Evapotranspiration of Cotton, Apocynum pictum, and Zyzyphus jujuba in the Tarim Basin, Xinjiang, China

Evapotranspiration is a crucial component of the water balance of ecosystems and landscapes, especially under arid climates. In the Tarim Basin, China, there is an increasing competition for water between irrigated agriculture, mainly cotton, and natural ecosystems, which results in periods of water shortage. Such water shortages also impact on cotton. Therefore, alternative crops have been searched for, which eventually withstand such periodical water shortages better than irrigated cotton, notably Zyzyphus jujuba and Apocynum pictum. The fruit tree Z. jujuba has been promoted from the previous decade onward in parts of the Tarim Basin. A. pictum is used as medicinal plant and in a small scale as fiber crop. A. pictum is a perennial herb, which is part of the natural riparian vegetation along the rivers of the Tarim Basin and which grows without irrigation. Therefore, the objective of this paper is to investigate the crop evapotranspiration (ETc) of those three plant species over the growing season. In this paper, the Penman-Monteith approach was employed. Daily means of stomatal resistance in June was 118 s/m, 222 s/m, and 927 s/m for cotton, A. pictum leafs, and Z. jujuba, respectively. In October, those daily mean stomatal resistance climbed to 885 s/m and 742 s/m for cotton and A. pictum leafs, respectively. ETc over the growing season was 514.7 mm for cotton, 217.2 mm for A. pictum, and 339 mm for Z. jujuba. The Kc value of Z. jujuba was in the range of other fruit trees. In this study cotton attained high yields compared to world average yields at a low ETc. This high water use efficiency was achieved through a shift from flood to drip irrigation, the utilization of plastic mulch, and breeding of cotton varieties.


Introduction
Evapotranspiration is an important component of the water balance of ecosystems and landscapes.Especially under arid climates accurate numbers of evapotranspiration are crucial for water resource management and planning.
Cotton is grown in many drylands of the world, e.g. in Central Asia (Uzbekistan, Turkmenistan, and Northwest China), Turkey, Texas (USA), or Australia [1].In all those areas, cotton depends on irrigation.The Tarim Basin in Xinjiang, Northwest China, has turned into the world's most important cotton production region with a total annual cotton lint production of 2.1 million t in 2010, i.e. 8.85% of the world production [2] [3].In 2011, the share of the cotton lint production in Xinjiang of the worldwide production climbed to 11% [3] [4].This increase in cotton production went along with a shift from flood irrigation [5] to drip irrigation [6] and application of plastic mulch [7].Despite of those production changes, the cotton production resulted in periodical water shortage along the Tarim River, which was the major river of the Tarim Basin [8] [9].Therefore, alternative crops have been searched for, which eventually withstand such periodical water shortages better than irrigated cotton.Among those alternative crops, the most prominent ones are Zyzyphus jujuba and Apocynum pictum.
The fruit tree Z. jujuba has been promoted from the previous decade onward along the southern rim of the Tarim Basin (personal communication, Xinjiang Forestry Administration).Today, in the oases of Qarqan and Qarklik, Z. jujuba has become the major crop and has replaced cotton as source of income to local people.A. pictumis harvests as medicinal plant and thus serves as additional income source for people in the Tarim Basin.Medicinal tea from A. pictum is traded all over Xinjiang.Furthermore, A. pictum yields fibers, which are used for small scale textile production.A. pictum is a perennial herb, which is part of the natural riparian vegetation along the rivers of the Tarim Basin [10].A. pictum, as many other plant species of the natural riparian vegetation, is a phreatophyte [11] [12].Phreatophytes do not need to be irrigated, but establish continuous contact to the groundwater through strongly developed root systems, in order to adapt to the arid climate [13].
The objective of this paper is to investigate the crop evapotranspiration of cotton, being the major crop in the Tarim Basin, and the two alternative crops Z. jujube and A. pictum.A widely applied approach to determining crop and plant evapotranspiration is the reference evapotranspiration approach as described in the FAO Irrigation and Drainage Paper No. 56 [14], which has been compared with other models and proven suitable for dryland conditions in e.g.Iran by [15] [16].After the Penman-Monteith Equation [14], the evapotranspiration is calculated as follows: ET 0 refers to a short grass vegetation of 100% vegetation coverage and 12 cm height, which is not water stressed.Crop evapotranspiration (ET c ) finally is calculated as [14]: where ET c : crop evapotranspiration (mm/d), ET 0 : Reference evapotranspiration (mm/d), K c : crop coefficient.Crop coefficients are given for the world's major crops, including cotton and fruit trees [14].Thus, there are no crop coefficients for A. pictum neither for Z. jujuba.The crop coefficients were developed primarily after sources from before 1989 [17]- [21].Therefore, the crop coefficient given for cotton reflects average cotton cultivation with a low, if any, proportion of drip irrigation.Therefore, in this paper we aim at determining ET c and crop coefficients of cotton, A. pictum, and Z. jujuba for representative sites of the Tarim Basin.

