Iran J Sci Technol Trans Civ Eng DOI 10.1007/s40996-016-0036-x

RESEARCH PAPER

Evapotranspiration Estimation Using Remote Sensing Technology Based on SEBAL Algorithm Mehrdad Abrishamkar1 • Azadeh Ahmadi1

Received: 14 November 2015 / Accepted: 15 May 2016 Ó Shiraz University 2016

Abstract The aim of this study is to obtain the actual evapotranspiration (ETa) values using Surface Energy Balance Algorithm for Land (SEBAL) algorithm, during the year 2008 in Nekooabad agricultural and drainage network and Ben-Saman region located in Gavkhouni basin in Iran. For this purpose, the time series of Moderate Resolution Imaging Spectroradiometer images from Terra satellite are prepared and ETa values of the regions are estimated for this year. In order to evaluate the SEBAL algorithm performance, the algorithm was executed on Nekooabad agricultural network located in Najafabad region with enough land information and three different methods were used to verify the results. (1) Epan-method: The correlation coefficient between obtained ETa by SEBAL algorithm and the reference ET values (ET0) from pan evaporation method is about 86.72 %. (2) Penman– Monteith method: This method is used for result comparison (the correlation coefficient is about 90.47 %). (3) Water balance method: Considering the water withdrawal of the wells, precipitation and inflow and outflow to the region to the water balance equation, the ETa is estimated as the unknown variable. The results are compared with the result of SEBAL algorithm that shows an error of 7.21 %. SEBAL measured the ETa of Ben-Saman region about 228.16 mm, while there is not enough information of land change in recent years about this region. The results indicate that more cooperation is needed for implementing the special plans for water resources allocation to prevent irregular withdrawals. & Azadeh Ahmadi [email protected] 1

Department of Civil Engineering, Isfahan University of Technology, Isfahan, Iran

Keywords Actual evapotranspiration  Remote sensing  SEBAL  MODIS  Penman–Monteith method

1 Introduction Water is increasingly becoming a scarce natural resource. The population increase puts great demand on water resources for urban, industrial and agricultural use. Unfortunately, water shortage has now turned to be the biggest threat for people in many parts of the world. The main cause of this problem is long-term imbalances between availability of water resources and water demands of different stake-holders in addition to climate change phenomena. Because water resources have become scarcer compared to demand, those countries that already plan allocation of water in an organized manner through the administration of water-right systems have become increasingly aware that water rights defined in terms of ‘‘rights to divert’’ are less useful than water rights defined as ‘‘rights to consume.’’ Although the two are intrinsically related, managing consumption seems to take priority for managing allocation (Bastiaanssen et al. 2005). Spatial information on consumption of water for large areas and particularly within irrigated agricultural areas is important for decision makers to assess where scarce water resources are wasted and where in a given region the water productivity can be improved (Thoreson et al. 2005; Zwart and Bastiaanssen 2007). Since water resource management strategies are usually implemented on a basin scale, understanding of the hydrological processes at this scale is a prerequisite for the formulation of these strategies. As pressure on water resources increases, sufficient knowledge on where, when and how much water is used is required. Evapotranspiration (ET) in this case becomes an important

123

Iran J Sci Technol Trans Civ Eng

factor for assessing water productivity and monitoring of irrigation performance (Mutiga et al. 2009). ET including water evaporation from soil surfaces and vegetation transpiration represents a fundamental process of hydrological cycle and is essential to both water and energy balance. Reference evapotranspiration (ET0) is a representation of the environmental demand for evapotranspiration and represents the evapotranspiration rate of a short green crop (grass), completely shading the ground, of uniform height and with adequate water status in the soil profile. Potential evapotranspiration (ETp) is a measure of the ability of the atmosphere to remove water from the surface through the processes of evaporation and transpiration assuming no control on water supply. Actual evapotranspiration (ETa) is the quantity of water that is actually removed from a surface due to the processes of evaporation and transpiration. There is a distinction made between ETp and ETa. ETp is the demand or maximum amount of water that would be evapotranspired if enough water was. ETa is how much water actually is evapotranspired and is limited by the amount of water that is available. ETa is always less than or equal to ETp. Accurate quantitative monitoring of the temporal and spatial distribution of ET is needed for improving our understanding of energy and hydrological partitioning between the land surface and atmosphere (Boegh et al. 2002; Cleugh et al. 2007; Jang et al. 2009). Traditionally, ETa is computed as a residual in water balance equations, from estimates of ETp or, indirectly, from field measurements at meteorological stations. Additionally, ET is subject to rapid changes in time and space due to the wide spatial variability of precipitation, hydraulic characteristics of soils, and vegetation types and densities. Therefore, it is nearly impossible to determine its spatial and temporal distributions for large areas from ground measurements alone. Remote sensing presents an indirect ET measurement technique; it involves using a set of equations in a strict hierarchical sequence to convert the spectral radiances measured by satellites or airplanes into estimates of ETa. The advent of the possibility to indirectly measure fundamental ET processes from satellites has radically changed our abilities in the realm of water resource management (Bastiaanssen et al. 2005). Remote sensing data provided by satellites are a means of obtaining consistent and frequent observation of spectral reflectance of the land surface radiation on micro- to macroscale (Bastiaanssena et al. 1998a). Numerous evapotranspiration algorithms have been developed making use of remote sensing data acquired from sensors on airborne and satellite platforms (Papadavid et al. 2013; Moazed et al. 2014; Karatepe and Ikiel 2013). SEBAL is a spatial ETa estimation method based on energy balance and a satellite remote sensing technique, which differentiates between sensible heat flux and latent heat of vaporization flux. SEBAL model was developed by Bastiaanssen et al.

