Diurnal pattern of the drying front in desert and its application for determining the ... · 2016-01-10 · Badain Jaran Desert, a ﬁeld campaign was conducted by the observations

Hydrol Earth Syst Sci 13 703ndash714 2009wwwhydrol-earth-syst-scinet137032009copy Author(s) 2009 This work is distributed underthe Creative Commons Attribution 30 License

Hydrology andEarth System

Sciences

Diurnal pattern of the drying front in desert and its application fordetermining the effective infiltration

Y Zeng12 Z Su2 L Wan1 Z Yang1 T Zhang3 H Tian3 X Shi3 X Wang1 and W Cao1

1School of Water Resources and Environment China University of Geosciences Beijing China2International Institute for Geo-information Science and Earth Observation Enschede Netherlands3Cold and Arid Regions Environmental and Engineering Research Institute Chinese Academy of SciencesLanzhou China now at International Institute for Geo-information Science and Earth Observation Enschede Netherlands

Received 14 January 2009 ndash Published in Hydrol Earth Syst Sci Discuss 24 February 2009Revised 11 May 2009 ndash Accepted 21 May 2009 ndash Published 3 June 2009

Abstract Located in western Inner Mongolia the BadainJaran Desert is the second largest desert in China and con-sists of a regular series of stable megadunes among whichover 70 permanent lakes exist The unexpected lakes indesert attracted research interests on exploring the hydrolog-ical process under this particular landscape however a veryfew literatures exist on the diurnal and spatial variation ofthe drying front in this area which is the main issue in thedesert hydrological process to characterize the movement ofwater in soil In order to understand the drying front in theBadain Jaran Desert a field campaign was conducted by theobservations of soil physical parameters and micrometeoro-logical parameters With the field data the performance ofa vadose zone soil water balance model the HYDRUS wasverified and calibrated Then the HYDRUS was used to pro-duce the spatial and temporal information of coupled waterwater vapour and heat transport in sand to characterize thevariation pattern of the drying front before during and afterthe rainfall Finally the deepest drying front was applied todetermine the effective infiltration which is defined as theamount of soil water captured by the sand beneath the deep-est drying front by infiltrating water of an incident rainfallevent

1 Introduction

Infiltration has long been regarded as one of the most difficultproblems for hydrological forecasting due to its relevanceto runoff soil moisture storage evapotranspiration ground-

Correspondence toY Zeng(yijianitcnl)

water recharge and unsaturated flow (Ghildyal and Tripathi1987 Mitkov et al 1998 Tami et al 2004) Its significanceon assessing the hydrological process in the desert environ-ment which is the main plant community shaping process inarid ecosystems is even magnified In desert the precipita-tion is often the solo source of water replenishment whilethe infiltration is the main process to understand how muchrainfall could be retained in sand for some time before be-ing either passed downward as percolation or returned to theatmosphere by the process of evapotranspiration

In water scarce regions demonstrating whether modernrainfall is recharging aquifers through infiltration and onwhat rate or scale is one of the main issues facing scientificsociety The Badain Jaran Desert with unexpected numer-ous groundwater-fed perennial lakes has attracted substantialresearch interest on the desert hydrological process and theinfiltration in this special landscape lakes in inter-dune areas(Gu et al 2006 Walker et al 1987) has been widely stud-ied Hofmann (1996 1999) stated there was an aquifer in thenorthern areas of the eastern lakes region recharged directlyby precipitation through infiltration process while he did notstudy the infiltration rate Based on hydrochemical and iso-topic analysis Gu (2006) points out that groundwater flowsfrom adjacent areas in the south to the northwestern part witha historic and fossil recharge (1000 to 30 000 BP) and withthe actual analytical results on groundwater chlorides he es-timates that 1 to 15 mmyr of groundwater recharge wouldoccur by direct infiltration from precipitation With the tracermethods more and more recharge rate data are reported andadded to a growing catalogue of tracer results for the BadainJaran Desert region which leads to the estimated averagerecharge rate in this region varies from 095 to 36 mmyr(Geyh et al 1996 Hofmann 1996 Ma et al 2003 Ma

Published by Copernicus Publications on behalf of the European Geosciences Union

704 Y Zeng et al Diurnal pattern of the drying front in desert

and Edmunds 2006 Yang 2006 Yang and Williams 2003bZhang and Ming 2006 Chen et al 2004 Gates et al 20062008b Jakel 1996)

Afore mentioned studies are mainly focused on the rateat which the infiltrated rainfall recharges the groundwaterwhich could be called the effective infiltration or the effectiverainfall (Wu et al 1997 Bonacci et al 2006) However theeffective infiltration or rainfall could also be defined as theamount of rainfall stored in the shrub root zone excludingthe fraction that runs off the soil surface or passes throughthe root zone that does not contribute to shrub growth andthe fraction that evaporates (Wang et al 2007) From abovetwo definitions obviously the effective infiltration or rain-fall could be generally defined as the portion of rainfall thatpenetrates and remains in a domain of interest With thisgeneral definition in this study the effective infiltration isthe amount of soil water remained in sand beneath the dry-ing front (marking the interface between the upwards anddownwards soil water fluxes) by infiltrating water of a rain-fall event and thus not evaporated and is evaluated based onthe vadose zone soil water balance method

As what mentioned above the estimates of infiltration ratein desert have been conducted a lot with the tracer methodwhich is mainly used to evaluate variations of flow and trans-port in thick desert vadose zones in response to paleoclimaticforcing and thus with the time scale of thousand-year (Scan-lon et al 2003) While with the soil water balance modelthe infiltration and soil water redistribution process in the un-saturated zone could be tracked at any meaningful time scaleof interest In order to do that quantifying changes in in-filtration properties by a reliable observation and measure-ment is needed Previous studies has indicated the feasibilityof using soil water content probes installed into the soil totrack the propagation of the wetting front during infiltration(Topp et al 1982 Noborio et al 1996 Timlin and Pachep-sky 2002 Melone et al 2006 Tami et al 2004 Yang 20042006 Wang et al 2007 2008 Geiger and Durnford 2000)and therefore to track the drying front during evaporation

In this study the measurements of soil water content soiltemperature and precipitation were used to verify and cali-brate the performance of a vadose zone soil water balancemodel the HYDRUS which was subsequently used to pro-duce the temporal and spatial information of coupled waterwater vapor and heat transport This enabled us to character-ize the variation pattern of the drying front before during andafter rainfall so that we could assess how much precipitationis evaporated and how much is conserved in sand Our goalwas to understand the capacity of the sand dune to capturecertain amount of rainfall during a single precipitation eventwhich is available to plants The first section of the paperthus describes the experiments and methods used Then fol-lows simulation results based on experimental observationsFinally we conclude the paper with a discussion of practicalresults

