Skip directly to search Skip directly to A to Z list Skip directly to site content


Airborne Sediment from Lake Roosevelt

(a/k/a COLUMBIA RIVER MILE 597-745)


In October 2000, the Confederated Tribes of the Colville Indian Reservation requested that the Agency for Toxic Substances and Disease Registry (ATSDR) review and comment on the proposed Air Toxics Study of airborne bed and bank sediments from Franklin D. Roosevelt Lake [1].

The purpose of the proposed study is to address the following questions:

  • What meteorological conditions drive entrainment of exposed sediments?

  • What areas have the greatest potential for entrainment and transport?

  • What airborne contaminants pose the greatest risk to public health?


The Franklin D. Roosevelt Lake, commonly known as Lake Roosevelt, was created by the construction of the Grand Coluee Dam in 1942. The lake extends 217 kilometers from the Grand Coluee Dam in northeastern Washington. The upper Columbia River and Lake Roosevelt form the southern and eastern boundaries of the Colville Reservation in Ferry County. Lake Roosevelt forms the western border of Stevens County and the Spokane Indian Reservation ( Figure 1).

The Lake Roosevelt pool height is lowered for a ten week period in the Spring to accommodate snow melt. Another draw down occurs for a two-week period in the Fall to accommodate the downstream migration of anadromous fish. When the pool height is lowered, thousands of acres of sediment and banks are exposed to wind erosion. Bank and river sediments are entrained, dispersed and deposited during periods of high winds [1]. Persons living near Lake Roosevelt may be exposed to airborne particulate from lake sediment during these wind events.

Approximately 5,000 American Indians live on this 1.3 million acre Colville Reservation in Ferry County, Washington and another 7,000 tribal members live off the reservation in the surrounding area [2]. The towns of Inchelium and Keller are the two population centers on the Colville Reservation that are located close to Lake Roosevelt. Population centers in Stevens County close to the Lake Roosevelt include Hunters, Gifford, Kettle Falls, Marcus, and Northport. The population of Ferry and Stevens Counties in 1999 were estimated at 7,188 and 40,137 persons, respectively [3].

A summary of the study's proposed activities are as follows [1]:

Year 1

  • collect background information for siting meteorological stations.
  • identify the potential source areas for airborne sediment
  • perform sediment sampling
  • perform preliminary air sampling
  • complete a progress report

Year 2

  • perform dispersion modeling of contaminant impact (based on Meteorological Data)
  • perfrom air monitoring
  • operate meteorological stations
  • complete a progress report

Year 3-5

  • perform air monitoring
  • operate meteorological stations.

DISCUSSION - General Comments

Wind erosion off Owens Lake (California) and Lake Koocanusa (Montana) lake beds have resulted in elevated PM10 levels [4][5]. PM10 is airborne dust that is less than 10 micrometers in diameter. Researchers at Washington State University and the U.S. Department of Agriculture have measured elevated PM10 levels resulting from the wind erosion of crop land as part of the Columbia Plateau PM10 study [6, 7]. This region lies immediately south of the Colville Reservation. These studies are summarized in Appendix P of the U.S. Army Corps of Engineer's draft Environmental Impact Statement of the Lower Snake River Juvenile Salmon Migration Feasibility Study [8]. The Army Corps' document explores the potential impact on air quality of drawing down four impoundments on the Lower Snake River.

In response to the study question 1, "What meteorological conditions drive entrainment of exposed sediments?", the Lake Koocanusa study may be particularly relevant to answering this question since the lake bed is exposed during a spring draw down and the lake also lies within the Upper Columbia watershed (Kootenay River). (The study's weather related findings are summarized in Appendix A.) Results of the Columbia Plateau PM10 Program may also be useful because of the study's relative close proximity to Lake Roosevelt.

The Agency for Toxic Substances and Disease Registry (ATSDR) does not see a need for an extensive network of meteorology stations along Lake Roosevelt during the initial phase of the investigation. One station should be sited on the north-south portion of the lake above Fort Spokane and another station on the lower section that runs east-west. The principle weather-related parameter for modeling wind-generated dust on Lake Roosevelt is the "fastest mile or maximum one hour wind speed" [8]. This parameter may be accurately estimated from data collected by these two meteorology stations.

In response to study question 2, "What areas have the greatest potential for entrainment and transport?", this question will be difficult to answer because of the number of parameters associated with wind erosion of exposed lake beds. Factors that influence the amount of dust generated from wind erosion of lake beds include: [8].

