Methodological problems of the soil-water retention curve

Abstract

Undisturbed soil samples of a "Terra Roxa Estruturada Latossólica" (Rhodic Kanhapludalf) were taken in Piracicaba,SP from the 25 cm depth, in order to study both methodological and spatial variabilities of the soil water retention curve. The sampling followed a 5 m x 5 m grid of 10 columns and 25 Unes, resulting 250 collecting points. With these samples, 250 retention curves were obtained using porous plate funnels (tensions of 5 x 10²; 1 x 10³; 6 x 10³ and 1 x 10(4) Pa) and Richards pressure chambers (tensions of 3 x 10(4); 8 x 10(4); 3 x 10(5), and 1.5 x 10(6) Pa). The methodological variability was analyzed mainly in the light of the pertinent thermodynamics, using several values of tension. Results have shown that the analytical problems are such that they can lead to erroneous interpret rations, affecting the quality of the experiment and also the comparison of results from several laboratories. For instance, 43% of the samples showed non consistant results, with values of volumetric soil-water content smaller at the tension of 3 x 10(4) Pa than at the tension of 1.5 x 10(6) Pa, which, from the practical point of view, could, lead to subestimated values of available water.