Application of HYDRUS-1D Model to Estimate Parameters of Soil Water Characteristic Curve Parameters

碩士 === 國立中央大學 === 應用地質研究所 === 98 === Infiltration test is an important technique to investigate the hydrologic properties in soils, while the unsaturated flow numerical models can be an efficient tool to understand flow mechanism when the physical properties and conditions of soils are changed. This...

Full description

Bibliographic Details
Main Authors: Yi-Hsien Li, 李奕賢
Other Authors: Chuen-Fa Ni
Format: Others
Language:zh-TW
Published: 2010
Online Access:http://ndltd.ncl.edu.tw/handle/09696158260471921317
id ndltd-TW-098NCU05503043
record_format oai_dc
spelling ndltd-TW-098NCU055030432016-04-20T04:18:02Z http://ndltd.ncl.edu.tw/handle/09696158260471921317 Application of HYDRUS-1D Model to Estimate Parameters of Soil Water Characteristic Curve Parameters HYDRUS-1D模式應用於入滲試驗推估非飽和土壤特性參數 Yi-Hsien Li 李奕賢 碩士 國立中央大學 應用地質研究所 98 Infiltration test is an important technique to investigate the hydrologic properties in soils, while the unsaturated flow numerical models can be an efficient tool to understand flow mechanism when the physical properties and conditions of soils are changed. This study employs HYDRUS-1D model and self developed Levenberg-Marquardt algorithm to inversely estimate parameters in the van Genuchten formula. The developed LM model was tested based on the synthetically generated data from HYDRUS-1D model. Additionally, the initial guesses of soil parameters in LM model were systematically analyzed to obtain general insight into the laboratory and field scale applications. The tested model was then used to estimate soil parameters of laboratory sand box under different infiltration scenarios. During the infiltration tests, the suction or pressure and water content were measured with tensiometers TDR(Time Domain Reflectometry). Such measured data can thus be used in the LM model to estimate the soil parameters. The estimated parameters were compared with those based on trial and error approach. Field scale double-ring infiltration was tested on the river bank of Zhuoshui river near His-Luo. The saturated infiltration rate(fc), initial infiltration rate (f0), and infiltration parameter by using the in Horton formula were estimated by the developed LM model. The objective of the field scale test is to obtain the saturated infiltration rate(f0) to compare with the results of laboratory experiment. Estimate Van Genuchten parameters by combining the laboratory infiltration experiment of field soil sample and saturated infiltration parameters. The results of the study reveal that the developed LM model associated with HYDRUS-1D model can accurately estimate van Genuchten soil parameters. The results of model test shows that the lower initial guess values would lead to better estimations of residual water content (θr) and α values. The saturated hydraulic conductivity (Ks) is the most sensitive parameter while the initial guess n values are relatively insignificant. The LM inversion of tests show that LM model associated with HYDRUS-1D model can estimate accurately van Genuchten for laboratory scale tests soil parameters. The field scale multiple depth double-ring infiltration tests reveals that the saturated infiltration rates(f0) of Zhuoshui river near His-Luo have one order of magnitude difference for different depths such depth difference of hydraulic conductivity (Ks) (0.0014 and 0.00015 cm/s) were also shown in laboratory scale tests. Based on the soil sample from Zhuoshui river the sand box experiment data were then used in LM model to estimate soil parameters. This simulation results showed that the value of α is from 0.01688 to 0.09289(cm-1), the value of n from 2.528 to 5.886(-), the material of the soil can be classified as sand and silt. Chuen-Fa Ni 倪春發 2010 學位論文 ; thesis 105 zh-TW
collection NDLTD
language zh-TW
format Others
sources NDLTD
description 碩士 === 國立中央大學 === 應用地質研究所 === 98 === Infiltration test is an important technique to investigate the hydrologic properties in soils, while the unsaturated flow numerical models can be an efficient tool to understand flow mechanism when the physical properties and conditions of soils are changed. This study employs HYDRUS-1D model and self developed Levenberg-Marquardt algorithm to inversely estimate parameters in the van Genuchten formula. The developed LM model was tested based on the synthetically generated data from HYDRUS-1D model. Additionally, the initial guesses of soil parameters in LM model were systematically analyzed to obtain general insight into the laboratory and field scale applications. The tested model was then used to estimate soil parameters of laboratory sand box under different infiltration scenarios. During the infiltration tests, the suction or pressure and water content were measured with tensiometers TDR(Time Domain Reflectometry). Such measured data can thus be used in the LM model to estimate the soil parameters. The estimated parameters were compared with those based on trial and error approach. Field scale double-ring infiltration was tested on the river bank of Zhuoshui river near His-Luo. The saturated infiltration rate(fc), initial infiltration rate (f0), and infiltration parameter by using the in Horton formula were estimated by the developed LM model. The objective of the field scale test is to obtain the saturated infiltration rate(f0) to compare with the results of laboratory experiment. Estimate Van Genuchten parameters by combining the laboratory infiltration experiment of field soil sample and saturated infiltration parameters. The results of the study reveal that the developed LM model associated with HYDRUS-1D model can accurately estimate van Genuchten soil parameters. The results of model test shows that the lower initial guess values would lead to better estimations of residual water content (θr) and α values. The saturated hydraulic conductivity (Ks) is the most sensitive parameter while the initial guess n values are relatively insignificant. The LM inversion of tests show that LM model associated with HYDRUS-1D model can estimate accurately van Genuchten for laboratory scale tests soil parameters. The field scale multiple depth double-ring infiltration tests reveals that the saturated infiltration rates(f0) of Zhuoshui river near His-Luo have one order of magnitude difference for different depths such depth difference of hydraulic conductivity (Ks) (0.0014 and 0.00015 cm/s) were also shown in laboratory scale tests. Based on the soil sample from Zhuoshui river the sand box experiment data were then used in LM model to estimate soil parameters. This simulation results showed that the value of α is from 0.01688 to 0.09289(cm-1), the value of n from 2.528 to 5.886(-), the material of the soil can be classified as sand and silt.
author2 Chuen-Fa Ni
author_facet Chuen-Fa Ni
Yi-Hsien Li
李奕賢
author Yi-Hsien Li
李奕賢
spellingShingle Yi-Hsien Li
李奕賢
Application of HYDRUS-1D Model to Estimate Parameters of Soil Water Characteristic Curve Parameters
author_sort Yi-Hsien Li
title Application of HYDRUS-1D Model to Estimate Parameters of Soil Water Characteristic Curve Parameters
title_short Application of HYDRUS-1D Model to Estimate Parameters of Soil Water Characteristic Curve Parameters
title_full Application of HYDRUS-1D Model to Estimate Parameters of Soil Water Characteristic Curve Parameters
title_fullStr Application of HYDRUS-1D Model to Estimate Parameters of Soil Water Characteristic Curve Parameters
title_full_unstemmed Application of HYDRUS-1D Model to Estimate Parameters of Soil Water Characteristic Curve Parameters
title_sort application of hydrus-1d model to estimate parameters of soil water characteristic curve parameters
publishDate 2010
url http://ndltd.ncl.edu.tw/handle/09696158260471921317
work_keys_str_mv AT yihsienli applicationofhydrus1dmodeltoestimateparametersofsoilwatercharacteristiccurveparameters
AT lǐyìxián applicationofhydrus1dmodeltoestimateparametersofsoilwatercharacteristiccurveparameters
AT yihsienli hydrus1dmóshìyīngyòngyúrùshènshìyàntuīgūfēibǎohétǔrǎngtèxìngcānshù
AT lǐyìxián hydrus1dmóshìyīngyòngyúrùshènshìyàntuīgūfēibǎohétǔrǎngtèxìngcānshù
_version_ 1718228362760028160