Total Suspended Solids: A Seemingly Simple Metric that Continues to Frustrate the Stormwater World

Nearly twenty years ago, when I was cutting my teeth in a stormwater research and development laboratory, the debate over the most appropriate method to measure the solids concentration in stormwater samples was front and center.  The great total suspended solids (TSS) vs. suspended sediment concentration (SSC) deliberation.  Compounding the problem was the fact that TSS represented both a regulatory benchmark and a specific analytical method, of which there are multiple variations (SM 2540D1 or EPA 160.22), for measuring the solids concentration in stormwater samples.  From a regulatory perspective, we need to evaluate and subsequently rate our BMP's ability to capture and retain suspended solids relative to benchmarks, such as 80% annual load reduction.  To prove compliance, we turned to the TSS analytical methods initially established for wastewater to analyze our samples.  Unfortunately, research by USGS3 and others demonstrated that the TSS methods often fail to reliable account for coarse solids in stormwater samples and that the SSC method (ASTM D 3977-974) typically yields more representative results.  

The ensuing debate lasted for years but mostly revolved around the SCC method being the more accurate measure of suspended solids in stormwater samples.  Fast forward to today, and a new debate has emerged in stormwater circles over suspended solids analysis that calls into question BMP evaluation as we know it.

Since the SSC method emerged as the more accurate measure of the suspended solids concentration in stormwater, it has been incorporated into a number of BMP monitoring protocols and deemed acceptable by a growing list of regulatory agencies as a means of determining the TSS load reduction BMPs are achieving.  Well respected inspection programs like the New Jersey Department of Environmental Protection’s (NJDEP) stormwater manufactured treatment device (MTD) certification program have long relied on the SSC method in their laboratory protocols in favor of TSS methods.  Problem solved, right?  

Unfortunately, stormwater sample collection and analysis continues to be an error-prone endeavor, particularly when it comes to suspended solids.  Given the dynamic nature of solids loads in stormwater runoff and the fact that stormwater isn’t always well mixed, obtaining a representative sample to analyze is easier said than done.  Error introduced during analysis only compounds the problem.  

Putting sampling concerns aside for the sake of this discussion, analytical error remains a significant problem during suspended solids analysis.  Want proof?  Spike a half dozen samples with a known concentration of solids and send them off to a handful of different labs for analysis.  More often than not, the reported results will underreport the actual concentration.  Our internal research into this issue suggests the problem lies in rinsing, or more specifically, a lack of thorough rinsing.  When a technician fails to thoroughly rinse a sample bottle to remove and account for the solids left inside of it, the reported results are not representative of the actual solids concentrations.  Our experience suggests this issue is common, even at accredited laboratories, unless instruction and, in some cases, training on thorough rinsing is provided.  

On a positive note, alerting analytical facilities to this issue does yield more representative results.  Once technicians are made aware of the problem, recovery rates in spiked samples begin to approach 100%.  In other words, with a bit of instruction, we can eliminate a potentially significant source of error.  

Despite potential solutions, concerns over our ability to collect representative samples paired with the potential for analytical error have some in the industry calling for an end to sampling.  This is easier said than done, especially in the field, but in a laboratory setting, a modified mass balance approach to BMP monitoring is being championed.  More accurately described as a mass capture method, BMPs are evaluated by injecting a known mass of sediment over time at typical concentrations, and then the mass retained in the unit is recovered and measured to determine BMP removal efficiency.  On the plus side, this approach eliminates the potential to collect non-representative samples and the potential for error during sample analysis.  However, the method is certainly not without its own potential sources of error.  If the captured mass is not fully recovered, BMP performance will be underestimated, and much like sample analysis, the magnitude of this type of error is primarily driven by the care taken during the sediment cleanout/recovery process.  The point being, the potential for significant errors is there.  Proponents argue that its ok since the error works against the technology, so it’s more conservative.  I’m all for being conservative, but I’m also for being as accurate as possible when it comes to evaluating BMPs.

The mass capture approach also adds significant time, cost, and complexity to laboratory testing compared to sampling.  While we know that the error associated with sample collection and analysis can be minimized, we do not yet have a handle on how closely the mass capture method gets to actual performance or a credible means of determining if the added cost and time are justified.  This approach also raises the question of comparability.  The majority of our stormwater data comes from the field where sampling remains the norm.  I recognize that field monitoring is ripe with its own sources of error and bias, but I’ll save that for another day!  That said, are we justified in moving away from sampling in favor of a method that still may introduce significant error, at added time and cost, and with less potential comparability to other datasets? 

My personal opinion is that we need more research into both approaches to quantify the potential error and determine whether it can be eliminated or reasonably minimized.  A true side by side comparison of the methods that hold other variables constant so we can determine whether the added cost and complexity is really worth it.   For now, that doesn’t seem to be in the cards.  NJDEP is already in the process of updating their laboratory protocol for hydrodynamic separators and the new version will mandate that only the mass capture approach be utilized moving forward.  This change will trigger a mandatory retest for all of the currently certified HDS technologies.  Assuming most are retested, the total cost will likely exceed a million dollars that could have otherwise been invested in new innovations, all without a quantifiable upside to water quality.  It’s not the path I would have chosen, but this is stormwater after all, where we tend to shoot first and seek answers later.  

1. SM 2540 D, Standard Methods for the Examination of Water and Wastewater, APHA-AWWA-WPCF, 21st Edition, 1997

2. “Residue, Non-Filterable – Method 160.2 (Gravimetric, Dried at 103 – 105ºC).” EPA Methods for Chemical Analysis of Water and Wastes. EPA publication 600/4-79/020. March 1983

3. Gray, J.R., Glysson, D.G., Turcios, M. L., and Schwarz, E.G.  (2000).  Comparability of Suspended-Sediment Concentration and Total Suspended Solids Data.  U.S. Geological Survey Investigations Report 00-4191.  Available Online: