Table 2.

Regression coefficients for mass reconstruction (Eq. D) using various regression methods for Interagency Monitoring of Protected Visual Environments (IMPROVE) network samples collected at urban Fresno supersite in CA from 3 September 2004 to 31 December 2010

	OLSb	OWLSc	EVd	Average ± standard deviatione	Minimum–maximum
Categorya
Coefficient a1 (SO4 =)	1.61	0.90	0.93
Coefficient a2 (NO3 −)	0.92	0.85	0.88
Coefficient a3 (OC)	1.67	1.74	1.71
Coefficient a4 (Other)	0.59	0.78	0.78
Species
Avg. SO4 = (μg/m3)				1.33 ± 1.26	0.079–25
Avg. NO3 − (μg/m3)				3.9 ± 4.9	0.138–38
Avg. OC (μg/m3)				3.2 ± 2.5	0.54–24
Avg. Other (μ/m3)				2.6 ± 1.8	0.53–26

^a http://views.cira.colostate.edu/web/. To ensure data quality, only samples with species concentrations exceeding their uncertainties were included for regression analyses

^bOrdinary least squares − no weighting

^cOrdinary weighted least squares − weighting depends on uncertainty of independent variable

^dEffective variance least squares − weighting depends on uncertainties of both the independent (i.e., SO₄ ⁼, NO₃ ⁻, OC, and Other) and dependent variables (Watson et al. 1984)

^eAverage and calculated ranges are as follows (number of samples in all averages = 708)