC + 74.7

R² = 0.73

100

0 1 2 3 4 5 6 7

DOC, mg/L

• Wildfire caused 50% increase in raw water THMs

• Mean source water THM increased 50% from 170 to 253 µg/L

• Concern for elevated THM levels in finished product water 14


2 months after Jesusita Fire (2-MO)

12 months Control after Jesusita (CTRL) Fire (12-MO)

2 years after Zaca Fire (2-YR)

Lake Casitas (CTRL)

15


Leaching Procedure

• Leaching Procedure

• Dry samples separated by US mesh #4 and #45

• 5 g of < #45 mesh, 2.5 g of < #4 mesh, 2.5 g of embers,

twigs, leaves per liter of 0.001M NaHCO 3 buffer

• Agitate and shake for 48 hours

• Filter through 200 and 0.45 µm filters

CTRL 2-MO 4-MO 7-MO

• Leachates diluted to 4, 9 and 25 mg/L-C for characterization 16


DOM Size Distributions

DOC, mg/L

SUVA, L/mg-M

100

90

80

70

60

50

40

30

20

10

0

8

6

4

2

0

Bulk

SUVA

Bulk DOC

> 10K Da

5 - 10K Da

1 - 5K Da

< 1K Da

2-MO 2-YR CTRL

• Relative DOC leaching potential

between 2-MO, CTRL, 2-YR, is

10:7:1

• Order of magnitude DOC

decrease over 2 years

• Fire-affected size distributions

dominated by large size DOC

which are least aromatic

relative to other fractions

• CTRL dominated by medium

sized fractions of intermediate

17

aromaticity (except >10K Da)

17


THM Formation

120

110

100

90

Fire

DOC, mg/L

THM, µg/L

Precipitation, cm

5000

4000

DOC, DOC, Precipitation

80

70

60

50

40

30

20

10

3000

2000

1000

THM, µg/L

0

May

CTRL

2009

July Sept Oct Nov Dec

2009

May

2010

• 120% increase in THM leaching potential 2 months after fire

• 50% increase in THM between pre-fire and first flush

• Indicates THM precursor degradation – photolysis (?)

• Timing of first flush after the fire is critical for water treatment

0

18


THM Formation

120

110

100

90

Fire

DOC, mg/L

THM, µg/L

Precipitation, cm

5000

4000

DOC, DOC, Precipitation

80

70

60

50

40

30

20

10

3000

2000

1000

THM, µg/L

0

May

CTRL

2009

• First flush caused a 60% decrease in THM

• Sudden 90 % increase in THM leaching formation potential after the first

flush:

• Drying process (?)

July Sept Oct Nov Dec

• Activation of biota or by biota (?)

May

2010

0

19


THM Formation

120

110

100

90

Fire

DOC, mg/L

THM, µg/L

Precipitation, cm

5000

4000

DOC, DOC, Precipitation

80

70

60

50

40

30

20

10

3000

2000

1000

THM, µg/L

• Second flush reduced THM leaching potential by 96% from the post-fire peak

• First 2 post-fire precipitation events transport the majority of DOC and THM

leaching potential away from the slopes and into the water column

• Springtime vegetation growth contributes to new THM leaching potential

• DOC-THM R 2 = 0.76

0

May

CTRL

2009

July Sept Oct Nov Dec

May

2010

0

20


Polarity Rapid Assessment Method (PRAM)

1

RC

C18 UV @ 254 nm

Sample

Diol

NH2

A/Ao

RC = 1 - A/A o

Solid Phase

Extraction Medias

0

Volume

Retention Coefficient (RC) = % of UV-absorbing material adsorbed by

solid phase extraction material

Benefits: Ambient pH, low sample volume, 10 min run time, parallel

or series configurations

Does not measure same polarity characteristics as XAD resins

(Rosario Ortiz, et al., 2009)

21


DOM Structure & Polar Affinities

Hydrophilic

(Diol)

Charged

(NH2)

