1: \begin{abstract}

2: In order to use strong gravitational lens time delays to measure

3: precise and accurate cosmological parameters the effects of mass along

4: the line of sight must be taken into account.  We present a method to

5: achieve this by constraining the probability distribution function of

6: the effective line of sight convergence $\kext$. The method is based

7: on matching the observed overdensity in the weighted number of

8: galaxies to that found in mock catalogs with $\kext$ obtained by

9: ray-tracing through structure formation simulations. We explore

10: weighting schemes based on projected distance, mass, luminosity, and

11: redshift.  This additional information reduces the uncertainty of

12: $\kext$ from $\sigmak\sim$0.06 to $\sim$0.04 for very overdense lines

13: of sight like that of the system \blens. For more common lines of

14: sight, $\sigmak$ is reduced to $\lesssim$0.03, corresponding to an

15: uncertainty of $\lesssim3$\% on distance. This uncertainty has

16: comparable effects on cosmological parameters to that arising from the

17: mass model of the deflector and its immediate environment.

18: Photometric redshifts based on g, r, i and K photometries are

19: sufficient to constrain $\kext$ almost as well as with spectroscopic

20: redshifts.  As an illustration, we apply our method to the system

21: \blens.  Our most reliable $\kext$ estimator gives

22: $\sigmak$=0.047 down from 0.065 using only galaxy counts.  Although

23: deeper multi-band observations of the field of

24: \blens\ are necessary to obtain a more precise estimate, we conclude

25: that griK photometry, in addition to spectroscopy to characterize the

26: immediate environment, is an effective way to increase the precision

27: of time-delay cosmography.

28: \end{abstract}

29: