The Characteristics Of Soil Water In Deep Loess Profile And The Process Of Groundwater Recharge On The Loess Tableland

Soil water is one of the most important water resources, deep soil water is crucial formaintaining the function as “soil reservoir” in the Loess Plateau. Groundwater is usuallythe important water resource for human livings; however, the recharge mechanism ofgroundwater has not been well understood in the Loess Tableland region. Soil water storedin thick loess soil is the link between precipitation and groundwater; therefore,understanding the dynamics of soil water in deep loess profile is of great importance tounderstand the groundwater recharge for the Loess Tableland. The main objectives of thisdissertation were:（1） to investigate the characteristics of soil water in deep loess profiles inthe Loess Tableland region, including characteristics of vertical distribution, soil waterresources, soil desiccation, soil water dynamics, and soil water balance;（2） to explore themechanism of groundwater recharge. In this dissertation, soil water content in deep loessprofiles under different land use patterns were measured by both field investigation andlong term observation in situ; the isotopic composition of hydrogen and oxygen inprecipitation, soil water and groundwater were analyzed. The main results are as follows:1Vertical distribution characteristic of soil water in deep loess profileThe soil texture in0²0m profile of Loess Tableland is mainly middle loam, exceptsome clayey profiles. The field capacity and wilting point is （21.39±0.13）%and（8.06±0.45）%, respectively. The content of physical clay in paleosol layers is2%⁶%higher than that in loess layers, and the porosity of paleosol layers are smaller than that ofloess layers, therefore, paleosol layers has a stronger water-holding capacity than loesslayers.The vertical distribution characteristic of soil water in deep loess profile is related tothe loess-paleosol sequences. Generally, one paleosol layer and one loess layer constitute an up-down humidity level and there is an increasing trend in soil water content withincreased profile depth.2Soil water resources and soil desiccationLong term experiments in situ showed that land use patterns can significantly affectsoil water content in deep loess profiles, and the influence depth is deeper than the depth of10m. The amount of soil water resources and the soil water availability in deep loessprofile of different land use patterns were different. Generally, the sequence is fallow land> low-yield cropland> natural grassland> high-yield cropland>4-yr planted alfalfagrassland>18-yr apple orchard>8-yr planted alfalfa grassland>23-yr planted alfalfagrassland.The dried soil layer was divided into two types, one is temporary dried soil layer,which is typically formed in soils for annual crop plants, the other is persistent dried soillayer, which is typically formed in soils for perennial artificial vegetations. The temporarydried soil layer usually disappears in the wet years, while the persistent dried soil layer ismore stable and persistent. It will take several decades to recover the soil water content ofthe persistent dried soil layer after the perennial plants are removed.3Characteristics of soil water dynamics in deep loess profileThe rainfall infiltration depth is linearly related to the precipitation during the rainyseason, the relationship can be described with the equation D=a· P–b，the coefficients（a, b） varied in different land use patterns. Generally, the sequence of rainfall infiltrationdepth is fallow land> low-yield cropland> apple orchard> high-yield cropland> alfalfagrassland.The temporal variation of soil water content was obvious in the upper part of the soilprofile; however, it decreased gradually with the depth increasing. Therefore, the shallowsoil layer can be regarded as changeable layer in soil water content. The depths of such soillayer differed with land use types, the layers in fallow land, high-yield cropland, low-yieldcropland, alfalfa grassland and apple orchard were0⁴,0⁴,0⁵,0², and0⁴m,respectively. The soil layers below these depths were regarded as relatively stable layer andhad little changes in soil water content.The seasonal variation of soil water storage can be divided in to three main periods,i.e., decreasing period （March to July）, increasing period （July to October）, and relatively stable period （October to March of next year）.The soil water storage in the0¹5m soil layer