| The contradiction between water supply and demand in the Heihe River Basin(HRB)is increasingly prominent,and the ecological environment in the lower reaches of HRB is seriously degraded.Unified dispatching of water resources has been implemented in the HRB since 2000 by the State Council.The agricultural water diversion from Heihe River will be reduced from 1.67 billion m3 to 1.4 billion m3.To solve the problems of water resources decreasing for agriculture,improving the agricultural water use efficiency is an important measure to ensure the grain production security in the middle oasis of HRB.Affected by climate variation,land use and water resources,agricultural water consumption is spatio-temporal dynamic.Besides,the scale for the agricultural efficient water regulation has been extended from farmland to region and even to basin.Therefore,the key to improve the efficiency of irrigation and regional water resources allocation is to further quantify the spatio-temporal distribution of main parameters in the irrigation districts under changing environment,comprehensively consider the synergistic effect of environmental factors and the impact of water-saving measures,and establish the distributed crop model and spatio-temporal optimization model of water demand.AquaCrop model was taken as the tool in this paper,and a distributed crop model,AquaCrop-RS model based on spatial heterogeneity,of different scales by combining field experiment and remote sensing data was established.On this basis,the production and distribution of water consumption of corn in the middle oasis was analyzed.A distributed irrigation decision support system(DIDSS),based on 3S technology,by further combining multi-objective optimisation model and the genetic algorithm NSGA-?,was constructed.The irrigation schedule was spatially optimized with DIDSS and the agricultural water use efficiency was improved in the middle of oasis of HRB.The WUE is the economic yield produced by per unit water,which was divided into WUEET with the ratio of the yield to the evapotranspiration and WUEI with the ratio of the yield to the irrigation volume.The main results are outlined as follow:(1)Remote sensing data was used in performing the inversion of parameters of fertility stress(mainly maximum canopy coverage CCx and relative biomass Brel),and the agronomic management parameter(sowing date).The distribution of CCx and Brei showed significantly spatial and temporal heterogeneity,and the spatial distribution of Brel was related to the distribution of soil texture to some extent.The inversion of sowing date showed relatively high accuracy with remote sensing.Statistics showed that the estimation error of sowing date within 5 days accounted for 64%,69%,and 68%of the sampling points for 2012,2015,and 2016,respectively.The Root-Mean-Square Error(RMSE)was approximately 5 days,the Mean Absolute Error(MAE)was approximately 4 days,and the maximum error was within 10 days of sowing date.The inversion accuracy of 2015 was the highest and that of 2012 was the lowest.which could be attributed to the small size of samples and the restriction of sampled areas for 2012.Statistics showed that the sowing date of maize was mostly distributed between 79th?122nd day of the year.(2)Distributed AquaCrop-RS model for different scale was constructed with the spatial crop parameters and agronomic management parameters incorporated into AquaCrop-GIS model.The simulation accuracies for evapotranspiration and yield of maize in the irrigation district and the middle oasis were all improved.Findings showed that the AquaCrop-RS model had an increasing simulation accuracy by 20%for evapotranspiration,and 26%?36%for yield in the irrigation district,respectively.In the middle oasis,the simulation accuracies for evapotranspiration were improved by 21%,5%,and 26%for AquaCrop-RS1?AquaCrop-RS2?AquaCrop-RS3,respectively.AquaCrop-RS1 model had a small effect on yield simulation accuracy;however,the AquaCrop-RS2 model and AquaCrop-RS3 had considerable improvements by 34%to 68%and 35%to 72%,respectively.Spatial crop parameters played an important role in improving the regional simulation accuracy of yield,and the introduction of spatial sowing date had a considerable effect on the simulation accuracy of evapotranspiration for the AquaCrop-RS model.(3)AquaCrop-RS model was used to analyze the production and distribution of water consumption of corn in the middle oasis.The leakage in the upper reaches of the middle oasis was between 58 and 281 mm.and that in the middle and lower reaches was between 282 and 534 mm.accounting for approximately 40%of the irrigation volume.The WUET of maize in the middle oasis was between 2.0 and 2.3 kg/m3,and the spatial difference was insignificant.However,the spatial variability of WUEI was obvious,it was generally between 2.09 and 3.35 kg/m3 in the upper reaches of the middle oasis,and between 1.24 and 2.08 kg/m3 in the middle and lower reaches.The irrigation volume decreased by 205?371 mm.yield increased by 1.0%?-7.7%,and WUEI increased by 39%?90%on the average in the middle oasis for deficit irrigation with surface irrigation;and that with drip irrigation,the irrigation volume decreased by 277?412 mm.yield increased by 1.1%?-8.1%.WUEI,increased by 62%?109%.The effects of the two deficit irrigation methods on WUEET were not significant.(4)A distributed irrigation decision model,based on 3S technology,was constructed and then was used to optimized the irrigation schedule,which realized the reasonable spatio-temporal distribution of agricultural water resources for maize cultivation area in the middle oasis,and improved the regional agricultural water use efficiency.The irrigation quota of maize in the middle oasis after optimisation decreased by 0?600 mm,and the leakage decreased by 50-450 mm on the average.The yield increased by 1%?5%except some areas with a decrease between 0 and 5%in the upper reaches.WUEl in the upper reaches increased by 24%?50%and that in the middle and lower reaches increased by 101%?150%.WUEET generally increased by 1%?6%in the middle oasis.(5)The distributed irrigation decision model was used to optimize the irrigation schedule for maize in the normal year,constructing optimized data sets of irrigation schedule.The opt:imized irrigation quota was 291?455 mm for different irrigation districts.Influenced by irrigation amount and climate,especially by wind speed,the evapotranspiration showed obvious regionality.The evapotranspiration in Ganzhou District was the largest,ranging from 626 to 650 mm,from 569 to 591 mm in Gaotai County and from 531 to 550 mm in Linze County.The spatial differences of WUEI and WUEET in the area were also significant,and WUEI was generally larger than WUEET-WUEI was mainly distributed between 2.79 and 4.53 kg/m3,and WUEET was mainly between 1.93 and 2.45 kg/m3. |
PDF Full Text Request