Study Area
This study focuses on cotton, A. pictum and Z. jujuba.The ET c measurements for those three plants were conducted on the sites Yingbaza, Qongaral, and Qarqan, respectively.All three sites are located in the Tarim Basin in South Xinjiang.The Tarim Basin covers an area of 1.02 million km 2 , and is home to a population of 9.02 million people.The area of irrigated land has increased all over the Tarim Basin, from 706,000 ha in 1949, over 1,330,000 ha in 1980 and 1,412,000 ha in 1990, in 2010 to 1,600,000 ha [2].The climate is extremely continental and arid with an annual precipitation ranging between 30 mm and 70 mm [2].In Korla and Qarqan, January mean temperatures are −10˚C and −7˚C, respectively, while July average temperatures are 26˚C and 24˚C, resüectively (http://www.tutiempo.net/clima/China/CN.html).
Yingbaza (41˚12'22'' N 84˚13'16'' E) is a village in Bügür County, about 200 km southwest of Korla at the middle reaches of the Tarim River.The study site Yingbaza is a cotton field 5 km north of the village Yingbaza.The field plot has a size of 100 m × 80 m.Cotton is planted under plastic mulch combined with drip irrigation as described by [22].Thereby, at both sides of each drip line one row of cotton seeds are planted.The two rows of cotton and the drip line are covered with one sheet of plastic foil as plastic mulch.Cotton is planted during April and harvested during September and October.The yields are around 1.5 t/ha lint cotton.Qongaral (41˚18'27'' N 85˚33'32'' E) is located in Korla City at the Layi River, a branch of the Tarim River in its inland delta (Figure 1). A. pictum forms natural mono-species stands along the Layi River.There, A. pictum grows up to a height of 1 m with an above ground biomass of 0.6 to 0.92 t/ha.The groundwater level fluctuates around 5 m. A. pictumis occasionally grazed by sheep and goats.The A. pictum stands are bordered by Tamarix shrub vegetation (unpublished data).
Qarqan (38˚08'49'' N 85˚29'47'' E) is an oasis town at the southern rim of the Tarim Basin (Figure 1).It is one of the first oasis towns which started to promote cultivating Z. jujuba ten years ago.The investigated site is an orchard surrounded by Populusalba tree rows as shelterbelt.The trees are planted in rows, which were 4 m apart from each other.Tree spacing within rows is 2 m.The tree height is 3 m.The soil is partly covered with grass vegetation, which is occasionally grazed by sheep and goats.

Methods
At each of the three sites, ET 0 and ET c were determined based on field measurements.Equation ( 5) was solved for K c so that K c was calculated based on ET 0 and ET c .
At each site, a mobile climate station was installed from the morning until the evening during the days of measurement (Table 1).This mobile climate station consisted of an air temperature/air humidity sensor (PASSRHT   from UMS, Munich), wind speed sensor (Davis anemometer), and two pyranometers (GLOBAL-pyra 03 from Delta Ohm), which measured incoming solar radiation and reflected radiation.Air temperature, air humidity, and wind speed were recorded as minutely averages by a data logger (EM50 from UMS, Munich).The radiations were recorded by a data logger (GP1 from UMS, Munich) and aggregated to minutely averages, too.The conversions of air humidity and air temperature into saturation vapor pressure deficit as well as the calculation of the net radiation based on the incoming solar radiation and reflected radiation followed the FAO guidelines by [14].These climate data and their conversion results were used for the ET 0 and ET c calculations.Soil heat flux was assumed as 0.1 * R n as suggested by [14].While the climate station was operated, the stomatal resistance was hourly with a porometer (SC-1, Decagon).On each site, three plants were selected randomly, from which three leafs each at the top, in the middle, and at the lower part of the crown were measured every hour.

