Spatio-temporal modeling of particulate air pollution in the conterminous United States using geographic and meteorological predictors

Yanosky, Jeff D; Paciorek, Christopher J; Laden, Francine; Hart, Jaime E; Puett, Robin C; Liao, Duanping; Suh, Helen H

doi:10.1186/1476-069X-13-63

Environmental Health

Table 2 Bias and precision statistics from cross-validation (CV) of PM _2.5 , PM ₁₀ , and PM _2.5–10 models from 1988–1998

From: Spatio-temporal modeling of particulate air pollution in the conterminous United States using geographic and meteorological predictors

Pollutant	Region^A	Monthly values									Spatial CV R² ^G
Pollutant	Region^A	N^B	N excluded^C	Model R^2D	CV R²	Intercept^E	Slope^E	NMBF (%)^F	CVMAE^F	NMEF (%)^F	Spatial CV R² ^G
PM_2.5 ^H	All	10,823	0	0.82	0.77	0.2	0.92	−0.8	1.81	14.8	0.88
	Northeast	2,455	0	0.77	0.72	0.2	0.93	−0.5	1.73	13.0	0.85
	Midwest	1,564	0	0.74	0.70	0.3	0.88	−1.3	1.64	12.2	0.78
	Southeast	2,385	0	0.76	0.73	0.1	0.97	−1.0	1.74	11.7	0.78
	Southcentral	1,446	0	0.74	0.67	0.4	0.86	0.2	1.67	15.6	0.77
	Southwest	1,205	0	0.84	0.77	0.2	0.93	1.3	2.49	22.8	0.89
	Northwest	809	0	0.75	0.56	0.4	0.81	−6.1	2.12	26.8	0.60
	Central Plains	959	0	0.77	0.67	0.1	0.95	−1.2	1.55	20.2	0.81
PM₁₀	All	145,398	12	0.71	0.58	0.5	0.82	−3.3	5.44	21.8	0.66
	Northeast	35,593	5	0.66	0.57	0.5	0.83	−2.5	4.71	18.2	0.57
	Midwest	16,276	0	0.65	0.51	0.7	0.78	−2.9	5.57	20.4	0.55
	Southeast	26,882	0	0.72	0.61	0.5	0.84	−1.7	4.04	15.6	0.57
	Southcentral	12,668	0	0.66	0.50	0.9	0.72	0.1	5.20	21.2	0.50
	Southwest	23,586	5	0.76	0.60	0.5	0.84	−5.1	7.52	27.4	0.68
	Northwest	8,874	2	0.63	0.52	0.9	0.72	−4.2	6.86	26.0	0.66
	Central Plains	21,069	0	0.64	0.50	0.6	0.76	−7.4	5.57	31.7	0.66
PM_2.5-10 ^I	All	4,032	205	0.61	0.45	0.7	0.70	−4.7	4.73	42.6	0.56
	Northeast	802	48	0.52	0.32	0.9	0.56	−14.6	4.28	46.5	0.37
	Midwest	378	21	0.64	0.47	0.8	0.66	−4.4	4.28	36.9	0.44
	Southeast	771	58	0.35	0.12	1.1	0.51	1.7	3.66	43.3	0.09
	Southcentral	453	2	0.58	0.43	0.8	0.68	−5.1	5.35	38.2	0.38
	Southwest	835	34	0.70	0.53	0.4	0.81	−8.4	6.15	44.4	0.70
	Northwest	271	15	0.60	0.54	0.6	0.85	27.0	4.14	47.5	0.63
	Central Plains	522	27	0.42	0.32	0.7	0.69	−4.8	4.81	45.9	0.42

^ACorresponds to regions shown in Figure 1.
^BIncludes data from CV sets one through nine; see text for details.
^C12 PM₁₀ values above 150 μg/m³ (>99.99th percentile) were excluded from CV statistics as outliers. Extreme values may have been due to local events such as wildland or other fires, dust storms, etc.
^DCalculated on the native rather than natural-log scale and among observations used for CV (for only the year 2000 for PM_2.5 and PM_2.5–10) for comparison to the CV R². For PM_2.5–10, predicted levels<=0 were removed.
^EFrom major axis regression of predictions on measurements (both are natural-log transformed monthly means); see text for details.
^FNMBF is normalized mean bias factor; CVMAE is cross-validation mean absolute error; NMEF is normalized mean error factor; see text for details.
^GSpatial CV R² calculated at 1,031 and 422 sites with >3 valid monthly-average measurements for PM_2.5 and PM_2.5–10, respectively, and at 1,502 sites with >35 valid monthly-average measurements for PM₁₀.
^HMeasured and predicted levels (rather than the natural-log of the PM_2.5 to PM₁₀ ratio) were compared.
^ICalculated as the difference between monthly PM₁₀ and PM_2.5 measurements and, separately, monthly PM₁₀ and PM_2.5 model predictions. Of the 207 values excluded as outliers, 8 were removed due to extreme PM₁₀ or PM_2.5 measurements; an additional 197 were due to measured or predicted PM_2.5–10 below the limit of detection of 0.57 μg/m³ (<3.6th percentile of measured and <1.5th of predicted PM_2.5–10).

Back to article page

ISSN: 1476-069X

Contact us

General enquiries: journalsubmissions@springernature.com