Hydrophobic

(C18)

Adapted from Stevenson, 1994


Polarity Trends

Hydrophobic Hydrophilic Charged Hydrophobic Hydrophilic Charged

Hydrophilic Hydrophobic Charged

• Charged DOM constituents are 25% lower in fire-affected DOM than in unburned control

• 2-YR contains less hydrophobic and hydrophilic constituents

• No size-dependent relationships for hydrophobic and hydrophilic constituents

23


NH2 RC

Charge to Mass Relationship

1.00

0.90

0.80

0.70

0.60

0.50

0.40

0.30

0.20

0.10

0.00

y = 0.0869x + 0.0466

R² = 0.9

0 1 2 3 4 5 6

DOM Mass, K Da

• Direct linear relationship

between charge and

mass (as measured by

UV)

• Small NOM fragments

carry less negative

charge than large NOM

fragments

• Wildfire generated large NOM fragments of high

negative charge

• Expect increased demand for cationic coagulant

during treatment

24

24


DOM Fluorescence

Bulk

2-MO 2-YR CTRL

Fluorescence Intensity

< 1 K Da

Humic Region: shorter wavelengths in fire-affected samples

Fluorescence Intensity

25


DOM Fluorescence

Bulk

2-MO 2-YR CTRL

Fluorescence Intensity

Fulvic acids: present in fire-affected samples only

< 1 K Da

Fluorescence Intensity

26


DOM Fluorescence

Bulk

2-MO 2-YR CTRL

Fluorescence Intensity

Aromatic proteins and microbial by-products: recently-burned leachate only

< 1 K Da

Fluorescence Intensity

27


Correlations – 1 st Year Wildfire Recovery

Line

No.

Parameters

R 2

General Water Quality

Remarks

1.

UV – DOC

B, >10, 5-10, 1-5

0.84 – 0.99

Positive relationships

Charge to Size

Positive relationship between

2. NH2 – Size 0.74 DOM molecular size and

negative charge

28


Correlations – 1 st Year Wildfire Recovery

Line

No.

Parameters

R 2

THM Formation Potential

Remarks

3.

DOC B

DOC


a

Objective 3:

Water Treatment

30


Pilot Treatment Goals - 10/08 to 3/09

• Santa Barbara’s Cater WTP– 37 MGD conventional

flocculation-sedimentation treatment

(Treatment Goal: 20% < THM MCL in 7-day SDS test 64 µg/L)

• Strategy that lowered THMs from wildfires

• THM formation

- Pre-ozonation instead of pre-chlorination –

- pH suppression - 8.2 to 7.5 after coagulation

• Precursor Removal

- Ferric-based coagulants vs. polyaluminum and PAC –

ferric sulfate removed most THM precursors. This may

be because it removed DOC most effectively in 5-10K

Da range

31

31


UV & DOC Removal by Size

-100%

UV UV

DOC DOC

-28%

-49% -72%

PAC Ferric + O 3 Sulfate

48%

71%

1%

276%

150%

100%

50%

0%

-50%

-70%

-86% -100%

-37%

0%

-50%

-100%

150%

100%

50%

0%

-50%

-100%

• PAC+Cl 2 – effective in removing 1-5K

Da DOM

Bulk 322% > 10K Da 5-10K Da 1-5K Da 10K Da 5-10K Da 1-5K Da 10K Da 5-10K Da 1-5K Da 10K > Da 5-10K Da Da 1-5K Da


Overall Conclusions

• Size, polarity, charge, and fluoroactivity were used

to evaluate the physicochemical characteristics of

NOM and their relationships to THM formation

potential

• These relationships were used to:

• select water sources for Fort Collins, CO during

snowmelt episodes

• track recovery of watershed after wildfires

• select ferric sulfate as the coagulant to remove

a significant part of NOM precursor in 5-10 K Da

size range, which was a major THM producer

33


clearer?

34

Questions ?

Can I make anything

More magazines by this user
Similar magazines