Table 1. Evidence for decelerating waves in WNV in NYC for 2000–2008.

	Year	Birds		Mosquito		Combined
		ρ	p-value	ρ	p-value	ρ	p-value
Convex hull	2000	−0.44	3.63×10−8	−0.236	0.0196	−0.461	9.16×10−9
	2001	−0.11	0.141	−0.225	0.0125	−0.260	3.98×10−3
	2002	−0.245	0.00353	−0.187	0.0353	−0.331	1.02×10−4
	2003	−0.374	5.93×10−5	−0.385	1.51×10−4	−0.497	4.66×10−8
	2004	−0.188	0.0397	−0.193	0.0238	−0.175	0.0336
	2005	−0.0033	0.486	−0.254	0.015	−0.123	0.103
	2006	−0.299	0.00596	−0.24	0.0101	−0.286	2.76×10−3
	2007	−0.125	0.119	−0.314	1.79×10−3	−0.371	1.94×10−4
	2008			−0.145	0.0713
Maximum distance	2000	−0.234	0.00316	−0.129	0.131	−0.26	0.00114
	2001	−0.174	0.0437	−0.259	0.00484	−0.331	0.000314
	2002	−0.173	0.0296	−0.2085	0.0219	−0.231	0.00535
	2003	−0.26	0.0044	−0.254	0.00988	−0.24155	0.006985
	2004	−0.188	0.04	−0.149	0.064	−0.171	0.0368
	2005	−0.0292	0.382	−0.0975	0.206	−0.04	0.3385
	2006	−0.281	0.0198	−0.2125	0.02	−0.2125	0.0204
	2007	−0.197	0.031	−0.147	0.09	−0.07575	0.238
	2008			−0.16	0.052
Boundary displacement	2000	−0.374	4.7×10−6	−0.317	0.00447	−0.401	8.39×10−7
	2001	−0.325	0.00197	−0.25	0.00699	−0.217	0.0163
	2002	−0.424	5.59×10−6	−0.297	0.00374	−0.441	9.73×10−7
	2003	−0.465	1.74×10−6	−0.336	0.001	−0.468	4.05×10−7
	2004	−0.174	0.0662	−0.3095	0.0015	−0.309	0.0012
	2005	−0.191	0.0442	−0.315	0.00503	−0.266	0.00607
	2006	−0.156	0.111	−0.218	0.0191	−0.269	0.005
	2007	−0.2435	0.0244	−0.416	8.325×10−5	−0.4415	1.31×10−5
	2008			−0.272	0.0059

P-values obtained from one-tailed hypothesis tests of Spearman's rank-order correlation between estimated wave speed and time elapsed. The extent of the infected area was estimated using three alternative methods: the convex hull encompassing all previous locations where WNV was detected, the average length of transects originating from the epicentre where they intersect these convex hulls, and the maximum Euclidean distance from the initial infected location (see Materials and Methods for details). The wave speed of WNV was calculated by dividing the increase of the extent of the infected area by the time elapsed since the last increase in the extent, for WNV-positive dead birds, WNV-positive mosquito pools, or the combination of both WNV-positive dead birds and mosquito pools. Values in bold italic are significant at the Holm-Bonferroni corrected level, while regular bold figures are significant at the nominal α = 0.05 level. The combined analysis, providing the greatest statistical power, provides evidence at the Holm-Bonferroni corrected level for a decelerating wave in 7 out of the 8 years studied, and for all years at the α = 0.05 level.