2 Materials and methods

21 Study site description

The study site is a relatively flat area at the foot of duneswith an area of about 100 square meters located in thesoutheast Badain Jaran Desert which lies in the northwestof the Alashan plateau in western Inner Mongolia of Chinabetween 39 20primeN to 41 30primeN and 100 E to 104 E TheBadain Jaran Desert is regarded as the second largest desertin China and covers an area of about 49 000 km2 (Yang2001) The desert landscape primarily consists of unvege-tated or sparsely vegetated aeolian sand dune fields as a coresurrounded by desert plains with pediments on the margin(Yang et al 2003a) The elevation ranges from 1500 m inthe southeast to 900 m above sea level in the northwest withthe mega dunes up to 400 m high in the southeast alignedin a SW-NE pattern (Dong et al 2004) The mean grainsize of the dune sand in this desert is between 021 mm and022 mm (Jakel 1996 Yang and Williams 2003b) Thereare over 70 lakes in the interdune areas of which the sur-face area vary from 02 km2 on the southeastern edge of thedesert to 16 km2 slightly northwards and the water depthfrom 2 m to 16 m (Yang and Williams 2003b) On the sur-face there is neither recharging into these lakes nor dis-charging from them and there is no river in the entire desert(Yang 2001 Wang and Cheng 1999) With such hydrolog-ical conditions vegetation in the study area is sparse Theplant or shrub communities depend mainly on the geomor-phology and on edaphic conditions The dominant plantspecies are Haloxylon ammodendron Psammochoa villosaPhragmites communis Artemisia ordosica while in the areaaround lakes and springs the denser groundcover observablecould be found (Polypogon monspeliensis Triglochin mar-itima Achnatherum splendens Carex sp Glaux maritima)(Gates et al 2008a)

The distribution of sparse vegetation in this area is mainlycaused by the dominated climate pattern in the AlashanPlateau which is characterized by a markedly arid and conti-nental climate During the winter months a well-developedhigh pressure system cold and dry continental air masseswith temperatures below zero influences the area whichcaused the mean monthly temperature in January becomeminus10C In the summer months the tropical air masses fromthe Pacific Ocean hits the Badain Jaran Desert with more thanhalf of the average annual precipitation fell in July The spa-tial and temporal distribution of the precipitation in the areais determined by the Asian summer monsoons at the presenttime and the mean monthly temperatures in July is 25Cwith a diurnal temperature variation up to more than 45CThe mean annual precipitation varies from 84 to 120 mm inthe southeast and from 37 to 40 mm in the northwest andthe potential evaporation is approximately 2600 mmyr be-ing the highest in China (Ma et al 2003 Ma and Edmunds2006 Gates et al 2008a)

Hydrol Earth Syst Sci 13 703ndash714 2009 wwwhydrol-earth-syst-scinet137032009

Y Zeng et al Diurnal pattern of the drying front in desert 705

22 Experimental design and data collection

In order to quantify the effective infiltration measurementsof soil water content temperature and matric potential wereconducted in the observation period from 3 June to 20 Junein 2008 The instruments used for measuring soil moistureand temperature are the profile sensors using numerous oftransducers combined in one probe with a certain distanceinterval between them It is especially necessary to use theprofile sensor in desert considering the sand profile wouldcollapse easily during digging As for the soil matric poten-tial there was not a profile sensor available and four sepa-rate transducers pF-meters (GeoPrecision GmbH EttlingenGermany) were installed at the corresponding depth of soilwater content monitoring

The soil water content profile sensor was called asEasyAG50 (Sentek Pty Ltd Stepny South Australia) theprinciples of operation and design of which were describedin detail by Fares and Polyakov (2006) and the manufacturersquoscalibration manual (Sentek Pty Ltd 2001) An individualEasyAG50 consisted of a printed circuit board five capac-itance sensors and a polyvinyl chloride (PVC) access tubewith external and internal diameters of 32 and 28 mm re-spectively Capacitance sensors were set at 10 20 30 40and 50 cm from the sand surface The reading range of thesecapacitance sensors was from oven dry to saturation Thecorrelation coefficient between the readings and actual soilwater content based on field calibration was 8939 witha resolution of 0008 and a precision ofplusmn006 Thetemperature effects on readings of MCP (Multi-sensor Ca-pacitance Probe) were conducted in the lab with the stan-dard procedures described in detail by Polyakov (2005) andFares (2007) The temperature effects for the capacitancesensors were 144 of readings from 12C to 45C at 10 cm139 from 11C to 50C at 20 cm 14 from 9C to 51Cat 30 cm 13 from 9C to 55C at 40 cm and 15 of read-ings from 8C to 55C at 50 cm respectively After the cal-ibration the temperature effects at the depth of 10 20 3040 and 50 cm were reduced by 92 93 938 88 and82 respectively

The soil temperature profiles were measured by STP01(Hukseflux Thermal Sensors BV Delft the Netherlands)which was designed to measure the soil temperature at spe-cific depths by determining the thermal gradients between acertain specific depth and the reference point It improved theaccuracy in positioning the uncertainty of which was usu-ally large when using a series of separate sensors This madethe temperature gradient measurement more correctly whichsubsequently improved the accuracy of the absolute tempera-ture measurement The principles of measurement by STP01had been described in detail by the manufacturersquos manual(Hukseflux Themal Sensors 2007) A single probe generallycontained five thermocouples a copper lead a Pt 100 a TCjunctions and a sensor foil with the thickness of 06 mm Thethermocouples were set at 2 5 10 20 and 50 cm from the

sand surface The measurement range of STP01 was fromminus30C to 70C with an accuracy ofplusmn002C

During installation a single MCP probe (soil water con-tent profile sensor) was inserted into a 60-cm-long accesswhich would be put into a pit with sufficient width and depthThe probe was inserted into the sand through the side of thepit that was unaltered and was placed vertically to the sandsurface At the same time about 30 cm away from MCPprobe on the same side of the pit the STP01 (soil temper-ature profile sensor) was installed vertically with the aid ofa steel ruler considering the sensor foil was too soft to beinserted in sand directly After the set up of MCP and STP01probe the pit was carefully refilled avoiding perturbations asfar as possible During the refilling the separate soil matricpotential sensors pF-meters were also inserted vertically onthe same side of the pit where the two probes were installedat the depth of 50 40 30 and 20 cm from the sand surfaceThe pF-meter was based on the registration of the molar heatcapacity of the ceramic cup which was in contact with thesoil for balance to record the matric potential over a largemeasurement range from pF 0 to pF 70 with an accuracy ofplusmn pF 005 an precision of less thanplusmn pF 0015 and a res-olution power of pF 001 The site was set up at the end ofMay 2008 and measurements were not begun until one weeklater to allow the repacking sand to settle down

The precipitation was measured using a tipping bucket raingauge model TR-5251 (RST Instruments Ltd CoquitlamCanada) with a precision ofplusmn001 mm and a resolution of001 mm At the same time other micrometeorological pa-rameters such as the air temperature relative humidity netradiation and wind velocity at a height of 2 m above the sur-face layer were also measured Two data collection and pro-cessing systems CR1000 and CR5000 (Campbell ScientificInc Logan Utah USA) were used to record the field dataat half-hourly intervals The first one was used to record thesoil physical parameters while the second one for the mi-crometeorological parameters

23 Soil water balance model

231 Model description

The HYDRUS1D code which refers to the coupled watervapour and heat transport in soil was applied in order to de-termine the temporal and spatial soil water fluxes The gov-erning equation for one-dimensional vertical flow of liquidwater and water vapour in variably saturated media is givenby the following mass conservation equation (Saito et al2006)

partθ

partt=minus

partqL

partzminus

partqv

partz(1)

WhereqLandqv are the flux densities of liquid water andwater vapour (cm dminus1) respectivelyt is time (d) z is thevertical axis positive upward (cm)

wwwhydrol-earth-syst-scinet137032009 Hydrol Earth Syst Sci 13 703ndash714 2009


The flux density of liquid waterqL is defined as (Philipand De Vries 1957)

qL = qLh + qLT = minusKLh(parth

partz+ 1) minus KLT

partT

partz(2)

Where qLh and qLT are respectively the isothermal andthermal liquid water flux densities (cm dminus1) h is the ma-tric potential head (cm)T is the temperature (K) andKLh

