Aa

MDT

a

A

R

R

2

A

P

K

A

R

S

D

M

0d

e c o l o g i c a l m o d e l l i n g 2 1 7 ( 2 0 0 8 ) 148–156

avai lab le at www.sc iencedi rec t .com

journa l homepage: www.e lsev ier .com/ locate /eco lmodel

model for Evaporation and Drainage investigationst Ground of Ordinary Rainfed-areas

ehmet Aydin ∗

epartment of Soil Science, Faculty of Agriculture, Mustafa Kemal University,R-31040 Antakya, Turkey

r t i c l e i n f o

rticle history:

eceived 25 December 2007

eceived in revised form

8 April 2008

ccepted 12 June 2008

ublished on line 18 July 2008

eywords:

rid environment

ainfed-areas

oil evaporation

rainage

odelling

a b s t r a c t

Quantification of water losses through evaporation and drainage from bare soils in arid

and semi-arid regions is very important for an effective management strategy to conserve

soil water. In this study, a model for Evaporation and Drainage investigations at Ground of

Ordinary Rainfed-areas (hereafter E-DiGOR) is presented. Daily potential evaporation (Ep)

from bare soils was calculated using the Penman–Monteith equation with a surface resis-

tance of zero. Actual soil evaporation (Ea) was computed according to Aydin et al. [Aydin,

M., Yang, S.L., Kurt, N., Yano, T., 2005. Test of a simple model for estimating evaporation

from bare soils in different environments. Ecol. Model. 182 (1), 91–105; Aydin, M., Yano, T.,

Evrendilek, F., Uygur, V., 2008. Implications of climate change for evaporation from bare

soils in a Mediterranean environment. Environ. Monit. Assess. 140, 123–130]. Deep drainage

(D) was simply calculated by the mass balance, taking field capacity into account. In order

to test the performance of the model mainly for drainage estimations, a micro-lysimeter-

experiment was carried out under field conditions. The experimental terrain was nearly

flat, with no appreciable slope. Estimated and measured water balance components such as

actual evaporation (R2 = 91.4%; P < 0.01), drainage (R2 = 88.5%; P < 0.01) and soil water storage

(R2 = 89.7%; P < 0.01) were in agreement. This agreement supported the model hypothesis,

thus rendering the model useful in estimating soil evaporation, drainage and water storage

in an interactive way with a few parameters.

Once the estimated and measured data from the experiment had been compared for vali-

dation, simulations were carried out continuously for the period of 1994–2006 in a semi-arid

environment of Turkey. Ep rates were lower during the winter season because of the lesser

evaporative demand of the atmosphere. However, Ea rates were mainly found to be functions

of the rainfall pattern, and presumably soil wetness in addition to atmospheric evaporative

demand. D volumes below 120 cm soil depth were high during rainy months, with a peak

in January. Annual Ep varied between 850.6 and 909.8 mm during the period of 13 years.

Ea ranged from 248.0 to 392.9 mm with a mean annual value of 302.5 mm. D substantially

varied inter-annually (150.5–757.4 mm) depending on the intensity and frequency of rainfall

events and especially antecedent soil wetness. The next logical step in model development

would be the inclusion of runoff for sloping lands.

∗ Tel.: +90 326 245 58 45; fax: +90 326 245 58 32.E-mail address: maydin@mku.edu.tr.

304-3800/$ – see front matter © 2008 Elsevier B.V. All rights reserved.oi:10.1016/j.ecolmodel.2008.06.015

© 2008 Elsevier B.V. All rights reserved.

g 2 1

1

TaretppSwmaroebrfiwt1aaRea1aeis

iosbmuFrteikwctesebroavTss

e c o l o g i c a l m o d e l l i n

. Introduction

he most important transport processes are characterized bysimultaneous change in the amount of energy and/or mate-

ial with time and place (Aydin and Huwe, 1993). Soil water andnergy balances interact since they are integrative aspects ofhe same processes within the same environment. For exam-le, soil water potential plays a role in water flow, and waterotential mainly depends on water content (Campbell, 1985).oil evaporation is an important driving force for changes inater content. There exist many methods for direct deter-inations of evaporation (Boast, 1986). However, instead of

dopting a method to directly measure evaporation, manyesearchers prefer a practical model to estimate actual evap-ration. In general, models of soil water evaporation havexpressed the rate of loss from cropped areas rather than fromare soils. However, in semi-arid regions, the soil evaporationepresents a large fraction of the total water loss from bareelds. In regions where summer fallow is practiced, the soilater evaporation accounts for most of the incoming precipi-

ation affecting directly soil water storage (Hanks, 1992; Hillel,998). Thus, many simple models are available for the reason-ble predictions of soil evaporation (Gardner, 1959; Gardnernd Hillel, 1962; Hanks and Gardner, 1965; Black et al., 1969;itchie, 1972; Van Keulen and Hillel, 1974; Hillel, 1975; Jacksont al., 1976; Campbell, 1985; Brisson and Perrier, 1991; Katulnd Parlange, 1992; Malik et al., 1992; Alvenas and Jansson,997). The evaporation from bare soils depends not only ontmospheric conditions but also on soil properties. Param-terization of evaporation from a non-plant-covered surfaces very important in the hierarchy strategy of modelling landurface processes (Mihailovic et al., 1995).

