Table 4. Parameter estimates (mean and 95% credible interval) for our two models (see text) fitted to different data sets (2007, 2008), and the mortality estimates (mean and range of the total number of collisions at all turbines during n nights) based both on the model (second last column), and on the conventional “corrected count” method.

data set and model	n nights	Intercept α0	acousticactivity α1	wind velocity α2	windvelocity2 α3	mortality estimate from model	correctedcount
2007, 1-level observation model	473	−2.7 (−3.2, −2.1)	0.4 (0.1, 0.7)	−0.2 (−0.9, 0.3)	−0.2 (−0.6, 0.2)	37 (28–49)	38 (27–59)
2007, 3-level observation model	473	−2.3 (−2.9, −1.7)	0.5 (0.1, 0.9)	0.4 (−0.4, 1.3)	−0.7 (−2.0, 0)	38 (29–49)	38 (27–59)
2008, 1-level observation model	1225	−3.5 (−4.1, −3.0)	0.4 (0.1, 0.7)	−1.8 (−3.0, −0.8)	−0.8 (−1.6, −0.3)	56 (46–70)	57 (42–89)
2008,3-level observation model	1225	−4.1 (−5.3, −3.3)	0.5 (0.2, 0.8)	−2.9 (−5.6, −1.2)	−1.3 (−2.6, −0.4)	57 (46–71)	57 (42–89)