Simulation of the hydrological cycle, including streamflow, in cold regions can be challenging because the melt of the seasonal snowpack forms important contributions to runoff and infiltration. Snowpacks are redistributed by wind and gravity and subject to variable melt energetics and sublimation, resulting in snowmelt with extraordinarily high spatial and temporal variability. Therefore, spatially distributed models with a strong physics base provide advantages for hydrological prediction. The Canadian Hydrological Model (CHM) uses a variable resolution, unstructured, triangular mesh to downscale meteorological forcings to model snow redistribution and ablation processes at snowdrift permitting scales. Warm season process representations, such as infiltration into frozen and unfrozen soils, soil moisture and depressional storage, overland and subsurface runoff, freezing and thawing of soils have been added to CHM to enable the calculation of the full hydrological cycle. Here, the new process algorithms will be presented and evaluated against detailed field observations, as well as the results of simulations showing the interplay of the dynamic spatial patterns of cold and warm season hydrological fluxes and the influence on the large scale water balance.