The evaporation and drainage from bare soils are stronglynterdependent, as they occur sequentially and simultane-usly. Unfortunately, it is almost impossible to measureeepage rate directly. Indirect experimental techniques maye used, but are cumbersome and require sophisticated equip-ent (Aydin, 1994). One can study water budget of a soil

nder natural conditions by theoretical means (Mwendera andeyen, 1997; Eilers et al., 2007; Moret et al., 2007). A majorequirement for studying vertical soil water flow is the solu-ion of the Richards’ equation (Yang and Yanful, 2002; Varadot al., 2006). This equation has a clear physical basis ands subject to specified initial and boundary conditions, withnown relations among the volumetric water content, soilater potential and hydraulic conductivity. In many practi-

al situations, however, the detailed information concerninghe hydraulic conductivity and water retention relations nec-ssary for the solutions is not readily available. In theseituations, much simpler but not necessarily less precise mod-ls are required (Aydin et al., 2005). If models have a soundasis in physical science, the processes involved are accu-ately depicted within the models (Kite et al., 2001). Dependingn the amount and quality of information available for inputnd required for output, either a very simple model or a

ery complex model may be appropriate (Aydin et al., 2005).here is no single way that is likely to be applicable to allituations. However, when the models are to be included inoil-water management systems, they need to be relatively

7 ( 2 0 0 8 ) 148–156 149

simple and require readily available parameters (Armstronget al., 1993).

In this study, a model for Evaporation and Drainage inves-tigations at Ground of Ordinary Rainfed-areas (called E-DiGORhereafter) is therefore presented and validated for the descrip-tions and predictions of the past, present and future dynamics.

2. Description of the model

Daily potential evaporation from bare soils can be calculatedusing the Penman–Monteith equation (Allen et al., 1994) witha surface resistance of zero (Wallace et al., 1999; Aydin et al.,2005):

Ep = �(Rn − Gs) + 86.4�cpı/ra�(�+ �)

(1)

where Ep is potential soil evaporation (kg m−2 d−1 ∼= mm d−1),� is the slope of saturated vapour pressure–temperature curve(kPa ◦C−1), Rn is the net radiation (MJ m−2 d−1), Gs is the soilheat flux (MJ m−2 d−1), � is the air density (kg m−3), cp is thespecific heat of air (kJ kg−1 ◦C−1 = 1.013), ı is the vapour pres-sure deficit of the air (kPa), ra is the aerodynamic resistance(s m−1), � is the latent heat of vaporization (MJ kg−1), � is thepsychrometric constant (kPa ◦C−1), and 86.4 is the factor forconversion from kJ s−1 to MJ d−1.

Initially evaporation from a wet soil proceeds at the poten-tial rate. With time, the soil surface becomes progressivelydrier, and drying front moves into the soil with a time-lag.Thus, soil water potential at the top surface layer declines. Inorder to estimate soil water potential at the top surface layer,Aydin et al. (2008) tested a model originally described by Aydinand Uygur (2006):

= −[

(1/˛)(

10∑

Ep)3

2(�fc − �ad) (Davt/�)1/2

](2)

where � is soil water potential (cm of water) at the top sur-face layer, ˛ is a soil-specific parameter (cm) related to flowpath tortuosity in the soil,

∑Ep is cumulative potential soil

evaporation (cm), and �fc and �ad are average-volumetric watercontent (cm3 cm−3) at field capacity and air-dryness, respec-tively. Field capacity is defined as the amount of water, whichthe soil can hold against gravitational forces. Dav is averagehydraulic diffusivity (cm2 d−1) determined experimentally, t isthe time since the start of evaporation (days), and � is 3.1416.

The term, 2(�fc − �ad)(Davt/�)1/2, in Eq. (2), represents watersupplied from deeper layers to the soil surface. This term issimilar to the equation given by Gardner (1959) for the cal-culation of cumulative actual soil evaporation. There are twooptions for starting the calculations: (1) calculations start ona day that a considerable soil depth (e.g. top 5 cm) is at nearlysaturation as is the case after heavy rainfall and (2) if soil waterretention curve is available, an initial quantity for

∑Ep, which

results in a � value corresponding to initial water content of

the top surface layer, can be used for the first day of the studyperiod. As can be seen from Eq. (2), it was assumed that (i)water potential near the soil surface decreases as a cubic func-tion of potential soil evaporation and (ii) upward transport of

i n g

150 e c o l o g i c a l m o d e l l

water to soil surface dried is capable of partial compensationfor such a decrease. During a drying period, the top surfacelayer of the soil may dry out eventually to air-dry wetness,and the soil surface approaches equilibrium with overlyingatmosphere. The soil then no longer evaporates at a consider-able rate, except for water transport by the slow process ofmoisture diffusion. The hydraulic diffusivity is a combinedterm including both water content-matric potential and watercontent-hydraulic conductivity relations. The main advantageis that the diffusivity variable is very useful as it accounts forchanges in both soil water storage and hydraulic conductiv-ity. The combination of average diffusivity and water contentis an estimate of the amount of water that the soil is capa-ble of supplying to the soil surface, assuming an infinite valueof potential evaporation rate at the soil surface (Hanks, 1992;Hillel, 1998).

