The soil moisture regimes of broad beens attach great importance to dry matter formation under the flowering and pods formation stage. It is meaningful to conduct research on the effects of different soil moisture conditions on the physical and ecological characters of broad beans, as well as the process of adaption to moisture based on the physiological characteristics of broad beans. A pot experiment was conducted to monitor the photosynthetic rate, transpiration rate, water use efficiency and stomatal conductance of broad beans. Following an analysis of the light response characteristics of photosynthetic efficiency of broad beans under different soil water conditions. The research reveals that under the optimum water content of broad beans, the photosynthetic parameters keep at a relatively high level so that they can grow in a normal condition. In contrast, when the water content is either high or low, the stomata conductance of the broad beans tends to decrease as well as the transpiration rate. Consequently, the normal growth of the board beans will be affected resulted from the decrease of photosynthetic rate. It is suggested that a normal soil moisture content of the board beans should be maintained in 28.06%~30.14%, and the best soil moisture content is approximately 29%. The suitable photosynthetically radiation intensity of photosynthesis is 600~1600μmol/(m2?s1).

Rice has been planted in our country for more than 7 000 years and is a staple food crop． Nearly half of world population feed on rice，but rice is a kind of hydrophillic crops，its irrigation water consumption reaches 18 000 m3 /hm2 ，the high water consumption causes serious water resources shortage，and the crop area is drastically reduced． Tongliao city agriculture still has a large proportion of area adopting the traditional irrigation methods． Rice irrigation quota is relatively high and there still exists serious water resources waste phenomenon，so agricultural water-saving has great potential． In order to solve the bottleneck of Tongliao city water resources shortage，taking the main water -consumption crop，rice，as the research object，through test analysis and data collection，the area of the soil bulk density，the ph value，the irrigation system and the crop yield are studied to seek the optimal irrigation schedule and water productivity for rice under drip irrigation system to provide guidance for hilly rice planting．

Based on the local irrigation and fertilization model，in order to explore the suitable irrigation and fertilization system for spring wheat in Hexi Corridor，the field experiment was carried out in Yongchang County，Gansu Province． The experiment was conducted with three irrigation levels ( 3 300，2 550，1 800 m3 /hm2 ) and fertilization levels ( 337．5，225，187．5 kg /hm2 ) ． The distribution law of NO－ 3 －N in soil among different treatments under sprinkler irrigation was studied，and the correlation between yield，water use efficiency and quantity of irrigation，fertilization was analyzed． The results show that the content of NO－ 3 －N in the soil is mainly affected by the amount of fertilizer applied; the NO－ 3 －N is mainly distributed in 0～ 40 cm soil layer under the condition of sprinkler irrigation，which is beneficial to the uptake and utilization of nitrogen in spring wheat roots． The yield and water use efficiency are linearly related to the amount of irrigation． However， the correlation between the yield and water use efficiency is not significant． The medium irrigation level is 5% higher than that of the flood irrigation，the water is saved by 32%，and the water and synergistic effect is remarkable． According to the yield regression results，the irrigation and fertilization system for irrigation and fertilization in Hexi Corridor is the middle irrigation and fertilization level．

The first drip irrigation cotton fields were studied in the shihezi area of xinjiang the water diffusivity in silt loam with different depths has been studied by horizontal column infiltration at Indoor laboratory．It was improved by the result that with the decrease of water content，Boltzmann parameters has been increased，and the phenomenon of growth largely restricted at 40~60cm depth soil．At the same water content，the highest level of soil water diffusivity could be found at 40~60cm depth and there has no?significant difference at other depth．The numerical relationship could be shown among three depth is that the 40-60cm one is greater than the 20~40cm one，the 20~40cm one is greater than the 0~20cm one．Furthermore, the parameters were fitted and verified with RETC shows the root mean square error has been shown significant difference between simulation values and measured values．It could be found in the result that BCM model show a great expression for the soil moisture diffusivity of different depth of silt loam．

