What is EPA Method 9?

If Method 9 is required in a Title V permit, the person who conducts the visible emission observations (VEO) is often referred to as the observer. The observer must be trained and currently certified in accordance with the Method, i.e., attending a hands-on smoke school and passing qualification testing periodically.

Because the Method is a subjective tool, two observers may report different opacity values on the same plume at the same time. If the observers both have valid certification, either observation is valid, even in a deposition or in court.

The frequency of Method 9 readings is governed by air permit/compliance plan for the facility.

A single Method 9 reading requires a total of 24 visible emissions observations taken at 15-second intervals, for a total of six (6) minutes. Those observations are averaged to calculate the opacity value.

Permits may require more than 24 observations – they may specify 12, 18, and 30 minutes, yielding a greater number of observations and requiring custom VEO forms. VEO forms must contain required Method 9 fields, but can be customized to fit a facility's requirements.

New Source Performance Testing (NSPT) typically requires three (3), 1-hour observations for each source.

Opacity values are expressed as a percent and are calculated using a running average.

All subjective methods have an associated error and Method 9 is no exception. The efficacy of the Method has been proven in field tests. The results of an EPA study that involved 769 sets of 25 observations for each type of smoke (black and white) yielded:

• Black plumes: 100% of the sets were estimated with positive error of less than 7.5 percent opacity. 99% were estimated with a positive error of less than 5% opacity.
• White plumes: 99% of the sets were estimated with a positive error (higher values) of less than 7.5 percent opacity. 95% were estimated with a positive error of less than 5% opacity.

Source:EPA Method 9 Document

Continuous opacity monitors (COMs) are available to track opacity measurements on stacks. So why don't we just use them? This is a complex subject and the following is a broad brushstroke of the issues related to COMs:

COMs are widely used, but are prone to drift offsets and downtime. Their reliability results in many facilities owning a backup COM to cover mechanical failures. COMs can be more accurate, but a visible emissions reader is considered more reliable in case law. Additionally, COMs operate on stacks only, therefore operations such as rock crushing, storage piles of material, handling areas, and transfer points require manual methods.

Method 9 allows for a percentage of error. At the time the Method was promulgated, acceptable error levels were determined by field trials (see the toggle field, above). The determination of a pass/fail on a field test run is determined using percentages of error and standard deviations.

To pass a field test run, you must achieve a score of 37 deviations or less. One field run consists of 25 black smoke point observations and 25 white smoke point observations. The number 37 is calculated as follows:

• Each point observation can be 15% on either side of the correct opacity level. For example, if a point plume with 25% opacity is shown, an observer can answer as low as 10% or as high as 40% and be within the 15% allowance.
• Each point observation is assigned a value called a deviation:
• Zero (0) for exact observations.
• One (1) for observations within 5% accuracy.
• Two (2) for observations within 10% accuracy.
• Three (3) for observations within 15% accuracy.
• If an observation is outside of the 15% accuracy, the observation is considered a failed observation. One failed observation constitutes a failed run.

At the end of the the 50 test point observations (25 white, 25 black), the scores are added together. If you are at 37 deviations or below, you pass.

If you are 38 or above, OR have failed one test point observation, you do not pass.

Note: You can certify if have two contiguous passed runs. For example, runs consist of 25 white followed by 25 black - you can pass if you fail white on Run #1, pass black on Run #1, then pass white on Run #2. This examples equates to 50 contiguous observations and if your deviations are within the passing parameters, you certify for Method 9.