Figure 5. Fits to the three different sets of data (columns) from each of the three different retrieval techniques (rows). The first scenario consists of the four IRAC photometry channels. The second scenario consists of the four IRAC photometry channels, ground-based H- and Ks-band photometry, and HST WFC3 spectroscopy. The third scenario is representative of a FINESSE-like future, spaceborne telescope. The best fits from each scenario and technique are shown in light blue. The light-blue circles with the black borders are the best fits binned to the data. The chi-squared per data point from the optimal estimation best-fit broadband scenario, multi-instrument scenario, and future telescope scenario are, respectively, 0.197, 0.686, and 0.955. The bootstrap Monte Carlo and the differential evolution Markov chain Monte Carlo approaches generate many thousands of spectra. The median of these spectra is shown in blue and the 1σ and 2σ spread in the spectra are shown in dark- and light-red, respectively. The best fit from the thousands of spectra are shown in light blue. The best-fit reduce-chi-squares from BMC and DEMC are of similar values to those from OE. The dotted curves at the bottom of each panel are the broadband filter transmission functions. The insets are a zoom in of a spectral region between 1 and 2 μm to better show the spread in the spectra. Note that there is virtually no spread in the spectra for the future telescope case.
...read more