Using Bayesian state-space models to understand the population dynamics of the dominant malaria vector, Anopheles funestus in rural Tanzania

Table 3 Model selection: Description of all models fitted with and without environmental covariates and their corresponding delta-Deviance Information Criterion ΔDIC

Model	Removed covariate(s)	Fitness measure	Penalized deviance (pD)/DIC	ΔpD/DIC
Model 1-Full	None		35,500	25,339
Model 2	Temperature	Larval survival	10,555	394
Model 3	Rainfall	Larval survival	10,277	116
Model 4	1 week cumulative rainfall*density dependency	Larval survival	11,886	1725
Model 5	Density dependency	Larval survival	11,011	850
Model 6	Temperature	Larval development period	10,873	712
Model 7^a	Temperature	Adult survival	10,161	0
Model 8	Model 7—temperature	Larval survival	10,299	138
Model 9	Model 7—rainfall	Larval survival	15,766	5605
Model 10	Model 7—1 week rainfall:density dependency	Larval survival	10,199	38
Model 11	Model 7—density dependency	Larval survival	10,383	222
Model 12	Model 7—Temperature	Larval development	10,497	336

Model 8–12 consists of model-7 minus one more environmental covariate. Model 4 involved the removal of the interaction term
^aThe best model (lowest DIC/Penalized Deviance) value-model-7 followed by model-10, * Interaction between 1 week cumulative rainfall and density dependency

ISSN: 1475-2875