Mulched drip irrigation cotton yield potential estimation based on large-scale water-nitrogen coupling model in Xinjiang, China with limits of water resources

Zhao Aiqin



Mulched drip irrigation cotton yield potential estimation based on large-scale water-nitrogen coupling model in Xinjiang, China with limits of water resources

Zhao Aiqin

100125

In arid Xinjiang of China, a main cotton producing area, irrigation is an important measure of agricultural production, and researches yield rich experimental data. This study aimed to establish a large-scale water-nitrogen coupling model based on comprehensive analysis of existing field experimental data to estimate mulched drip irrigation cotton yield potential in Xinjiang. A total of 172 datasets on Xinjiang from the year of 1998 to 2016 were collected through literature retrieval. They included 19 known cotton varieties withas a main variety. Considering data acquisition and climatic conditions, Xinjiang was divided into northern and southern Xinjiang. In each, a water-nitrogen coupling model was established based on random classification of data by different cotton varieties. The model validation showed the model was reliable with determination coefficient (2) of 0.57, normalized root mean square of error of 11%, and concordance index of 0.85 in the northern region, and2of 0.84, normalized root mean square of error of 8.3%, and concordance index of 0.95 in the southern Xinjiang, respectively. Based on the model, the optimal irrigation was 604 and 552 mm in the southern and northern Xinjiang, respectively, and the optimal fertilizer-N application rate was 325 and 354 kg/hm2in the southern and northern Xinjiang, respectively. Available irrigation amount was 494 mm in 2020 according to Xinjiang 2014—2016 and future planning (2020). Assuming fertilizer-N supply was optimally supplied, based on the established model, the yield potential of the northern and southern Xinjiang was 5 900 and 6 695 kg/hm2in 2020 under the limits of water sources, respectively. The total lint yield should be about 5.2×106t in 2020 in Xinjiang. The actual lint yield in 2014 in Xinjiang was about 3.7×106t, about 71% of the potential under the limits of water resources. The study can provide valuable information for application of mulched drip irrigation and agricultural water resources planning in Xinjiang.

cotton; irrigation; fertilizers; water-nitrogen model; mulched drip irrigation; Xinjiang,China

0 Introduction

Meeting future demand for crop yield has been a hotspot of studies in agriculture[1-2]. Xinjiang with superior sunlight and heat resources and few meteorological disasters is the largest cotton production area in China. According to data from, the Xinjiang cotton yield accounts for 67% of national total in 2016. In Xinjiang, agriculture heavily depends on irrigation due to a lack of rainfall and strong evaporation. Mulched drip irrigation makes for a great contribution to the crop yield. It saves water by 18%-42%, increases water use efficiency by 35%-103% and cotton lint yield by 10%-19%[3].Sufficient water supply is required for maintaining and increasing cotton yield in Xinjiang. However, according toissued by the government in 2011, the water supply ratio for the agriculture will decrease to 90% and the total water supply will decrease by 10 billion m3by 2020 in Xinjiang. The Xinjiang agriculture will face a serious water shortage by 2020. It is important to clarify the cotton yield potential in Xinjiang under the reduced water supply to agriculture.

Following the definition of yield potential defined by Evans[4], the cotton yield potential in Xinjiang is the yield of an adapted cotton cultivar when grown without water and nutrient limitations and kept free of biotic stresses in irrigated area. Crop growth models have been used for cotton yield estimation and good predictions have been reported by using AquaCrop model, Cotton Growth Simulation model (CGSM), Crop-Syst, COTTON2K, WOFOST et al[5-11]. However, these models usually need complex input parameters. In most places, some parameters required by the models are difficult to acquire[12], especially for large-scale region such as the whole Xinjiang. Thus, only a few of models have been used for cotton yield prediction in Xinjiang[13].