(cm dminus1) and KLT (cm2 Kminus1 dminus1) are the isothermal andthermal hydraulic conductivities for liquid-phase fluxes dueto gradients inh andT respectively

Using the product rule for differentiation and assuming therelative humidity in soil pores being constant with tempera-ture (Philip and De Vries 1957) the flux density of watervapourqv can be written as

qv=qvh+qvT =minusKvh

parth

partzminusKvT

partT

partz(3)

Whereqvh and qvT are the isothermal and thermal watervapour flux densities (cm dminus1) respectivelyKvh (cm dminus1)and KvT (cm2 Kminus1 dminus1) are the isothermal and thermalvapour hydraulic conductivities for water vapour respec-tively Combining Eqs (1) (2) and (3) we obtain the gov-erning liquid water and water vapour flow equation

[partθ

partt=

part

partz

[KLh

parth

partz+KLh+KLT

partT

partz+Kvh

parth

partz+KvT

partT

partz

]]

=part

partz

[KT h

parth

partz+KLh+KT T

partT

partz

](4)

Where KT h (cm dminus1) and KT T (cm2 Kminus1 dminus1) are theisothermal and thermal total hydraulic conductivities respec-tively and

KT h = KLh + Kvh (5)

KT T = KLT + KvT (6)

For the sake of brevity detailed description of the HY-DRUS1D code is not given here But any interested readersare referred to Saito et al (2006)

232 Material properties

For the hourly time scale the most active layer for the cou-pled liquid water water vapor and heat in sand was nearthe surface and usually was limited in the depth of less than50 cm (Zeng et al 2008) which was also the frequent in-fluence depth of a typical rainfall in the desert environment(Wang et al 2007 Wang et al 2008) In the field site wesampled sand at the depth of 10 20 30 40 and 50 cm re-spectively and all samples were used for basic soil tests andsoil water characteristic curves determinations (see later fordetails) It was found that the test results for these five sam-ples were close to each other and the deviation from the av-eraged value was less thanplusmn1 Then the final results of

Fig1 Particle size distribution curve

(a)

(b) Fig2 The field measurement of the soil matric potentials (a) and the fitted results (b)

Fig 1 Particle size distribution curve

the material properties were derived by averaging the resultsof these samples

The field site consisted of fine type sand the grain-sizedistribution of which was determined and classified follow-ing the Specification of Soil Test (SST) using sieve anal-ysis method (NJHRI 1999) The fine sand has a coeffi-cient of uniformity of 2072 and a coefficient of curvatureof 0998 The particle-size distribution curve of the fine sandwas shown in Fig 1

The soil water retention curve was one of the most fun-damental hydraulic characteristics to solve the flow equa-tion of water in soils describing the relationship betweenvolumetric water content and matric potential (ie the suc-tion head) in the soils According to the field measurementthe volumetric water content and matric potential data pairs(θminush pairs) at all measured depths were pooled in Fig 2aFor every single depth there was a different pattern espe-cially at the depth of 50 cm The data points highlighted bythe oval-shaped curve were recorded after the occurrence ofrainfall These points indicated that there were phenomenaof hysteresis which meant that the soil undergoing dryingprocesses such as evaporation or gravity drainage generallytends to retain a greater amount of water than for the samemagnitude of suction during wetting processes such as infil-tration or capillary rise (Lu and Likos 2004) The data pointsin the oval-shaped curve represented the wetting process ofthe sand after the rainfall which did not follow the dryingprocess (data points outside the oval-shaped curve) beforethe rainfall At the depth of 50 cm the hysteresis phenom-ena were more conspicuous than those at other depths Itwas because hysteresis was less pronounced near the resid-ual water content where pore water retention felt within thependular regime (Lu and Likos 2004) Due to the compli-cated hysteresis pattern which was beyond the scope of thisstudy it was difficult to use the inverse method to get thevan Genuchtenrsquos parameters using the in situθminush data pairs



(Zeng et al 2008) Instead the pressure extraction chambermodel 1600 (Soil Moisture Equipment Corp Santa BarbaraCA) was used to determine the water-holding characteris-tics with the soil water retention equation which was givenby (van Genuchten 1980)

θ(h)=

θr+

θsminusθr

[1+|αh|n]m

θs

h le 0h gt 0

(7)

whereθ is the volumetric water content (cm3 cmminus3) at pres-sure headh (cm) θ r and θ s are the residual and saturatedwater contents respectively (cm3 cmminus3) α(gt0 in cmminus1) isrelated to the inverse of the air-entry pressuren (gt1) is ameasure of the pore-size distribution affecting the slope ofthe retention function (m=1minus1n)

The comparison between the in situθminush data pairs andthe van Genuchten model curve (determined by the pres-sure extraction chamber) was shown in Fig 2b The goodagreement was found at low water content level while themodel fitted only part of the data at the depth of 50 cmwhich was due to the pronounced hysteresis in moist sand(Lu and Likos 2004) The part of unfitted data representedthe drying process of sand at the depth of 50 cm Consider-ing the scope of this study was on the determination of ef-fective rainfall the data of interest atminus50 cm depth wouldbe located at the wetting process which were fitted wellby the van Genuchten model withθ r=0017 cm3 cmminus3 θ s

=0382 cm3 cmminus3 α=000236 cmminus1 andn=36098 (Fig 2b)The hysteretic behaviour in the soil water retention curveshould be understood due to its consequent impact on thestress strength flow and deformation behaviour of unsatu-rated soil systems However this issue was beyond the scopeof this study and will not be discussed here and any inter-ested readers are referred to Lu and Likos (2004)

233 Initial and boundary conditions

Considering the groundwater table was at a depth of morethan 2 m (Gates et al 2008b) the soil profile was consid-ered to be 100 cm deep and the bottom condition for thisprofile was regarded as free drainage and the nodes locatedat depths of 2 cm 5 cm 10 cm 20 cm 30 cm 40 cm and50 cm were selected for comparing calculated temperaturesand volumetric water contents with measured values Thespatial discretization of 1 cm was used leading to 101 nodesacross the profile The calculations were performed for a pe-riod of 176 days from 3 to 20 June in 2008 Discretization intime varies between a minimum and a maximum time-stepcontrolled by some time-step criterion (Saito et al 2006)Except for the aforementioned geometry and time informa-tion it is necessary to specify initial conditions for tempera-ture and soil water content in order to solve this problem byHYDRUS1D

The initial soil water content and temperature distributionsacross the profile were determined from measured values on


(a)


Fig 2 The field measurement of the soil matric potentials(a) andthe fitted results(b)

3 June by interpolating the measured values between two ad-jacent depths Boundary conditions at the soil surface forliquid water water vapor and heat were determined from thesurface energy balance equation (eg van Bavel and Hillel1976)

RnminusHminusLEminusG=0 (8)

whereRn is the net radiationH is the sensible heat flux den-sity LE is the latent heat flux densityL is the latent heatEis the evaporation rate andG is the surface heat flux densityNet radiation (eg Brutsaert 1982) and sensible heat flux(eg van Bavel and Hillel 1976) were defined as

Rn=εsεaσT 4a minusεsσT 4

s H=Ca

TsminusTa

rH(9)

whereεa is the emissivity of the atmosphereεs is the emis-sivity of the bare soilσ is the Stefan-Bolzmann constantTa

is the air temperatureTs is the soil surface temperatureCa

is the volumetric heat capacity of air andrH is the aerody-namic resistance to heat transfer The surface evaporation iscalculated as

