Table 3.
Model | AICc | ΔAICc | AICc weight | K | Deviance |
---|---|---|---|---|---|
(a) Boreal chorus frogs | |||||
R m(evap + depth + veg) | 6,784.40 | 0.00 | 0.94 | 55 | 6,672.25 |
R m(evap + veg) | 6,789.89 | 5.48 | 0.06 | 52 | 6,683.96 |
(b) Spotted frogs | |||||
R m(RO + depth + RO × depth) | 5,786.03 | 0.00 | 0.46 | 55 | 5,673.87 |
R m(RO + depth) | 5,786.97 | 0.94 | 0.29 | 52 | 5,681.04 |
R m(ROave + Depthave + t) | 5,788.8 | 2.76 | 0.12 | 76 | 5,632.66 |
R m(ROave + t) | 5,789.18 | 3.15 | 0.09 | 73 | 5,639.37 |
R m(RO) | 5,790.85 | 4.81 | 0.04 | 49 | 5,691.13 |
Models include the best supported structure for wetland dynamics, ψm(Precipave + t), and detection probability, δ(Vegave + t). States (m) were modelled independently and covariates are abbreviated as follows: veg represents site‐specific vegetation cover that varies annually, depth is the maximum wetland depth, evap is total April to June evapotranspiration in mm, and RO is annual runoff in millimeters. The terms Depthave, Precipave, ROave, and Vegave represent the average value of these respective variables at sites over years and thus represent spatial variation only. Model selection information including Akaike's Information Criterion corrected for small sample size (AICc), the difference in AICc from the top ranked model (ΔAICc), model weights, number of parameters (K), and deviance are shown for each model.