Assuming that the water potential at dry soil surface is atequilibrium with the atmosphere, the minimum water poten-tial can be derived from the Kelvin equation (Brown andOosterhuis, 1992; Kirby and Ringrose-Voase, 2000; Aydin et al.,2005):

ad = RgT

mglnHr (3)

where �ad is the water potential for air-dry conditions(cm of water), T is the absolute temperature (K), g is theacceleration due to gravity (981 cm s−2), m is the molecularweight of water (0.01802 kg mol−1), Hr is the relative humid-ity of the air (fraction), and Rg is the universal gas constant(8.3143 × 104 kg cm2 s−2 mol−1 K−1).

Several simplified models have been advocated for the rela-tion between soil evaporation and soil water potential (Ehlersand van der Ploeg, 1976; Beese et al., 1977; Huwe and van derPloeg, 1990; Aydin, 1994; Zhu and Mohanty, 2002). Recently,Aydin et al. (2005) validated a simple model originally pro-posed by Aydin (1998) for estimating actual soil evaporationusing matric potential at the top surface layer, neglecting theinfluence of the hydraulic gradient. Consequently, it is possi-ble to incorporate Eqs. (1)–(3) into Aydin model (Aydin et al.,2008):

Ea = Log| | − Log| ad|Log| tp| − Log| ad|

Ep (4)

If | | ≤ | tp|, then Ea = Ep or Ea/Ep = 1.0

If | | ≥ | ad|, Ea = 0.0

Remain that Ep ≥ 0.

where Ea and Ep are actual and potential evaporation rates(mm d−1), respectively, |� tp| is the absolute value of soil waterpotential (matric potential) at which actual evaporation startsto drop below potential one, |�ad| is the absolute value of soilwater potential at air-dryness, and |� | is the absolute value of

soil water potential determined by Eq. (2). The values of all �are in cm of water.

In practice, soil water storage between the soil surface (0)and a given depth (Z) is calculated by integrating water content

2 1 7 ( 2 0 0 8 ) 148–156

of individual soil layers:

S =∫ z

0

� dz (5)

where S is soil water storage with dimension length (mm)allowing to compare it with precipitation or evaporation andto facilitate the calculations, � is volumetric water content(cm3 cm−3) of the individual soil layers, and Z is the depth ofsoil zone (mm) considered. If S = �adZ, then Ea = 0. Consider-ing the initial water content (�i = cm3 cm−3) of the soil layers,similar definitions hold for initial water amount

(∫ z

0�i dz

).

Aydin et al. (2008) reported that Eq. (2) did not have a clearphysical definition for resetting after rainfall events and needsfurther studies to determine the minimum amount of dailyrainfall based on a ratio of rainfall to potential evaporation inorder to reset the model. When there is a day with rainfall (onthe day number, j) during the drying cycle, calculation of Ea ismodified based on the magnitude of the amount of rainfall (P).Three cases are here identified:

Case 1. No substantial re-wetting (P(j) < E(j)a ). In this case, the

calculation continues without any modification.

Case 2. Slight re-wetting (E(j)a ≤ P(j) ≤ E

(j)p ). In this case, the

small amount of rain only slightly reduces the increase in|� |, and contributes to Ea. The evaporation demand on thatday is met by the rainfall. Thus, E(j)

a is equated to P(j) (with

mathematical term (E(j)a = P(j)).

Case 3. Moderate or complete re-wetting (P(j) > E(j)p ). The evap-

oration demand on the day j is met by the rainfall, and aproportion of rainfall is retained near the soil surface for evap-oration on the following day(s). In this case, both (� ) and (Ea)are reset. Thus, calculation of Ea has to be updated for thefollowing conditions:

Condition 1. For P(j) < �fcZ − S(j−1), counting from the day j:P(j) − (E(j)

a + E(j+1)a + E

(j+2)a + · · · + E

(j+n)a ) ∼= 0. If |� (j+n)| < |� (j−1)| and

|� (j+n)| < |�ad|, then it is assumed that |� (j+n+1)| =� (j−1). Forthe calculation of � (j+n+2), the assumption is that adjusted[∑

Ep](j+n+2) =

[∑Ep

](j−1) + E(j+n+2)p .

Condition 2. For P(j) ≥ �fcZ − S(j−1), Ea is calculated with thestandard procedure given in Eqs. (2) and (4) on the days fol-lowing reset.

The soil water storage on any day can be imposed on thedifference between rainfall (in case) and actual evaporation onthe consecutive day. Symbolizing this produced variable as W,and assuming a negligible runoff from nearly level soils, thefollowing expression can be written:

W(j) = S(j−1) + P(j) − Ea(j)

If W(j) < �fcz, then S(j) = W(j) (6)

If W(j) ≥ �fcz, then S(j) = �fcz.

For the first day of simulation period, the initial wateramount is used as S(j−1) in Eq. (6). When the water content

g 2 1

ostcbawOftas

cl

[

ws(TdD

ceu

3

3

Sevaso(

iol1cmtw(

3

Fficd

e c o l o g i c a l m o d e l l i n

f the soil reaches field capacity, water can drain through theoil profile. At field capacity, any excess water drains throughhe soil zone. Although, in practice, a small amount of wateran drain from the soil zone when the water content is slightlyelow field capacity, the assumption that recharge only occurst field capacity leads to minor errors (Eilers et al., 2007). Itas also assumed that impervious barriers are not in play.n the other hand, the assumption does not exactly allow

or water repellency. Similarly, possible by-pass or rapid flowhrough or from the soil zone is ignored in the present model,lthough these processes may be active in the dried clayoils.

Drainage is simply calculated by the mass balance. Theumulative drainage until the day j can be expressed as fol-ows:

∑D

](j)=

∫ z

0

�idz+[∑

P

](j)−

[∑Ea

](j)− S(j) (7)

here∑

D is cumulative drainage (mm) out of storage depthince the first day of simulation period,

∑P is total rainfall

mm), and∑

Ea is cumulative actual soil evaporation (mm).hus, from the differences between the consecutive days,rainage rates (D = mm d−1) can be easily calculated, if any:(j) = [

∑D](j) − [

∑D](j−1).

The upward flux from deeper layers into the profile zoneonsidered is neglected. For comparison, all quantities arexpressed in terms of volume per unit area (equivalent depthnits).

. Materials and methods

.1. Study area

tudy site – Adana (36◦59′N, 35◦18′E) – is located in a semi-aridnvironment of Turkey. A typical Mediterranean climate pre-ails in the study area with the long term (1975–2006) meannnual temperature, precipitation and potential evapotran-piration of 19.0 ◦C, 650 and 1320 mm, respectively. About 87%f precipitation occurs during the period of November to May

Yano et al., 2007).The soil at the study site has no water table and salin-

ty problems and is fine-textured soil throughout the profilef 120 cm (e.g. sand: 331, silt: 122, and clay: 547 g kg−1 at the

ayer of 0–40 cm). Dry bulk density varied between 1.20 and.27 g cm−3. On average, volumetric water content at fieldapacity was measured as 0.35 cm3 cm−3. Only a brief sum-ary for the soil is provided here. The physical properties of

he soil were described in detail by Aydin and Huwe (1993),ith some further information on the soil given by Dinc et al.

1991).

.2. Experiment for validation of the model

or validation, a micro-lysimeter-experiment was carried out

or a period of 5 months under field conditions. The exper-mental plot with no appreciable slope of 10 m × 10 m washosen. The micro-lysimeters were composed of PVC cylin-ers with 57.0 cm in length, 10.2 cm in internal diameters and

7 ( 2 0 0 8 ) 148–156 151

a thickness of 0.4 cm. The cylinders, along with a steel cut-ter, were inserted into wetted soil and then carefully takenout as described by Aydin et al. (2005). Wire-gauze sieveswere affixed to the base of the cylinders in order to retainsoil particles without disturbance of drainage (Boast andRobertson, 1982; Evett et al., 1995; Yang et al., 2002; Aydinet al., 2008). A seepage collector as a close-ended segmentwas also connected to the bottom of micro-lysimeters. Threereplicated micro-lysimeters with the seepage collector wereplaced in open-ended PVC columns installed in the experi-mental plot, which served as soil-retaining walls. In order toprevent possible runoff, the upper part of the micro-lysimeterswas left 1 cm above inside and surrounding soil levels at thebeginning of the experiment. The soil was allowed to dryby evaporation while being kept free of weeds during theexperiment. A rain gauge was located near micro-lysimeters.During a re-wetting event, the water drained from the bot-tom of micro-lysimeters was measured 1 day after rainfall isstopped, and the drainage rate was obtained. During a dry-ing cycle, micro-lysimeters were weighed every 2–3 days atthe same time with an electronic balance. Evaporation wasdetermined from the change in weight by taking seepageand rainfall amount into consideration. The recorded datafrom three micro-lysimeters were averaged to represent themeasurements.

3.3. Application of the model

Climate data for the study area were obtained from TurkishState Meteorological Service (DMI). Daily potential evapora-tion from bare soil was calculated using the Penman–Monteithequation (Eq. (1)). In this study, albedo of the bare soil wasassumed to be 0.15 (Van Dam et al., 1997; Ács, 2003). In the cal-culations of soil water potential by Eq. (2), ˛ for the clay soil wastaken as 1.1 cm. �ad and Dav were assumed to be 0.05 cm3 cm−3

and 95 cm2 d−1, respectively. The data necessary for Eq. (4) areeasily obtainable from Eqs. (1)–(3). We used 60.0 cm of wateras � tp for the clay soil as suggested by Aydin et al. (2005).Soil water storage and drainage were estimated by Eqs. (6)and (7), respectively. A relational diagram of the equations isillustrated in Fig. 1. As previously explained, 56 (=57 − 1) cmcolumns were used for the in situ experiment. Measurementsfrom these columns were compared to those calculated for thedepth of 56 cm. Once the estimated and measured data fromthe experiment had been compared for validation, the simu-lation of soil water components was carried out continuouslyfor the entire period of 1994–2006. However, in order to soundlyconstruct the soil-water budget of the study area, a more real-istic soil depth of 120 cm was used for the computations of fieldwater balance. It is because this soil depth plays an importantrole in crop production and soil-water conservation. For exam-ple, the effective root depth of many field crops grown locally isabout 120 cm. Initial water content measured gravimetricallywas inferred from unpublished data of the author for the firstsimulation year. Initial water amount with dimension lengthat the beginning of any year was assumed to be the same as

that of the end of the previous year. From estimates of soilwater storage, soil water deficit (SWD), defined as the amountof water required to bring the soil up to the field capacity, wascalculated.

