Resumo

MEDINA GONZALEZ, Hanoi; HERNANDEZ PEREIRA, Yunay; SANTIAGO PILOTO, Alberto B.  e  LAU QUAN, Andrés. Modeling a soil moisture profile using a Monte-Carlo Kalman filter. Rev Cie Téc Agr [online]. 2015, vol.24, n.2, pp.31-37. ISSN 2071-0054.

The objective of this work is to show the performance of an Ensemble Kalman Filter, the most disseminated among the Monte-Carlo based filters, for improving the estimation of the soil moisture profiles, by assimilating of near surface measurements of this state variable in a model based on the Richards equation. As part of this general purpose we also show the impact of the size of the ensemble of the state variables on the assimilation efficiency. In this study we estimate the “actual” dynamics of a moisture profile, generated from a model realization using synthetic data, considering perturbed values of the inputs and parameters adopted in the “actual simulation”. The assimilated “measurements” are obtained by disturbing the actual daily value of pressure heads in the first node of the discretized profile. The application of the filtering tool with an ensemble size of 50 members allows retrieving the actual profile after only ten assimilations, despite of the high incorporated uncertainty. The study shows that the method is able of accurately predicting the actual moisture dynamics even with only ten members of the ensemble. The advantages of this method over other analogue ones, as well as the great capabilities for its introduction in the water management of Cuban soils are also depicted.

Palavras-chave : soil moisture; ensemble Kalman filter; Monte-Carlo method; Richard equation; uncertainty.

        · resumo em Espanhol     · texto em Espanhol     · Espanhol ( pdf )