Purpose
The purpose of this site inspection was to...
- Test Peter Gascoyne's salt distribution photos, providing a first look at whether they can quantify the amount of salt applied to an area.
- Evaluate what variables might affect the accuracy of estimates.
- Start the development a method, and protocol, that would maximize accuracy, speed, and documentation of amounts of salt.
This link displays all photos, with captions, in order. Throughout this report, clicking once on a Link takes you to a supporting photo on Flickr.
To enlarge any photo shown in this report, click on it.
Methods
Inspection location and people involved
I inspected numerous sidewalks along streets and walkways leading to parking at the University of Wisconsin Extension, a large, off-campus building. All walks had been salted the morning before sampling, but there was no evidence that the parking pavement had been salted. Sampling lasted from about 1:00 to 5:00 pm on Feb. 20, 2026. I had sampled the site earlier, so I knew today's salting was typical here, and that rain had removed all previous salt. Link.
I spoke to the Building Manager, Ms. Stevie Seltman, stevie.seltman@wisc.edu., whom I met while sampling. She knows the site is oversalted and has complained. She said she would file a complaint with Building Inspection.
The contracts for salting off-campus are managed by U.W. Groundskeeping. According to their website: "Grounds is responsible for maintaining a safe and beautiful outdoor environment across the 939-acre UW campus. With a crew of more than forty full-time professionals, Grounds performs a myriad of activities including... snow removal to maintaining concrete, asphalt, and signs." "Grounds is continually striving to improve its ecological footprint by actively engaging in sustainable lawn care and road salt reduction initiatives."
The contractor salting the sidewalks is Maple Leaf Landscaping, 608-845-2203. The parking areas are salted by a different contractor. Private contractors, rather than Groundskeeping, spread salt at 3500 because this location is separated from the university, but U.W. Groundskeeping still controls the contracts for work at 3500. If there's ongoing snow or some other problem, the Custodian at 3500 may become involved. So, there are 4 entities who may become involved with snow and ice removal at 3500. The onsite Building Manager has control of only the Custodian.
Weather
Moderate rain fell for several hours on the evening of 2/19, followed by temperatures dropping below freezing, then about half an inch (or less) of snow. Salt stains I had noticed on previous days were now gone.
When I arrived at 1:09 pm, all snow on the pavement had melted, although there were a few small patches of snow on the grass. The areas of pavement with more than light salt were still damp (dark in tone) but quickly dried within an hour or two during sampling. Wind prior to and during sampling was 10-15 mph, probably not enough to blow salt from the pavement, and about 1/3 of the salt was lightly adhering to the pavement anyway.
Conditions for sampling at this site and time were near ideal because...
- Foot traffic is very low (about one person every two hours), reducing any pulverizing or dispersal of salt by feet.
- Sidewalk concrete is smooth, uniform, and relatively new.
- Prior salting was removed entirely by rain the previous night.
- Light snow triggered the salting, but all snow melted before sampling.
- There was not enough snow or meltwater to remove much salt.
- Meltwater mostly dried before sampling.
- There was too little wind or time to blow salt off the pavement.
Observations and Results
Overall distribution trends
Salt distribution along the sidewalks was highly non-random or clumped. Some squares had barely any salt, while others had much more salt. I took many photos showing half or more of all the pedestrian walks. There was one substantial salt spill near the back entrance consisting of multiple small piles. Link.
A distinctive pattern of salt distribution (right) suggests "sowing" by hand or from a scoop. The pattern resembles a "comet" with a blurry head, connected to a longer tapering tail
Besides clumping, another visible trend was for salt application to be heavier towards the entrances, on the one stair, and on other areas of concentrated traffic, such as where sidewalks meet the streets.
How much salt was dissolved by meltwater?
I don't believe there had been enough snow--with resulting meltwater--to significantly reduce the amount of salt I could see or collect. There were just a few places where significant dissolution of salt may have occurred. Link.
However, the salted areas remained slightly damp (wet) for an hour or so after I arrived, so a small amount of salt could be distributed in a fine solid film on the concrete surface or have been absorbed a short distance into the concrete.
In two places where I collected salt, about 1/3 of the salt particles were lightly adhering to the pavement, but I was still able to use the dust bin to easily scrape all particles towards the center for final collection with a fine brush. This "adhesion" suggests a small amount of water surrounded some particles, then dried--indicating partial solution of these particles. But it was probably not enough to substantially reduce the salt collected.
How can salt disappear before samples are collected and weighed?
- Dissolution by rain, meltwater, or adsorption of humidity. See above.
- Wind. After a windy day, nearly all of the salt at 3500 Univ. Ave disappeared.
- Bouncing and scattering during application.
- Disturbance by foot traffic grinds to smaller particles, disperses salt, and can track out salty footprints.
- Collecting is incomplete, because some particles scatter or adhere, or wet salt binds to surface.