152 e c o l o g i c a l m o d e l l i n g 2 1 7 ( 2 0 0 8 ) 148–156

agram

Fig. 1 – Schematic definition di

4. Results and discussion

4.1. Model performance

The data from micro-lysimeter-experiment were pooled, andthe estimated components of soil water balance were plottedagainst the measured ones (Fig. 2). As can be seen in Fig. 2, thesimulated and measured evaporation rates compared reason-ably well (R2 = 91.4%; P < 0.01). Similarly, the scatter of drainagedata clearly shows that the measured drainage rates and thepredicted ones by the model are in close agreement (R2 = 88.5%;P < 0.01). In addition, the simulations of water stored inthe micro-lysimeters reflected the measurements sufficiently(R2 = 89.7%; P < 0.01). Consequently, results indicate that theE-DiGOR model is useful in quantifying the components ofsoil water balance with only a few parameters. However, thesystematic errors in the model parameters should be consid-ered. For example, in a drying system, a single and constant“weighted mean diffusivity” is itself an oversimplification(Baver et al., 1972). For a more realistic treatment, the dif-fusivity values for different ranges of water content mustbe determined experimentally. On the other hand, the com-plexity of water regime of clay soils should be explored. Theinterpretation of water transport in clay soils is quite difficultrelative to that of non-clay soils (Ritchie et al., 1972; Hasegawaand Sato, 1987; Aydin and Huwe, 1993; Aydin et al., 2005, 2008).

4.2. Quantification of water balance components

Simulations of daily Ep, Ea, S and D for the bare soil were con-ducted for each of the years during the period of 1994–2006

of the computational model.

and revealed similar patterns. Graphical illustration of dailyP, Ep, Ea, D and S values for the entire period would haverequired a lot of space. For this reason, daily changes in thevariables for Adana in 1997, similar to the long-term rainfalland climate regimes of the study area, were given as a rep-resentative example in Fig. 3. Ep rates were lower during thewinter season because of the lesser evaporative demand of theatmosphere. However, Ea rates were mainly found to be func-tions of the amount and timing of rainfall, and presumablysoil wetness in addition to atmospheric evaporative demand.During much of the rainy seasons, evaporation from bare soilwas at or close to the potential rate. During a drying period, Ea

began to decrease continuously. Similarly, during the dry sum-mer months, with a dry layer at the soil surface, Ea rates werevery low or zero. Ea outputs of the model are consistent withthe results of Aydin (2004) and Aydin et al. (2008). Drainageoccurred on the days with rainfall (or on the following days)if W(j) > �fcZ. Drainage rates below 120 cm soil depth were highduring rainy seasons with a maximum value of 27 mm d−1.The drainage was affected by rainfall and increased with ahigher amount of rainfall and soil water content. Soil waterstorage varies daily depending on the intensity and frequencyof rainfall events and on evaporation rates. The water storedin the soil reaches field capacity during the wet seasons withlesser evaporative demand of the atmosphere. However, thewater storage decreased continuously during the dry season.Hillel (1998) has also reported that the process of evaporationfrom a bare soil can be divided into three stages: (a) when

the soil is wet, evaporation occurs at the potential rate, hencethe only limitation is the atmospheric demand; (b) when thesoil becomes dryer, water cannot be supplied to the soil sur-face fast enough to meet the evaporative demand, and so

e c o l o g i c a l m o d e l l i n g 2 1 7 ( 2 0 0 8 ) 148–156 153

Fw

tls

fEhO

Fig. 3 – Comparison of potential (Ep) and actual (Ea)

evaporation from bare soil, drainage rate (D) below 120 cm

ig. 2 – Estimated versus measured components of soilater balance for micro-lysimeter-experiment.

he rate of evaporation decreases as the thickness of the dryayer increases; finally (c) the rate of evaporation becomes verymall compared to the potential demand.

Monthly variations of Ep, Ea, D and SWD along with rainfall

or a period of 13 years (1994–2006) in Adana are given in Fig. 4.

p represents evaporative demand of the atmosphere and wasigher during the summer months than the winter months.n the contrary, Ea values were very low in the summer season

depth and soil water storage in the profile along withrainfall in Adana in 1997.

and high in the spring months, depending on the rainfall pat-tern and presumably soil wetness. Drainage occurred duringrainy months, from November to May, with a peak in January.Results emphasize that D is a component, which should not beneglected when dealing with water conservation even on deepclay soils. There is a strong seasonal pattern of SWD. AverageSWD with dimension length varied between 88 mm month−1

in the summer season to about 2 mm month−1 in the winterperiod. These results may be promising in terms of preven-tion of water losses through evaporation and drainage frombare soils and adoption of an effective management strategyfor soil water, particularly, in rainfed-areas.

