Random temperature and humidity models in Atlantic environment

Moisture and temperature variations have a high influence on building materials behavior, like creep and durability in timber or activity of corrosion in reinforced concrete. Improving the prediction of these behaviors requires a better modeling of these variations. Variations of temperature and humidity recorded during several years at an hourly time step have been analyzed so as to make the part between a deterministic signal (explained by astronomical reasons) and a stochastic part. The variographical analysis for temperature has revealed a correlation structure of about seven days. The humidity is considered as a random variable whose variations are constrained by temperature values and physical laws (air saturation). The combined simulation enables to build synthetic climates, which present the same patterns, in terms of statistical distribution and lane dynamics than real data records. These synthetic signals will be used for studying the building materials response under long-term environmental loading. An application to timber durability will be presented in a companion paper.