E=ρvsminusρva

rv+rs(10)



whereρvs ρva are the water vapor density at the soil surfaceand the atmospheric vapor density respectivelyrv rs arethe aerodynamic resistance to water vapor flow and the soilsurface resistance to water vapor flow that acts as an addi-tional resistance along with aerodynamic resistance respec-tively (Camillo and Gurney 1986) As for the heat transportdomain the upper boundary condition was determined bythe measured surface temperature while the lower boundarycondition was considered as zero temperature gradient (Saitoet al 2006)

3 Results and discussion

The physics of infiltration has been developed and improvedduring the last almost seven decades The numerical mod-els play a critical role in evaluating this physical processthat governing soil heating spatial distribution of water andgaseous exchange between the soil and the atmosphere (Mel-one et al 2006 Geiger and Durnford 2000 Mitkov et al1998 Shigeru et al 1992) The most widely used modelis the Richardsrsquo equation which has been modified into theHYDRUS1D considering the coupled liquid water vapor andheat transport With the soil hydraulic properties and theinitial and boundary conditions introduced above the HY-DRUS1D was calibrated with the field data which was sub-sequently used to calculate and analyze the detailed informa-tion of the soil water flow processes under an incident rainfallevent

31 Model verification

In this section the measured water contents and soil tempera-tures were compared to those calculated by the HYDRUS1Dcode which refers to the coupled liquid water water vaporand heat transport in soil The predicted and observed soiltemperatures at depths of 2 5 10 20 and 50 cm were shownin Fig 3 The simulationrsquos goodness of fit was quantifiedwith the following relative root mean square error measure

RRMSE=

radicNWsumi=1

(MiminusCi)2Nw

Max(M1 M2 MNw )minusMin (M1 M2 MNw )(11)

where Nw is the number of the measurementsMi and Ci are measurements and calculationsrespectively Max(M1 M2 MNw ) andMin (M1 M2 MNw ) are the maximum andminimum value of the measurements The RRMSE isdimensionless and RRMSE=0 indicates the best fit Thesmaller is the RRMSE the better the fit of simulation

The RRMSEs of the temperatures at depths of 2 5 1020 and 50 cm were respectively 0063 0061 0096 0183and 017 Although the RRMSEs of the fitness at the depthof 20 and 50 cm were larger than 01 and close to 02 which

Fig3 The calculated and measured soil temperature at all depths

Fig4 The calculated and measured soil moisture content at the depth of 10cm and 20cm

Fig 3 The calculated and measured soil temperature at all depths

means the largest deviation wereminus6C and 23C at respec-tive depth the calculated and measured temperature gener-ally agreed at all five depths and the typical sinusoidal diurnalvariation of soil temperature were captured

During the field monitoring an incident precipitation wascaptured The evaporation after rainfall was an exponentialfunction of elapsed time (Li et al 2006) which was usedin the calculation of soil moisture variation in HYDRUS1DThe simulated and measured soil water content at the depthsof 10 and 20 cm was depicted in the Fig 4 while the sim-ulated results for the soil moisture at the depths of 30 40and 50 cm were not shown here which kept almost the samevalue during the period of field measurement The RRMSEsof the soil water content at the depths of 10 cm and 20 cmwere 0048 and 019 respectively Both temperature and soilwater content were simulated well by using HYDRUS1D

32 Determination of the drying front

The drying front is defined as the interface between the up-wards and downwards soil water fluxes In order to under-stand the variation of drying front in soil the detailed soilwater content and temperature variability were shown as the





Fig 4 The calculated and measured soil moisture content at thedepth of 10 cm and 20 cm

projections of the relative altitude coordinates (Z-axis) toconstant depth coordinates (Y -axis) and elapsed time coordi-nates (X-axis) The interpolating and smoothing procedurefor projected soil water and temperature data were carriedout by using the Surfer plotting software (Golden Software1990) The interpolation was done using the kriging optionin Surfer The fake three dimensional fields consisting ofa space-time field (two-dimensional field) and a dependentaltitude variable (eg specific flux temperature or soil ma-tric potential) (Zeng et al 2008) were plotted to understandthe time series information of specific flux soil temperatureand soil matric potential for the whole soil profile which pre-sented a clear overview on the physical flow processes in soil

321 The driving force

The driving forces causing flow in soil are inter-dependentfor example as temperature gradients are accompanied bygradients of surface tension at the air-water interface thereis a possibility of thermal capillary flow and thermal capil-lary film flow of water in liquid phase (Ghildyal and Tripathi1987) There is a need to identify the soil matric potential andtemperature pattern for the coupled flow processes

Figure 5a shows daily time series of soil matric poten-tial profile during the calculation period (from 3 June to20 June in 2008) On the day before the rainfall the poten-tial varied radically in the surface layer with a thickness ofabout 2 cm fromminus10 5564 cm tominus104989 cm water col-umn While in the sand layer between the depth of 5 cmand 20 cm a low potential zone developed with the lowestvalue ofminus31 932 cm water column at the depth of 15 cm Itmeant that the soil water above and below this zone wouldbe transported toward this low potential layer due to the con-vergent potential gradients because the potential gradientsabove and below certain depth were directed to this depthand formed a convergent zero soil matric potential gradient

(a)

(b)

Fig5 The time series of soil matric potential profile (a) and its gradient profile (b) Fig 5 The time series of soil matric potential profile(a) and itsgradient profile(b)

plane (Fig 5b) The potential gradients were decided by1h=(hi+1minushi)(cmcm) wherehi represents the soil matricpotential at a depth ofi cm

After rainfall the pattern of the soil matric potential wastotally changed The rainfall broke the original low poten-tial zone and pushed it down to the layer between the depthof 20 cm and 30 cm with lowest value ofminus3989 cm wa-ter column The rainfall increased the soil matric poten-tial almost 9 times At the mean time the zero soil matricpotential gradient planes moved downwards and this wasshown clearly in Fig 5b According to the pattern of thesoil matric potential and its gradient the soil moisture would



transport upwards from the bottom of the soil profile dur-ing the whole period which was accordant with Walvoordrsquosstatements (Walvoord et al 2002)

Figure 6a shows the daily time series of soil temperatureprofiles where the interval of isolines was 2C The densi-ties of isolines indicated how strongly the soil temperature atcertain depth fluctuated The dense and sparse isolines repre-sented the rapid and slow variation of soil temperature withtime respectively It could be found that the isolines nearthe surface were the densest while the isolines at the bottomof soil profile were the sparsest It indicated that the ampli-tude of soil temperature variation was reduced with depthDuring the observation period the surface temperature var-ied from 916C to 63C with a range of daily maximumsurface temperature between 39C and 63C and a rangeof daily minimum surface temperature between 916C and196C While at the depth of 50cm the soil temperature var-ied from 251C to 3066C with a variation span of 436Cand 335C for the maximum and minimum soil temperature

Figure 6b shows the space-time temperature gradient fieldwhich clearly shows how heat transport in soil controlsthe dependence of the temperature gradient profiles in timeand space The temperature gradients were derived fromT =(Ti+1minusTi) (Ccm) whereTi represented the soil tem-perature at a depth ofi cm The variation of isolines inFig 6b was similar to that in Fig 6a The isolines experi-enced alternatively the sparseness and denseness with timeelapsed on the surface at the mean time developed down-wards from denser to sparser with depth According to thepattern of the soil temperature and its gradient the thermaldriven soil water transport was more complicated than thematric potential driven one The positive temperature gra-dient could distribute throughout the whole soil profile nearthe twelfth day of period this meant that the thermal fluxescould transport soil moisture to the surface from the bottomof the soil profile On the other hand the negative tempera-ture gradient could distribute throughout the whole soil pro-file at most of days of period However during the wholecalculation period most of downward thermal fluxes due tothe negative temperature gradient would not transport the soilmoisture on the surface to reach the bottom of the soil profilewhile were retarded by the upward potential driven fluxesThe soil water fluxes are discussed in more detail in the fol-lowing section