123

(1998a) to map the spatial variation of ET across a range of land uses (Bhattarai et al. 2011). Satellite data are ideally suited for deriving spatially continuous fields of ET using energy balance techniques (Allen et al. 2007). It is worth clarifying here that the use of different satellite images does not create any problem, because a radiometric calibration is performed and the respective calibration factors are available with each satellite image; therefore, different satellite images provide the same remotely sensed data in terms of reflectance (Papadavid et al. 2013). But the availability of standard MODIS data products (surface temperature, surface reflectance) from the EOS Data Gateway at no cost makes them a viable alternative for operational estimation of evapotranspiration (Patel et al. 2006). MODIS images have been archived since 1999 and are freely available through the National Aeronautics Space Administration (NASA) and have a daily coverage with pixel resolution of 1 km, with some bands having a resolution of 250 and 500 m as well. Such a pixel resolution is well suited for monitoring fluxes at regional scale spanning hundreds of square kilometers and is thus considered appropriate for meso- and river basin scales as was in this study (Kongo et al. 2011). The SEBAL method has been validated under various land uses in several countries under different climatic conditions including Spain, Italy, Turkey, Pakistan, India, Sri Lanka, Egypt, Niger, China and the USA for estimation of ET (Sun et al. 2011). In order to assess SEBAL method results, they could be compared to results obtained through lysimeter, empirical equations, water balance equation or evaporation pan. If lysimeter data are unavailable, it might be helpful to utilize empirical equation, water balance method and evaporation pan. Finding a certain coefficient as the pan coefficient (kPan) that displays the real condition of the region is the first step in determining the ET0. A simple method for estimating ET0 is using the pan evaporation; ET0 is thus resulted from multiplying pan evaporation by the pan’s coefficient. Changes in the pan’s coefficient are due to the pan’s specific location and climate conditions specified by Doorenbos and Pruitt (1977) and Allen et al. (1998). Snyder et al. (2005) compared the evapotranspiration rate of the reference plant to that estimated through the pan evaporation in which the pan’s coefficient was calculated using Allen and Pruitt, Cuenca, Snyder and Orang methods. Noshadi and Sepaskhah (2005) compared the evaluation of three geostatistical interpolation methods including ordinary kriging, residual kriging and cokriging for the interpolation of long-term monthly and yearly reference crop evapotranspiration. Akbari et al. (2007) assessed performance of irrigation systems in the Zayandehroud Basin, Esfahan, Iran, using NOAA satellite images, the SEBAL algorithm to obtain ETa and ETp and biomass production. The advantages of the

Iran J Sci Technol Trans Civ Eng

methodology over the traditional assessments are addressed: (1) Most data are readily available, (2) all water users are included, (3) groundwater extraction can be estimated, and (4) a real time assessment can be set up using this approach. Noodehi (2010) used four methods, namely Allen and Pruitt, Cuenca, Snyder and Orang methods in order to measure the pan’s coefficient. The comparison of the obtained evapotranspiration with lysimeter evapotranspiration indicates the suitability of the Orang method. Hong et al. (2011) applied various downscaling methods to combine MODIS and Landsat imagery in order to obtain both high temporal and spatial resolutions. The SEBAL was used to derive daily ET distributions from Landsat 7 and MODIS images. Two downscaling procedures were evaluated: input downscaling and output downscaling utilizing subtraction and regression method. They found that all of the four proposed downscaling methodologies can generate reasonable spatial patterns of the disaggregated ET map. Bhattarai et al. (2011) processed sixteen Landsat 5 Thematic Mapper (TM) images from April 2000 to September 2006 and compared the results with energybudget eddy covariance (EBEC) ET estimates from four US Geological Survey (USGS) stations. The results support the applicability of the modified SEBAL model in the rapidly growing southeastern USA as a tool for estimating consumptive water use via remotely sensed methods. The purpose of the present study is to estimate the ETa of Ben-Saman region due to its logistic location being the first major location downstream Zayandehruod dam and the existence of ongoing conflicts between Isfahan and Chaharmahal–Bakhtiari provinces on water consumption allocations. The SEBAL algorithm was employed because of its accuracy and economic viability compared to the other methods. To carry out this method, MODIS image was used because of its daily repetition period and the fact that it was free. While its sensor resolution is relatively lower compared to LANDSAT and ASTER SENSORS, it was nonetheless put into use because of easy accessibility. In order to assess the results of SEBAL method, the algorithm was assessed in Nekooabad region due to the easy accessibility to the required data. For this purpose, Penman–Monteith method, Water balance method and evaporation pan method were used. Finally, the SEBAL algorithm was carried out in BenSaman region which was the main purpose of the study and the results were discussed in order to solve the problems.

