ESCANABA — The amount of DNA from the virus responsible for COVID-19 in Escanaba’s sewer water appears to be dropping, but the two most recent samples submitted to the state’s surveillance program offer very different pictures of the virus.

According to the Sentinel Wastewater Epidemiology Evaluation Project, commonly known a “SWEEP,” the most recent sample submitted by Escanaba Waste Water Treatment Plant on Jan. 10 contained more viral DNA than 75% of samples collected at the wastewater treatment plant since the city began participating in the project in August of 2021. The last sample taken before that date, which was dated Jan. 7, contained more viral DNA that 98% of samples.

SWEEP samples are collected twice a week to minimize inconsistencies caused by sample or reporting errors, which makes the gap between the Jan. 10 and Jan. 7 samples significant. When compared to the Jan. 3 sample, which had more viral DNA than 90% of samples, the Jan. 7 sample suggests an increase of 8 percentiles in viral activity. However, comparing the Jan. 7 and Jan. 10 samples suggests a 23-percentile drop in viral DNA in just three days.

There are two possible explanations for the sudden shift: either the amount of detectable virus rose and then dramatically dropped or there was a reporting error in one of the two samples.

SWEEP does not report a margin of error for samples, and calculating one would be tricky. The program monitors 20 different water treatment plants across the state, and each site’s data is only compared to other data from the same site.

This means that if, for example, an error was reported from Traverse City’s samples, there would be no impact on the data from Petoskey.

It also means that the each individual site must be consistent in their sampling methods and labs must take care with testing, as there is no universal benchmark for sample data.

So far, only one sample has strongly suggested any error in the Escanaba dataset, a 6th percentile sample from Nov. 21, 2023, that was sandwiched between a 97th percentile sample on Nov. 20 and a 98th percentile sample on Nov. 27. It’s impossible to know from SWEEP’s reported data if there was an error in sample collection, laboratory testing, or even typo when data was entered into the system.

Still, the Nov. 21, 2023 sample had a significant effect on SWEEP’s 15-day trend data, which uses the average viral load to present increases and decreases as percentage ranges, based on exponential growth or decay. In the case of the Nov. 21 sample, the sudden drop translated into both a substantial decrease in 15-day trend data for one period and a substantial increase in the next, when the samples once again were reported in percentiles in the upper-90s.

Looking at the most recent data, the last two samples have also resulted in a drop. As of Jan. 13, viral load is reported to have dropped between 10% and 99% — the second lowest range for decreasing viral activity on a scale that goes as high as 1,000% or more for increase and decreases.

If there was an error in the Jan. 7 sample, the city would be on a more consistent on a downward trajectory of viral activity from the 93% and 90% samples of Dec. 28, 2023 and Jan. 3, 2024. However, because the late December and early January samples marked an increase in average viral activity that started just after Christmas, an error in the Jan. 10 sample would mean the amount of detectable virus had continued to climb.

Even perfect sample data would not necessarily reflect how many people in Escanaba are actually sick with COVID-19. Viral load varies greatly even among individuals, and many people who are COVID-19 positive never show symptoms. Others who may be carrying and spreading the virus but appear healthy may be in the ten-day incubation period and will develop the disease later.

Shifts to at-home testing have made knowing the exact number of ill people almost impossible. However, historically, and increase in viral load in the sewer system has correlated to increases in illness in the community.