  • area of exposed dry sediments
  • amount of fine particulate matter in the sediments
  • sediment moisture content
  • frequency that the surface is disrupted, providing fresh material for wind erosion
  • frequency and duration of winds strong enough to lift erodible particles
  • roughness of the exposed surface (a smooth surface versus one impregnated with rocks and other obstacles).
  • the amount of vegetation covering the banks and sediment

The highly varied terrain surrounding Lake Roosevelt will make predicting the airborne dust levels from wind erosion an even more difficult task. For this reason, ATSDR concurs with proposed initial emphasis on sediment sampling and air monitoring.

The initial air monitoring should be performed in areas where people, particularly the most sensitive persons (children and elders), are most likely to be exposed, rather than attempting to site the monitoring stations at locations containing the suspected maximum levels [9]. Villages close to Lake Roosevelt that contain schools and housing developments are suitable locations for siting the initial air monitoring stations. Possible locations for monitoring stations include Keller, Two Rivers - Fort Spokane area, Hunters, Inchelium, and Kettle Falls - Marcus area.

If the initial air monitoring results indicate communities are exposed to contaminants at levels of health concern, then additional efforts should focus on the magnitude and extent of the exposure. However, if the initial air monitoring results do not show exposure; no further action may be warranted.

Researchers should also evaluate how the meteorological conditions during the monitoring compare to long term averages. This is done comparing data from the on-site meteorological stations and the existing meteorological stations. If sufficient correlation exists, the existing meteorological stations could be used instead of additional onsite stations. If the maximum wind conditions are low compared to historical levels, additional sampling or modeling may be warranted.

Study question 3 states"What airborne contaminants pose the greatest risk to public health?" Characterizing the type and amount of contamination in the sediments are essential to answering this question. ATSDR concurs with the plan to perform this task during the first year of the study. Sediment samples should be sieved to separate the particles that are capable of being entrained. Specific chemical analysis should be performed for metals, pesticides, crystalline silica, PCBs and dioxins/furans.

Initial air monitoring should be performed for PM10, monitored continuously (for 24-hour periods), during the 1st year's Spring and Fall draw down periods. PM10 is one measure of fine airborne particulate; PM2.5 is the other measure of even finer particulate. (PM2.5 is airborne dust that is less than 2.5 micrometers in diameter.) PM2.5 is believed to be the most harmful to human health [10]. ATSDR is recommending PM10 monitoring rather than PM2.5 monitoring because of the following: (1) particulate from crustal sources (e.g., sediment) are typically found in the PM10- size range whereas PM2.5-size particles primarily result from combustion sources, (2) most of the wind-generated dust studies to date have measured PM10 rather than PM2.5, and (3) PM10 is an enforceable air quality standard whereas PM2.5 is not. ATSDR concurs that PM10 monitoring should begin 30 days prior to draw down in the Spring and 15 days prior to the drawn down in the Fall to compare PM10 data collected during periods with and without exposed lake bed and bank sediments.

Air monitoring should be performed for total suspended particulate (TSP) and specific metal/pesticide contaminants if sediment analyses indicates metal and pesticide exposures could occur at the anticipated levels of PM10. (This determination can be performed immediately upon receiving results of the sediment samples so that TSP monitoring may be performed in the later portion of the same Spring draw down period, if necessary).

The following example illustrates the calculation of airborne lead concentrations from anticipated PM10 levels and soil concentrations.

Assuming a PM10 level of 1,500 micrograms per cubic meter (g/m3)(10-times the EPA's 24-hour PM10 standard of 150 g/m3) and using the National Ambient Air Quality Standard (NAAQS) of lead in air (1.5 g/m33), the ratio of lead to PM10 is:


This ratio can be directly applied to the sediment. Therefore, the sediment with 0.1% ( 1,000 mg/kg) or less of lead yields an air concentration of 1.5 g/m3 or less of lead particulate, at a PM10 concentration of 1,500 g/m3. Analyzing the air samples for lead would not be necessary if the lead in the sediment is below 1000 mg/kg and the maximum PM10 value doesn't exceed 1,500 g/m3. This calculation should be completed with actual sediment sample and PM10 data. This example assumes that the lead particulate will be airborne the same fraction that is found in the sediment. Also note that the PM10 levels (1,500 g/m3) used in this example would represent a health hazard for prolonged periods of exposure.

Performing initial PM10 monitoring allows for immediate exposure measurement and evaluation. Once the PM10 data are collected, the use of air models can be investigated and validated with the collected data, if necessary. Model validation is important because of the complex terrain and potentially complex wind patterns in the valley. Once the model is validated, the model can then be applied to other areas such as individual houses. Modeling may not be necessary if the monitor locations are judged to be at the most impacted areas and/or the data indicates that a hazard is not present.

ATSDR concurs with using passive samples (for total settled dust) to investigate the potential for ingesting settled dust. Passive samplers should be co-located with the PM10 monitors at points of exposure and should be performed during the first year of the study.

Air Modeling

These comments on air modeling concern the Study Plans, Year 2, Task 1: Dispersion Modeling of Contaminant Impact Base on Meteorological Data.

This section refers to using an "EPA-approved" or approval pending model found in 40 CFR Appendix W" and that is consistent with study objectives and limitations such as ISC3, ISC-Prime, or AERMOD." However, there is not an "EPA approved" (or approval pending) model for fugitive dust modeling. Appendix W states:

"Due to the difficult nature of characterizing and modeling fugitive dust and fugitive emissions, it is recommended that the proposed procedure be cleared by the appropriate Regional Office for each specific situation before the modeling exercise is begun." [11]

This indicates that site specific models are selected with EPA. This also means that models other than "EPA approved" may be relevant. The ISC-Prime model is not appropriate for this study because it is primarily intended to address building downwash. AERMOD may be the most appropriate EPA model because of its ability to use irregularly-shaped area sources. AERMOD comprehensively addresses dispersion in the presence of complex terrain better than ISCST and other models [12]. However, AERMOD presently does not include a deposition algorithm. The California Meteorological Model (CALMET) may be used to account for the terrain if the initial meteorological data indicate that the terrain significantly influences the local wind patterns. The California Puff Model (CALPUFF) can be used to estimate airborne dust concentration based results of the CALMET modeling [13].

Since this study is not an EPA regulated activity, the selected model does not have to be approved by EPA. One such model that could be considered is the Midwest Research Institute's MRI Regional Dispersion Model ( [14]. Since determining emission rates may be difficult, the EPA's Fugitive Dust Emission Estimation Software (MECH) may be appropriate for use [15].

The key input parameters for these wind erosion models typically include: the dust emission rate, frictional velocity of the lake bed sediment, the fastest mile (wind speed), proportion of time that the wind speed exceeds the threshold velocity, and the area of exposed sediment [16]. (This is also summarized starting on page 46 of Appendix P of the U.S. Army Corps of Engineer's draft Environmental Impact Statement of the Lower Snake River Juvenile Salmon Migration Feasibility Study [8].)

Additional Sources of Environmental Data

The study should review existing soil and sediment databases to determine if data in sediment samples in the study area already exist. Several databases containing national datasets include the following:

Existing air sampling data for PM10, metals, pesticides, and dioxins and furans may be obtained from U.S. EPA's Aerometric Information Retrieval System (AIRS) at Meteorological data from existing stations can be obtained from the National Climatic Data Center at


ATSDR supports the Confederated Tribes of the Colville Indian Reservation's plan to assess the public health impact of wind-generated dust from Lake Roosevelt sediments.

ATSDR concurs with the approach outlined in the initial phase of the study. A multi-year study may not be required if the results of initial PM10 monitoring indicate that PM10 exposure is below levels of health concern in the most populated areas close to the lake.


  1. Focus initial air monitoring efforts to measure levels of PM10 at points of exposure. This should include exposure to sensitive populations such as children.

  2. Seek input from the EPA regarding type and scope of air modeling.

  3. Seek input from the Spokane Nation since their members also may be affected by wind-generated dust from Lake Roosevelt.

At the request of the Confederated Tribes of the Colville Indian Reservation, ATSDR will provide on-going technical support in the development of this study plan.


  1. Hurst, Donald. Draft Design Study: Air Toxics Study of Airborne Bed and Bank Sediments Prepared for the Lake Roosevelt Water Quality Council and the Confederated Tribes of the Colville Indian Reservation. October 24, 2000.

  2. Indian Health Service. Portland Area IHS. Colville Service Unit. February 1996. Available at

  3. United States Census. State and County Quick Facts. July 1999. Available at

  4. Great Basin Unified Air Pollution Control District. 1998. Owens Valley PM10 Planning Area Demonstration of Attainment State Implementation Plan. Bishop, California. November 16, 1998.

  5. Environalysis. 1996. Lake Koocanusa, Montana Fugitive Dust Study -- Final Report. Prepared for U.S. Army Corps of Engineers, Seattle District, Seattle, Washington. September 1996.

  6. WEAQP (Northwest Columbia Plateau Wind Erosion Air Quality Project). 1995. An Interim Technical Report. Keith E. Saxton, editor. U.S. Department of Agriculture, Agricultural Research Service, Pullman, Washington. February 1995.

  7. Claiborn, C., B. Lamb, A. Miller, J. Beseda, B. Clode, J. Vaughan, L. Kang, and C. Newvine. 1998. Regional Measurements and Modeling of Windblown Agricultural Dust: The Columbia Plateau PM10 Program. Journal of Geophysical Research, Vol. 103(16):753-767.

