Bolometric corrections based on the optical-to-ultraviolet continuum spectrum of quasars are widely used to quantify their radiative output , although such estimates are affected by a myriad of uncertainties , such as the generally unknown line-of-sight angle to the central engine .
In order to shed light on these issues , we investigate the state-of-the-art models of Hubeny et al .
( 2000 ) that describe the continuum spectrum of thin accretion discs and include relativistic effects .
We explore the bolometric corrections as a function of mass accretion rates , black hole masses and viewing angles , restricted to the parameter space expected for type-1 quasars .
We find that a nonlinear relationship \log L _ { bol } = A + B \log ( \lambda L _ { \lambda } ) with B \leq 0.9 is favoured by the models and becomes tighter as the wavelength decreases .
We calculate from the model the bolometric corrections corresponding to the wavelengths \lambda = 1450 \AA , 3000 \AA and 5100 \AA .
In particular , for \lambda = 3000 \AA we find A = 9.24 \pm 0.77 and B = 0.81 \pm 0.02 .
We demonstrate that the often-made assumption that quasars emit isotropically may lead to severe systematic errors in the determination of L _ { bol } , when using the method of integrating the “ big blue bump ” spectrum .
For a typical viewing angle of \approx 30 ^ { \circ } to the quasar central engine , we obtain that the value of L _ { bol } resulting from the isotropy assumption has a systematic error of \approx 30 \% high compared to the value of L _ { bol } which incorporates the anisotropic emission of the accretion disc .
These results are of direct relevance to observational determinations of the bolometric luminosities of quasars , and may be used to improve such estimates .