Results
The air temperature, air humidity, and radiation (Table 1) were in the same range at all three sites.Only the wind speed was much lower in the Z. jujuba orchard (below 0.25 m/s) compared to the cotton field and the A. pictum stand investigated.At the latter two sites the wind speed frequently was above 1.5 m/s.LAI active differed considerably between the three plant species.Z. jujuba, being a tree, had an LAI active between 0.42 and 0.46 throughout the measurement period (Table 1).The LAI active of A. pictum leafs and stems did not exceed 0.13 and 0.06 respectively, throughout the growing season.The LAI active of cotton was 0.04 during the initial stage until the measurement in mid-June.Afterwards, it increased sharply up to 1.75 in August (Table 1).
The stomatal resistance of cotton and A. pictum leafs increased during the growing season.Cotton leafs additionally showed an increasing stomatal resistance from measurements in the morning to measurements in the afternoon/evening.This trend was most pronounced during the measurements in August and October.In contrast, Z. jujuba did not show such a clear trend.Though, in June and July the stomatal resistances of Z. jujuba increased from 764 s/m and 348 s/m in the morning to 1446 s/m and 833 s/m between 12:00 and 13:00 local time, respectively.At 18:00 local time, the stomatal resistances of Z. jujuba were 539 s/m in June and 708 s/m in July (Figure 2).
While the stomatal resistance of cotton in June (daily average of June 15 th was 118 s/m) was lower than the corresponding values of A. pictum leafs, stems, and Z. jujuba (daily averages of the measurements in June were 222 s/m, 470 s/m, and 927 s/m, respectively), cotton attained higher daily mean stomatal resistances (885 s/m) than A. pictum leafs (742 s/m) and A. pictum stems (419 s/m) in October (Figure 2).Corresponding to the low LAI active in June, the hourly ET c of cotton was lower than that of A. pictum and Z. jujuba (Figure 3).During the July measurements, the hourly ET c of cotton exceeded the ET c of the other two plant species.Cotton and Z. jujuba showed a peak of the ET c around noon, while A. pictum showed such a peak only during the measurements in May, July, and October (Figure 3).
The ET 0 values in the A. pictum stand and the cotton field were in the same range throughout the growing season (Table 2), e.g. between 6.1 mm/d and 8.1 mm/d from June to August, while the ET 0 of the Z. jujuba orchard was less than half of those values, i.e. 2.8 mm/d to 3.5 mm/d from June to August.This result corresponds with the considerably lower wind speed in the Z. jujuba orchard compared to the other two sites.Corresponding to the LAI development, the ET c of cotton was the lowest of the three plant species in June (0.6 mm/d), but increased sharply until August reaching 6.3 mm/d.Accordingly, the K c values also increased sharply from 0.09 in June to 0.79 in August.The K c values of A. pictum fluctuated between 0.13 and 0.26 from May to August and decreased to 0.1 in October (Table 2).The corresponding ET c values ranged between 1 mm/d and 1.8 mm/d from May to August and dropped to 0.3 mm/d.The ET c of Z. jujuba was between 2.4 mm/d and 2.7 mm/d during the time period from June to August.So, it showed the highest ET c during spring and was between the ET c of cotton and A. pictum from June onwards.The estimated ET c of the whole growing season for cotton, A. pictum, and Z. jujuba were 524.7 mm, 217.2 mm, and 339 mm, respectively (Table 2).