was significantly increased in fallowland and low-yield cropland but significantly decreased in the alfalfa grassland comparedwith that in the initial experiment. No significant change in the soil water storage wasobserved in the high-yield cropland. This result suggested that deep soil water and groundwater could be replenished in fallow land and low-yield cropland.4The selection of soil depths for soil water balance calculationsThe selection of soil thickness is very important for the calculation of soil waterbalance on the Loess Tableland, which is determined by both the land uses and the growthstages of the vegetations. The calculated soil thickness should be at least the depth ofrainfall infiltration in fallow land. The calculated soil thicknesses should be at least4m incropland, and15m in alfalfa grassland and apple orchard during the vigorous growth stageof planted woodland and grassland. Because of the difficulty in collecting samples in deepsoil layers, the calculated soil thicknesses should be at least10m under planted woodlandand grassland. The calculated evapotranspiration of0¹0m soil layer account for94.9%（2010） and99.0%（2011） of0¹5m soil layer in apple orchard （planted in1993）,respectively. The calculated evapotranspiration of0¹0m soil layer account for93.2%（2010） and95.2%（2011） of0¹5m soil layer in alfalfa grassland （planted in1993）,respectively. The depth of rainfall infiltration in the study region can be estimated by thelinearly relationship between the depth and rainfall during rainy season.5The isotope compositions of precipitation, soil water and groundwaterThe δD values of precipitation range from142.0‰to2.0‰, with the arithmeticmean of55.4‰. The δ¹⁸O values of precipitation range from19.6‰to1.2‰, withthe arithmetic mean of8.1‰. The equation of the Local Meteoric Water Line （LMWL）for the tableland was δD=7.39δ¹⁸O+4.34（R2=0.94，n=71）. The seasonal variation of theisotope compositions of precipitation was obvious, with high values in winter and springand low values in summer and autumn. The amount effect of isotopes in precipitation wasobvious, and the lighter isotopes values were observed in heavy rainfall events or afterlong duration rainfall.The δD values of soil water range from126.5‰to46.7‰, with the arithmeticmean of75.1‰. The δ¹⁸O values of soil water range from16.6‰to4.3‰, with the arithmetic mean of9.6‰. The values of stable isotopes in soil water fall in the right-sideof the LMWL, implying that soil water originates from precipitation, and evaporationoccurs during the rainfall infiltration. The isotope composition in soil water changesgreatly with time in shallow soil layers, with the increase of soil depth, the temporalchange in the isotope contents was decreased.The δD values of well water range from72.3‰to69.1‰, with the arithmeticmean of71.3‰. The δ¹⁸O values of well water range from10.5‰to10.1‰, with thearithmetic mean of10.3‰. The δD values of spring water range from72.4‰to68.4‰,with the arithmetic mean of70.7‰. The δ¹⁸O values of spring water range from10.5‰to10.0‰, with the arithmetic mean of10.2‰. The seasonal variation of the isotopecompositions in groundwater was not obvious, and the variation of isotopic compositionsin groundwater is much smaller than those of precipitation and soil water.6Recharge of groundwaterThrough comparing the variations of isotopic composition in precipitation, soil waterand groundwater, it can be found that the piston-type flow occurred in soil layer in theprocess of rainfall infiltration. Meanwhile, part of the rainwater can move to deeper soillayers quickly by the preferential type flow. There was a seasonal groundwater recharge,and the recharge rates during July to October should be greater than other time of the yearon the Loess tableland. The groundwater was likely recharged by piston flow andpreferential flow through the unsaturated zone, and the recharge was dominated by thepreferential flow. However, the occurrence of preferential flow is not a commonphenomenon spatially, and it has a certain relationship with land use patterns. Generally,soil desiccation caused by negative water balance can reduce the probability of preferentialflow occurrence in apple orchard and alfalfa grassland, whereas preferential flow easilyoccurs in farmland or nature grassland.The result indicates that the conversion from a large area of farmland to apple orchardcan affect the natural mode of water cycle and reduce the recharge of groundwater on theLoess Tableland.