Evaluating the ability of dual domain models to capture the water content evolution observed in the macropore and matrix regions of a soil by x-ray CT imaging during an infiltration experiment
Abstract
We evaluate whether one-dimensional flow models can capture the pattern of vertical water content evolution observed with x-ray CT scanning during infiltration experiments conducted in a column of homogenized soil and a macroporous soil containing desiccation cracks. For both experiments we compare the data to models representing homogeneous, dual-porosity, and dual-permeability models simulated with the software HYRDUS-1D. The soil parameters for each model were estimated by calibration to the water content data derived from the CT scans during the infiltration experiments, where the soil matrix parameters were obtained from the experiment performed in the homogeneous soil. Three methods are used to compare how well the models were able to fit the data: visual comparison of the vertical water content profiles versus accuracy of bottom of wetting front depth (BWFD) and mean wetting front depth (MWFD) estimates. For the homogenous column, the standard Richard's equation was able to fit to both measures of the wetting front over time (R2= 0.997 for BWFD and 0.995 for MWFD). While the general pattern of the wetting front profile was well represented, the magnitudes estimated by the calibrated model were in poor agreement with the data. The model overestimated the observed water content by up to 20% (vol/vol), likely because a high porosity was required by the model to match the movement of the wetting front. For the macroporous soil it was possible to use the structural image of the crack network to delineate changes in water content occurring within the macropores versus the soil matrix. The dual domain models are likewise able to report water content associated with the macropore domain versus the matrix domain. For both dual domain models, we found a poor fit between the water content profiles estimated for the data and model in the macropore domain, though again the measures of depth to the wetting front were similar (R2=0.940 for BWFD and 0.899 for MWFD for the dual porosity model and R2=0.989 for BWFD and 0.865 for MWFD for the dual permeability model). In contrast, the overall water content profile and depth to the wetting front for the matrix domain provided a good match to the data for both models (R2= 0.957 for BWFD and 0.916 for MWFD for the dual porosity model and R2=0.990 for BWFD and 0.958 for MWFD for the dual permeability model).
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2020
- Bibcode:
- 2020AGUFMH202...10M
- Keywords:
-
- 1843 Land/atmosphere interactions;
- HYDROLOGY;
- 1847 Modeling;
- HYDROLOGY;
- 1865 Soils;
- HYDROLOGY;
- 1875 Vadose zone;
- HYDROLOGY