2 Materials and Methods 2.1 MODIS Data MODIS sensor on TERRA platform has a spectral resolution featuring 36 bands ranging from 2.50- to 1000-m

spatial resolution (Hafeez et al. 2002). In this study, 15 Archived MODIS Level 1B data with 36 channels of visible and near-infrared reflectance and radiance, as well as thermal infrared radiance, were used. Cloud-free images covering the study area for 2008 were downloaded from LANCE-MODIS data system operated by the Goddard Space Flight Center (GSFC) Terrestrial Information Systems Branch. LANCE-MODIS provides certain Aqua and Terra MODIS L0, L1, L2, L2G and L3 products on a nearreal-time basis. One such system is the MODIS Terra and Aqua satellite system (Mutiga et al. 2009; Allen et al. 2007; Hafeez et al. 2002; Member of MOCST 2009). The MODIS Terra satellite, activated in 2000 by NASA, is a large format polar orbiting satellite that acquires shortwave and thermal images for all parts of the earth’s surface at least once per day. MODIS Aqua was activated in 2002. Although the temporal frequency of MODIS is much vaunted by some, it is seriously compromised by large view angles on many days, causing blurring, and the spatial resolution of MODIS is insufficient to provide the same level of information as Landsat. MODIS images on at least two of every 3 days for a specific location are acquired from such a large, lateral distance (i.e., from a large view angle), where the pixel size for short-wave information stretches from the specified size of 0.5 9 0.5 km to an effective 1 km 9 1.5 km size and thermal pixels stretch from the specified size of 1 9 1 km to an effective 2 km 9 3 km size (Allen et al. 2007). The purpose of the Level 1B software system, developed by the MODIS Characterization and Support Team (MCST), is to provide calibrated MODIS data to the MODIS Science Team (MST) for many applications in the area of Earth science. One such application is the construction of images from Level 1B output products. The Level 1B products do not directly contain images. Rather, they contain the calibrated data used by other software applications to construct the images. The images were processed to provide the necessary data (emissivity, vegetation index, surface albedo and land surface temperature), required for the estimation of an ETa using SEBAL approach (Mutiga et al. 2009). 2.2 SEBAL Algorithm The ET is governed by the energy and heat exchanges at the land surface. SEBAL (Bastiaanssena et al. 1998a) was used to estimate ETa and ETp. SEBAL algorithm requires spatially distributed visible, near-infrared and thermal infrared input data, from satellite imageries (Bastiaanssena et al. 1998a). SEBAL parameterization is an iterative and feedback-based semiempirical procedure, which deduces the radiation, heat and evaporative fluxes. The algorithm computes the most essential hydro-meteorological

123

Iran J Sci Technol Trans Civ Eng

parameters and requires little field information (only incoming solar radiation, air temperature and wind speed data are required) (Bastiaanssena et al. 1998a, b; Bastiaanssen 2000). The theoretical and computational basis of SEBAL is described in Bastiaanssen et al. (1998a, 2005) and Bastiaanssen (2000). SEBAL is an image-processing model comprised of 25 computational steps resulting in ETa and ETp as well as other energy exchanges between land and atmosphere (Thoreson et al. 2005). The SEBAL is a relatively simple parameterization of the energy balance and surface fluxes based on spectral satellite measurements. The energy-budget equation is expressed by Eq. 1: kET ¼ Rn  G  H

ð1Þ 2

where kET is the latent heat flux (W/m ), Rn is the net radiation flux at the surface (W/m2), G is the soil heat flux (W/m2), and H is the sensible heat flux to the air (W/m2). 2.2.1 Net Radiation Flux (Rn) Rn represents the actual radiant energy available at the surface. It is computed by subtracting all outgoing radiant fluxes from all incoming radiant fluxes. This is given in the surface radiation balance equation (Li et al. 2013): Rn ¼ ð1  aÞRs# þ RL#  RL"  ð1  e0 ÞRL#

ð2Þ

where Rs; is the incoming shortwave radiation (W/m2), a is the surface albedo (dimensionless), RL; is the incoming long-wave radiation (W/m2), RL: is the outgoing longwave radiation (W/m2), and e0 is the surface thermal emissivity (dimensionless). 2.2.2 Soil Heat Flux (G) Soil heat flux is the rate of heat storage into the soil and vegetation due to conduction. SEBAL first computes the ratio G/Rn using the following empirical equation developed by Bastiaanssen (2000) representing values near midday:  G Ts  ¼  0:0032  a þ 0:0062  a2 Rn a  ð3Þ  1  0:978  NDVI4 where Ts is the surface temperature (°C) and NDVI is the Normalized Difference Vegetation Index (Li et al. 2013). 2.2.3 Normalized Difference Vegetation Index (NDVI) The NDVI is the ratio of the differences in reflectivity for the near-infrared band (qNIR) and the red band (qR) to their sum: q  qR NDVI ¼ NIR ð4Þ qNIR þ qR

