Intercombination and allowed transitions in O IV

We report on large-scale ab initio multiconfiguration Hartree-Fock calculations for important multiplets including the UV 0.01 multiplet, 2s(2)2p(2)P(J-)2s2p(24)P(J'), in O IV. The resulting transition probabilities should be accurate to +/- 5%-10% for intersystem lines and 1% or better for permitted lines. We present new calculations of line emission coefficients using these transition probabilities and collisional rate coefficients published by Zhang, Graziani, & Pradhan in 1994. We readdress the use of these lines as diagnostics of electron densities, paying particular attention to uncertainties in density determinations. We find (1) the absolute uncertainties in derived densities are difficult to assess, but are at least +/-40%, controlled by uncertainties in collision strengths; (2) our new calculations bring observed and computed line ratios into better agreement (earlier papers yielding systematically different densities make the agreement worse); (3) there is additional evidence for strong blends in the 1404.8 emission feature, as argued by other authors, and/or evidence for inaccurate laboratory wavelengths; and (4) the computed branching ratios are in good agreement with observed ratios to within observational uncertainties of +/-7%. We determine electron densities in a variety of solar features from HRTS data from the first and second flights of this instrument. The derived electron densities vary remarkably little between quiet Sun network and active regions, but resonance line intensities vary dramatically, and we discuss reasons for this. Finally, we discuss how new high-quality data from the SUMER instrument to be flown on Solar and Heliosphereic Observatory (SOHO) could be used to address outstanding problems concerning blends and heating of the solar transition region.