  8. U.S. Army Corp of Engineers. Walla Walla District. 1999. Draft Lower Snake River Juvenile Salmon Migration Feasibility Study/ Environmental Impact Statement. Appendix P. (Air Quality) Available at

  9. Agency for Toxic Substances and Disease Registry. Environmental Data Needed for Public Health Assessments. Atlanta: US Department of Health and Human Services. Public Health Service. 1994. Available at

  10. United States Environmental Protection Agency. Air Quality Criteria for Particulate. EPA600/AP-45/0001b. Research Triangle Park, NC: U.S. Environmental Protection Agency. 1995.

  11. Unites States of America, 40 Code of Federal Regulations Part 51, Appendix W.

  12. Support Center for Regulatory Air Models. Officer of Air Quality Planning and Standards. Washington, DC: U.S. Environmental Protection Agency. 2001. Available at

  13. EachTech Atmospheric Studies Group. Concord MA. Available at

  14. Midwest Research Institute. Kansas City, MO. Available

  15. EPAs Fugitive Dust Emission Estimation Software. Available at

  16. Environmental Protection Agency. AP-42, Emission Factors, Chapter 13.2.5. Industrial Wind Erosion. U.S. Environmental Protection Agency 1995.


Peter Kowalski
Environmental Health Scientist
Exposure Investigations and Consultations Branch
Division of Health Assessment and Consultation

Brian Kaplan
Environmental Health Scientist
Exposure Investigations and Consultations Branch
Division of Health Assessment and Consultation

Greg Zarus
Atmospheric Scientist
Exposure Investigations and Consultations Branch
Division of Health Assessment and Consultation

Reviewed by:

Daphne Moffett, PhD
Office of Tribal Affairs
Division of Health Assessment and Consultation

Susan M. Moore
Chief, Health Consultations Section
Exposure Investigations and Consultations Branch
Division of Health Assessment and Consultation

Richard W. Robinson
Regional Representative
Office of Regional Operations


The Lake Koocanusa Fugitive Dust Study (Excerpted from U.S. Army Corps Report 1999. Draft Lower Snake River Juvenile Salmon Migration Feasibility Study/ Environmental Impact Statement. Appendix P. (Air Quality) Available at

A recent similar study has illuminated the meteorological conditions associated with high fugitive dust events (Environalysis, 1996). PM10 monitoring was conducted at Lake Koocanusa, the reservoir formed by Libby Dam on the Kootenai River in northwestern Montana. Lake Koocanusa refills with snowpack melt in the late spring and summer. Two years of monitoring (May 1994 through June 1996) included meteorological conditions, continuous PM10 concentrations at the lake and in the nearby town of Eureka, and passive dust settling measurements. The following meteorological conditions are associated with entrainment of fugitive dust:

  • High dust events are preceded by several hours of increasing wind speeds from a constant direction.

  • The wind speeds that initiate a dust event are not unusually high. A minimum threshold of wind energy is required to initiate the dust event.

  • The high wind events last up to 9 hours.

  • Background levels may significantly contribute to the measured concentrations.

  • Dust levels rapidly fall when the wind speed drops below about 5 meters per second (m/sec)(10 miles per hour [mph]).

  • Dry lake banks appear to provide turbulent conditions that enhance emissions and produce more emissions.

The geography and micro-topography of the dry lake bed sediments can be an important factor. Different wind patterns can affect different areas of the lake bed. Steep slopes, which allow heavier particles to roll downslope, make smaller particles available for entrainment. Although some 1-hour PM10 concentrations measured during the Lake Koocanusa study were very large, all 24-hour concentrations were less than the air quality standard.

On several occasions measured maximum 1-hour PM10 concentrations exceeded about 500 g/m³. In all cases the average 24-hour PM10 concentrations were below 100 g/m³, or about 67 percent of the AAQS. The dust events were characterized by moderate persistent winds and dry conditions, and lasted from between 3 and 9 hours. The largest measured 1-hour concentration was associated with winds that blew over large areas of exposed lake banks.

Details Map
Figure 1. Details Map

Table of Contents The U.S. Government's Official Web PortalDepartment of Health and Human Services
Agency for Toxic Substances and Disease Registry, 4770 Buford Hwy NE, Atlanta, GA 30341
Contact CDC: 800-232-4636 / TTY: 888-232-6348

A-Z Index

  1. A
  2. B
  3. C
  4. D
  5. E
  6. F
  7. G
  8. H
  9. I
  10. J
  11. K
  12. L
  13. M
  14. N
  15. O
  16. P
  17. Q
  18. R
  19. S
  20. T
  21. U
  22. V
  23. W
  24. X
  25. Y
  26. Z
  27. #