Table 3.

Statistical modeling results between N1 and N2 gene concentrations and total COVID-19 cases during the 5-week lag time study period in city of Detroit, as well as Wayne, Macomb, and Oakland counties (* is shown in Fig. 5.)

Lag time	Model	N1-based results			N2-based results
		RMSE			RMSE
Unit of N1/N2 gene		gc/L	gc/d	gc/L of sanitary flow	gc/L	gc/d	gc/L of sanitary flow
3 week	Linear	7.22	135.76	11.78	12.40	926.30	5.62
	Autoregression	135.65	780.00	250.52	341.27	901.23	700.34
	Autoregression+ time effect	10.18	10.18	10.97	11.34	12.34	15.33
	Vector Autoregression	8.32	7.85	8.97	8.89	9.90	9.90
4 week	Linear	7.26	123.56	9.18	16.37	104.45	8.33
	Autoregression	182.92	234.90	635.69	132.35	730.74	500.62
	Autoregression+ time effect	7.50	7.47	7.20	9.75	7.39	7.33
	Vector Autoregression	8.00	7.99	8.62	6.88	8.31	7.62
5 week	Linear	1.83	48.97	2.62	13.95	36.19	2.36
	Autoregression	105.81	417.57	642.83	548.14	570.56	100.95
	Autoregression+ time effect*	1.47*	1.60	1.60	3.21*	1.60	1.42
	Vector Autoregression*	0.35*	0.53	4.44	7.57*	4.37	1.03