123

where qNIR and qR are reflectivities for bands 2 and 1. The NDVI is a sensitive indicator of the amount and condition of green vegetation. Values for NDVI range between -1 and ?1. Green surfaces have a NDVI between 0 and 1 and water and cloud are usually less than zero. 2.2.4 Sensible Heat Flux (H) Sensible heat flux is the rate of heat loss to the air by convection and conduction, due to a temperature difference. It is computed using the following equation for heat transport: H¼

qair  Cp  dT rah

ð5Þ

where qair is air density (kg/m3), Cp is air specific heat (1004 J/kg/K), dT (K) is the temperature difference (T1 - T2) between two heights (z1 and z2), and rah is the aerodynamic resistance to heat transport (s/m) (Li et al. 2013). 2.2.5 Instantaneous ET An instantaneous value of ET in equivalent evaporation depth is computed as: ETinst ¼ 3600

kET k

ð6Þ

where ETinst is the instantaneous ET (mm/h), 3600 is the time conversion from seconds to hours, and k is the latent heat of vaporization or the heat absorbed when a kilogram of water evaporates (J/kg) (Li et al. 2013). 2.2.6 Annual ET In order to calculate the annual evapotranspiration, first the actual daily evapotranspiration rate of satellite models is calculated using the following equation: ETa24 ¼ ET0 F  ET024

ð7Þ

where ET0–24 is the 24-hour evapotranspiration reference value (mm/day), ET0F the reference plant evapotranspiration value, and ETa–24 the actual daily evapotranspiration (mm/day). In order to generalize the results of the SEBAL model in any given day with an existing image (ETa–24) to the period covered in each image (ETa–period) and specifying the actual evapotranspiration rate of all days in any given water year (ETa–seasonal), the following equations were employed: ðETaperiod Þi ¼

k ðETa24 Þi X ðET0 Þj ðET024 Þi j

ð8Þ

Iran J Sci Technol Trans Civ Eng

ETaseasonal ¼

n  X

ETaperiod



ð9Þ

i

i¼1

where ET0–24 is the daily reference evapotranspiration value for the different days of the period (from j to k) and ETa–24 is the actual calculated evapotranspiration value for the period midday (day i). As such, ETa maps for different time periods (ETa–period) are obtained and the sum of which (ETa–seasonal) indicates the ETa of the basin during the water year (Li et al. 2013). 2.3 Validation Methods 2.3.1 Pan Evaporation The evaporation rate from pans filled with water is easily obtained. In the absence of rain, the amount of water evaporated during a period (mm/day) corresponds with the decrease in water depth in that period. Pans provide a measurement of the integrated effect of radiation, wind, temperature and humidity on the evaporation from an open water surface. Notwithstanding the difference between pan evaporation and the evapotranspiration of cropped surfaces, the use of pans to predict ET0 for periods of 10 days or longer may be warranted. The pan evaporation is related to the reference evapotranspiration by an empirically derived pan coefficient: ET0 ¼ kpan  EPan

ð10Þ

where ET0 is reference evapotranspiration (mm/day), kpan is pan coefficient, and EPan is pan evaporation (mm/day) (Li et al. 2013).

other crops can be related. The FAO Penman–Monteith method is selected as the method by which the evapotranspiration of this reference surface (ET0) can be unambiguously determined, and as the method which provides consistent ET0 values in all regions and climates. ET0 ¼

0:408DðRn  GÞ þ c T þ900273 u2 ðes  ea Þ D þ cð1 þ 0:34u2 Þ

ð12Þ

where T is mean daily air temperature at 2 m height (°C), u2 is wind speed at 2 m height (m/s), es is saturation vapor pressure (kPa), ea is actual vapor pressure (kPa), es - ea is saturation vapor pressure deficit (kPa), D is slope vapor pressure curve (kPa/°C), and c is psychrometric constant (kPa/°C) (Li et al. 2013). 2.3.4 Water Balance Equation Estimation of water balance component would help to understand and manage the catchment’s water properly. It also addresses current and future water use scenarios. A water balance model forms an important tool for analysis of hydrological behavior of a catchment. Also, it is used to assess the system sensitivity to natural or imposed impacts such as climate and land use change (Dastorani and Poormohammadi 2012). The common form of water balance equation would be as follows: P ¼ ETaseasonal þ ðROut  RIn Þ  W

ð13Þ

where P is mean annual precipitation (m3/s), ROut is output runoff (m3/s), Rin is input runoff (m3/s), and W is extraction of wells (m3/s).

2.3.2 Orang Method

3 Study Area

Orang (1998) presented the following equation to estimate Kpan:

The Zayandehruod basin located in the central part of Iran covers an area of 26,917 square kilometers between latitudes of ETa–seasonal to 52°450 and longitudes of 31°150 to 33°450 north. It encompasses a major part of Gavkhouni wetland closed basin. This basin is itself part of the larger central wetland basin in Iran. The basin is a major source of water tension issues in the area because of its unique complexities, differences in climate, differences in water priorities and water rights. Existence of major water-consuming industries, water transfers, urban water extractions and irrigation networks puts a stress on the Zayandehruod which is the only permanent river in the region.