The triangular profile weir is often used in rectangular and trapezoidal channel，so the numerical simulation analysis of its use in U-shape-channel is not deep． In order to study the applicability of U channel triangular profile weirs and the accuracy of the numerical simulation，the AUTOCAD and GAMBIT software are used to establish U channel triangular profile weirs physical model，the FLUENT6．3 is used for simulation calculation and the Tecplot is used for post-processing so as to obtain the water level，the velocity distribution and the relationship between water level and discharge under different conditions． The results show that the actual values is close to the simulation values，the average relative error of flow rate is 0．21%，the average relative error of level is 6．02%，the relationship between water level and discharge are in line with the exponential function． Therefore，the flow measurement of U channel triangular profile weir has good accuracy， certain reliability and practicality． The study can provide scientific basis for irrigation water distribution to improve the scientific management level of water resources．

At present, there are many implementation departments of farmland water conservancy projects, these departments can not carry out unified planning in the process of project renovation, and lack of program optimization and comparison, and it is difficult to achieve an optimal status in terms of engineering design and planning, which results in waste of water resources. At the same time, the unsound formation mechanism for agricultural water pricing and the low overall pricing of water resource, lesd to the extensive use of agricultural water and disrepair of water conservancy projects. In order to address this issue, this study proposes a nonlinear optimization model for irrigation channels involving rectangular, trapezoidal and U-shaped sections. In this new model, the variables are dependent on several factors including section size, channel length, the objective function is established by the lowest cost of irrigation water supply, the constraints are defined using the quantities of flow, flow rate and aspect ratio. The implementation of this model is based on the optimized programming of a genetic algorithm. The proposed method is validated using realistic project examples, and results indicates that the new model can efficiently and effectively determine the overall optimization scheme for meeting the requirements of farmland irrigation and water saving, as well as reducing the cost of water supply and providing reference for setting reasonable water pricing.

A dissolved and mixed water fertilizer integrated device is designed，which consists of a water supply module，a control module，a dissolving module and the auxiliary equipment． So the solid fertilizer dissolving in water and the solution outflowing from the device can be operated simultaneously． The experimental study on key structures of the device and the effects of fertilization methods on the fertilization uniformity is conducted． The results show that the device runs in good condition，and the operations of feeding，dissolution and fertilization can be performed simultaneously; under the experiment conditions，more than 90% of the fertilization can be completed within 40 minutes; the uniformity coefficient of outlet solution is 41．5% when 1 kg fertilizer is directly added into the storage barrel and 1．5 kg fertilizer is added into control model，which is more than that of other fertilizer distribution mode by 3．2% and 10．8%，respectively． Therefore the device can realize quantitative and uniform fertilization by changing fertilizer distribution mode．

In order to study the reference crop evapotranspiration ET0 of Hebei Province，based on Extreme Learning Machine Model ( ELM) and Generalized Regression Neural Network Model GRNN，the ET0 values of 1961 to 2015 are estimated in this paper． By using the Penman -Monteith model P－M as a standard，the model calculation results are compared with the Hargreaves model H－S． The result shows that: the accuracy of ELM model and GRNN model for ET0 daily and monthly value is higher than that of H－S model; the RMSE values of different models show GRNN＜ELM＜H－S as follows． It indicates that the computational accuracy of the machine learning model is higher than that of the H－S model，while the GRNN model has the highest computational accuracy.

Aiming at solve the problem of pipe diameter selection in a mountainous gravity drip irrigation pipe network，a mathematical model for optimization of the gravity tree irrigation pipe network was established，which regarding the cost of pipe network as objective function and the pipe standard diameter as decision variable，at the same time，it must be satisfied these constraints: irrigation water volume，water pressure，and flow rate etc． After that，sought for a pipe diameter design with minimum investment and maximum reliability when the pipe layout had been determined． By using genetic algorithm based on integer coding to solve the model． Then the simulated annealing penalty function was used to deal with these constraints，so that the good local optimization ability of simulated annealing and the global search ability of genetic algorithm were organically combined to realize the improvement of algorithm． Compared the optimized results of genetic algorithm with the results of economic flow method，the investment of pipe networks decreased by 27 280 yuan，just account for 86．9% of the pre－ optimization，and the utilization rate of the pipe head was increased from 65．38% to 97． 61%． The example results show that: improved genetic algorithm are superior to empirical methods both in terms of economy and gravity head utilization; the mathematical model and the genetic algorithm are simple，easy to implement and have good optimization performance and efficiency，it can be used for actual production． 

