Fig. 26.

The change in the $g$ and $u$ PDFs at NNLO for $Q={10}^4\mathrm{G}e{V}^2$, in going from the MSTW values to those in the present global NNLO fit [with default ${\mathit{\alpha }}_S(M_Z^2)=0.118$], which includes the LHC data. Also shown is the NNLO fit for the optimal value ${\mathit{\alpha }}_S(M_Z^2)=0.1172$