Climate variability, in terms of the climatic fluctuations in precipitation and potential evapotranspiration, impacts the variability of runoff at different timescales. This paper developed a new daily water balance model which unifies the probability distributed model and the SCS curve number method, and provides a unified framework for water balances across different timescales. The model uses a daily step but can be forced with climate inputs varying at different timescales. The model is applied to 82 MOPEX catchments, and the runoff at a coarser timescale is aggregated from the daily runoff. For runoff at each timescale, the relative role of each climate variability (daily, monthly, or inter-annual variability) is evaluated by comparing the modeled runoff forced with the climate variability at two consecutive timescales. It is found that the runoff variability at the daily, monthly, and annual scale is primarily controlled by the climate variability at the same timescale. The monthly climate variability significantly contributes to both the daily and inter-annual runoff variability. However, both daily and inter-annual climate variability play much smaller roles in monthly runoff variability. Besides monthly climate variability, mean annual runoff receives considerable contribution from the inter-annual climatic variability, which is often disregarded in previous studies. The quantitative evaluation of the roles of climate variability reveals how climate controls runoff across different timescales.