KPan ¼ 0:51206  0:000321  u2 þ 0:002889  RH þ 0:031886  lnðFÞ  0:000107  RH  lnðFÞ ð11Þ where RH is relative humidity (%), u2 is wind speed at 2 m above ground surface (km/day), and F is fetch or distance of the identified surface type (Orang 1998). 2.3.3 Penman–Monteith Equation The FAO Penman–Monteith equation determines the evapotranspiration from the hypothetical grass reference surface and provides a standard to which evapotranspiration in different periods of the year or in other regions can be compared and to which the evapotranspiration from

3.1 Nekooabad Irrigation and Drainage Network Considering the classification conducted by Agricultural and Economic Planning Research Center, Najafabad is one

123

Iran J Sci Technol Trans Civ Eng

of the most important sub-basins of the Zayandehroud basin due to its water resources quality and agriculture. Najafabad study area is located between 50°570 and 51°440 east longitude and 32°200 to 32°430 north latitude. Nekooabad water transfer network which encompasses a main canal with a length of 95.4 km transfers water from Nekooabad diversion dam to Nekooabad region. Generally speaking, the construction of Nekooabad irrigation system has led to the improvement of 30,000 ha and the development of 35,000 more hectares of cultivated land. Due to the accessibility of data such as the rate of pan evaporation, the amount of water input and output in the region, uniformity in the region’s cultivation and vegetation, and data on the region’s well water withdrawals, the area was selected as the suitable area for assessing the results and outputs of SEBAL algorithm. 3.2 Ben-Saman Region Considering groundwater studies, the Gavkhouni basin is divided into 21 units. Ben-Saman study area is located in the central region of Chaharmahal-Bakhtiari watershed basin. The region covers a total area of 817 square kilometers with heights covering 626.8 square kilometers and the remaining 190.2 square kilometers consisting of plains. The highest point in the area is located in the western

elevations with 2917 m height and the lowest in the eastern plain being 1781 m high. Figure 1 shows the location of Ben-Saman sub-basin in relation to the Gavkhouni basin and the neighboring provinces. A large part of the area is located in Chaharmahal-Bakhtiari province. Water released from the Zayandehruod dam initially enters this part of Chaharmahal-Bakhtiari province before entering the Isfahan province. During recent years, there have been some disagreements regarding water consumption in the region, mainly due to the inaccessibility to adequate data on cultivation. Ground data used in SEBAL algorithm were provided using the data sets collected by Najafabad synoptic meteorological station located at 32°360 north latitude and 51°230 east longitude and Saman synoptic meteorological station located at 32°270 north latitude and 50°560 east longitude. This study made use of MODIS sensor imagery stationed on Terra satellite including Level 1B products. Level 1B software system was mainly developed by the MCST team for the goal of providing calibrated MODIS data in order to carry out administrative tasks (Member of MOCST 2009). The images were provided from the MODIS internet site (http://lancemodis.eosdis.nasa.gov/ cgibin/imagery/realtime.cgi). One determining factor in selecting these images was cloud-free condition. Also, image timing was as such to

Fig. 1 Location of Ben-Saman sub-basin and Nekooabad network in relation to neighboring provinces and Gavkhouni basin

123

Iran J Sci Technol Trans Civ Eng

Fig. 2 a Albedo, b soil heat flux, c sensible heat flux d NDVI, e net radiation flux and f surface temperature maps resulted from the SEBAL algorithm

123

Iran J Sci Technol Trans Civ Eng Table 1 Numerical data due to one pixel and a certain date achieved from the SEBAL Coordinate of the pixel Lat

51.49305175E

Lon

32.52502549N

Sample

85

Line

73

Date and time of image Date

2008.09.30

Julian Date

274

Time

07:00 UTC

Parameters

Values

kET

384.806872

e0 Ts

0.954139 302.114685

Rs;

788.546224

RL:

450.675861

Rn

497.655472 350.588806

NDVI

0.344428

H

38.918974

G

73.929627

ETa–period

98.613269

ETa–24

5.633426

ETa

0.568914

dT

1.159085

a

0.221556

Reference Evapotranspiration Obtained from Pan evaporation Method- mm

RL;

300 y = 1.5253x R² = 0.7521

250 200 150 100 50 0 0

25

50

75

100

125

150

175

Actual Evapotranspiration- mm Fig. 3 Scatter diagram of evapotranspiration values obtained from the SEBAL and pan evaporation method

cover a whole water year. Therefore, 15 images were selected from the area where the satellite had previously made its course above Iran. Digital Elevation Model (DEM) maps, slope, aspect, latitude and shape files of the area were among other data used in implementing SEBAL algorithm. A major part of the model coding process,

123

implementation, analysis, comparison, image cropping and information processing were incorporated to ArcMap 9.3, Erdas Imagine 9.1, ENVI 4.2 and Excel.