Field experiments are traditionally used to determine crop yield. Abundant valuable data from field experiments are available since the mulched drip irrigation is used in Xinjiang[14-16]. Cotton production is mainly affected by many factors such as climate, soil, irrigation, management and variety. Among these factors, climatic conditions, irrigation level, crop varieties, fertilization and soil type are the main influencing factors, while crop varieties, irrigation and fertilization are considered as controllable factors[17]. Studies have showed a parabolic relationship between crop yield and irrigation water[18]. The crop yield changes with the amount of nitrogen applied in a quadratic curve[19]. Nitrogen and water are the main limiting factors for high yield of mulched drip irrigation in cotton, and there is synergistic effect between them, and the effect of water is higher than that of fertilizer[19-20].Therefore, many studies used a statistical model (water-nitrogen coupling model) based on field data to describe the relationship between water, nitrogen fertilizer and yield. The water-nitrogen coupling model has been regarded as the classic empirical model. This kind of model cannot be promoted across regions and species. However, it has the advantages of simple structure, simple calculation, less data requirement and low application threshold[17,21].

There are a large number of experimental data on water-nitrogen coupling of cotton yield under mulched drip irrigation in Xinjiang. However, due to different research sites, field management and experimental design, the maximum cotton yield values obtained by these studies vary, which is difficult to guide the planning and development of large-scale regional agriculture. This study aimed to integrate these data, established a large-scale comprehensive model of cotton yield, and estimated the potential of cotton yield under mulched drip irrigation in Xinjiang, so as to provide valuable information for water resources plan and regulation in Xinjiang.

1 Materials and methods

1.1 Cotton production overview in Xinjiang

The distance in Xinjiang (73°40'-96°23'E, 34°25'-49°10'N) is 2 200 km from east to west and 1 500 km from south to north. Xinjiang is in the arid and semiarid inland and has a typical temperate continental climate characterized by strong evaporation and rare rainfall. Surrounded by mountains, Xinjiang is divided by the Tianshan mountain into Tarim Basin in the south and Junggar Basin in the north. Eastern Xinjiang is on the east of Tianshan, which mainly include Kumul and Turpan Basin. The cotton production is mainly distributed in the southern Xinjiang, the northern Xinjiang, and the eastern Xinjiang, especially in the southern Xinjiang[22]. Based on data issued by(http://www.xjtj.gov. cn/sjcx/tjnj_3415/), the sowing area and yield of cotton in the eastern Xinjiang accounts for only 2.8% and 3.1% of the total of Xinjiang, respectively. The southern and northern Xinjiang could represent the main cotton production condition in the Xinjiang. In addition, limited by data acquisition, I didn’t include the eastern Xinjiang in this study. The rainfall in the northern Xinjiang is higher than the southern Xinjiang and the air temperature in the southern Xinjiang is higher than the northern Xinjiang. Table 1 shows difference of climatic factors in the southern and northern regions of Xinjiang. In this area, agriculture heavily depends on irrigation, and thus the area is a pure irrigation agricultural region[23].

Table 1 Difference of climatic factors in different cotton growing regions of Xinjiang

Note: Data on frost-free days, >10 ℃ cumulative temperature, and radiation are from literature[24]. Data on air temperature, sunlight, and relative humidity are obtained from 42 meteorological stations during 1961—2016 provided by China Meteorological Administration. Data on precipitation and evaporation are from literature[25].

Irrigation water is from underground water, surface water and glaciers in Tianshan mountain. Irrigation methods in Xinjiang include conventional flooding irrigation, spray irrigation, and micro-irrigation (mulched drip irrigation, subsurface drip irrigation and et al.), mostly water-saving irrigation. Since 1990, the cotton variety has been dominated byin the northern Xinjiang, andandin the southern Xinjiang[26-27]. The growing season of the cotton is from April to October.

1.2 Data sources

A total of 64 published papers were collected by reference retrial method in main databases of China National Knowledge Infrastructure (CNKI), Wanfang, SCI direct and others. Among them, 20 papers[10,28-47]provided clear information on place, year, fertilizer-N application rate, irrigation water amount etc. From the papers, 172 groups of data were derived including 121 groups of data from 13 papers for the northern Xinjiang and 51 groups of data from 6 papers for the southern Xinjiang (Table 2). The experimental areas were Bozhou (Jinghe, Bole), Tacheng (Usu, Shawan), Changji (Manas), Shihezi, Bazhou (Korla, Yuli), Aksu (Awati), which were the preferentially developed areas according to. Thus, data from these areas were representative of future cotton planting development. A total of 19 known varieties were involved.andwere the dominant varieties respectively in the northern and the southern Xinjiang, which is consistent with the actual condition in Xinjiang. Soil texture of these locations ranged sand to clay (Table 3). During the process of data collection and pretreatment, data quality was controlled based on the strict criteria. Data were collected from the main cotton production areas, involved the main cotton variety, and could represent cotton production condition in Xinjiang.