- Accidental loss at all stages, including weighing. Salt is slightly sticky; small amounts (1%?) tend to stay behind during each transfer (from brush to dust bin to transport container to scale.)
- Salt before application may contain rocky material as a contaminant. I did find a rocky particle about 1/8 in diameter in sample #9. I know it was rocky because it didn't dissolve in water. But I don't know if it was in the salt applied, or was introduced later.
Summary: Clues to look for indicating amount of salt lost...
- Salt stains indicating dissolution, movement, then drying. Link.
- Substantial wind prior to sampling, especially if surface is dry. Link.
- No-salt areas with sharp borders indicate removal by flowing water.
- Salty footprints suggest salty puddles that disappeared without leaving stains. Link.
- Salt particles, on close inspection, have smooth edges and glassy appearance indicating partial solution. Link.
- Salt particles gathering at edges like steps, curbs or walls indicate scatter by wind or traffic. See below.
"Edge effects" can distort salt spreading and salt measurement
When particulate salt is sowed by hand or machine, it bounces when it lands on pavement. Any that bounces and moves off pavement is wasted. This is one reason why city managers prefer brine over solid salt.
If a salt particle first lands towards the center of pavement, it may bounce, but it's still likely to remain on the pavement where it can be counted. Some of the salt sowed near the edge of pavement may bounce off the pavement, where it can't be counted. Hence the few inches near the edge should (in theory) have less salt. I did observe salt located off pavement, but within a few inches of the sidewalk.
Someone spreading salt by hand will aim for the pavement, leading to more salt towards the center of a sidewalk square. Link.
The opposite can happen when sowing salt into boxes in order to create "field standard" photos. The particles aren't going to fall straight down--instead some may have some sideways velocity. On first bounce, they may acquire more sideways velocity. Some particles moving sideways will encounter the wall of a box and fall downward, leading to an increased density of particles near the edge of the box.
The "box edge effect" is easily observed in many parking lots. Over time after salting, wind, pedestrians, or vehicles scatter the salt, leading to a very noticeable buildup at the edges near the curb. Link.
I didn't try to measure a salt decrease near the edge of sidewalks. However, the edge effect is well known to ecologists. It explains why soil development is poor near a cliff edge or on small islands. In those cases, wind rather than bouncing is what removes soil particles. Prior to this study, I have seen clear evidence that wind was removing salt from sidewalks.
Contaminants
Various forms of dirt, mud, organic matter, or excrement can make visual estimates more difficult. Especially, avoid sampling near mud trapped on a sidewalk because it indicates water has pooled in that spot.
I was surprised to find what appears to be bird excrement in some samples--white and pasty. It likely persists because it's insoluble, plus too sticky to blow away. One sample had up to 6 spots of excrement. While it appears white like salt particles, I don't believe it will affect estimates much. Link.
I picked contaminants like sticks or pebbles from swept samples. Be alert for small, white pebbles if you are sweeping up samples. White pebbles can be distinguished by testing to see if they dissolve in water. Link. The only serious problem I can think of would be if crushed concrete or marble particles had been scattered on the pavement.
Overall, contaminants aren't a problem as long as you can see a portion of the sidewalk.
Pedestrian Traffic
Since I studied salt at 3500 over about 8 hours on three days (2/16, 2/18, and 2/20), I could observe the number of pedestrians. The pathways were always clear. Along the front sidewalk parallel to University Av., I saw about 1 pedestrian per hour. On the E-W walkway between parking and the building front, saw only one pedestrian during the eight hours. I never saw pedestrians use the N-S walkways parallel to streets bordering the property. Most foot traffic goes between the rear parking lot and the rear entrance. There are almost no students or customers visiting.
One of the more dangerous slip/fall locations was a spot where water puddles at the property's SW corner, seen during two of three visits. The danger depends on how quickly the puddle drains, but salt cannot reduce the hazard. Link. Link.
Quantification of salt at 10 sample sites
To estimate salt on a site, I used PG's photos (Draft 1, right), calibrated by volume or weight (right), trying to match the pattern I saw on the pavement to the pattern displayed in one of the photos. Link to PG's latest salt standard, Draft 2, page 1 and page 2.
I call the collage of photos on the right the "field standard." It's used to estimate salt on a "field sample." The "field sample" may be a 1 sq ft patch of sidewalk visually observed, or a photo of that spot, also about 1 sq ft.
The weights listed for each photo on PG's field standard represent the total weight of salt, scattered at various rates, on 250 square feet (or 10 sidewalk squares measuring 5 x 5 feet). Therefore, to calculate how much salt you could ideally collect in the field from a 1 sq ft patch of sidewalk: Multiply its "weight in lbs" printed by the photo on the field standard, x 16 oz, all divided by 250 sq ft.