4 Results As it was pointed out in the introduction to the SEBAL model, the estimation of ETa levels is one of the expected outputs of SEBAL algorithm which could be presented daily and on a millimeter basis. One of the most important advantages of the SEBAL algorithm is its ability in calculating the evapotranspiration rates continuously, if satellite imagery is provided daily. In addition to evapotranspiration, the SEBAL algorithm produces other outputs such as surface temperature, NDVI, albedo, pure solar radiation, soil temperature flux and perceptible heat flux, all emphasizing its strength. A sample of these outputs could be observed in Fig. 2. In Table 1, the numerical data of one pixel at a certain date achieved from the SEBAL (running) are presented 4.1 Results Evaluation Since the aforementioned tools could not be accessed in the studied region, a method for comparing the results with values resulted from the evapotranspiration from the evaporation pan using the Orang equation was employed in order to analyze the results of actual annual evapotranspiration calculations. As demonstrated in Fig. 3, there is a correlation value of 86.72 (r2 = 75.21) percent between the ETa resulted from the SEBAL algorithm and the ET0 obtained from the pan evaporation method. In Table 2, the ETa is obtained using the SEBAL algorithm and pan evaporation method in various time periods (Gregorian calendar/dates). In order to assess the results of actual annual evapotranspiration calculations, a method for comparing the results with values obtained from the Penman–Monteith method was employed. This method shows a determination coefficient value of 90.47 (r2 = 81.86) percent between the ETa resulted from the SEBAL algorithm and the ET0 obtained from the Penman–Monteith method. In Table 2, the obtained evapotranspiration values from the SEBAL algorithm and the ET per time periods (Gregorian calendar) are specified. In Fig. 4, one could observe the dispersion of the ETa obtained from the SEBAL model and ET0 using the Penman–Monteith method. As shown in Fig. 5, the comparison between the ET0 obtained through pan evaporation method and Penman– Monteith method and ETa obtained from the SEBAL algorithm reveals a similar changing pattern between them. While the pan evaporation and Penman–Monteith methods

Iran J Sci Technol Trans Civ Eng Table 2 Evapotranspiration values obtained using the SEBAL algorithm and pan evaporation method and Penman–Monteith method ET0—Penman–Monteith method (mm)

Kpan

Epan(mm)

23.9–10.10

100.4

0.73

125.4

91.9

72.8

1.9–22.9

138.9

0.75

171.7

130.0

78.9

11.8–31.8

141.5

0.72

181.5

131.6

76.8

20.7–10.8

163.2

0.73

212.6

157.1

91.3

26.6–19.7

191.6

0.76

263.2

200.3

117.5

2.6–25.6

193.5

0.72

251.2

180.7

111.2

26.4–1.6 26.3–25.4

288.9 182.6

0.75 0.73

330.9 216.4

249.6 159.1

159.6 147.9

Reference Evapotranspiration Obtained from Penman-Moneith Method- mm

Calculation period Gregorian calender (2008)

-25

ETa (mm)

(Dastorani and Poormohammadi 2012) and the evapotranspiration was calculated as the unknown in the following equation:

300 y = 1.655x R² = 0.8407

250

ET0—Pan evaporation method (mm)

200 150

ET ¼ P þ W þ RIn  ROut ET ¼ 36:42 þ 705:4 þ 422:58  278:15 ¼ 886:25 ðMCMÞ

100

ð14Þ The calculated evapotranspiration value was 886.25 million cubic meters (MCM) which is acceptable once compared with the value obtained from the SEBAL algorithm. The difference between the evapotranspiration values calculated using the SEBAL algorithm and the water balance method was thus divided by the evapotranspiration value obtained through the water balance method. The result was then multiplied by 100 to calculate the percentage of relative error. The value of relative error 7.21 is deemed acceptable.

50 0 25

75

125

175

Actual Evapotranspiration- mm Fig. 4 Scatter diagram of evapotranspiration values obtained from the SEBAL and Penman–Monteith method

calculate the ET0, the SEBAL method estimates the ETa thus resulting in differences among these two groups of values. 4.2 Water Balance Method

4.3 The ETa in the Ben-Saman Region

In order to assess the rate of evapotranspiration, the water balance method is used. In this method, the values of ground (well) water withdrawals, precipitation volume and input and output runoffs were inserted in Eq. (13)

Using SEBAL algorithm, the ETa in the Ben-Saman region was obtained in a 15-day span in 2008 where the satellite images were clean, that is, without any clouds. Once the middle layers were obtained similar to the study conducted

350

ET0 resulted from the Penman- Monteith method

300

ET(mm/period)

Fig. 5 Comparing the results of ETa obtained from the SEBAL method and the ET0 resulted from the pan evaporation and Penman–Monteith methods

ET0 resulted from the Pan evaporation method

250

ETa obtained from the SEBAL method

200 150 100 50 0

0

1

2

3

4

5

6

7

8

9

Period

123

Iran J Sci Technol Trans Civ Eng

Fig. 6 Annual actual evapotranspiration map of Ben-Saman region (millimeter) plus DEM map

