NEHA April 2024 Journal of Environmental Health

FIGURE 5

Time Series Graph of the Soil and Water Assessment Tool (SWAT) Simulated and Observed Concentrations for Total Nitrogen (TN) and Ortho-Phosphate (Ortho-P) Using the Method for Estimating Attenuation of Nutrients From Septic Systems (MEANSS) to Estimate TN and Ortho-P Loading From Onsite Wastewater Treatment Systems to Prickly Pear Creek

1 x 10 1

1 x 10 3

1 x 10 2

1 x 10 0

1 x 10 1

1 x 10 -1

1 x 10 0

1 x 10 -2

1 x 10 -1

1 x 10 -3

1 x 10 -2

9 9

Simulated TN

Measured TN Simulated Ortho-P Measured Ortho-P Simulated Streamf low

Discussion The model performance as compared with the four sites where nitrate was measured in the groundwater indicates that the MEANSS anal- ysis is within ≤24% of field measured values. This finding is a good result for a screening tool, particularly when accounting for the high degree of uncertainty associated with measur- ing and partitioning nitrate loads in ground- water from OWTS discharges. MEANSS pro- vided nearly identical results to ArcNLET for the estimated hydrogeologic conditions. Comparing vadose zone reductions in MEANSS to a mechanistic model (e.g., STUMOD) showed large percent dierences in denitrification for some soils; however, the percent denitrification rates were generally low (<15%), which amplified small dierences in the two models when those dierences were compared in terms of relative percent dier- ences. Some of the error might be caused by the estimated HSG categories used in MEANSS based on the STUMOD soil descriptions.

in this comparison is less than the MEANSS estimated load for comparison to the Miller (1996) data for this same site (see Site 2 results). The estimates are dierent because the nitrogen load for this comparison includes all three of the MEANSS scoring cat- egories (Table 1). The Miller (1996) data esti- mated the groundwater load before entering the riparian; therefore, the MEANSS estimate for that comparison (Site 2) did not include the riparian soil scoring category. STUMOD The nitrogen reduction for 12 soil types using STUMOD and MEANSS is shown in Figure 3. The soil types in Figure 3 are those used by McCray et al. (2010). The HSG val- ues in parenthesis (A, B, C, and D) in Figure 3 are approximations of the HSG because the HSG classifications are not defined for the generic soil types used. MEANSS both underestimates and overestimates the nitro- gen reduction compared with STUMOD. The

STUMOD percent reduction for sandy clay soils (60%) was 4 times higher than the soils with the next-highest reduction (15% for 3 other soils). The cause of the reduction rate spike is unknown and was not matched by the MEANSS reductions for sandy clay soils. SWAT Model (Prickly Pear Watershed, Montana) Adjusting the MEANSS results to account for seasonal variation (see Methods) resulted in good calibration to the measured data for both TN and ortho-P as shown in Figures 4 and 5. Statistical summaries of the simulated and observed concentration data are based on 23 sample dates (19 dates for the load val- ues) and are provided in Figure 4. Time series plots of the observed versus simulated results are presented in Figure 5. The simulated median TN and total phosphate loads for Prickly Pear Creek using SWAT and MEANSS were 37% and 38% larger, respectively, than the measured loads.

April 2024 • Journal of Environmental Health