Table 2 Cotton varieties and correspondingdata groups in this study

Table 3 Summary of data-related soil, climate and cotton planting management

1.3 Model establishment and validation

Typical water-N coupling model is

=1+2+3+42+52+6(1)

Where,is cotton seed yield, kg/hm2;is irrigation amount, mm;is fertilizer-N application rate, kg/hm2;1-6are undefined coefficient. It is an empirical formula. The coefficient is obtained by regression of independent variable (and) and dependent variable ().

Model fitting and data analysis were conducted in Statistix 9.0. The established model was evaluated by index usually used including determination coefficient (2), root mean square of error (RMSE), normalized root mean square of error (nRMSE) and concordance index ()[48-49]. The high2value andvalues closer to 1 show good agreement between the simulated and measured values. The nRMSE value indicates relative difference between the measured and simulated values. Model is considered as excellent, good, normal and bad when the nRMSE value is <10%, 10%-20%, >20%-30% and >30%[48].

2 Results and analysis

2.1 Establishment and validation of water-N coupling model for cotton yield

Variety is an important influential factor and variety update contributes to 43.9% of yield increase[27]. In addition, correlation analysis in this study showed that if ignoring variety, the irrigation amount and fertilizer-N application rate were correlated with the cotton yield withof 0.43 and 0.36 (<0.01), respectively, which could not meet the model fitting precision. Thus, I classified the data by the cotton variety. Then, taking into account of data acquisition and climatic (radiation, temperature and et al.) conditions, thedata were classified into 2 subsets for the southern and northern Xinjiang, respectively. Each subset was randomly divided into 2 groups. One group was for model establishment and the other for validation. The data is shown in Table 4.

Varietyandwere used for model establishment in the northern and southern Xinjiang, respectively. The simulation results based on water-nitrogen coupling model for cotton yield are shown in Table 5 and Fig.1. As shown in Table 5 and Fig.1, thecotton yield model in the northern Xinjiang had the2of 0.57-0.62, nRMSE of 11%-21%, andof 0.85-0.86 while thecotton yield model in the southern Xinjiang had the2of 0.83-0.84, nRMSE of 8.3%, andof 0.95. The nRMSE lied between normal and excellent. Thevalue about 0.9 indicates a good agreement of simulated and experimental values. It indicates that the water-nitrogen coupling models are reliable for cotton yield simulation in Xinjiang. Based on the model, the maximum cotton yield was 5 961 and 6 773 kg/hm2for the northern and the southern Xinjiang, respectively (Fig.2). The maximum cotton seed yield of the southern Xinjiang was 1.1 times as that of the northern Xinjiang. To obtain the maximum yield, the optimal irrigation amount and fertilizer-N application rate were 552 mm and 354 kg/hm2in the northern Xinjiang, respectively, while the optimal irrigation amount and fertilizer-N application were 604 mm and 325 kg/hm2in the southern Xinjiang, respectively (Fig.2).

When the fertilizer-N application rate was fixed to the mean (245 and 249 kg/hm2, respectively for the southern and northern Xinjiang), the model could describe the relationship between irrigation and yield (Fig.3a and Fig.3c). When the irrigation amount was fixed to the mean (465 and 343 mm, respectively for the southern and northern Xinjiang), the model reflected the change of yield with fertilizer-N application rate (Fig.3b and Fig.3d). The yield increased firstly and then decreased after a certain critical point with both the irrigation and fertilizer-N application rate.

Table 4 Data for model simulation and validation

Note:is irrigation amount, mm;is fertilizer-N application rate, kg·hm-2;is cotton seed yield, kg·hm-2; same as below.

Table 5 Water-nitrogen coupling model simulation result under mulched drip irrigation

Note:is data points;is concordance index; nRMSE is normalized root mean square of error;2is coefficient of determination. Same as below.

Fig.1 Water-nitrogen coupling model simulation results for different cotton varieties under mulched-drip irrigation

Note: Values on contour lines are cotton seed yield, kg·hm-2.

Note: In Fig.a and Fig. c, the fertilizer-N application rate is 245 and 249kg·hm-2; in Fig.b and Fig.d, the irrigation amount is 465 and 343 mm.

