Abstract. A local scale, time dependent three-dimensional model of the neutral thermosphere was used to test the applicability of two previously published empirical relations between thermospheric vertical wind and velocity divergence, i.e., those due to Burnside et al. (1981) and Brekke (1997). The model self-consistently solves for vertical winds driven by heat and momentum deposited into the neutral atmosphere by high latitude ion convection. The Brekke condition accurately mimicked the overall "shape" of the three-dimensional model vertical wind field although, as written, it consistently overestimated the vertical wind magnitude by a factor of approximately 5/3, for the heating scenarios that we considered. This same general behavior was observed regardless of whether the forcing was static or rapidly changing with time. We discuss the likely reason for the Brekke condition overestimating the magnitude of our vertical winds, and suggest an alternative condition that should better describe vertical winds that are driven by local heating. The applicability of the Burnside condition was, by contrast, quite variable. During static heating, both the magnitude and the sign of the model vertical winds were predicted reliably at heights above those of maximum energy and momentum deposition per unit mass. However, below the thermal forcing, the Burnside condition predicted vertical winds of the wrong sign. It also introduced significant artefacts into the predicted vertical wind field when the forcing changed suddenly with time. If these results are of general applicability (which seems likely, given the way these relations are derived) then the Burnside condition could usually be used safely at altitudes above hmF2. But it should be avoided below this height at all times, and even at high altitudes during periods of dynamic forcing. While the Brekke condition (or our modified version of it) could likely be used in all circumstances, there are few experimental scenarios for which this would be useful. This is because evaluation of the Brekke condition would not usually be possible unless the vertical wind was already known in advance.