6533b862fe1ef96bd12c6b7e

RESEARCH PRODUCT

Characterization and Modelization of Surface Net Radiation through Neural Networks

subject

description

Artificial neural networks have shown to be a powerful tool for system modeling in a wide range of applications. In this chapter, the focus is on neural network applications to obtain qualitative/quantitative relationships between meteorological and soil parameters and net radiation, the latter being a significant term of the surface energy balance equation. By using a Multilayer Perceptron model an artificial neural network based on the above mentioned parameters, net radiation was estimated over a vineyard crop. A comparison has been made between the estimates provided by the Multilayer Perceptron and a linear regression model that only uses solar incoming shortwave radiation as input parameter. Self-Organizing Maps, another type of neural model, made it possible to get knowledge in an easy way on how the input variables are related to each other in the data set. The results achieved show the potential of artificial neural networks as a tool for net radiation estimation using more commonly measured meteorological parameters.