2.2 Cotton yield estimation based on water-nitrogen coupling model in 2020

2.2.1 Available irrigation water estimation for cotton in 2020

By investigation, I obtained micro-irrigation quota and sowing area for cotton in 2014 (Table 6). Based on the sowing area and micro-irrigation quota, the irrigation water for cotton based on water-saving technologies was 8.43×109m3in 2014. The(http:// www.xjslt.gov.cn/2017/12/11/slgb/60050.html) reported that the agricultural water allocation in Xinjiang was 5.51× 1010m3, and thus the irrigation water for cotton accounted for 15.3% of the total agricultural water allocation in 2014 (Table 6). Therequires agriculture water decrease to 90% of the total by 2020. So, the available agricultural water in 2020 would be 4.95×109m3. If the cotton sowing area was stable from 2014 to 2020, the irrigation water amount would be 7.57×109m3in 2020. Then, the available water amount per unit area for the cotton would be 494 mm.

2.2.2 Cotton yield potential in 2020 under limits of water resources

Since the chemical fertilizer-N is easy to obtain, I assumed that the fertilizer-N could be sufficiently applied to the optimal levels in 2020. Based on the 494 mm irrigation amount and the established models, the cotton yield would be 6 695 and 5 900 kg/hm2for the southern and northern Xinjiang, respectively (Table 7). The total cotton seed yield would be about 1.2×107t in 2020. The lint percent is usually 40% and 42% for theand[27], respectively. Thus, under the limited water supply in 2020, the cotton lint yield would be about 5.2×106t in 2020 in Xinjiang. The actual lint yield in 2014 in Xinjiang is about 3.7×106t[5], reaching 71% of the cotton lint yield in 2020 under the limits of water resources. It indicates that the cotton production in 2014 hasn’t reach the potential with the limits of water resources and the cotton yield could be improved by 29% if optimal fertilizer-N application rate, cotton variety, irrigation and management method were used.

Table 6 Irrigation amount for cotton in Xinjiang in 2014

Table 7 Potential cotton yield with optimal water and N supply estimated by large-scale water-nitrogen coupling model based on data in 1998—2016 and predicted cotton yield in 2020with limits of water resources under mulched-drip irrigation

3 Discussion

During the model establishment, I considered the influence of region, irrigation, fertilizer and variety. The southern and northern Xinjiang are characterized by distinctly different climatic conditions (Table 1). Thus, I indirectly considered the influence of climatic factors on cotton production. Based on the model, the yield increased firstly and then decreased after a certain critical point with both the irrigation and fertilizer-N application rate. The trend was consistent with previous studies in field experiments[18-20]. Different from previous studies, the model in this study evidenced a large-scale regional application of the water-nitrogen coupling model in Xinjiang, which has not been reported so far.

The sunlight and temperature in the southern Xinjiang are better than that in the northern Xinjiang, and thus the yield in the former higher than that in the latter is consistent with practical situation. When fertilizer-N rate and irrigation amount are optimal, the maximum cotton seed yield of the southern and northern Xinjiang can theoretically represent difference of climatic and soil conditions. According to Xu et al.’s study on thermal production potential in Xinjiang[50], the highest value was about 1.3 times the lowest value. In our study, the maximum cotton seed yield of the southern Xinjiang was 1.1 times as that of the northern Xinjiang.

Gao et al.[51]recommended the irrigation amount of 675 mm to obtain maximum cotton yield in Korla; Wang et al.[18]suggested the optimal irrigation amount 280-307 mm for the northern Xinjiang based on CHAIN_2D model validated by using data from fields experiments in Shiheizi; Xing et al.[52]reported the irrigation amount of 420 mm for maximum cotton yield in the Tarim basin. He et al.[53]recently found that the optimal irrigation amount and fertilizer-N application rate were 523 mm and 500 kg/hm2in Shiheizi. However, Guo[54]showed that optimal irrigation amount and fertilizer-N application rate were 419 mm and 224 kg/hm2in Shihezi, respectively. Suggested irrigation amount and fertilizer-N application rates from field experiments various greatly. It is because the results are not only dependent on different soil and climatic conditions but also on experience of researchers in management in different fields experiments. In this study, optimal irrigation amount and fertilizer-N application rate were respectively determined for the southern and the northern Xinjiang based on models. In water resource planning, the determined value could be used directly, which avoids the subjectivity during selecting results from different experiments at regional scale. In this respect, the model here is useful in solving regional problems at regional scale. However, the model here ignores the difference of plots and thus is of limitations in guiding field production.

