The version in the Kent Academic Repository may differ from the final published version. Users are advised to check http://kar.kent.ac.uk for the status of the paper. Users should always cite the published version of record.

In a companion work (Paper I) we detected a large population of highly variable Young Stellar Objects (YSOs) in the Vista Variables in the Via Lactea (VVV) survey, typically with class I or flat spectrum spectral energy distributions and diverse light curve types. Here we present infrared spectra (0.9-2.5 µm) of 37 of these variables, many of them observed in a bright state. The spectra confirm that 15/18 sources with eruptive light curves have signatures of a high accretion rate, either showing EXorlike emission features (∆v=2 CO, Brγ) and/or FUor-like features (∆v=2 CO and H 2 O strongly in absorption). Similar features were seen in some long term periodic YSOs and faders but not in dippers or short-term variables. The sample includes some dusty Mira variables (typically distinguished by smooth Mira-like light curves), 2 cataclysmic variables and a carbon star. In total we have added 19 new objects to the broad class of eruptive variable YSOs with episodic accretion. Eruptive variable YSOs in our sample that were observed at bright states show higher accretion luminosities than the rest of the sample. Most of the eruptive variables differ from the established FUor and EXor subclasses, showing intermediate outburst durations and a mixture of their spectroscopic characteristics. This is in line with a small number of other recent discoveries. Since these previously atypical objects are now the majority amongst embedded members of the class, we propose a new classification for them as MNors. This term (pronounced emnor) follows V1647 Ori, the illuminating star of McNeil's Nebula.

Time-varying phenomena are one of the most substantial sources of astrophysical information and their study has led to many fundamental discoveries in modern astronomy. We have developed an automated tool to search and analyze variable sources in the near infrared K s -band, using the data from the Vista Variables in the Vía Láctea (VVV) ESO Public Large Survey. This process relies on the characterization of variable sources using different variability indices, calculated from time series generated with Point Spread Function photometry of sources under analysis. In particular, we used two main indices: the total amplitude ∆K s and the eta index, η, to identify variable sources. Once variable objects are identified, periods are determined with Generalized Lomb-Scargle periodograms, and the Information Potential Metric. Variability classes are assigned according to a compromise between comparisons with VVV Templates and the period of the variability. The automated tool is applied on VVV tiles d001 and d002 and led to discovery of 200 variable sources. We detected 70 irregular variable sources and 130 periodic ones. In addition nine open cluster candidates projected in the region are analyzed, the infrared variable candidates found around these clusters are further scrutinized by cross-matching their locations against emission star candidates from VPHAS+ survey H α color cuts.