Development and functional evaluation of pedotransfer functions for estimating soil hydraulic properties in Portuguese soils: Implications for soil water dynamics

• Autor(a) principal: Ramos, Tiago B.
• Data de Publicação: 2023
• Outros Autores: Darouich, Hanaa, Gonçalves, Maria C.
• Tipo de documento: Artigo
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10400.5/29667
• topic: hierarchical pedotransfer functions; HYDRUS-1D; PROPSOLO; soil water balance; multiple soil classes

Accurate knowledge of soil hydraulic properties is essential for effectively simulating soil water dynamics in field plots. However, determining these properties directly can be challenging, especially when a large amount of data is required. Therefore, the primary objective of this study was to develop four hierarchical pedotransfer functions (PTFs) to estimate the soil hydraulic parameters of the Mualem-van Genuchten model. The PTFs that were developed ranged from utilizing a limited set of variables to those comprising a more comprehensive set of variables. These included particle size distribution, soil bulk density, soil depth, and soil water contents at matric potentials of − 33 and − 1580 kPa. The functional performance of these PTFs was also assessed in simulations of the soil water balance in ten agricultural fields over two growing seasons using the HYDRUS-1D model. The simpler PTFs exhibited moderate accuracy in estimating soil water contents at different matric potentials, while the more complex PTFs with a greater number of variables yielded increasing accuracy. However, estimation errors were more noticeable in the drier range of the soil water retention curve. The inclusion of bulk density as a predictor only affected the wetter range of the same curve. Estimation of log Ks improved from the simpler to the more complex PTFs, but the statistical indicators were consistently modest. The functional performance of the PTFs showed increased errors in simulating soil water contents in the ten study sites compared to reference simulations. The more complex PTFs with the greatest number of variables provided the best predictions in most cases. The PTFs primarily affected flow conditions in the soil profile, resulting in increased soil water storage and impacting predictions of deep percolation the most. Water availability in the soil profile then influenced actual transpiration, alleviating stress in some cases but causing waterlogging in one situation. The study emphasizes the cautious use of the PTFs and the need for additional information on clay-rich soils and hydraulic conductivity data to enhance their performance.