4 Conclusions

Based on the cotton variety and planting region, I established large-scale water-N coupling models of cotton yield for the southern and the northern Xinjiang. The models were reliable with the concordance index about 0.9,2of 0.57-0.84, and normalized root mean square of error between normal and excellent range. Based on the established model, for the southern and the northern Xinjiang, the optimal irrigation was 604 and 552 mm, respectively; the optimal fertilizer-N application rate was 325 and 354 kg/hm2, respectively. Correspondingly, the maximum cotton yield was 6 773 and 5 961 kg/hm2, respectively. In 2020 with the limits of water resources, assuming the fertilizer-N application rate was optimally supplied, the potential cotton yield in 2020 would be 6 695 and 5 900 kg/hm2for the southern Xinjiang and the northern Xinjiang, respectively, and the total lint yield would be about 5.2×106t. The actual lint yield in 2014 reached 71% of the cotton lint yield potential in 2020. The cotton yield has the great potential to be improved.

[1] Timsina J, Wolf J, Guilpart N, et al. Can Bangladesh produce enough cereals to meet future demand?[J/OL]. Agricultural Systems, 2016:02209. http://dx.doi.org/10.1016/j.agsy.2016. 11.003

[2] Zhang W, Cao G, Li X, et al. Closing yield gaps in China by empowering smallholder farmers[J]. Nature, 2016, 537(7622): 671.

[3] Ibraginmov N, Evett S R, Esanbekov Y, et al. Water use efficiency of irrigated cotton in Uzbekistan under drip and furrow irrigation[J]. Agric Water Manag 2007, 90: 112－120.

[4] Evans L T. Crop Evolution, Adaptation and Yield[M]. Cambridge: Cambridge University Press, 1996.

[5] Hussein F, Janat M, Yakoub A. Simulating cotton yield response to deficit irrigation with the FAO AquaCrop model[J]. Spanish Journal of Agricultural Research, 2011, 9(4): 1319－1330.

[6] Liang Xinzhong, Xu Min, Gao Wei, et al. A distributed cotton growth model developed from GOSSYM and its parameter determination[J]. Agronomy Journal, 2012, 104(3): 661－674.

[7] Pan Xuebiao, Long Tongfang, Dong Zhanshan. Studies on simulation model of cotton growth and development and yield composition (CGSM)[J]. Acta Gossypii Sinica, 1992, 4(Supp): 11－20.

[8] Sommer R, Kienzler K, Conrad C. Evaluation of the Crop-Syst model for simulating the potential yield of cotton[J]. Agron Sustainable Dev, 2008, 28: 345－354.

[9] Yang Y, Ouyang Z, Yang Y. Simulation of the effect of pruning and topping on cotton growth using COTTON2K model[J]. Field Crops Research, 2008, 106(2): 126－137.

[10] Wu Lifeng. Coupling Effect of Water and Fertilizer Under Fertigation and Simulation Growth of Cotton in Xinjiang[D]. Yangling: Northwest A&F University, 2015.

[11] Zhu Xiufang, Li Yizhan, Pan Yaozhong, et al. A review on the research and application of AquaCrop model[J]. Chinese Agricultural Science Bulletin, 2014, 30(8): 270－278.

[12] Pan Xuebiao, Li Yu’e. Study on cotton production regional assessment system of Xinjiang[J]. Scientia Agricultura Sinica, 2003(1): 37－43.

[13] Cai Huanjie, Shao Guangcheng, Zhang Zhenhua. Water demand and irrigation scheduling of drip irrigation for cotton under plastic mulch[J]. Journal of Hydraulic Engineering, 2002, 33(11): 119－123.

[14] Wang Feng, Sun Jingsheng, Liu Zugui, et al. Effects of irrigation scheduling on growth, yield and fiber quality of cotton under mechanical harvest cropping model[J]. Cotton Science, 2014, 26(1): 41－48.

