1: \section{Cross section determination}

2:  The measured cross section in a particular kinematic bin, for example in $d\sigma/dQ^{2}$, was determined from

3: 

4:  \begin{equation}

5:  \frac{d\sigma_{\rm Born}}{dQ^{2}}=\frac{N_{\rm data}-N_{\rm bg}}{N_{\rm MC}}\cdot\frac{d\sigma_{\rm Born}^{\rm SM}}{dQ^{2}}, \nonumber

6:  \end{equation}

7: 

8:  where $N_{\rm data}$ is the number of data events, $N_{\rm bg}$ is the number of background events estimated from the MC simulation and $N_{\rm MC}$ is the number of signal MC events. The SM prediction $d\sigma_{\rm Born}^{\rm SM}/dQ^{2}$ was evaluated in the on-shell scheme using the PDG~\cite{pl:b592:1} values for the electroweak parameters and the CTEQ5D PDFs~\cite{epj:c12:375}. Consequently, the acceptance, as well as the bin-centring and 

9: radiative corrections were all taken from the MC simulation. The radiative corrections define the measured cross section to have only tree-level QED and electroweak contributions.

10: A similar procedure was used for $d\sigma/dx$ and $d\sigma/dy$.

11: 

12: The major sources of systematic uncertainty

13: in the CC cross sections come from the uncertainties in calorimeter energy scale and

14:  the parton-shower scheme. 

15: The former was estimated using a method detailed in previous publications~\cite{pl:b539:197,epj:c32:1} for the NC data sample. 

16: The resulting shifts in the cross sections 

17: were typically less than

18: $10\%$, but increased to $20\%$ in the highest $Q^2$ bin and

19: $30\%$ in the highest $x$ bin. 

20: 

21: To estimate the sensitivity of the results to the details of the simulation of

22: the hadronic final state,

23: the {\sc Lepto} {\sc Meps} model was used instead of the 

24: {\sc Ariadne} model for calculating the acceptance corrections.

25: The largest effects of $\sim 5\%$ were observed in the highest $Q^2$ and $x$ bins. 

26: 

27: The uncertainty in the small contribution from photoproduction was estimated by fitting a 

28: linear combination of the $\PTM /E_{T}$ distributions of the 

29: signal and the background MC samples to the corresponding distribution in the data, 

30: allowing the normalisation of the photoproduction MC events to vary. No cut 

31: on $\PTM /E_{T}$ was applied for this check. Varying the normalisation of the photoproduction events by 

32: the uncertainty in the fit of $\pm 30\%$ resulted in changes of the measured cross sections within $\pm 3\%$.

33: 

34: The systematic uncertainties of the selection cuts were estimated by varying

35: the threshold value of each selection cut independently   

36: by around 10\%, which is a reasonable match to the resolution.

37: The resulting shifts in the cross sections were typically within $\pm 5\%$.

38: 

39: A major source of systematic uncertainty

40: in the NC cross section came from the uncertainty in the parton-shower scheme, which gave changes in the 

41: cross section of typically within $\pm 2\%$ but up to 4\% at high $Q^{2}$. Uncertainty in the electromagnetic 

42: energy scale was estimated by varying the energy scale by $\pm 1\%$. However, due to the use of the 

43: double-angle reconstruction, the resulting shifts in the cross section were typically $<0.5\%$.

44: The systematic effects of the selection cuts were estimated by varying

45: the threshold value of each selection cut independently   

46: by values commensurate with the resolutions.

47: The resulting shifts in the cross sections were typically within $\pm 1\%$.

48: 

49: The individual uncertainties were added in quadrature separately

50: for the positive and negative deviations from the

51: nominal cross-section values to obtain the total systematic uncertainty. 

52: The uncertainty in the measured polarisation, $\delta P_{e}/P_{e}$, was 1.6\% using the LPOL 

53: and 3.5\% using the TPOL. The choice of polarimeter measurement was made to maximise 

54: the available luminosity for the analysis, while minimising the uncertainty in the measured polarisation, on a run-by-run basis.

55: 

56: The relative uncertainty in the measured luminosity

57: of 3.5\% was not included in the total uncertainty shown in the differential cross-section figures.

58: 

59: 

60: 

61: 

62: 

63: 

64: