[1] In the paper ‘‘Three-dimensional nonlinear evolution of equatorial ionospheric bubbles with gravity wave seeding and tidal wind effects’’ by M. J. Keskinen and Sharon L. Vadas (Geophysical Research Letters, 36, L12102, doi:10.1029/2009GL037892), Keskinen and Vadas [2009, hereafter KV09] presented the three-dimensional nonlinear evolution of equatorial ionospheric bubbles using a realistic lower atmospheric gravity wave (GW) source. There are a number of misrepresentations and errors by KV09 in regards to the presentation and discussion of the GW source function that was used in the plasma model. The figures and discussion of KV09 refer to a different GW specification, despite the fact that the correct GW specification was used in the plasma model, results of which are presented in Figure 4 of KV09. As a result, the model data shown in Figures 1, 2, and 3 of KV09 were not used to seed the 3D plasma model described by KV09. Correct Figures 1–3 are supplied in this erratum. Additionally, part of paragraph 8 and all of paragraphs 9, 10, and 11 of KV09 should be disregarded and replaced with the text below. [2] The dispersion relation discussed in the latter portion of paragraph 8 in KV09 was not the GW dispersion relation, but was rather the full complex dispersion relation derived from the compressible fluid equations by Vadas and Fritts [2005, hereafter VF2005]. Due to formula errors in KV09, readers interested in the compressible dispersion relation should refer directly to VF2005. The last half of paragraph 8 in VF09 should be replaced by the following: For anelastic GWs with horizontal phase speeds much less than the speed of sound, the GW dispersion relation is given by equation (26) from VF2005 or by equation (47) from Vadas and Fritts [2009, hereafter VF2009]. [3] Vadas and Liu [2009, Figure 1] shows a GOES-12 satellite image on 01 October, 2005, at 21:22 UT, in Brazil. For the KV09 study, we simulated the GWs excited by the energetic convective cluster at 59.0 W and 13.5 S. This cluster contained several tightly-clumped convective plumes, each having an approximate full-width horizontal diameter of 20 km and updraft velocities of 40 m/s. We defined a convective cluster to be composed of 3 convective plumes in an equilateral triangle configuration, with a separation of 50 km between the plume centers (VF2009). Figure 1 shows the GW spectrum excited by this cluster. Figure 1 replaces KV09’s Figure 1. This cluster excites GWs with horizontal wavelengths of lH > 100 km and vertical wavelengths of lz > 50 km, which are capable of penetrating well into the thermosphere prior to dissipating [Vadas, 2007; Fritts and Vadas, 2008]. [4] The ray trace model we used was described by VF2009. The hyperbolic tanh zonal wind and temperature models shown in Figure 2 of KV09 were not used in either the ray trace or plasma models. Instead, the wind and temperature models we used were determined mainly from the Thermosphere-Ionosphere-Mesosphere-Electrodynamics General Circulation Model. Vadas et al. [2009] describe the background wind and temperature models used. Because they include semidiurnal and diurnal tides, both zonal and meridional background wind components were present. [5] The GWs whose spectrum is plotted in Figure 1 were located at 59.0 Wand 13.5 S at an altitude of z = 13.6 km at 21:22 UT. Those GWs were then ray traced into the middle and upper atmosphere through the spatially and temporallyvarying background winds and temperatures. We then reconstructed the GW wind, temperature, and density perturbations as a function of (x, y, z, t) using the method described by VF2009. We extracted the vertical profiles (cuts) of this solution at 10 different locations and times. These profiles were inputted into the 3D plasma model