[15] Wu Lifeng, Zhang Fucang, Zhou Hanmi, et al. Effect of drip irrigation and fertilizer application on water use efficiency and cotton yield in North of Xinjiang[J]. Transactions of the Chinese Society of Agricultural Engineering(Transactions of the CSAE), 2014, 30(20): 137－146.

[16] He Jinyu, Tian Juncang. Model of coupling water with fertilizer and optimum combination scheme of rice cultivated in aerobic soil with drip irrigation under plastic film[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2015, 31(13): 77－82.

[17] Li Zhongkai, Liu Hu, Zhao Wenzhi. Revisiting crop water production functions in terms of cross-regional applications[J]. Chinese Journal of Eco-Agriculture, 2018, 26(12): 1781－1794.

[18] Wang Jun, Li Jiusheng, Guan Hongjie. Modeling response of cotton yield and water productivity to irrigation amount under mulched drip irrigation in North Xinjiang[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2016, 32(3): 62－68.

[19] Hu Mingfang, Tian Changyan, Lu Zhaozhi, et al. Effects of N rate on cotton yield and nitrate-N concentration in plant tissue and soil[J]. Journal of Northwest Sci-Tech University of Agriculture and Forestry: Natural Science Edition, 2006, 34(4): 63－68.

[20] Wang Zhenhua, Zhu Yankai, Zhang Jinzhu, et al. Effects of water and nitrogen fertilization on physiological characteristics and yield of cotton under drip irrigation in mildly salinized soil[J]. Transactions of the Chinese Society for Agricultural Machinery, 2018, 49(6): 296－308.

[21] Clempner G, Solomon K H. Accuracy and geographic transferability of crop water production functions[C]// Proceedings of Congr Irrigation Systems for the 21stCentury. Portland, OR: American Society of Civil Engineers, 1987.

[22] Pan Wei, Yang Degang, Yang Li, et al. Spatial-temporal dynamics and optimal development scale of cotton industry in Xinjiang[J]. Chinese Journal of Eco-Agriculture, 2011,19(2):415－420.

[23] Wang Zhenhua, Yang Peiling, Zheng Xurong, et al. Soil salinity changes of root zone and arable in cotton field with drip irrigation under mulch for different years[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2014, 30(4): 90－99.

[24] Wang Chunhua, Wu Qifeng, Zhou Ying, et al. The influence of climate characteristics on overall arrangement of facility agriculture in Xinjiang[J]. Heilongjiang Agricultural Sciences, 2013(11): 58－64.

[25] Chen Chunyan, Zhao Keming, Ablimitian Ablikim, et al. Temporal and spatial distributions of hourly rain intensity under the warm background in Xinjiang[J]. Arid Land Geography, 2015, 38(4): 692－702.

[26] Diao Ming, Chu Guixin, Li Shaokun, et al. Studies on the relationship of upland cotton varieties in the course of replacement of varieties in the North Xinjiang over the past fifty years[J]. Sientia Agricultura Sinica, 2002, 35(12): 1456－1460.

[27] Tursunjan, Li Xueyuan, Tian Changyan, et al. Cultivar evolution and breeding potential of cotton in region of Southern Xinjiang[J]. Journal of Plant Genetic Resources, 2012, 13(4): 535－541.

[28] Chen Baoyan, Ma Xingwang, Yang Tao, et al. Effects of different N strategies on cotton under mulch drip irrigation with SPAD values and yield[J]. Chinese Agricultural Science Bulletin, 2011, 27(3):124－129.

[29] Deng Zhong, Zhai Guoliang, Zong Jie, et al. Effects of water and nitrogen regulation on root and shoot growth characteristics and yield of cotton in arid area[J]. Soil and Fertilizer Sciences in China, 2015(6):57－59.

[30] Yi Pengfei, Hudan Tumaerbai, Wang Yimin, et al. Schedule optimization of under-plastic-mulch drip irrigation for cotton in arid areas[J].Bulletin of Soil and Water Conservation, 2011, 31(1): 53－57.

[31] Lei Anjian, Feng Jinxiang, Duan Xiaoyan. Optimization and exploration of irrigation and fertilization methods of mulched drip irrigation cotton[J]. Gansu Agriculture, 2006(5):283.

