The molecule is non-linear albeit close to a linear one. The measurements were taken from
(1) Y. Sumiyoshi, H. Takada, and Y. Endo, 2004, Chem. Phys. Lett. 387, 116.
After omission of one line, the remaining data fit to well within the assumed 5 kHz. Only K_a = 0 transitions have been observed in the laboratory so far. Therefore, the rotational constants are known only approximately. Transitions involving K_a = 1 cannot be calculated reliably predominantly because the size of ε_aa is not known. Moreover, since the barrier to linearity may be small it is conceivable that knowledge of the K_a = 1 energies will not help much in locating higher energy levels. Therefore, only transitions involving K_a = 0 are provided. Nevertheless, the calculation may well be sufficient for all observational purposes. Transition frequencies with calculated uncertainties exceeding 0.2 MHz should be viewed with substabtial caution. Some deviation may also occur in the hyperfine structure splitting at higher rotational quantum numbers.
The partition function takes into account higher K_a values as far as possible at present.
The total dipole moment is from a quantum chemical calculation mentioned in that work.