322 The drying fronts

With above driving forces the soil water flux pattern couldbe determined which was subsequently used to identify thedrying fronts Figure 7a depicted the variation of the soilwater flux profiles with time elapsed The darker color indi-cated the negative fluxes which meant that the soil moisturelocated in the darker area would be kept in soil and not evapo-rated at that time According to Fig 7a there were two typesof darker areas the first type of darker area occurred in the

(a)

(b)

Fig6 The time series of soil temperature profile (a) and its gradient (b) Fig 6 The time series of soil temperature profile(a)and its gradient(b)

shallow layer limited in the depth of 45 cm close to surfacewith small time intervals and isolated shapes (A-type) thesecond type occurred below the depth of 45 cm with radicalvariation of the upper borderline while the darker area keptcontinuity (B-type)

Almost all of the A-type and B-type darker areas werelocated in the negative temperature gradient zones except



for the darker areas triggered by the rainfall event Theisolated shapes of A-type darker areas were activated whenthe absolute value of the positive matric potential gradientsdriven fluxes were greater than the absolute value of the neg-ative temperature gradient driven fluxes The space gaps be-tween the A-type and B-type areas were likewise contributedby the dominance of the positive matric potential gradientsdriven fluxes Between the eighth and tenth day of the obser-vation period the negative temperature gradients were overthe positive matric potential gradients which lead to the con-nection of the A-type and B-type areas After the rainfallthe negative matric potential gradient above the low matricpotential zone (Fig 5b) kept dominant for almost 3 dayswith the propagation of the zero convergent matric poten-tial planes from the depth of 15 cm to 26 cm The rainfallevent also caused the positive temperature gradient throughalmost the whole depth of profile between the twelfth andfourteenth day of the observation period However the soilwater fluxes driven by the summation of the positive tem-perature gradient and matric potential gradient during thesethree days were hold back by the zero convergent matric po-tential planes

The drying front could be identified by the borderlines ofdarker areas On the first day of the calculation period thedrying front started from the surface due to a small A-typedarker area near the surface As time elapsed the dryingfront dropped to the second A-type darker area at the depthof about 12 cm Since there was a gap between the second A-type darker area and the third one which developed from thedepth of 2 cm to 24 cm The drying front fell to the B-typedarker area at the depth of 60 cm and varied in the course oftime which was corresponding to the time course occupiedby the gap between the second and third A-type darker areaAfter this course of time for the gap the drying front wouldjump to the third A-type darker area from the B-type darkerarea Following the processes described above the dryingfront fluctuated sharply from the surface to the upper border-lines of A-type and B-type darker areas It should be notedthat the upward fluxes underneath the lower borderlines of A-type darker areas and above the upper borderlines of B-typedarker areas (the grey color area between A-type and B-typedarker areas) would not be evaporated due to the existenceof convergent zero flux planes (Zeng et al 2008) at the lowerborderlines of A-type areas

The sharp fluctuation of the drying front was interruptedby the occurrence of precipitation which could be observedclearly in Fig 7a After the rainfall of 6604 mm the dry-ing front was limited at the depth of 20 cm for 3 days fromthe eleventh to fourteenth day of the period Then the dry-ing front experienced the wide fluctuation between the A-type and B-type darker areas again It meant that the upwardsoil moisture flux above the upper borderlines of the B-typedarker areas was kept at the lower borderlines of the A-typedarker areas during these three days and increased the mois-ture content near the borderlines

(a)

(b)

Fig7 The time series of soil moisture flux pattern (a) and volumetric soil moisture content pattern (b) Fig 7 The time series of soil moisture flux pattern(a) and volu-metric soil moisture content pattern(b)

The corresponding variation of soil moisture profile wasshown in Fig 7b On the day before the rainfall due to thesharp fluctuation of the drying front the moisture content atthe depth of 10 cm and 20 cm were kept between the value of0017 and 0021 cm3cm3 and 00201 and 00216 cm3cm3respectively After the rainfall the interruption of the fluctu-ation of drying front increased the moisture at the depth of



10 cm and 20 cm to 0065 and 00256 cm3cm3 respectivelyHowever the moisture content at the depth of 30 40 and50 cm were not influenced directly by the rainfall event Af-ter the stop of the rainfall the moisture content at the depthsof 10 cm and 20 cm decreased with time due to the propaga-tion of the drying fronts When the connected A-type areaswere separated due to increasing evaporation the moisturecontent at the depths of 10 cm and 20 cm would go back tothe value before the rainfall gradually

From above discussion the 6604 mm of rainfall could notbe remained in sand and would be evaporated to the atmo-sphere after some course of time But the rainfall interruptedthe fluctuation of the drying front and made the isolated A-type darker areas connected for 3 days which subsequentlylimited the upwards soil water flux from the deeper sandlayer All these kept the sand moister than before the rain-fall in the shallow layer close to the surface (surface to thedepth of 20 cm) Furthermore at the end of the calcula-tion period seven days after the rainfall the moisture con-tent at the depth of 10 cm was 0036 cm3cm3 which was0015 cm3cm3 higher than its daily averaged value for thedays before the rainfall However the 6604 mm of rain-fall was not enough to penetrate into the B-type darker areawhich would keep the soil moisture in soil and form the ef-fective infiltration It should be noted that the effective in-filtration requires the rainfall penetrating below the deepestdrying front for example in this case under the depth of94 cm (Fig 7a)

33 Determination of effective infiltration

After the rainfall the infiltrating process will form a down-ward wetting front in soil When the rainfall stops the evap-oration on the surface would dry the wetted soil and form adrying front During the redistribution of the infiltrated rain-fall the drying front chases the wetting front If the infiltratedrainfall could not penetrate into the deepest drying front be-fore the stop of redistributing process and be caught up bythe drying front there would be no effective infiltration insand On the other hand if the rainfall does penetrate belowthe deepest drying front the volume of this part of rainfallcould be determined by the calculation of the cumulative in-filtration below the deepest drying front which can be iden-tified by the HYDRUS with field data With the HYDRUSthe volumetric soil moisture could be retrieved with a spatialresolution of 1 cm which helps us to know exactly when andwhere the redistribution of infiltrated rainfall stops The cu-mulative infiltration below the deepest drying front could bedetermined by the difference of the volumetric soil moistureprofiles before and after the redistribution which could bedescribed as

Reff=θwettedminusθoriginal

θ=

nsumi=1

(θi+θi+1

2)1z i=123 n (12)

Fig8 The schematic map of the calculation of the effective infiltration

Fig 8 The schematic map of the calculation of the effective infil-tration

whereReff is the effective infiltrationθwetted is the averagevolume of soil moisture after the redistributionθoriginal is theaverage volume of soil moisture before the redistributionθi

andθi+1 are the volumetric soil moistures below the deep-est drying front with an interval of1z andn is determinedby the difference between the deepest drying front and theinfiltrated depth [(the depth of deepest drying front the in-filtrated depth)1 z] For example in the case of the deepestdrying front of 94 cm if the infiltrated depth was 98 cm andthe interval (1z) was set as 1 cm then would be equal to4 ((98 cmminus94 cm)1 cm) whileθ1 θ2 θ3 θ4 andθ5 wouldbe the volumetric moisture content at the depth of 94 95 9697 and 98 cm (Fig 8)

