Table 2.

Posterior estimates of the Bayesian spatio-temporal model parameters^§

		Ae. aegypti			Ae. albopictus
	2.5th percentile	50th percentile	97.5th percentile	2.5th percentile	50th percentile	97.5th percentile
Forest cover (within 300 m radius buffer)	− 0.140	− 0.074	− 0.008	− 0.018	0.031	0.080
Water cover (within 300 m radius buffer)	− 0.009	0.059	0.128	− 0.006	0.046	0.098
Grass cover (within 300 m radius buffer)	− 0.135	− 0.065	0.005	− 0.075	− 0.022	0.031
Managed vegetation cover (within 300 m radius buffer)	0.072	0.144	0.215	0.037	0.091	0.146
Drain line density (within 300 m radius buffer)	− 0.152	− 0.064	0.023	− 0.235	− 0.167	− 0.100
Distance to waterway	− 0.242	− 0.165	− 0.088	− 0.068	− 0.010	0.048
Distance to water area	0.072	0.152	0.232	− 0.076	− 0.014	0.047
Building age	0.351	0.421	0.490	0.270	0.323	0.376
Max temperature lag 1	− 0.037	− 0.027	− 0.018	− 0.027	− 0.016	− 0.004
(Max temperature lag 1)2	− 0.008	− 0.004	0.000	− 0.010	− 0.005	− 0.001
Mean relative humidity lag 1	− 0.020	− 0.011	− 0.002	− 0.032	− 0.020	− 0.008
Precipitation lag 1	0.003	0.010	0.016	− 0.018	− 0.010	− 0.002
Max temperature lag 2	− 0.040	− 0.031	− 0.021	− 0.029	− 0.017	− 0.005
(Max temperature lag 2)2	− 0.008	− 0.005	− 0.001	− 0.010	− 0.005	0.000
Mean relative humidity lag 2	− 0.015	− 0.006	0.003	− 0.021	− 0.009	0.003
Precipitation lag 2	0.006	0.012	0.018	− 0.003	0.005	0.013
Max temperature lag 3	− 0.032	− 0.023	− 0.013	− 0.025	− 0.013	− 0.002
(Max temperature lag 3)2	− 0.009	− 0.006	− 0.002	− 0.010	− 0.005	0.000
Mean relative humidity lag 3	− 0.016	− 0.007	0.002	− 0.021	− 0.008	0.004
Precipitation lag 3	0.007	0.013	0.020	0.007	0.015	0.023
Unstructured spatial effect (site)*	0.603	0.642	0.686	0.443	0.475	0.508
Unstructured spatial effect (planning area)*	0.383	0.514	0.708	0.240	0.337	0.487
Temporally structured effect**	0.048	0.058	0.071	0.076	0.092	0.113
Negative binomial size parameter	6.681	6.863	7.049	12.877	13.679	14.597

All the environmental and anthropogenic variables were standardized to zero mean and unit variance prior to model fitting, and a quadratic term for each of the standardized temperature variables was also created to introduce nonlinear effects. Estimates are posterior median and equal tailed 95% credible intervals

^*Refers to the posterior estimate of the standard deviation of the spatial random effect

^** Refers to the posterior estimate of the standard deviation of the independent second-order increment in the temporally structured effect

^§Each regression parameter (i.e. intercept and the coefficients of the fixed effects) was assigned a normal prior $\text{N}\left(0,5^2\right)$. We assumed a $\text{logGamma}\left(1,0.01\right)$ prior on the logarithm of the precision of the spatial random effects and independent second-order increment in the temporally structured effect. The default penalized complexity prior in R-INLA was specified for the logarithm of the negative binomial size parameter