Discussion
The daily ET 0 determined for cotton and A. pictum (Table 2) is similar to ET 0 reported by [23] for Ejina in the Heihe river basin in Inner Mongolia, China, which has a similar climate as the Tarim Basin.The ET c over the growing season of cotton (Table 2) is below corresponding values given by [1] or [24], but similar to [25] and in the same range as the norms for water consumption from Soviet times (i.e.550 -600 mm) [26] and values given by [27] for Turkmenistan.
The crop coefficients of cotton after [24] are: K c ini = 0.35, K c mid = 1.15 -1.2, and K c end = 0.5 -0.7.The K c values of cotton of this study follow this pattern, as there is an increase from K c in June to K c in July and August followed by a decrease to K c in October (Table 2).Though, the K c values of this study are lower than those of [14] throughout the growing season.Especially the K c of June, during the initial stage of cotton, of 0.09 is much lower than the corresponding K c ini of 0.35 listed by [24].This difference can be explained through the application of plastic mulch combined with drip irrigation, because the plastic mulch covers all soil, which is wetted during irrigation from the drip lines so that evaporation from the soil surface is inhibited.The drip irrigation may also explain the differences of further K c values of this study compared to the K c listed by [14] as well as the low ET c in general.
Z. jujuba showed K c values in the range of the Rosaceae fruit trees listed by the FAO irrigation and Drainage Paper 56 [14].The low hourly and seasonal ET c of Z. jujuba compared to cotton (Figure 3 and Table 2) is due to the low ET 0 of Z. jujuba (Table 2).The low ET 0 in the Z. jujuba orchard is explained by the wind speed, which is much lower than on the cotton field and the A. pictum stand.The orchard is embedded into a narrow web of shelterbelt tree lines, which reduce the wind speed, while there is no obstacle, e.g.vegetation, for wind at the investigated cotton field and the A. pictum site.
The ET c over the growing season of A. pictum is in the same range as the water use of Alhagi sparsifolia and Tamarix ramosissima.Alhagi sparsifolia and Tamarix ramosissima are plant species of the natural riparian vegetation along the rivers of the Tarim Basin like A. pictum [28].The K c values of A. pictum are lower than all crop coefficients listed by [14].This is due to the low LAI active , which corresponds with the anatomy of A. pictum and the low vegetation coverage of that natural vegetation.A. pictum, though, is part of the natural riparian vegetation of the Tarim Basin and thus is adapted to the arid climate there, it does not have a higher stomatal resistance than the other two crops investigated.A. pictum is adapted to the arid climate by establishing contact to the groundwater being a phreatophyte [11] [12].Therefore, A. pictum survives under the arid climate by securing the water supply rather than saving water as other desert plants.So, if A. pictum grew with higher vegetation coverage (i.e. higher plant density) or the A. pictum plants grew larger, then the ET c and K c would increase.
In this study cotton attained a high yield compared to world average yields [3] at a rather low ET c .In Turkmenistan at similar ET c rates cotton yields and thus the water use efficiency is significantly lower [25].This high water use efficiency was achieved through a shift from flood to drip irrigation, the utilization of plastic mulch, and breeding of cotton varieties.The issue of garbage pollution through drip line and plastic mulch as well as their recycling is beyond the scope of this paper.

Figure 1 .
Figure 1.Map of the Tarim Basin with the investigation sites.

Table 1 .
Climate station and crop data at days of measurement of stomatal conductance.For the climate station data average values of hourly mean values are given.Only for solar radiation the sum of radiation per hour during the time of measurement is given.For A. pictum, two LAI values are given: LAI leafs/LAI stems.Crop and date Temp.(˚C) Relative air humidity (%) Wind speed (m/s) Solar radiation (MJ/m 2 ) Albedo Crop height (m) LAI active

Figure 2 .
Figure 2. Hourly mean values and standard deviation of the stomatal resistance (s/m) during the days of measurement.All x-axes refer to a time span of 07:00 to 19:00 Xinjiang time, i.e. geographical time, which is equivalent to 09:00 to 21:00 Bejing time.

Figure 3 .
Figure 3. Hourly ET c during the days of measurement of cotton, A. pictum, and Z. jujuba.The time at the x-axis is the local geographical time (Xinjiang time), which is two hours later than Beijing time.

Table 2 .
Daily ET 0 (mm), ET c (mm), and K c of cotton, Apocynum pictum, and Zyzyphus jujuba during the months May to October 2012.