The precipitation of 6604 mm was not enough to pene-trate into the deepest drying front in this case In order tohave an idea on how much volume of rainfall could pen-etrate into sand and became the effective infiltration thevolume of precipitation was set as 3302 mm with a rain-fall rate of 66 mmhr (Li et al 2006 Wang et al 2007)According to the calculation result the rainfall penetratedto the depth of 98 cm The original volumetric soil mois-ture content at the depth of 94 95 96 97 and 98 cm were00429 0043 00432 00433 and 00434 cm3cm3 respec-tively while they became 00465 00459 00455 0045 and00445 cm3cm3 after the stop of redistribution Accordingto the Eq (12) the effective infiltration was 00925 mm

4 Conclusions

The effective infiltration is defined here as the amount ofrainfall remained in sand beneath the deepest drying frontfrom an incident rainfall event In order to understand theeffective infiltration the coupled liquid water vapor and heat



transport in sand was analyzed to determine the drying frontby using the HYDRUS code According to the calculationresults the drying fronts were mainly determined by theshapes of the zero flux isolines There were two kinds ofshapes of zero flux isolines A-type which occurred in theshallow layer limited in the depth of 45 cm close to surfacewith small time intervals and isolated shapes and B-typewhich occurred below the depth of 45cm with a radical vari-ation (from the depth of 45 to 94 cm) and kept the continuityConsequently the drying front fluctuated sharply during thewhole calculation period The rainfall interrupted the fluctu-ation of the drying front and kept it in the depth of 20 cm for 3days Although the 6604 mm of rainfall could not penetrateinto the deepest drying front during the whole period it didkeep sand moister than before rainfall in the shallow layerclose to the surface (surface to the depth of 20 cm) Based onthe HYDRUS code a simple method was used to estimatethe effective infiltration With the artificially set rainfall of3302 mm and the rainfall rate of 66 mmhr the effective in-filtration was estimated as 00925 mm

Considering the maximum annual precipitation of 120 mmin the Badain Jaran Desert simply multiplying the estimatedeffective infiltration with a factor of 36 (120 mm3302 mm)the value of 0336 mmyr was lower than the historic recordsbased on hydrochemical and isotopic methods which wasfrom 095 to 36 mmyr In fact the annual precipitation of120 mm would not be simply characterized by about four sin-gle rainfalls each of which was of 3302 mm with the rain-fall rate of 66 mmhr The actual annual effective infiltrationwas possibly lesser in the Badain Jaran Desert The charac-teristics of individual rainfall events such as the amount thedensity and the duration and the variation pattern of the dry-ing front would affect the infiltrating and redistributing pro-cess in soil and thus the effective infiltration Furthermorethe annual spatial and temporal variation pattern of the pre-cipitation played a critical role in determining the effectiveinfiltration In order to understand the effective infiltration inthe Badain Jaran Desert a long-term observation should beestablished

AcknowledgementsThe research projects on which this paper isbased were funded by the National Natural Science Foundation ofChina (Grant Nr 90302003) the CAS International PartnershipProject ldquoThe Basic Research for Water Issues of Inland River Basinin Arid Regionrdquo (Grant Nr CXTD-Z2005-2)rdquo and ldquoThe ESA-MOST Dragon Programmerdquo and by the European Commission(Call FP7-ENV-2007-1 Grant nr 212921) as part of the CEOP ndashAEGIS project (httpwwwceop-aegisorg) coordinated by theUniversite Louis Pasteur We thank the anonymous referees verymuch for their helpful and constructive comments and suggestionsfor improving the manuscript We have a deep desire of acknowl-edgement for the help of Hirotaka Saito and Masaru Sakai on theguide of using HYDRUS

Edited by J Wen

References

Bonacci O Jukic D and Ljubenkov I Definition of catchmentarea in karst case of the rivers Krcic and Krka Croatia Hy-drolog Sci J 51 682ndash699 2006

Brutsaert W Evaporation into the atmosphere Theory historyand applications D Reidel Publ Dordrecht the Netherlands299 pp 1982

Camillo P J and Gurney R J Resistance Parameter for Bare-SoilEvaporation Models Soil Science SOSCAK 141 95ndash105 1986

Chen J S Li L Wang J Y Barry D A Sheng X F GuW Z Zhao X and Chen L Water resources ndash Groundwatermaintains dune landscape Nature 432 459-460 2004

Dong Z B Wang T and Wang X M Geomorphology of themegadunes in the Badain Jaran desert Geomorphology 60 191ndash203 2004

Fares A and Polyakov V Advances in crop water managementusing capacitive water sensors Adv Agron 90 43ndash77 2006

Fares A Hamdhani H and Jenkins D M Temperature-dependent scaled frequency Improved accuracy of multisensorcapacitance probes Soil Sci Soc Am J 71 894ndash900 2007

Gates J B Edmunds W M and Ma J Z Groundwater rechargerates to the Badain Jaran Desert preliminary results from envi-ronmental tracer studies the 34th congress of international asso-ciation of hydrogeologists Beijing China 9ndash13 October 2006311 2006

Gates J B Edmunds W M and Ma J Z Estimating groundwa-ter recharge in a cold desert environment in northern China usingchloride Hydrogeol J 16 893ndash910 2008a

Gates J B Bohlke J k Edmunds W M Ecohydrological fac-tors affecting nitrate concentrations in a phreatic desert aquiferin northwestern China Environ Sci Technol 42 3531ndash35372008b

Geiger S L and Durnford D S Infiltration in homogeneoussands and a mechanistic model of unstable flow Soil Sci SocAm J 64 460ndash469 2000

Geyh M A Gu W Z and Jakel D Groundwater rechargestudy in the Gobi Desert China Geowissenschaften 14 279ndash280 1996

Ghildyal B P and Tripathi R P Soil Physics Rajkamal ElectricPress India 1ndash656 1987

Gu W Z Seiler K P Stichler W and Lu J J Groundwaterrecharge in the badain jaran shamo inner mongolia PR Chinathe 34th congress of international association of hydrogeologistsBeijing China 9ndash13 October 2006 507 pp 2006

Hofmann J Lakes in the SE part of Badain Jaran Shamo theirlimnology and geochemistry Geowissenschaften 14 275ndash2781996

Hofmann J Geo-ecological investigations of the waters in thesouth-eastern Badain Jaran Desert in Status und spatquartareGewasserentwicklung Berliner Geogr Abh 64 Habilitationss-chrift Berlin 1ndash164 1999

Jakel D The Badain Jaran Desert its origin and developmentGeowissenschaften 14 272ndash274 1996

Li S Z Xiao H L Cheng G D Luo F and Liu L C Me-chanical disturbance of microbiotic crusts affects ecohydrologi-cal processes in a region of revegetation-fixed sand dunes AridLand Res Manag 20 61ndash77 2006

Lu N and Likos W J Unsaturated soil mechanics John Wi-leyampSons Inc Hoboken USA 584 pp 2004



Ma J Z Li D Zhang J W Edmunds W M and PrudhommeC Groundwater recharge and climatic change during the last1000 years from unsaturated zone of SE Badain Jaran DesertChinese Sci Bull 48 1469ndash1474 2003

Ma J Z and Edmunds W M Groundwater and lake evolution inthe Badain Jaran desert ecosystem Inner Mongolia HydrogeolJ 14 1231ndash1243 2006