In order to study the comprehensive water saving technology based on ET management，combining engineering water -saving， agricultural water saving and management water saving，four different combination irrigation modes for corn in Yaotou village of Baiwen Town，Lin County are selected，the regional water balance method is adopted to analyze the change of ET of different water deprivation，crop yield and crop water productivity and benefit under different modes． The results show that the selected“straw counters-field coated by furrow” mode is an ideal planting mode． Under the ET management comprehensive water -saving measures，the “straw counters -field coated by furrow”mode can be widespread used in Shanxi and other appropriate area． At the same time，it has important guiding effect to promote regional agriculture demonstration and sustainable economic and social development．

The agriculture irrigation water use efficiency is an important basis for evaluating the utilization of agriculture irrigation water. The improvement of irrigation water use efficiency is the key to enhance the comprehensive strength of national agriculture and ensure national food security. It is also the basic requirement and focus of building water saving society. In this paper, we selected the data of 13 regions in Heilongjiang Province from 2013 to 2015, and we estimated and analyzed the agriculture irrigation water use efficiency by DEA and Tobit model. The results show that: (1) DEA model with weight constraint is more in line with the actual situation of agricultural irrigation water in Heilongjiang Province. On the one hand, it reduces the large scale high efficiency problem in the traditional DEA model. On the other hand, it avoids the problem that the input element weight value is zero; (2) Agricultural irrigation water use efficiency and precipitation, capacity, grain crop area accounted for the proportion of the total sown area, per capita GDP was negative effect, and it and the temperature, the number of reservoirs, waterlogging area was positive effect, according to this,we make a suggestion about improving the agriculture irrigation water use efficiency in Heilongjiang Province.

Based on the time series data of Xinjiang, this paper uses Tapio model to study the decoupling relationship between agricultural water resources utilization and agricultural economic growth, and researches the effect decomposition of water resources based on the LMDI decomposition model. The relationship between agricultural water resources utilization and economic growth in Xinjiang displayed weakly decoupling stage, the unstable stage and the strongly decoupling stage, and there were some differences in the decoupling state in Xinjiang each basin. The comprehensive effect of agricultural water resources utilization in Xinjiang was positive effect, the contribution of scale effect wasnot obvious, the benefit effect was positive effect, the efficiency effect was negative as a whole, which indicated that the increase of the output value per unit area expanded the utilization of agricultural water resources, the reduction of water consumption per unit of agricultural output greatly conserved the utilization of agricultural water resources. Therefore, improving the utilization efficiency of agricultural water resources, and reducing the water consumption per unit of agricultural output were the main objectives of water conservancy work in Xinjiang.

Cooperation in the utilization of water resources restricts the implementation of Beijing -Tianjin -Hebei region coordinated development policy． According to the regional economic development level and the cost of compensation for protecting water ecological environment，the countermeasures for solving Beijing-Tianjin-Hebei regional problem of water environmental imbalance is studied in this paper． Based on virtual water footprint model，the water footprint and spatial change of water surplus/deficit of Beijing-Tianjin-Hebei region are studied and an ecological compensation standard model is built to quantitatively calculate ecological compensation amount between the three provinces of the Beijing-Tianjin-Hebei region． the urgency of reducing the water footprint ( i． e．，the total virtual water footprint) is evaluated and the focus is adjusted to provide basis for establishing water security system of Beijing-Tianjin-Hebei． On this basis，the Beijing -Tianjin-Hebei water resources cooperation compensation mechanism is put forward to provide the reference for the co -construction and sharing of ecological environment of Beijing-Tianjin-Hebei．

