Résumé
In mountainous areas of Continental South East Asia, double cropping in the irrigable valleys and terraces is often promoted as a way to increase farmers' income while alleviating the pressure of agriculture on fragile slopes. However, cold temperature and low global radiation may constrain this strategy. Lethally cold events may occur, and increases in crop cycle length may jeopardize the correct timing of the main rice crop, taking place in summer. The model PYE (potential yield estimator) simulating the impact of temperature and radiation on the development and yield of annual crops was adapted to account for the range of temperature occurring in the area under study. It was calibrated against experimental data for the three crops that are most often considered as spring crops in the irrigable land of the northern mountains of Vietnam: rice, soybean and maize. Then a virtual experiment was designed in order to simulate various scenarios combining crop species and sowing date with climatic data accounting for variability of climate across years, location and elevation. Completed with a sensitivity analysis, it allowed to define favourable 'sowing windows' for non-water limited environments, based on the three following criteria: high average yield, low incidence over years of lethally cold events and low incidence over years of delays in the maturity of the spring crop. The length of this sowing window varied greatly across the scenarios tested. The widest was obtained for the case of soybean whatever the location and elevation, which makes it the less risky of the tested options. The approach followed proved effective to identify favourable and unfavourable environments in order to help better targeting the policy in support to the introduction of a spring crop in the mountains of Vietnam.