Fig. 7 Zoning map of annual actual evapotranspiration rates of Ben-Saman map plus DEM map

123

Iran J Sci Technol Trans Civ Eng

in NekooAbad, a number of fifteen maps showing daily evapotranspiration were generated. The ENVI 4.2 software was then employed to obtain the annual evapotranspiration map of the basin through bundling all the daily maps. The calculated mean value of annual evapotranspiration, the maximum and the minimum values were 228 mm (186 million cubic meters), 2039 and 0 mm, respectively. As demonstrated by the evapotranspiration and DEM maps in Fig. 6, in the west region, one could observe that while the DEM map of the region includes high-altitude points, the evapotranspiration rate is quite significant which in turn signifies the fact that farming lands have expanded to include higher grounds and thus in order to irrigate these farms, a considerable amount of water is pumped to these areas. A similar pattern is also observed on the right bank of the river. The maximum evapotranspiration rate takes place on the southern field of the region. Within this region, as observed in the altitude map, lack of mountainous areas results in higher evapotranspiration rates and the abundance of farming lands leads to water consumption rates far greater than the other areas. Other high-consuming regions, i.e., regions with higher levels of evapotranspiration, are those located on the Zayandehruod river banks which have all necessary prerequisites for agriculture such as sufficient arable land and water. The zones, mapped according to the actual evapotranspiration, are presented in Fig. 7. The ranges are expanded from 0 to 500 mm as low, 500–1000 mm as average, 1000–1500 mm as high and 1500–2040 mm as very high.

5 Conclusion Data obtained through remote sensing could be employed in the estimation of ETa of various regions since it is easily and quickly accessible, it is cheap, up-to-date, has high spatial coverage and there is no need to quantify other complex hydrological products. In this study, after obtaining energy balance parameters, ETa was calculated on a daily basis and finally after calculating the seasonal evapotranspiration values, the annual values were obtained. The output of this model is one of the important features of this model which could be expanded over plains, regions, basins and even larger areas and is not limited to specific points. However, the use of this model’s outputs in estimating ETa in the region needs more accurate check points. In order to assess the performance of SEBAL algorithm, the algorithm was executed in NekooAbad region in NajafAbad basin and three different methods were employed in order to verify the results. Using pan evaporation (pan evaporation method): indicating a correlation of 86.72 (r2 = 75.21) percent between ETa obtained through the SEBAL algorithm and the ET0 through the pan

evaporation method. Using Penman–Monteith method: indicating a correlation of 90.47 (r2 = 81.86) percent between ETa obtained through the SEBAL algorithm and the ET0 through the Penman–Monteith method. Balance method: In order to assess the rate of evapotranspiration in this method, the values of ground water withdrawals (well), precipitation volume and input and output runoffs were inserted in balance equation. The evapotranspiration parameter was calculated as the unknown. This indicates a relative error of 7.21 % when compared to the output of SEBAL algorithm. The SEBAL method has calculated the ETa of BenSaman region to be 228.16 mm (186 million cubic meters annually). This region is a main conflict point of water consumption between two provinces. Moreover, actual annual evapotranspiration map indicates that water is being excessively extracted in high volumes, specially by pumping from the Zayanehroud river to the elevated lands located in Chaharmahal–Bakhtiari province, which is the source of conflicts between upstream and downstream riparian. Acknowledgments This study was sponsored by research committee of Isfahan Regional Water Authority, Research Grant No. IRWA92.15.

References Akbari M, Toomanian N, Droogers P, Bastiaanssen W, Gieske A (2007) Monitoring irrigation performance in Esfahan, Iran, using NOAA satellite imagery. Agric Water Manag 88(1–3):99–109 Allen RG, Pereira LS, Raes D, Smith M (1998) Crop evapotranspiration: guidelines for computing crop water requirments. Rome: FAO. Irrig Drain, 56 Allen RG, Tasumi M, Morse A, Trezza R, Wright J, Bastiaanssen WGM, Kramber W, Lorite I, Robison C (2007) Satellite-based energy balance for mapping evapotranspiration with internalized calibration (METRIC)-applications. J Irrig Drain Eng 133(4):395–406 Bastiaanssen WGM (2000) SEBAL-based sensible and latent heat fluxes in the irrigated Gediz Basin, Turkey. J Hydrol 229(1–2):87–100 Bastiaanssen WGM, Noordman E, Pelgrum H, Davids G, Thoreson B, Allen R (2005) SEBAL model with remotely sensed data to improve water-resources management under actual field conditions. J Irrig Drain Eng. 131(1):85–93 Bastiaanssena WGM, Menentia M, Feddesb RA, Holtslagc AAM (1998a) A remote sensing surface energy balance algorithm for land (SEBAL) 1. Formulation. J Hydrol 212–213:198–212 Bastiaanssena WGM, Pelgruma H, Wangb J, Mab Y, Morenoc JF, Roerinka GJ, Wala T (1998b) A remote sensing surface energy balance algorithm for land (SEBAL) 2. Validation. J Hydrol 212–213:213–229 Bhattarai N, Dougherty M, Marzen LJ, Kalin L (2011) Validation of evaporation estimates from a modified surface energy balance algorithm for land (SEBAL) model in the south-eastern United States. Int J Remote Sens 3(6):511–519 Boegh E, Soegaard H, Thomsen A (2002) Evaluating evapotranspiration rates and surface conditions using Landsat TM to estimate

