The experimental data up to 40 GHz have been reported by
(1) L. Halonen and I. M. Mills, 1978, J. Mol. Spectrosc. 73, 494.
The low symmetry of the molecule causes a small c-dipole moment component. Several transitions obeying c-type selection rules were reported, facilitating predictions to higher frequencies.
Transition frequencies were reported in part to kilo-Hertz digits and to ten kilo-Hertz digits, respectively. The assigned uncertainties are 10 kHz for the former, as stated in the article. The latter data appear to be less accurate, and 50 kHz were assigned. These uncertainties yielded a balanced fit, though the latter may be slightly conservative.
With respect to the first entry from May 2011, additional extensive millimeter and submillimeter data have been reported by
(2) L. Nguyen, A. Walters, L. Margulès, R. A. Motiyenko, J.-C. Guillemin, C. Kahane, and C. Ceccarelli,
2013, Astron. Astrophys. 553, Art. No. A84.
Some of the higher K_a data were already excluded from the fit in (2) because of suspected perturbations with one or more low-lying vibrational states. Here, all data above K_a = 15 were omitted, resulting in a somewhat different parameter set.
Predictions should be reliable throughout, at least up to K_a = 15 or 16, well beyond the needs of astronomers.
Note: Non-trivial spin-statistics do not matter for this low symmetry (C_S) isotopolog.
The dipole moment has been derived from the main isotopolog, see e041505.cat, taking into account the rotation of the inertial axes. The magnitude of the c-component should be taken with some care.