The 33S(p,\gamma)34Cl reaction is important for constraining predictions of certain isotopic abundances in oxygen-neon novae. Models currently predict as much as 150 times the solar abundance of 33S in oxygen-neon nova ejecta. This overproduction factor may, however, vary by orders of magnitude due to uncertainties in the 33S(p,\gamma)34Cl reaction rate at nova peak temperatures. Depending on this rate, 33S could potentially be used as a diagnostic tool for classifying certain types of presolar grains. Better knowledge of the 33S(p,\gamma)34Cl rate would also aid in interpreting nova observations over the S-Ca mass region and contribute to the firm establishment of the maximum endpoint of nova nucleosynthesis. Additionally, the total S elemental abundance which is affected by this reaction has been proposed as a thermometer to study the peak temperatures of novae. Previously, the 33S(p,\gamma)34Cl reaction rate had only been studied directly down to resonance energies of 432 keV. However, for nova peak temperatures of 0.2-0.4 GK there are 7 known states in 34Cl both below the 432 keV resonance and within the Gamow window that could play a dominant role. Direct measurements of the resonance strengths of these states were performed using the DRAGON recoil separator at TRIUMF. Additionally two new states within this energy region are reported. Several hydrodynamic simulations have been performed, using all available experimental information for the 33S(p,\gamma)34Cl rate, to explore the impact of the remaining uncertainty in this rate on nucleosynthesis in nova explosions. These calculations give a range of ~ 20-150 for the expected 33S overproduction factor, and a range of ~ 100-450 for the 32S/33S ratio expected in ONe novae.