Photos stored on Flickr show all photos taken on 2/20 at 3500 Univ. Ave. These photos provide the context of all sample sites, plus verticals and closeups of each site. Obviously, sampling only the heaviest spots will bias results towards "oversalting." These context photos allow an observer to judge how representative the 10 sampling locations are.
When selecting sample locations, I selected places of heavier salting-- the head of the "comet" described above, which were found on all walkways. Sites were selected in part because they were close to my vehicle (as you might expect an inspector to do). But I did not seek out the very heaviest clumps of salt. So I feel the sample locations are broadly representative of the many spots where a handful of salt landed. But not representative of salt in between these "landing spots."
Having picked a sidewalk square with fairly heavy salt, I then focused on a smaller area, roughly 18" square, that had the heaviest concentration.
For each spot, I picked a square (or photo) on PT's "field standard" representing a "weight" of salt that I thought matched best. Then I picked an adjacent square on the standard I thought matched second-best.
I found myself using how many particles touched one another as a helpful cue.
Since this was my first test of PG's photo standards in the field, my method changed from one sample to the next.
At these spots (#8 & #9) I made a visual estimate and also collected salt from within the template.
These are the most useful samples because they are standardized and can be compared to the actual salt weight I collected there. Link to vertical photos of all 10 sample locations. Go there to see how your estimates compare to mine.
I'm not showing the photos of the other 8 spots where I made estimates using the "field standard." That's because I wasn't using a standard method for indicating the exact spot, a standard method of taking the photo, or a length scale. The estimates for these spots are probably less accurate and the photo documentation less useful.
But if you want to see all 10 photos and even test the accuracy of your own estimates, go to these links on Flickr:
- All photos of 3500 Univ. Av. for context, taken on 2/20. Paper labels showing the location of the 1-10 sites of estimation can be seen in some of the photos.
- Closeup vertical photos of sites 1-10. Use these photos to test yourself, to see how close you come to my estimates, using PG's field standard. My estimates are given in the description of the album on Flickr. Don't peek there until you make your own estimates!
Results for sample #8
- Best estimate: 10.5 lb. per 250 sq ft
- Second-best estimate: 9.0 lb.
- Salt swept up=0.4 0z.
The large 2 white spots surrounded by a dark halo are likely bird excrement. On close inspection, about 4 additional smaller spots of excrement can be seen. I don't think these spots significantly affect the salt estimate. The blue template was cut from poster paper.
Results for sample #9
- Best estimate: 16.5 lb. per 250 sq ft.
- Second-best estimate: 13.5 lb.
- Salt swept up=0.9 oz.
- At #8, I collected 0.4 oz, or 60.0% of the amount of salt indicated by the first-choice visual estimate.
- At #9, I collected 0.9 oz. This actual salt was 105% of the salt indicated by my first-choice estimate, or 85% of the amount of salt indicated by the second-choice estimate. Measuring more salt than estimated indicates a bad estimate. Therefore, the second-choice estimate must be the more accurate one. This indicates of a small degree of inaccuracy in the first-choice estimate at #9--a one-step overestimation.
Discussion
- Degree of unevenness in tone, or mottling, of the background concrete.
- Degree of roughness. Depressions can trap salt or create shadows.
- Contrast of the photo. Easily manipulated after the photo is taken.
- Lightness/darkness of the concrete background. The darker the background, the easier it is to perceive the bright salt particles.
- Moisture makes the concrete darker, but if the salt also appears darker, then moisture provides no visual benefit.
- Lighting, whether diffuse or bright sun. I don't know whether bright sun would promote more accurate estimates. It's possible that the observer might perceive the long shadow of a salt particle as a second particle.
- Foot or vehicle traffic grinds down particles, leading to fewer (or smaller) distinct particles, plus a lighter background (where salt dust has accumulated).
Placing a 1 square foot template around a sample (#8 & #9) proved extremely helpful for documentation. When taking the documentation photo, just filling the camera viewfinder with the 1 ft square creates a length scale, while the template frames exactly the spot you are going to estimate, then gather salt from. Next, the template facilitates accurate removal of the salt by sweeping.
- More trials indoors with salt scattered on plain gray backgrounds. Measure accuracy of comparison of field standards to field samples.
- Trials indoors with mottled backgrounds similar to concrete.
- Trials indoors assessing how other variables degrade accuracy.
- Trials using ImageJ to create and compare "fingerprints" of salt samples.
- Trials in the field (in good weather any time of year) under varying conditions, comparing visual estimates to known amounts of salt spread.
- Develop protocols for how to handle extreme variability in salt distribution, and how to minimize bias.
- Decide on trigger level of salt for citations.
- Develop agreed on protocols for inspection of sites, and for training and regularly testing inspectors.
- Create citizen science opportunities and encourage volunteer inspectors to file complaints.
.jpg)
Posts
