Calculating the turbulent fluxes in the atmospheric surface layer with neural networks - Example workflow

This dataset is a small example of the workflow for calculating the turbulent fluxes in the atmospheric surface layer with neural networks. You can find some example data of the meteorological tower at KIT-Campus Nord. Data is given in 1-hourly resolution for the period from 2016-05-19 to 2016-05-23. Also provided is a matlab script "run.m" that uses given the net (6-3-2 feedforward net) to estimate the turbulent parameters u* and T* from given inputs. The net is provided in binary (.mat) and readable (.m) format.