[32] Zheng Zhong, Ma Fuyu, Mu Zixin, et al. Effects of factors of water and fertilizers under mulch drip irrigation on cotton canopy structure and yield[J]. Agricultural Research in the Arid Areas, 2001, 19(2): 42－47.

[33] Shen Xiaojun, Zhang Jiyang, Liu Zugui, et al. Effects of different water treatments on yield and water use efficiency of cotton with drip irrigation under film mulch[J]. Agricultural Research in the Arid Areas, 2012, 30(2): 118－124.

[34] Shen Xiaojun, Chen Hongmei, Sun Jingsheng, et al. Response of different water deficit on cotton growth and water use efficiency and yield under mulched drip irrigation[J]. Journal of Irrigation and Drainage, 2010, 29(1): 40－43.

[35] Deng Zhong, Zhai Guoliang, Wang Xiaosen, et al. Application schedule of N-P-K in drip fertigation affects abscission of cotton bolls[J]. Journal of Irrigation and Drainage, 2018, 36(8): 1－6.

[36] Deng Zhong, Bai Dan, Zhai Guoliang, et al. Effects of water and nitrogen regulation on the yield and water and nitrogen use efficiency of cotton in south Xinjiang, Northwest China under plastic mulched drip irrigation[J]. Chinese Journal of Applied Ecology, 2013, 24(9): 2525－2532.

[37] Wang Xiaojuan, Wei Changzhou, Zhang Jun, et al. Effects of irrigation methods and N application level on cotton growth and nitrogen use efficiency[J]. Cotton Science, 2012, 24(6): 554－561.

[38] Wu Zhengguang, Hudan Tumarbay, Zhang Jinzhu, et al. Study on effect of drip irrigation under membrane on the change of soil moisture and yield in cotton field[J]. Xinjiang Agricultural Sciences, 2010, 47(8): 1574－1579.

[39] Liu Xinyong, Tian Changyan, Ma Yingjie, et al. Water consumption characteristics and scheduling of drip irrigation under plastic film for cotton in south Xinjiang[J]. Agricultural Research in the Arid Areas, 2006, 24(1): 108－112.

[40] Gong Jiang, Li Jun, Liu Mao, et al. Water effect in cotton under film drip irrigation in Shihezi, Xinjiang[J]. China Cotton, 2011(2): 11－13.

[41] Yao Qingqing, Yang Tao, Ma Xingwang, et al. Effects of water and fertilizer regulation strategies on cotton canopy structure and yield under mulched drip irrigation[J].Cotton Science, 2013, 25(1): 73－80.

[42] Yu Mei, Yang Jinsong, Liu Meixian, et al. Effects of mulched drip irrigation frequency on soil water-salt transport and cotton yield[J]. Agricultural Research in the Arid Areas, 2011, 29(3): 18－24.

[43] Gao Shan, Zhang Yu, Wang Jichuan. Effects of drip irrigation under mulch on canopy apparent photosynthesis and yield structure of hybrid cotton Zhao feng 1 in South Xinjiang[J]. Journal of Tarim University, 2011, 23(4): 46－54.

[44] Wang Haidong. Water-fertilizer Coupling Effects of Field Cotton in Xinjiang under Fertigation[D]. Yangling: Northwest A&F University, 2015.

[45] Fu Hengyang, Ji Guobin. Experimental research on cotton water requirement and irrigation program under mulched drip irrigation in arid areas[J]. Yellow River, 2016, 38(1): 144－148.

[46] Gong Jiang, Wang Haijiang, Xiehaixia, et al. Effects of water-nitrogen coupling on the development and yield of cotton in drip irrigation under film[J]. Journal of Irrigation and Drainage, 2008, 27(6): 51－54.

[47] Zhang Yan, Hou Xiuling, Wang Xiaojing, et al. Effect of optimized nitrogen management on seal island cotton yield and fiber quality under plastic mulch by drip irrigation[J]. Acta Agriculturae Boreali-Occidentalis Sinica, 2006, 15(4): 122－125.

[48] Marco Dettori, Carla Cesaraccio, Andrea Motroni, et al. Using CERES-Wheat to simulate durum wheat production and phenology in Southern Sardinia, Italy[J]. Field Crop Research, 2011, 120: 179－188.

