Underestimation of Sulfur Concentrations During High Loadings and Humidity Conditions in the Eastern US

 

Issue

The sulfur concentration measurement from the Teflon filter in the IMPROVE Module A sampler can be underestimated by over a factor of 2 (Figure 1). This bias occurred only at Eastern US sites and was most severe during the summers of 1992-94 on days with the highest sulfur loadings coinciding with high humidity.  Table 1 list the sites where this bias was found to occur.

Cause

The exact cause of the sulfur underestimation has never been fully explained.  However, analysis of data from IMPROVE and from a special study at Great Smoky Mountains in the of summer 1994 indicates that the problem is produced by a combination of very high relative humidity, sulfur present predominantly as hygroscopic sulfuric acid, and high filter face velocities.  With these conditions, some of the sulfate may migrate away from the center of the filter and perhaps even be lost from the filter.

Long Term Solution

  In 1995, the filter mask was removed from most Eastern US sites.  This increased the filter size from 2.2 cm2 to 3.5 cm2 and reduced the filter face velocity by 35%.  This has nearly eliminated the problem, but several significant sulfur and sulfate differences were observed at Washington DC even with unmasked filters.  On 8/16/95, significant differences were also observed at four sites after the mask was removed.

The filter size was also increased at sites which had no indication of a sulfur underestimation, but have high mass loadings (Table 2).  This was done to prevent any possible sulfur losses and minimize clogging of the Teflon filter.

Recommendations for Data Analysis

It is recommended that data analysts use the sulfate ion measurements, divided by 3, from the IMPROVE Module B nylon filter instead of the sulfur mass measurement from module A for time periods before the Teflon filter size was increased at the sites listed below (Table 1)

Further Information

Recommendations  by J. Sisler  - Detailed recommendations and site by site analysis of the problem.

Sulfur-Sulfate History by B. Eldred - Detailed information on the potential causes of the sulfur underestimation.

Sulfur-Sulfate Trends by B. Eldred Further information on sulfur and sulfate comparisons and changes over time. 

Figure 1.  Scatter plots of Great Smoky Mountain, TN, National Park sulfate (SO4) against sulfur scaled by 3 (3*S) for each season during 1993.  The SO4 and 3*S should fall along the 1 to 1 line as they do during spring, autumn and winter.  During the summer months the 3*S severely underestimates the SO4 at concentrations above 5 micro-g/m3.


 

Table 1.  Monitoring sites impacted by the sulfur underestimation problem and dates when the Teflon filter was increased from 2.2 sq. cm to 3.5 sq. cm.

Site

Code

Site Name

Sulfur Bias

Nylon Filter clogging
 during 1998

Filter Size
 Increase Date

LYBR

Lye Brook, VT WA

YES

NO

4/95

ACAD

Acadia, ME NP

YES

NO

4/95

BRIG

Edwin B. Forsyth, NJ NWR

YES

YES

5/95

DOSO

Dolly Sods, WV, WA

YES

YES

4/95

SHEN

Shenandoah, VA, NP

YES

YES

4/95

JEFF

James River Face, VA, WA

YES

YES

4/95

GRSM

Great Smoky Mnt, TN, NP

YES

YES

4/95

SHRO

Shinning Rock, NC, WA

YES

YES

4/95

UPBU

Upper Buffalo, AR, WA

YES

NO

5/95

MACA

Mammoth Cave, KY, NP

YES

YES

4/95

SIPS

Sipsey, AL, WA

YES

YES

5/95

ROMA

Cape Romain, SC, NWR

YES

YES

8/98

OKEF

Okefenokee, GA, NWR

YES

YES

8/98

CHAS

Chassahowitzka, FL, NWR

YES

YES

8/98

EVER

Everglades, FL, NP

 

 

8/98

Table 2.  Monitoring sites with a 3.5 sq cm filter and no evidence of sulfur underestimation in their time series.

Site

Code

Site Name

Filter Size
 Increase Date

MOOS

Moose Horn, ME, WA

7/98

GUMO

Guadalupe MNT, TX, NP

4/01

BIBE

Big Bend, TX, NP

4/01

SEQU

Sequoia, CA, NP

5/98

SAGO

San Gorgonio, CA, WA

4/95

CACR

Caney Creek, AR

4/01

HEGL

Hercules-Glades, MO

4/01

WIMO

Wichita Mountains, OK

5/01

PUSO

Puget Sound, WA

3/98