Annual quantities of water balance components for 13years are summarized in Table 1. Annual precipitationdenoted noticeable inter-annual variations. Potential evapora-

tion from bare soil varied between 850.6 and 909.8 mm. Actualsoil evaporation ranged from 248.0 to 392.9 mm with a meanannual value of 302.5 mm. As mentioned before, actual soilevaporation depends not only on the atmospheric evapora-

154 e c o l o g i c a l m o d e l l i n g

Fig. 4 – Monthly mean rainfall, potential (Ep) and actual (Ea)evaporation from bare soil, drainage (D) below 120 cmdepth and soil water deficit in the profile for a period of 13years as of 1994 in Adana.

Table 1 – Annual quantities of water balance components in Ad

Year Rainfall Potential soilevaporation (mm)

Actual soilevaporation (m

1994 650.3 884.6 281.11995 497.5 874.4 279.01996 612.1 906.2 317.71997 628.0 865.3 392.91998 766.3 873.8 321.01999 507.9 896.5 281.32000 700.9 850.6 248.02001 1036.4 865.0 279.62002 484.4 864.7 333.52003 705.6 882.9 278.92004 599.9 891.5 250.02005 524.7 894.1 322.42006 579.9 909.8 346.9

Average 638.0 881.5 302.5

Initial water amount at the beginning of 1994: 380.0 mm/120 cm. Equivalen

2 1 7 ( 2 0 0 8 ) 148–156

tive demand but also on rainfall pattern, and consequentlysoil wetness. Drainage varied substantially inter-annually(150.5–757.4 mm) depending on the intensity and frequency ofrainfall events and especially SWD from the preceding dry sea-son. Similar tendency for drainage was also reported by Eilerset al. (2007). In order to give an idea to the readers, at the end ofthe 6-month periods (late June and late December), water stor-age of the 120 cm deep soil profile is also provided in Table 1.Soil water storage at the end of years may not make sensebecause the study area has rainfall, generally, in the monthsof November and December. Consequently, water content of120 cm soil profile at the end of December, is close to fieldcapacity values. But giving total water content of the profile atthe beginning of dry summer season (late June) may be mean-ingful. This case can also be seen on examining the 13-yearaverage monthly rainfall and SWD shown in Fig. 4.

The advantage of this model quantifying the componentsof soil water balance is that actual soil evaporation, drainageand soil water storage are quantified in an interactive waysince these components are strongly interdependent. Anotheradvantage is that input parameters of the model are simpleand readily obtainable such as climatic data to compute thepotential soil evaporation, average diffusivity for the dryingsoil, and volumetric water content at field capacity. On theother hand, the model accounts for changes in soil matricpotential at the top surface layer, very important for seedgermination, and seedling establishment after sowing. In prin-ciple, the model is based on a single soil-water store. Thiskind of models requires readily available parameters, whereasmulti-store (multiple soil layers) models require many inputs,which are often not available, as also emphasized by Eilerset al. (2007). However, the modellers should not ignore runoffin sloping areas and in nearly flat terrain if the soil has

low infiltration rates or low permeability layers. Neverthe-less, the inclusion of runoff would be a simple modificationif so desired. For the present model, the portion of rain-fall not lost by runoff should be taken into consideration.

ana for a period of 13 years

m)Drainage (mm) Soil water storage

(mm/120 cm) at the end of

June December

329.2 327.1 420.0220.5 333.4 418.0293.1 352.0 419.3234.7 326.3 419.7445.0 353.9 420.0301.7 338.6 344.9377.8 357.7 420.0757.4 362.8 419.4150.5 308.7 419.8427.4 334.2 419.1351.3 317.7 417.7202.1 369.4 417.9251.3 318.3 399.6

334.0 – –

t depth of water at field capacity: 420.0 mm/120 cm.

g 2 1

Oesffwmmwtr

5

InaeTtottgE

A

Tt

r

Á

A

A

A

A

A

A

A

e c o l o g i c a l m o d e l l i n

n the other hand, possible preferential flow is not consid-red in the present model. In many soils and environments,uch processes are important, and their inclusion could use-ully represent a further development of the model. Apartrom micro-lysimeter-experiment conducted in this study,hat can be done is to compare model outputs to field-basedeasurement techniques although physical credibility of theodel is quite high. For this purpose, the weighing lysimetersith a drainage reservoir would be useful. Additionally, in fur-

her studies, sensitivity analysis should be carried out with aange of variables involved into calculations.