[49] Wang Jun, Huang Guanhua, Zhan Hongbin, et al. Evaluation of soil water dynamics and crop yield under furrow irrigation with a two-dimensional flow and crop growth coupled model [J]. Agricultural Water Management, 2014, 141: 10－22.

[50] Xu Wenxiu, Niu Xinxiang, Bian Xiuju. The calculations and analyses on thermal production potential of cotton in Xinjiang[J]. Cotton Science, 2007, 19(6): 455－460.

[51] Gao Long, Tian Fuqiang, Ni Guangheng, et al. Expeirmental study on soil water-salt movement and irrigation scheduling for cotton under mulched drip irrigation condition[J]. Journal of Hydraulic Engineering, 2010, 41(12): 1483－1490.

[52] Xing Xiaoning, Yao Baolin, Sun Sanmin. Effects of different irrigation regimes on cotton growth and yield with drip irrigation under plastic film in oasis region of South Xinjiang [J]. Acta Agriculturae Borealic-occidentalis Sinia, 2016, 25(2): 227－236.

[53] He Huaijie, Wang Zhenhua, Zheng Xurong, et al. Effects of water-nitrogen coupling on growth and yield of cotton under mulch drip irrigation[J]. Xinjiang Agricultural Sciences, 2017, 54(11): 1983－1989.

[54] Guo Jingqiang. Study on Water-nitrogen Coupling Mechanism on Cotton Plant in Condition of Film Mulch and Drip Irrigation[D]. Shihezi: Shihezi University, 2005.

基于大尺度水肥耦合模型估算新疆膜下滴灌棉花生产潜力

赵爱琴

（农业部规划设计研究院，农业部规划设计研究院博士后科研工作站，中国农业工程学会，北京 100125）

在中国棉花主产区新疆，灌溉是农业生产的重要措施，前人研究获得了丰富的试验数据。该研究综合现有田间试验数据，建立了大尺度水-氮耦合模型，估算新疆膜下滴灌棉花的产量潜力，并根据建立的模型，预测有限水资源供应下2020年新疆的棉花产量。从文献收集1998—2016年新疆地区共172组数据。考虑棉花品种、灌水量和施氮量是影响棉花产量的主要因素，首先根据数据可获性和气候条件，将新疆分为南疆和北疆，然后基于棉花品种，分别建立了南疆和北疆大尺度水氮耦合模型。模型验证表明模型是可靠的，北疆地区模型的2为0.57，归一化均方根误差为11%,一致性指数0.85，南疆地区2为0.83，归一化均方根误差为8.3%，一致性指数为0.95。根据模型，南疆最佳灌溉和氮肥施用量分别为604 mm和325 kg/hm2，相应产量潜力为6 773 kg/hm2；北疆最佳灌溉和氮肥施用量分别为552 mm和354 kg/hm2，相应产量潜力为5 961 kg/hm2；根据新疆未来规划（2020），2020年新疆单位面积棉花可获得灌溉水量为494 mm，假设2020年肥力最优供应，根据建立的模型，2020年受水资源限制下，北疆和南疆棉籽单产分别为5 900和6 695 kg/hm2，皮棉总产可达5.2×106t。2014年皮棉总产约3.7×106t，为2020年受水资源限制下皮棉产量潜力的71%。该研究可为新疆膜下滴灌的推广和农业水资源规划提供参考。

棉花；灌溉；肥料；水氮模型；膜下滴灌；新疆

2019-02-19

National Natural Science Foundation of China (41601604); Elite Journal of CAST (Contributing to Research Subject Development in Agricultural Engineering)

Zhao Aiqin, Ph.D, research on soil and water, agricultural engineering. Email: zhaoaiqin2002@126.com

10.11975/j.issn.1002-6819.2019.05.013

S275.6

A

1002-6819(2019)-05-0111-08

Zhao Aiqin. Mulched drip irrigation cotton yield potential estimation based on large-scale water-nitrogen coupling model in Xinjiang, China with limits of water resources[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2019, 35(5): 111－118. (in English with Chinese abstract) doi：10.11975/j.issn.1002-6819.2019.05.013 http://www.tcsae.org

赵爱琴.基于大尺度水肥耦合模型估算新疆膜下滴灌棉花生产潜力[J]. 农业工程学报，2019，35(5)：111－118. doi：10.11975/j.issn.1002-6819.2019.05.013 http://www.tcsae.org