Surface flux modelling using ARIMA technique in humid subtropical monsoon area

In this paper, we attempt to fit sensible heat fluxes obtained by the profile method between 1–2 and 1–4 m height into the autoregressive integrated moving average (ARIMA) family. In both the profiles, fluxes are found to fit into the first-order autoregression [AR (1)] model, which is the continuous analog of the first-order Markov chain. The ARIMA (1, 0, 0) model was able to simulate the heat flux value for 1 h in advance for both the profiles (1–2 and 1–4 m height) generating a prediction error of 18.28% and 4.95%, respectively.