While optical imaging surveys are potentially powerful probes of dark energy, their effectiveness depends critically on the control of a number of systematic errors. I will describe tools and techniques we have developed for mitigating several sources of error in future galaxy clustering and cosmic shear surveys. In particular, we have proposed a method to calibrate photometric redshift errors using the cross-correlation of galaxies at different inferred redshifts and are developing improved shear power spectrum estimators and statistical models of the non-linearly evolved mass-density field with the goal of obtaining unbiased and accurate parameter constraints. The latter work makes use of a new statistical framework for using limited numbers of N-body simulations to build a model for the sample variance distribution of the nonlinear matter power spectrum. The performance of these tools with preliminary models is promising and a demonstration using the Deep Lens Survey data is in the works.