World Sweeper Logo

Environmental Information for Sweeping Professionals

A Proposed Method For Performance Testing of Street Sweepers

At last! Pacific Water Resources has developed a procedure that emphasizes using a large enough statistical sample.

by Roger Sutherland, P.E.

One of the looming problems in sweeping is the lack of a testing procedure specifically designed to show how effective particular sweepers are at picking up different types of debris. What usually happens in a demonstration for a road sweeper, for example, is that a very large amount of material is laid down. Then, while an assortment of municipal officials watch, each competitive sweeper takes a few passes. The winner is often the one which, subjectively, seems to leave the least on the ground. Of particular significance is that there is currently no way to evaluate which sweeper has done the best job at picking up fine sediment, the material known to produce the highest level of heavy metals and other pollutants in stormwater. It is widely recognized that the only way sweeping will be able to enter the 21st century as a Best Management Practice (BMP) for stormwater control is if a uniform testing procedure is developed.

Roger Sutherland, P.E., president of Pacific Water Resources, is an Oregon-based stormwater consultant who took part in the National Urban Runoff Program's (NURP) sweeper evaluations in the early 1980s. He is also one of the foremost authorities on sweeping as it applies to stormwater pollutant control. Here are his thoughts on the topic of sweeping methods and tracking:

The big mistake with most of the sweeping demos that municipalities and others put on is that they're simply not real. Most of the time, what I see being put down onto the pavement is an amount of material that no sweeper is going to pick up in one pass.

For example, they designed and conducted a sweep-off in Palm Springs. It was a great idea, but was based upon sweepers going over the same territory five times. Who in the real world is going to get five shots at sweeping the same area?! Then, the final test to see how well each sweeper did was based upon hand vacuuming one square foot of pavement. That, alone, ensured no empirical comparison could possibly be made from a stormwater perspective. The minimum that should have been hand vacuumed after each sweeper is about 400 to 500 square feet of pavement.

When I worked with Gary Minton of Resource Planning Associates in Seattle, Washington, who did the sweeper effectiveness testing field work, our hand vacuumed test plots were 2,000 square feet each. Everyone was real sick and tired of hand vacuuming by the time we got done. But, when you monitor that size of area, you then have a shot at getting significant results. Still, you can't really fault the Palm Springs people, because they were at least trying to come to some realistic conclusions. The real problem is there is not yet a sanctioned testing procedure that anyone testing a road sweeper can go by.

In terms of stormwater pollutants, if a street looks dirty, it probably has a sediment load of thousands of pounds per curb mile.

Through the multiple studies we have done, I have become convinced that 'stormwater sweeping' can be an effective BMP component for stormwater pollutant control. It will take a different way of looking at sweeping, however; it can no longer be done just for cosmetic reasons. When a street actually appears to be dirty to someone with an untrained eye, it is probably holding thousands of pounds per curb mile, instead of a more classic loading of just a couple hundred pounds per that distance. If we want to make an impact on stormwater runoff pollutants, then we can't let the debris loading get that high between sweepings. And, if the sweeping program you are implementing is not doing an effective job of keeping the volume of dirt down, then what's out there on any given day is a very significant loading.

A realistic testing procedure must also be developed, one that can be adapted to make sense in virtually every part of the country. It must be one that takes into consideration other variables which can dramatically affect the performance of the sweeper. By this, I'm talking about pavement condition: gutter design, pitches, etc. If a sweeper has a gutter broom that can vary its pitch, for example, used correctly it makes a substantial difference. Variable gutter brooms are an example of a new technology on sweepers, something that didn't exist when the NURP studies were performed. Back then, if you wanted to change the pitch of a gutter broom, you had to get out of the truck with a wrench and spend half an hour - and the same to set it back to the previous position. You can imagine how often operators would actually take the time to do that.

Actually, enough previous studies have been done that the information exists to allow us to come up with formulas for these kinds of issues. The only thing stopping it, really, is money and a sanctioning body. I agree with the American Sweeper view that the American Public Works Association (APWA) may be the organization which needs to get the ball rolling, along with some oversight from perhaps the EPA and state environmental agencies such as Oregon state's Department of Environmental Quality. Collating the studies and developing a standard testing procedure is certainly possible with the information that now exists. Until the process is developed, however, there is really no way for public works directors to have a standard. There is no way for them to test for sweeping effectiveness. Nor is there a way for such officials, especially in smaller communities, to be able to evaluate the water quality benefits between purchasing one brand of sweeper and another.

The two components needed are methodology - the way to test - and what debris to use for the testing. For the latter, if there is no 'representative' street dirt available, procedures need to be developed which will describe how to produce a simulant (or substitute material) for street dirt. This is nothing more than a guideline for making something that mimics the physical characteristics of street dirt. Actually, that part's not simple. Sand is easy; however, you also need to include coarse fractions, silt fractions, extremely fines and more. Essentially, base material needs to be obtained, then sieved into fractions, and then a certain weight of these different fractions is measured and they all get mixed up. Then, the resulting mixture could be laid down as street dirt. The location in the country, though, would determine the types of soils and base material that would need to be included to put the simulant together easily and correctly.

Pacific Water Resources Logo We have come up with a fairly well developed method for how to conduct the actual testing procedure. This is backed up by our SIMPTM computer modeling. The key is having ample area, so you can collect material from a sample area before sweeping, as well as find out what gets left on the ground by each sweeper after it sweeps. Without going into details, you collect your samples with a vacuum that's attached to a push broom. That way when the broom is pushed and dust kicks up, the dust is captured, too. As a control, dirt is collected with that device in an adjacent plot prior to any sweeping. Then, each sweeper collects what it can, and the push broom/vacuum follows after it in the same area. With that methodology you end up with a 'before sample' (from the adjacent plot), and an 'immediately after sample.' For that last part, you must hand vacuum at least 400-500 square feet. And, the tests shouldn't be run if there are winds of more than a mile or two per hour.

When you hand vacuum an area after the sweeper has gone by, depending upon how effective your sweeper is, the area has to be 3 or 4 times bigger than the 'before' test. Otherwise there won't be enough sample material picked up to be representative, or to sieve and weigh with accuracy. That is a synthesis of the very specific methodology we've developed on how to conduct the process.

Unless the big players -- the EPA, Ecology, APWA and others -- come into the process, we won't have the impact needed to get sweeper testing going.

I think it's quite important that correct, sanctioned procedures be developed for doing this sort of testing. Such a procedure will also provide something on which manufacturers can base their own testing. Unless some of the big players come into the process, however, like the EPA, Ecology, APWA and probably some others, we won't have the impact we need to get sweeper testing going around the US.

Stormwater runoff pollution is a big problem, right now, in many parts of the country. And it's going to get nothing but more important. That, you can count on. If sweeping can show that it is an effective alternative to other methods of non-point pollution control, which it appears to be, it could usher in a whole new era in how sweeping is conducted throughout the country.


To contact Roger Sutherland at Pacific Water Resources, call 503-671-9709. The company may also be reached via email at:

This article is reprinted from American Sweeper magazine, Volume 6 Number 1.
World Sweeper Logo

© 2005 - 2021 World Sweeper
All rights reserved.

Back to Environmental Information
Site Map / Table of Contents