Spatial And Temporal Variation Of Farmland Water And Soil Environment And Study Of Suitable Groundwater Table Depth In The Lower Reaches Of Hetao Irrigation District

The shortage of fresh water resources and soil secondary salinization were the key factors restricting the agricultural development of Hetao Irrigation District（HID）in Inner Mongolia.The policy of water diversion from the Yellow River in HID had dropped sharply,and the contradiction between supply and demand of agricultural water has become increasingly prominent.Shallow buried underground saline water can recharge soil salt,and unreasonable irrigation can easily cause soil secondary salinization.This thesis took the Wulate Irrigation Area（WIA）in the lower reaches of HID as the research object,to explore the evolution characteristics and internal relationship of farmland water and soil environment,and the HYDRUS-MODFLOW coupling model of saturated-unsaturated zone was constructed,the suitable groundwater table depth（GTD）and irrigation amount considering spatial variation were proposed.Based on artificial neural network technology,the suitable diversion and drainage ratio of different hydrological years with the limitation of suitable GTD was predicted,in order to provide scientific basis and technical support for the efficient utilization of agricultural water resources,the safety of farmland ecological environment and the control of soil salinization.The main research contents and conclusions are as follows:（1）Based on 2a of field investigation sampling collection and monitoring test,combined with data collection,the chemical composition and genesis of groundwater in WIA were qualitatively and quantitatively studied by using statistical analysis and hydrogeochemical analysis methods.The results showed that Na⁺and Cl^-in groundwater were the main contribution indexes of TDS.The main hydrochemical types were Cl-Na type and Cl·SO-Ca·Mg type and the average TDS of groundwater increased by 22.20%after autumn irrigation.The contribution rate of each factor was evaporite（56.72%）>silicate（21.44%）>carbonate（15.80%）>atmospheric input（3.94%）>human activities（2.10%）.In the study area,the Wulashan uplift fault zone（S1）,the water-salt alternating slow zone（S2）,and the piedmont plain yellow irrigation area（S3）were mainly controlled by evaporation dominance,the groundwater in the well-canal double irrigation area（S4）and the channel recharge area（S5）in the front of the Wulashan Mountains were obviously affected by rock dominance,and ions exchange affected the chemical composition of groundwater to a certain extent.（2）Based on GIS technology,combined with 2a in-situ monitoring test,the composition of soil salinization and its correlation with groundwater environment in WIA were clarified by using cluster analysis,grey correlation analysis and information statistics comprehensively.The results showed that the study area was mainly weakly salinized soil and moderately salinized soil.The salt content in different periods was before spring sowing(2.74 g·kg^-1)>before harvest(2.65 g·kg^-1)>growth period(2.26g·kg^-1).Cl^-,Na⁺and SO₄^2-were the highest ions in the soil.The main types of soil profile were surface accumulation type,bottom accumulation type and uniform distribution type,and their salt contents were 3-9 g·kg^-1,1-3 g·kg^-1and<2 g·kg^-1,respectively.The correlation between K⁺+Na⁺,Cl^-,SO₄^2-of groundwater and topsoil salinity was high,and the correlation decreases with the increase of GTD.The influence coefficients of groundwater depth,TDS,Na⁺,Cl^-and SO₄^2-on soil salinity were 27.07%,26.13%,22.14%,20.27%and 18.34%,respectively.When considering the spatial variability,the single index of groundwater has no obvious effect on soil salinity.（3）Taking the water flux and solute flux at the junction of unsaturated zone and saturated zone as the coupling boundary,the HYDRUS-2D and MODFLOW models were coupled to construct a regional distributed soil-groundwater flow and solute transport coupling model（HYDRUS-MODFLOW）considering the spatial variation of water and soil environment.After calibration and verification,the coupling model simulation accuracy meet the requirements,and can better simulate the solute and water transport of soil and groundwater in different zones.Based on the coupling model,the suitable GTD ranges of the three typical test zones in the typical test area separately are area I(soil total salt 0.53～2.73 g·kg^-1,groundwater TDS 1.92 g·L^-1)was 156～236 cm.area II(soil total salt content 0.60～3.71 g·kg^-1,groundwater TDS 1.27 g·L^-1)was135～235 cm;area III(soil total salt 0.80～5.93 g·kg^-1,groundwater TDS 2.49 g·L^-1)is181～241 cm.（4）Based on the theory of phreatic influence layer,the critical GTD of salinization in WIA was 132 cm,which was suitable for areas with non-salinized soil and freshwater groundwater in the short term,and the critical groundwater depth of crop wilting was 229 cm,which was suitable for natural conditions without external water supply.Based on the HYDRUS-MODFLOW regional distributed coupling model,the groundwater depth ranges in the southwestern,northwestern,central,eastern and channel recharge areas of WIA were 202～242 cm,193～239 cm,206～246 cm,159～239 cm and181-249 cm,respectively.In the range of suitable GTD,0.8IR（92 mm,104 mm and 88mm）could meet most areas irrigation demand,but when the soil salinity more than 6g·kg^-1in the western and central areas of WIA,the irrigation amount should be at least1.0IR（115 mm,130 mm and 110 mm）.（5）Based on the artificial neural network model,the nonlinear relationship between the GTD and the ratio of drainage and diversion,evaporation and precipitation was constructed,and the constructed machine learning model was screened.The BP neural network model,support vector machine model and random forest model with better simulation results were used to determine the suitable drainage ratio of WIA in different hydrological years based on the suitable GTD（206 cm）of the WIA:wet year3.24,normal year 5.96,dry year 7.34.