. Conclusions

n conclusion, the model is useful in quantifying compo-ents of soil water balance with a few parameters. The modeldequately represents the important physical processes tostimate actual evaporation and deep drainage from bare soils.he next logical step in model development would be estima-

ion of runoff, taking account of rainfall intensity, the slopef study area, and soil properties. In addition, there is a needo develop software for a fast and precise simulation sincehe volume of involved calculations is considerably high, thusiving rise to a versatile and functional implementation of the-DiGOR model.

cknowledgement

he author would like to thank Dr. F. Evrendilek for improvinghe English text.

e f e r e n c e s

cs, F., 2003. A comparative analysis of transpiration and baresoil evaporation. Boundary-Layer Meteorol. 109, 139–162.

llen, R.G., Smith, M., Perrier, A., Pereira, L.S., 1994. An update forthe definition of reference evapotranspiration. ICID Bull. 43(2), pp. 92.

lvenas, G., Jansson, P.-E., 1997. Model for evaporation, moistureand temperature of bare soil: calibration and sensitivityanalysis. Agric. For. Meteorol. 88, 47–56.

rmstrong, A.C., Matthews, A.M., Portwood, A.M., Addiscott, T.,Leeds-Harrison, P.B., 1993. Modelling the Effects of ClimateChange on the Hydrology and Water Quality of StructuredSoils, NATO Advanced Research Workshop, Silsoe, September20–24.

ydin, M., 1994. Hydraulic properties and water balance of a claysoil cropped with cotton. Irrig. Sci. 15, 17–23.

ydin, M., 1998. A new model for predicting evaporation frombare field soil. In: Proceedings of the International Symposiumand 2nd Chinese National Conference on RainwaterUtilization, Xuzhou-Jiangsu, China, pp. 283–287.

ydin, M., 2004. Comparison of two different models forestimating soil evaporation under the conditions of CukurovaRegion. In: Proceedings of the International Workshop for theResearch Project on the Impact of Climate Change on

Agricultural Production System in Arid Areas (ICCAP),Cappadocia, Turkey, pp. 140–143 (Publication No. 4).

ydin, M., Huwe, B., 1993. Test of a combined soil moisture/soilheat simulation model on a bare field soil in Southern Turkey.Z. Pflanzenern. Bodenk. 156, 441–446.

7 ( 2 0 0 8 ) 148–156 155

Aydin, M., Yang, S.L., Kurt, N., Yano, T., 2005. Test of a simplemodel for estimating evaporation from bare soils in differentenvironments. Ecol. Model. 182 (1), 91–105.

Aydin, M., Uygur, V., 2006. A model for estimating soil waterpotential of bare fields. In: Proceedings of the 18thInternational Soil Meeting (ISM) on Soils Sustaining Life onEarth, Managing Soil and Technology, Sanliurfa, pp. 477–480.

Aydin, M., Yano, T., Evrendilek, F., Uygur, V., 2008. Implications ofclimate change for evaporation from bare soils in aMediterranean environment. Environ. Monit. Assess. 140,123–130.

Baver, L.D., Gardner, W.H., Gardner, W.R., 1972. Soil Physics. JohnWiley and Sons, Inc., New York.

Beese, F., van der Ploeg, R.R., Richter, W., 1977. Test of a soil watermodel under field conditions. Soil Sci. Soc. Am. J. 41, 979–984.

Black, T.A., Gardner, W.R., Thurtell, G.W., 1969. The prediction ofevaporation, drainage and soil water storage for a bare soil.Soil Sci. Soc. Am. Proc. 33, 655–669.

Boast, C.W., 1986. Evaporation from bare soil measured with highspatial resolution. In: Klute, A (Ed.). Methods of Soil Analysis.Part 1. Physical and Mineralogical Methods. Agron. Monogr.No. 9, pp. 889–900.

Boast, C.W., Robertson, T.M., 1982. A microlysimeter method fordetermining evaporation from bare soil: description andlaboratory evaluation. Soil Sci. Soc. Am. J. 46, 689–696.

Brisson, N., Perrier, A., 1991. A semi-empirical model of bare soilevaporation for crop simulation models. Water Resour. Res. 27(5), 719–727.

Brown, R.W., Oosterhuis, D.M., 1992. Measuring plant and soilwater potentials with thermocouple psychrometers: someconcerns. Agron. J. 84, 78–86.

Campbell, G.S., 1985. Soil Physics with Basic: Transport Modelsfor Soil–Plant Systems. Elsevier, Amsterdam, Oxford.

Dinc, U., Senol, S., Kapur, S., Sari, M., 1991. Catenary soilrelationships in the Cukurova region, southern Turkey. Catena18, 185–196.

Ehlers, W., van der Ploeg, R.R., 1976. Evaporation, drainage andunsaturated hydraulic conductivity of tilled and untilledfallow soil. Z. Pflanzenern. Bodenk. 139, 373–386.

Eilers, V.H.M., Carter, R.C., Rushton, K.R., 2007. A single layer soilwater balance model for estimating deep drainage (potentialrecharge): an application to cropped land in semi-aridNorth-east Nigeria. Geoderma 140, 119–131.

Evett, S.R., Warrick, A.W., Matthias, A.D., 1995. Wall material andcapping effects on microlysimeter temperatures andevaporation. Soil Sci. Soc. Am. J. 59 (2), 329–336.

Gardner, W.R., 1959. Solutions of the flow equation for the dryingof soils and other porous media. Soil Sci. Soc. Am. Proc. 23,183–187.