Melone F Corradini C Morbidelli R and Saltalippi C Labo-ratory experimental check of a conceptual model for infiltrationunder complex rainfall patterns Hydrol Process 20 439ndash4522006

Mitkov I Tartakovsky D M and Winter C L Dynamics ofwetting fronts in porous media Phys Rev E 58 R5245ndashR52481998

Noborio K McInnes K J and Heilman J L Measurements ofcumulative infiltration and wetting front location by time domainreflectometry Soil Sci 161 480ndash483 1996

NJHRI Specification of soil test (SL 237-1999) China WaterPower press Beijing China 757 pp 1999

Philip J R and De Vries V D Moisture movement in porous ma-terials under temperature gradient Trans Am Geophys Union38 222ndash232 1957

Polyakov V Fares A and Ryder M H Calibration of a capaci-tance system for measuring water content of tropical soil VadoseZone Journal 4 1004ndash1010 2005

Saito H Simunek J and Mohanty B P Numerical analysis ofcoupled water vapor and heat transport in the vadose zone Va-dose Zone Journal 5 784ndash800 2006

Scanlon B R Keese K Reedy R C Simunek J and An-draski B J Variations in flow and transport in thick desert va-dose zones in response to paleoclimatic forcing (0ndash90 kyr) Fieldmeasurements modeling and uncertainties Water Resour Res39(7) 1179 doi1010292002WR001604 2003

Hukseflux Themal Sensors Soil temperature profile sensor usermanual (version 0606) Delft the Netherlands available athttpwwwhuksefluxcom 2007

Sentek Pty Ltd Calibration of Sentek Pty Ltd soil moisture sen-sors Stepney SA Australia available athttpwwwsentekcomauhomedefaultasp 2001

Shigeru O Yosuke K and Ayumi Y Studies on the infiltration-discharge of rain water and translation penomena in soil J Hy-drol 132 1-23 1992

Golden Software Surfer Reference Manual Version 4 ednGolden COhttpwwwgoldensoftwarecom 1990

Tami D Rahardjo H and Leong E C Effects of hysteresis onsteady-state infiltration in unsaturated slopes J Geotech Geoen-viron 130 956ndash967 2004

Timlin D and Pachepsky Y Infiltration measurement using a ver-tical time-domain reflectometry probe and a reflection simulationmodel Soil Sci 167 1ndash8 2002

Topp G C Davis J L and Annan A P Electromagnetic de-termination of soil water content using TDR I applications towetting fronts and steep gradients Soil Sci Soc Am J 46 672ndash678 1982

van Bavel C H M and Hillel D I Calculating potential and ac-tual evaporation from a bare soil surface by simulation of concur-rent flow of water and heat Agric For Meteorol J 17 453ndash4761976

van Genuchten M T A closed-form equation for predicting thehydraulic conductivity of unsaturated soils Soil Sci Soc Am J44 892ndash898 1980

Walker A S Olsen J W and Bagen The Badain Jaran DesertRemote sensing investigations Geogr J 153 205ndash210 1987

Walvoord M A Plummer M A Phillips F M and WolfsbergA V Deep arid system hydrodynamics 1 Equilibrium statesand response times in thick desert vadose zones Water ResourRes 8(12) 1308 doi1010292001WR000824 2002

Wang G X and Cheng G D Water resource development and itsinfluence on the environment in arid areas of China ndash the case ofthe Hei River basin J Arid Environ 43 121ndash131 1999

Wang X P Li X R Xiao H L Berndtsson R and Pan YX Effects of surface characteristics on infiltration patterns in anarid shrub desert Hydrol Process 21 72ndash79 2007

Wang X P Cui Y Pan Y X Li X R Yu Z and Young M HEffects of rainfall characteristics on infiltration and redistributionpatterns in revegetation-stabilized desert ecosystems J Hydrol358 134ndash143 2008

Wu J Q Zhang R D and Yang J Z Estimating infiltrationrecharge using a response function model J Hydrol 198 124ndash139 1997

Yang H Rahardjo H Wibawa B and Leong E C A soil col-umn apparatus for laboratory infiltration study Geotech Test J27 347ndash355 2004

Yang H Rahardjo H and Leong E C Behavior of unsaturatedlayered soil columns during infiltration J Hydrol Eng 11 329ndash337 2006

Yang X P Landscape evolution and palaeoclimate in the desertsof northwestern China with a special reference to Badain Jaranand Taklamakan Chinese Sci Bull 46 6ndash11 2001

Yang X P Liu T S and Xiao H L Evolution of megadunes andlakes in the Badain Jaran Desert Inner Mongolia China duringthe last 31 000 years Quatern Int 104 99ndash112 2003a

Yang X P and Williams M A J The ion chemistry of lakesand late Holocene desiccation in the Badain Jaran Desert InnerMongolia China Catena 51 45ndash60 2003b

Yang X P Chemistry and late Quaternary evolution of ground andsurface waters in the area of Yabulai Mountains western InnerMongolia China Catena 66 135ndash144 2006

Zeng Y Wan L Su Z Saito H Huang K and Wang XDirnal soil water dynamics in the shallow vadose zone EnvironGeol doi101007s00254-008-1485-8 2008

Zhang H C and Ming Q Z Hydrology and Lake Evolution inHyperarid Northwestern China and the Mystery of MegaduneFormation in Badain Jaran Desert Advances in Earth Science21 532ndash538 2006 (in Chinese)



and Edmunds 2006 Yang 2006 Yang and Williams 2003bZhang and Ming 2006 Chen et al 2004 Gates et al 20062008b Jakel 1996)

Afore mentioned studies are mainly focused on the rateat which the infiltrated rainfall recharges the groundwaterwhich could be called the effective infiltration or the effectiverainfall (Wu et al 1997 Bonacci et al 2006) However theeffective infiltration or rainfall could also be defined as theamount of rainfall stored in the shrub root zone excludingthe fraction that runs off the soil surface or passes throughthe root zone that does not contribute to shrub growth andthe fraction that evaporates (Wang et al 2007) From abovetwo definitions obviously the effective infiltration or rain-fall could be generally defined as the portion of rainfall thatpenetrates and remains in a domain of interest With thisgeneral definition in this study the effective infiltration isthe amount of soil water remained in sand beneath the dry-ing front (marking the interface between the upwards anddownwards soil water fluxes) by infiltrating water of a rain-fall event and thus not evaporated and is evaluated based onthe vadose zone soil water balance method

As what mentioned above the estimates of infiltration ratein desert have been conducted a lot with the tracer methodwhich is mainly used to evaluate variations of flow and trans-port in thick desert vadose zones in response to paleoclimaticforcing and thus with the time scale of thousand-year (Scan-lon et al 2003) While with the soil water balance modelthe infiltration and soil water redistribution process in the un-saturated zone could be tracked at any meaningful time scaleof interest In order to do that quantifying changes in in-filtration properties by a reliable observation and measure-ment is needed Previous studies has indicated the feasibilityof using soil water content probes installed into the soil totrack the propagation of the wetting front during infiltration(Topp et al 1982 Noborio et al 1996 Timlin and Pachep-sky 2002 Melone et al 2006 Tami et al 2004 Yang 20042006 Wang et al 2007 2008 Geiger and Durnford 2000)and therefore to track the drying front during evaporation