The selection of sample irrigation is of great significance to the statistic of water consumption of agricultural irrigation． According to the technique methods mentioned in the“the statistical scheme for total water consumption ( try out) ”，the water consumption of lager users are measured one by one，and general water users consumption are measured by sampling investigation，and then the irrigation area water consumption is predicted comprehensively． Combining with the basic situation of each provincial irrigation area，the method and layout of sample irrigation selection about irrigation water use statistics are determined． The selection and layout of sample irrigation of various provinces include the number determine of sample irrigation at the provincial level，the number allocation of sample irrigation and the layout of sample irrigation in the irrigation subarea． The number of sample irrigation at the provincial level is determined according to“the statistical scheme for total water consumption ( try out) ”and the related statistical theories． The number of sample irrigation is distributed proportionally according to the number weight of irrigation areas in the subarea． According to the information system for irrigation district，the sample irrigation layout is determined through the non-random sampling method．

Bayesian theory is a mathematical statistical method based on probability． In this paper，the Bayesian theory is introduced into the groundwater quality assessment system to establish Bayesian groundwater quality evaluation model． A total of 8 water quality data of the zone from Dushanzi to Shawan counties are selected for example application． At the same time，the evaluation result of Bayesian model is compared with those of the single factor evaluation method，the F value method and the fuzzy comprehensive evaluation method to verify its accuracy． The results show that the evaluation results of Bayesian model are similar to those of the fuzzy comprehensive method，and the calculation process is simpler． So it is worthy of further application and verification．

Based on water quality monitoring data in three periods ( abundant water period，smooth water period，dry water period) of Xinghai lake in14 sampling stations，the Bayesian evaluation method is used to evaluate water quality status and analyze the main influencing factors in order to provide advices for the reasonable development and utilization，water environmental protection and comprehensive control of Xinghai lake． The results show that: ① the water quality of 79．9% of monitoring points reaches Ⅴ grade; the maximum probability value， Pk，is high and greater than 50% except for D7 and D14; the results are feasible，the water pollution is serious and forecasted to be increasing，so it is needed to strengthen the comprehensive control． ② TP，TN and CODMn are the main pollution factors of Xinghai Lake， so they are the governing factors for the control of Xinghai lake water quality．

Based on the theory of water footprint，this paper constructs the accounting formula of livestock water footprint and analyzes the related data of livestock industry of 2016 in Xinjiang． It is found that the order of the water footprints of livestock and poultry production per unit quantity from high to low is: big livestock ( 14．15 m3 /kg) ，sheep ( 8．73 m3 /kg) ，goats ( 6．23 m3 /kg) ，poultry ( 3．76 m3 /kg) ，pigs ( 3．08 m3 /kg) ; the total water footprint of Xinjiang animal husbandry is 185．27×108 m3 ，in which the blue water，green water，grey water footprint is 12．42×108 m3 ，106．86×108 m3 ，65．99×108 m3 ，respectively; the proportion of grey water footprint is 35．62%，and the stress of animal husbandry on environmental protection is 3 times that of farming; the water footprint of livestock and poultry production is: big livestock ( 85．79water footprint; water for animal husbandry; Xinjiang) ，sheep ( 62．88×108 m3 ) ，poultry ( 16．44×108 m3 ) ，pigs ( 15．27× 108 m3 ) ，goats ( 4．88×108 m3 ) ，respectively; according to the proportion of grey water footprint，the order of the pressure of livestock and poultry on environmental protection from high to low is large livestock，pigs，poultry，sheep，goats．

Under the background of the sharp contradiction between supply and demand of water resources, the current research focus will be the ultimate carrying capacity of water resources and the equilibrium point between economic development and water resources. Yulin City was selected as the research object. The forecast of its water supply and demand made an analysis of its maximum economic scale that can be carried and its maximum population size. They are 552.8 billion yuan and 6.03 million people in 2020, and 1,924.4 billion yuan and 7.44 million people in 2030. The multi-objective optimization model is constructed on this basis to optimize the industrial structure in 2020 and 2030. The total national economic output and total water consumption of Yulin City are 486,0 billion yuan and 1,142 million m3 in 2020 and 1065,9 billion yuan and 168.34 million m3 in 2030. The proportion of industrial structure were 4.6: 53.7: 41.7, 3.6: 47.9: 48.5 in 2020 and 2030 respectively. The water resources can meet water requirements.