Methods for sampling street sweeping material - Standard Operating Procedures Revision as of 21:35, 2 December 2020 by Mtrojan (talk | contribs) (Created page with "{{alert|The methods described on this page were developed by University of Minnesota researchers as part of a study to develop a street sweeping credit methodology.|alert-info...")(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff) Methods for sampling street sweeping material - Standard Operating Procedures

Information: The methods described on this page were developed by University of Minnesota researchers as part of a study to develop a street sweeping credit methodology.

This page describes methods for sampling street sweeeping material and calculating statistical values (e.g. median, mean, confidence intervals) for sampled material. The methods described on this page were developed by University of Minnesota researchers as part of a study to develop a street sweeping credit methodology (see Acknowledgements for street sweeping credit method). Stormwater practitioners and permittees choosing to use the street sweeping credit to estimate pollutant reductions associated with street sweeping should apply the following sampling methods.

Subsampling Sweeper Load

The following subsampling should be done as soon as possible after the sweeper load has been collected and dumped, within 24 hours, and before any precipitation occurs.

Materials needed:

• Gallon-size Ziplock bag
• Permanent marker or label for bag
• Trowel

Method

1. Visually assess the sweeper load (e.g., in the hopper or on the pile), examining how much of the load is composed of soil and plant debris.
2. Before sample collection, label a gallon-sized ziplock bag with sample identification information (e.g., street sweeping route, date).
3. Wearing gloves, use a trowel to scoop at least five small amounts of sample into the gallon bag. Walk around the pile, scooping from various points. Make sure to collect a sample that accurately reflects the composition of the sweeper pile. If the outside of the pile appears to have dried, scrape away the outer layer to collect samples that have accurate moisture contents.
4. Collect sufficient sample to fill the ziplock bag ¾ full.
5. Samples should be stored in a refrigerator until moisture determination. If moisture cannot be determined within a day, the sample may be frozen.

Determining Moisture Content on Street Sweeping Loads

Materials needed

• Drying oven
• Balance (0.01 g)
• Aluminum tin (small bowl, small bread pan, soil tin, or other small tin)

Method

• If frozen, thaw sample overnight.

1. Using gloved hand, homogenize sample inside of the ziplock bag.
2. Weigh a tin. Record the tin weight.
3. Place 50-100 g of sample into the dish. Record the weight of the tin + sample.
4. Place in a drying oven at 105°C for 24-48 hours, or until the mass of the sample does not change.
5. Weigh the tin + dried sample. Record the weight.
6. Calculate the dry mass/wet mass ratio by subtracting the tin weight from each mass

\( (dry mass)/(wet mass) = ((dry masssample + tin - tin mass))/((wet masssample + tin - tin mass)) \)

7. If needed, calculate the percent moisture (on a dry mass basis) as

\( percentmoisture dry mass basis = ((wet masssample + tin -tin mass) - (dry masssample + tin - tin mass))/((dry masssample + tin - tin mass)) x 100 \)

Determining organic matter (OM) concentration of sweepings

We expected that percent organic matter (% OM) would be related to percent total phosphorus (% TP) in street sweepings because % OM should be a good indicator of the content of leaves and other plant-derived material in sweepings, and thus, insofar as % TP is relatively constrained in leaf litter, of % TP. We therefore evaluated the relationship between % OM and % TP using data from the Prior Lake Street Sweeping Study (Kalinosky et al. 2014, Kalinosky 2015). In that study we determined the OM concentration of sweepings using the Loss-on-Ignition (LOI) method, whereby a sample is combusted in a muffle furnace and the percentage of material combusted is assumed proportional to the percentage OM and/or organic carbon (OC) in the sample (see methods below). The LOI method has fallen out of favor for determining the OC concentration in soils and sediments because it can lead to overestimates of OC concentration when high temperatures drive off bound water from clay minerals and/or cause release of CO2 from carbonates (Goldin 1987). Thus the LOI-OC relationship depends on clay content and mineralogy (Howard and Howard 1990, Santisteban et al. 2004). However, LOI is more accurate for measuring OC in highly organic soil horizons (e.g., forest floor, or O horizon materials) with low mineral content (Westman et al. 2006).

Despite the limitations of using LOI for estimating OC concentrations in soils and sediments, we suggest that evaluating the use of % OM determined by LOI to predict % TP in street sweeping samples is appropriate for several reasons. First, we present empirical relationships directly between % OM determined by LOI and %TP and thus make no assumptions about the relationship between % OM determined through LOI and % OC. Second, LOI has distinct advantages over other methods of measuring either %OC or %TP in street sweepings because, unlike those methods, it can be done on relatively large samples and thus overcomes the need to do laborious fractionations of sweeping samples in order to obtain representative samples from heterogeneous sweepings. Third, it is relatively inexpensive and simple to do, so cities may be able to afford having samples analyzed for LOI in a commercial laboratory or they might choose to conduct the analyses themselves.

Methods for determining Loss-on-Ignition (LOI)

We recommend analyzing 3 subsamples of a sweeping sample and obtaining an average % OM value (see Appendix 1, Section 1.1 for methods for sampling a sweeping load). If a commercial laboratory is used for analysis, we recommend selecting a laboratory that combusts samples at 600-650°C for at least 3 hours (Hoogsteen et al. 2015).

Materials Needed

• Porcelain crucibles
• Dessicator with dessicant
• Tongs
• Muffle furnace

Methods
1. Samples should be dried at 105°C in a drying oven for 48 hours and allowed to cool in a dessicator with dessicant.
2. Weigh a labeled crucible and record weight. This is the “crucible mass”.
3. Weigh approximately 20 g dry subsample into crucible and record weight of sample + crucible. This is the “pre-combusted masssample+crucible”. Subsample should be as representative of the entire sample as possible.
4. Combust in muffle furnace at 600°C for 6 hours.
5. Remove from furnace and allow to cool in dessicator with dessicant.
6. Weight combusted sample and crucible and record weight. This is the “combusted masssample+crucible”.

Calculate %OM as follows

\( percent organic matter = ((precombusted masssample + crucible - crucible mass) - (combusted masssample + crucible - crucible mass)) / (precombusted massample + crucible - crucible mass)) X 100 \)

Calculating Confidence Intervals

For calculating the 95% confidence intervals around the mean, University of Minnesota researchers used a bootstrapping method due to the log-normal distribution of the variables, as the normal method can lead to biased estimates (Olsson 2005). On each bootstrap sample (R = 10000), we calculated the mean (µ) and variance (σ2) of the log-normal distribution, and used these parameters to return the sample mean on the untransformed scale using established equations (sample mean = exp(µ + σ2/2)). We report the percentile confidence intervals calculated using the boot.ci function in R package boot v. 1.3-24, R version 4.0.0 (R Core Team 2020).