Premium
Analysis and Field Evaluation of the Ceres Models Water Balance Component
Author(s) -
Gabrielle Benoît,
Menasseri Safya,
Houot Sabine
Publication year - 1995
Publication title -
soil science society of america journal
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.836
H-Index - 168
eISSN - 1435-0661
pISSN - 0361-5995
DOI - 10.2136/sssaj1995.03615995005900050029x
Subject(s) - loam , hydraulic conductivity , soil water , water balance , soil science , soil texture , water content , environmental science , drainage , water retention , sensitivity (control systems) , water potential , hydrology (agriculture) , mathematics , geotechnical engineering , geology , engineering , ecology , electronic engineering , biology
The soil water status partly determines the N losses from soilcrop systems. With the ultimate objective of estimating N losses, the capacity‐based water balance module of the Ceres models was tested against field data collected from various pedoclimatic regimes in France. A process‐oriented analysis of initial simulation results for a loamy soil prompted introduction of Darcy's law in the drainage and capillary rise parts of the model. As a result, a more accurate prediction of the soil water storage and surface water content was achieved. This was confirmed by comparing model output against independent data from bare or maize ( Zea mays L.)‐cropped conditions and for silt loam or sandy loam soils. For a 1‐yr period, the mean square error between modeled and measured water storages was in the range 1.9 to 3 cm 2 water for the modified model, in contrast with 4 to 12 cm 2 using the original model (which performed best on well‐drained soils). A unidimensional sensitivity analysis was conducted with regard to the three new parameters introduced in the revised model: the saturated hydraulic conductivity and two texture‐dependent constants used in simple analytical representations of the moisture retention and hydraulic conductivity curves. The sensitivity analysis proved that this more physical approach in capacity‐based models required less rigorous parameterization than mechanistic models. Moreover, the accuracy of the simulations performed with the modified model fell within the experimental error in the measurements.