Table 3.

Parameter estimates for models with G and Q also assumed unknown and estimated: the block on the left fits the model assuming independence between the fields (i.e. τ_NF = 0), while the block on the right corresponds to a dependent model that estimates τ_NF

	Independent model (τNF = 0)			Dependent model (τNF estimated)
	Estimates	2.5%	97.5%	Estimates	2.5%	97.5%	True value
β	5.21	3.86	6.39	5.61	4.69	6.12	5.00
τN	1.01	0.77	1.36	0.92	0.66	1.40	1.00
GSD(τN)	2.73	2.40	3.21	2.61	2.25	3.26	2.72
τF	0.61	0.47	0.82	0.57	0.40	0.86	0.64
GSD(τF)	2.18	1.98	2.47	2.13	1.88	2.53	2.23
τNF	0	0	0	0.45	0.25	0.82	0.50
G	351.23	298.17	408.55	384.84	320.89	419.10	351.48
Q	12.58	11.07	16.42	14.87	11.18	16.79	13.78
log()	4.52	2.51	6.50	4.55	1.96	7.22	4.72
log()	3.28	1.70	4.85	3.27	−0.19	6.62	4.20
β	Interzonal airflow rate (m3 min−1)
G	Contaminant's mass emission rate (mg min−1)
Q	Room supply and exhaust airflow rate (m3 min−1)
τN, τF	Variance for measurement error at near and far field (log scaled)
τNF	Covariance between near and far field for measurement error (log scaled)
GSD(τN)	Geometric standard deviation for measurement error at near field
,	Steady-state concentration at near and far field (mg m−3)

The simulated data set was generated using true parameter values indicated in the rightmost column.