|其他摘要||Water resources is a major constraint on agricultural development in arid areas. In recent years, with the development of water-saving irrigation technology, the non-balance condition between supply and demand of water resources was alleviated and the irrigation area was expanded. Understanding the evapotranspiration process of cropland under non-full irrigating has great scientific significance to reveal the water cycle process and guide water-saving practices. Manas river basin is the largest oasis agricultural area in Xinjiang, the area of mulched drip irrigation nearly 75% of the total drainage area of irrigated area, making history of large-scale application of Drip Irrigation in agricultural production. It has a strong typical in Xinjiang even if the whole country that water usage process and agricultural water-saving mode of the Manas river basin. By eddy covariance system and remote sensing retrieval model for evapotranspiration (SEBAL), this pater measures evapotranspiration of the regional main crops and analyzes the important spatial-temporal variation of regional evapotranspiration and major factors. Conclusions below are drawn from paper:
(1) The top three dominant crops that with the largest planting area are cotton, corn and wheat, with planting areas accounting for 51.9%, 6.3%, and 1.4% of cultivate area in Manas river basin, respectively. Those three crops have distinctive characteristics on spatial distributions: cotton mainly distributed in the middle and lower reaches of plain where the illumination is adequate and temperature is high, and corn mainly distributed in mountainous alluvial plain and piedmont alluvial and diluvial plains, and wheat dispersed distributed in mountainous alluvial plain and the middle and lower reaches of diluvial plains.
(2) Evapotranspiration and its rate of cotton field under mulched drip irrigation in oasis were reached peak in the flowering stage, the phasic evapotranspiration is 248.51 mm, average diurnal evapotranspiration rate is 3.94 mm·d-1; this was followed by the values in the budding stage, the phasic evapotranspiration is 98.34mm, average diurnal evapotranspiration rate is 3.18 mm·d-1; and the minimum values were occurred in the sowing and seeding stages, the phasic evapotranspiration is 10.70 mm, average diurnal evapotranspiration rate is 1.07 mm·d-1; they are basically in line at different growth stages. In summary, evapotranspiration is 487.14 mm during the whole growing period, and average crop coefficient is 0.42. Net solar radiation is the dominant influence factor of cotton evapotranspiration over the entire growth stages. Soil temperature changes have great influence on cotton evapotranspiration at seedling stage and boll-opening stage, and temperature and wind speed at bud stage and boll-forming stages. The increase of temperature and wind speed can result in the increase of cotton evapotranspiration.
(3) Using SEBAL model retrieve crop evapotranspiration on oasis farmland in arid region is feasible. The model inversing evapotranspiration values are 1.27 mm·d-1, 2.25 mm·d-1, 4.34 mm·d-1, 4.52 mm·d-1, 3.55 mm·d-1, 3.66 mm·d-1, 3.26 mm·d-1, the ground monitoring evapotranspiration values are separately 1.23mm·d-1, 2.35 mm·d-1, 3.43 mm·d-1, 5.17 mm·d-1, 2.98 mm·d-1, 3.59 mm·d-1, 2.92 mm·d-1, relative error are between 1.9% to 26.6%; the evapotranspiration values of the whole growth period by the model inversing is 508.30 mm and the evapotranspiration values of the whole growth period by monitored is 478.03 mm, relative error is only 6.3%, which means that the SEBAL model has a higher retrieval accuracy and it can applicate in evapotranspiration relate study in this area.
Due to the restrictions of measured data of observation stations and other reasons, we only had used one year data of eddy covariance to calculate and analysis the evapotranspiration of cotton field under mulched drip irrigation, failed to carry out analysis of longer sequences. The specific coefficient of surface retrieval parameters of SEBAL model is by SEBAL model itself provided rather than recalculated by measured data, which need to strengthen confirmatory analysis by more ground measured data and strive to make the model retrieval results more accurate and stable.|