The objective of this study is to develop sets of 300-year daily time-series of climatic data for the baseline period (1961-1990) and for changed climate scenarios (2040-2069 period) at a 0.5° grid scale for current and potential agricultural areas of Canada. Most climate impact models (e.g. crop growth models) require daily climate data on a fine spatial resolution. Daily climate scenarios are available from state-of-the-art climate models (GCMs or General Circulation Models) on a much coarser resolution (typically on a 3.75°x3.75°grid); moreover, their surface climate variables are not as reliable as large-scale atmospheric circulation data. Therefore, daily climate scenarios from GCMs are not appropriate as direct inputs in impact models. Because of their simplicity, stochastic weather generators have been adopted in this study to develop daily climate scenarios on a 0.5°x0.5° grid covering the Canadian agricultural region for the baseline period and future. The reliability of stochastic weather generators as a tool to generate daily scenarios with identical statistical properties to the observations for the baseline climate and a future changing/changed climate is evaluated. Two weather generators respectively developed at Long Ashton Research Station and Agriculture and Agri-Food Canada, LARS-WG and AAFC-WG, are compared for their capabilities of reproducing statistical properties of observed daily data for diverse Canadian climates. Different schemes for modifying weather generator parameters are tested and assessed in relation to their capability to reproduce statistical properties of daily climate data for a changing/changed climate by evaluating them on long-term historical data. For the test, the period 1911-1940 is selected as the baseline climate and 1971-2000 is selected as the changed climate period. The results confirm the validity of AAFC-WG as a tool to develop daily climate scenarios. As a reference to a changing/changed climate, baseline climate is required in impact studies. Daily climate data for maximum and minimum temperatures, precipitation and solar radiation from climate stations across Canada for 1961-1990 are used to compute monthly statistics. These statistics were then interpolated on 0.5°x0.5° grids. A total of 2046 grids are involved for the agricultural area across Canada. The AAFC-WG requires standardized empirical distributions and correlation matrices for multivariate autoregressive models to generate synthetic daily climate data. A total of 31 stations are selected to cover the agricultural areas across the country to estimate these inputs. Using AAFC-WG calibrated at the nearest station with statistics of daily climate data interpolated to the grid, daily climate scenarios are generated for baseline climate of 1961-1990 for a 300 years period. Using techniques similar to those used in generating daily climate scenarios for the baseline climate, future daily climate scenarios are generated at each 0.5°x0.5°grid for 2040-2069 by perturbing the inputs to AAFC-WG with climate change projections generated by HadCM3 A2a and CGCM1 GHG+A simulations. The climate scenarios are generated for 300 years to facilitate risk analysis, but they only represent possible climate for 2040-2069 based on climate change projections generated by the two GCM simulations.