Gardner, W.R., Hillel, D., 1962. The relation of externalevaporative conditions to the drying of soils. J. Geophys. Res.67, 4319–4325.

Hanks, R.J., 1992. Applied Soil Physics: Soil Water andTemperature Applications, 2nd ed. Springer-Verlag, New York,Berlin.

Hanks, R.J., Gardner, H.R., 1965. Influence of differentdiffusivity–water content relations on evaporation of waterfrom soils. Soil Sci. Soc. Am. Proc. 29, 495–498.

Hasegawa, S., Sato, T., 1987. Water uptake by roots in cracks andwater movement in clayey subsoil. Soil Sci. 143, 381–386.

Hillel, D., 1975. Evaporation from bare soil under steady anddiurnally fluctuating evaporativity. Soil Sci. 120, 230–237.

Hillel, D., 1998. Environmental Soil Physics. Academic Press,Amsterdam, New York, p. 771.

Huwe, B., van der Ploeg, R.R., 1990. Modelle zur Simulation desStickstoffhaushaltes von Standorten mit unterschiedlicherlandwirtschaftlicher Nutzung. Eigenv. Inst Wasserbau Univ.Stuttgart. Heft, 69.

i n g

156 e c o l o g i c a l m o d e l l

Jackson, R.D., Idso, S.B., Reginato, R.J., 1976. Calculation ofevaporation rates during the transition from energy limitingto soil phases using albedo data. Water Resour. Res. 12, 23–26.

Katul, G.G., Parlange, M.B., 1992. Estimation of bare soilevaporation using skin temperature measurements. J. Hydrol.132, 91–106.

Kirby, J.M., Ringrose-Voase, A.J., 2000. Drying of some Philippineand Indonesian puddle rice soils following surface drainage:numerical analysis using a swelling soil flow model. Soil Till.Res. 57, 13–30.

Kite, G., Droogers, P., Murray-Rust, H., Voogt de, K., 2001.Modeling scenarios for water allocation in the Gediz Basin,Turkey. Research Report 50. International Water ManagementInstitute, Colombo, Sri Lanka.

Malik, R.S., Anlauf, R., Richter, J., 1992. A simple model forpredicting evaporation from bare soil. Z. Pflanzenern. Bodenk.155, 293–299.

Mihailovic, D.T., Rajkovic, B., Dekic, L., Pielke, R.A., Lee, T.J., Ye, Z.,1995. The validation of various schemes for parameterizingevaporation from bare soil for use in meteorological models: anumerical study using in situ data. Boundary-Layer Meteorol.76, 259–289.

Moret, D., Braud, I., Arrue, J.L., 2007. Water balance simulation ofa dryland soil during fallow under conventional andconservation tillage in semiarid Aragon Northeast Spain. SoilTill. Res. 92, 251–263.

Mwendera, E.J., Feyen, J., 1997. Tillage and evaporativity effects on

the drying characteristics of a silty loam: evaporationprediction models. Soil Till. Res. 41, 127–140.

Ritchie, J.T., 1972. Model for predicting evaporation from a rowcrop with incomplete cover. Water Resour. Res. 8,1204–1213.

2 1 7 ( 2 0 0 8 ) 148–156

Ritchie, J.T., Kissel, D.E., Burnett, E., 1972. Water movement inundisturbed swelling clay soil. Soil Sci. Soc. Am. Proc. 36,874–879.

Van Dam, J.C., Huygen, J., Wesseling, J.G., Feddes, R.A., Kabat, P.,van Walsum, P.E.V., Groendijk, P., van Diepen, C.A., 1997.Theory of SWAP version 2.0. Simulation of water flow, solutetransport and plant growth in thesoil–water–atmosphere–plant environment. TechnicalDocument 45, DLO Winand Staring Centre, Report 71,Department of Water Resources, Agricultural University,Wageningen. 167 pp.

Van Keulen, H., Hillel, D., 1974. A simulation study of the dryingfront phenomenon. Soil Sci. 118, 270–273.

Varado, N., Braud, I., Ross, P.J., Haverkamp, R., 2006. Assessmentof an efficient numerical solution of the 1D Richards’ equationon bare soil. J. Hydrol. 323, 244–257.

Wallace, J.S., Jackson, N.A., Ong, C.K., 1999. Modelling soilevaporation in an agroforestry system in Kenya. Agric. For.Meteorol. 94, 189–202.

Yang, M., Yanful, E.K., 2002. Water balance during evaporationand drainage in cover soils under different water tableconditions. Adv. Environ. Res. 6, 505–521.

Yang, S.L., Yano, T., Aydin, M., Kitamura, Y., Takeuchi, S., 2002.Short term effects of saline irrigation on evapotranspirationfrom lysimeter-grown citrus trees. Agric. Water Manage. 56(2), 131–141.

Yano, T., Aydin, M., Haraguchi, T., 2007. Impact of climate change

on irrigation demand and crop growth in a Mediterraneanenvironment of Turkey. Sensors 7 (10), 2297–2315.

Zhu, J., Mohanty, B.P., 2002. Upscaling of soil hydraulic propertiesfor steady state evaporation and infiltration. Water Resour.Res. 38 (9), 1–13.