petitCODE cloudy model interpolated to the parameters best describing the data according to the posterior probability distribution (black line), as well as the SPHERE spectrophotometric data, GPI H-band spectrum, and L′ data point from Macintosh et al. (2015). For photometric data points, the x-error bar reflects the filter width rather than uncertainties. The gray lines represent 32 randomly drawn samples from the posterior probability distribution to reflect the spread of plausible model parameter combinations that fit the data. Photometric points describe the average flux in the respective filter, whereas the orange points describe the average flux in the respective filter for the best fitting model. The residuals are shown in multiples of 1σ uncertainties of the data.