In this study the measurements of soil water content soiltemperature and precipitation were used to verify and cali-brate the performance of a vadose zone soil water balancemodel the HYDRUS which was subsequently used to pro-duce the temporal and spatial information of coupled waterwater vapor and heat transport This enabled us to character-ize the variation pattern of the drying front before during andafter rainfall so that we could assess how much precipitationis evaporated and how much is conserved in sand Our goalwas to understand the capacity of the sand dune to capturecertain amount of rainfall during a single precipitation eventwhich is available to plants The first section of the paperthus describes the experiments and methods used Then fol-lows simulation results based on experimental observationsFinally we conclude the paper with a discussion of practicalresults

2 Materials and methods

21 Study site description

The study site is a relatively flat area at the foot of duneswith an area of about 100 square meters located in thesoutheast Badain Jaran Desert which lies in the northwestof the Alashan plateau in western Inner Mongolia of Chinabetween 39 20primeN to 41 30primeN and 100 E to 104 E TheBadain Jaran Desert is regarded as the second largest desertin China and covers an area of about 49 000 km2 (Yang2001) The desert landscape primarily consists of unvege-tated or sparsely vegetated aeolian sand dune fields as a coresurrounded by desert plains with pediments on the margin(Yang et al 2003a) The elevation ranges from 1500 m inthe southeast to 900 m above sea level in the northwest withthe mega dunes up to 400 m high in the southeast alignedin a SW-NE pattern (Dong et al 2004) The mean grainsize of the dune sand in this desert is between 021 mm and022 mm (Jakel 1996 Yang and Williams 2003b) Thereare over 70 lakes in the interdune areas of which the sur-face area vary from 02 km2 on the southeastern edge of thedesert to 16 km2 slightly northwards and the water depthfrom 2 m to 16 m (Yang and Williams 2003b) On the sur-face there is neither recharging into these lakes nor dis-charging from them and there is no river in the entire desert(Yang 2001 Wang and Cheng 1999) With such hydrolog-ical conditions vegetation in the study area is sparse Theplant or shrub communities depend mainly on the geomor-phology and on edaphic conditions The dominant plantspecies are Haloxylon ammodendron Psammochoa villosaPhragmites communis Artemisia ordosica while in the areaaround lakes and springs the denser groundcover observablecould be found (Polypogon monspeliensis Triglochin mar-itima Achnatherum splendens Carex sp Glaux maritima)(Gates et al 2008a)

The distribution of sparse vegetation in this area is mainlycaused by the dominated climate pattern in the AlashanPlateau which is characterized by a markedly arid and conti-nental climate During the winter months a well-developedhigh pressure system cold and dry continental air masseswith temperatures below zero influences the area whichcaused the mean monthly temperature in January becomeminus10C In the summer months the tropical air masses fromthe Pacific Ocean hits the Badain Jaran Desert with more thanhalf of the average annual precipitation fell in July The spa-tial and temporal distribution of the precipitation in the areais determined by the Asian summer monsoons at the presenttime and the mean monthly temperatures in July is 25Cwith a diurnal temperature variation up to more than 45CThe mean annual precipitation varies from 84 to 120 mm inthe southeast and from 37 to 40 mm in the northwest andthe potential evaporation is approximately 2600 mmyr be-ing the highest in China (Ma et al 2003 Ma and Edmunds2006 Gates et al 2008a)













partθ

partt=minus

partqL

partzminus

partqv

partz(1)






partz+ 1) minus KLT

partT

partz(2)





parth

partzminusKvT

partT

partz(3)


[partθ

partt=

part

partz

[KLh

parth

partz+KLh+KLT

partT

partz+Kvh

parth

partz+KvT

partT

partz

]]

=part

partz

[KT h

parth

partz+KLh+KT T

partT

partz

](4)








(a)









θ(h)=

θr+

θsminusθr

[1+|αh|n]m

θs

h le 0h gt 0

(7)








(a)







s H=Ca

TsminusTa

rH(9)




E=ρvsminusρva

rv+rs(10)








RRMSE=

radicNWsumi=1

(MiminusCi)2Nw




















(a)

(b)











(a)

(b)









(a)

(b)










θ=

nsumi=1

(θi+θi+1

2)1z i=123 n (12)






4 Conclusions







Edited by J Wen

References



































































partθ

partt=minus

partqL

partzminus

partqv

partz(1)






partz+ 1) minus KLT

partT

partz(2)





parth

partzminusKvT

partT

partz(3)


[partθ

partt=

part

partz

[KLh

parth

partz+KLh+KLT

partT

partz+Kvh

parth

partz+KvT

partT

partz

]]

=part

partz

[KT h

parth

partz+KLh+KT T

partT

partz

](4)








(a)









θ(h)=

θr+

θsminusθr

[1+|αh|n]m

θs

h le 0h gt 0

(7)








(a)







s H=Ca

TsminusTa

rH(9)




E=ρvsminusρva

rv+rs(10)








RRMSE=

radicNWsumi=1

(MiminusCi)2Nw




















(a)

(b)











(a)

(b)









(a)

(b)










θ=

nsumi=1

(θi+θi+1

2)1z i=123 n (12)






4 Conclusions







Edited by J Wen

References



























































partz+ 1) minus KLT

partT

partz(2)





parth

partzminusKvT

partT

partz(3)


[partθ

partt=

part

partz

[KLh

parth

partz+KLh+KLT

partT

partz+Kvh

parth

partz+KvT

partT

partz

]]

=part

partz

[KT h

parth

partz+KLh+KT T

partT

partz

](4)








(a)









θ(h)=

θr+

θsminusθr

[1+|αh|n]m

θs

h le 0h gt 0

(7)








(a)







s H=Ca

TsminusTa

rH(9)




E=ρvsminusρva

rv+rs(10)








RRMSE=

radicNWsumi=1

(MiminusCi)2Nw




















(a)

(b)











(a)

(b)









(a)

(b)










θ=

nsumi=1

(θi+θi+1

2)1z i=123 n (12)






4 Conclusions







Edited by J Wen

References


























































θ(h)=

θr+

θsminusθr

[1+|αh|n]m

θs

h le 0h gt 0

(7)








(a)







s H=Ca

TsminusTa

rH(9)




E=ρvsminusρva

rv+rs(10)








RRMSE=

radicNWsumi=1

(MiminusCi)2Nw




















(a)

(b)











(a)

(b)









(a)

(b)










θ=

nsumi=1

(θi+θi+1

2)1z i=123 n (12)






4 Conclusions







Edited by J Wen

References






























































RRMSE=

radicNWsumi=1

(MiminusCi)2Nw




















(a)

(b)











(a)

(b)









(a)

(b)










θ=

nsumi=1

(θi+θi+1

2)1z i=123 n (12)






4 Conclusions







Edited by J Wen

References
































































(a)

(b)











(a)

(b)









(a)

(b)










θ=

nsumi=1

(θi+θi+1

2)1z i=123 n (12)






4 Conclusions







Edited by J Wen

References






























































(a)

(b)









(a)

(b)










θ=

nsumi=1

(θi+θi+1

2)1z i=123 n (12)






4 Conclusions







Edited by J Wen

References




























































(a)

(b)










θ=

nsumi=1

(θi+θi+1

2)1z i=123 n (12)






4 Conclusions







Edited by J Wen

References






























































θ=

nsumi=1

(θi+θi+1

2)1z i=123 n (12)






4 Conclusions







Edited by J Wen

References




























































Edited by J Wen

References



























































































Documents

Diurnal pattern of the drying front in desert and its application for determining the ... · 2016-01-10 · Badain Jaran Desert, a ﬁeld campaign was conducted by the observations