123

Iran J Sci Technol Trans Civ Eng atmospheric resistance and surface resistance. Remote Sens Environ 79:329–343 Cleugh HA, Leuning R, Mu Q, Running SW (2007) Regional evaporation estimates from flux tower and MODIS satellite data. Remote Sens Environ 106:285–304 Dastorani MT, Poormohammadi S (2012) Evaluation of water balance in a mountainous upland catchment using SEBAL approach. Water Resour Manag 26(7):2069–2080 Doorenbos J, Pruitt WO (1977) Guidelines for predicting of crop water requirements. Irrigation and drainage paper no 24, FAOONU, Rome Hafeez MM, Chemin Y, Van DeGiesen N, Bouman BAM (2002) Estimation of crop water deficit through remote sensing in Central Luzon, Philippines. In: Geosci Remote Sens Symp IGARSS ‘02. 2002 IEEE International, 5:2778–2780 Hong SH, Hendrickx JMH, Borchers B (2011) Down-scaling of SEBAL derived evapotranspiration maps from MODIS (250 m) to Landsat (30 m) scales. Int J Remote Sens 32(21):6457–6477. doi:10.1080/01431161.2010.512929 http://lancemodis.eosdis.nasa.gov/cgibin/imagery/realtime.cgi. Accessed 15 Feb 2013 Jang K, Kang S, Kim H, Kwon H (2009) Evaluation of shortwave irradiance and evapotranspiration derived from Moderate Resolution Imaging Spectroradiometer (MODIS). Asia Pac J Atmos Sci 45(2):233–246 Karatepe A, Ikiel C (2013) Analyzing land cover changes of Osmancik (Corum, Turkey) basin with landsat TM images. Iran J Sci Technol 37(A2):141–146 Kongo MV, Jewitt GWP, Lorentz SA (2011) Evaporative water use of different land uses in the upper-Thukela river basin assessed from satellite imagery. Agric Water Manag 98(11):1727–1739 Li Z, Liu X, Ma T, Kejia D, Zhou Q, Yao B, Niu T (2013) Retrieval of the surface evapotranspiration patterns in the alpine grassland–wetland ecosystem applying SEBAL model in the source region of the Yellow River, China. Ecol Modell 270:64–75 Member of MOCST (2009) MODIS Level 1B product user’s guide Moazed H, Ghaemi AA, Rafiee MR (2014) Evaluation of several reference evapotranspiration methods: a comparitive study of

123

greenhouse and outdoor conditions. Iran J Sci Technol Trans Civ Eng 38(C2):421–437 Mutiga ZK, Su Z, Woldai T (2009) Using satellite remote sensing to assess evapotranspiration: case study of the upper Ewaso Ng’iro North Basin, Kenya. Int J Appl Earth Obs Geoinf 12:S100–S108 Noodehi DA (2010) Evaluation of pan evaporation coefficients for estimating evapotranspiration (case Study: sari synoptic station). J Agron Sci Res 2(7):65–74 (In Farsi) Noshadi M, Sepaskhah AR (2005) Application of geostatistics for potential evapotranspiration estimation. Iran J Sci Technol Trans B Eng 29(B3):343–355 Orang M (1998) Potential accuracy of the popular non-linear regression equations for estimating pan coefficient values in the original and FAO-24 tables. Unpublished California Department of Water Resources Report, Sacramento, California Papadavid G, Hadjimitsis DG, Toulios L, Michaelides S (2013) A modified SEBAL modeling approach for estimating crop evapotranspiration in semi-arid conditions. Water Resour Manag 27(9):3493–3506 Patel NR, Rakhesh D, Mohammed AJ (2006) Mapping of regional evapotranspiration in wheat using Terra/MODIS satellite data. Hydrol Sci J 51(2):325–335 Snyder RL, Orang M, Matyac S, Grismer ME (2005) Simplified estimation of reference evapotranspiration from pan evaporation data in california. J Irrig Drain Eng 131(3):249–253 Sun Z, Wei B, Wei S, Wenming S, Changzuo W, Daian Y, Liu Z (2011) Evapotranspiration estimation based on the SEBAL model in the Nansi Lake Wetland of China. Math Comput Modell 54(3–4):1086–1092 Thoreson B, Matyac S, Hawkins T, Roberson M (2005) Remotely sensed ET computed by SEBAL-implications for water resource planning in California. Irrig Adapt Chang Water Supplies. doi:10.1061/40792(173)554 Zwart SJ, Bastiaanssen WGM (2007) SEBAL for detecting spatial variation of water productivity and scope for improvement in eight irrigated wheat systems. Agric Water Manag 89(3):287–296