Spatial Variation And Autocorrelation Analysis Of Soil Moisture Physical Characteristics In The Yellow River Irrigated Area Of North Henan

As one of the high-standard breadbasket in China,the area in the middle and downstream of the Yellow River plays a vital role in ensuring food security.The area is located in the hinterland of North China,with insufficient annual precipitation and uneven temporal and spatial rainfall distribution.Therefore,irrigation is the primary measure to ensure a high grain yield in this area.To achieve precision irrigation in the farmland and make appropriate agricultural water management decisions regionally,the spatial heterogeneity and correlation of soil moisture physical characteristics is an inevitable factor to consider.To quantify the spatial relationship of soil moisture physical characteristics in the irrigation areas of the middle and downstream of the Yellow River,we did a case study at the typical farmlands in the yellow river irrigated area of north Henan.The spatial variability,the correlation and distribution pattern of soil clay,silt,sand,saturated hydraulic conductivity,soil aggregates,soil water retention curve and soil moisture at field scale were analyzed and discussed using classical and spatial statistics.Also,the reasonable sampling number of each variable was calculated with the theory of interval estimation.The results were as follows:(1)The spatial variation: Soil physical properties such as soil clay,silt,sand,bulk density,soil moisture,and soil water retention curve mainly showed weak variation in the study area,while saturated hydraulic conductivity showed substantial spatial variability.The spatial variation of soil aggregates was particularly medium,and the spatial distribution of aggregates with small particle size showed equilibrium with less heterogeneity,But the spatial distribution showed significant variability for the soil aggregates with large particle size,The saturated hydraulic conductivity followed a lognormal distribution in space,while the other variables fit the normal distribution.In addition,most of the soil physical properties in the study area showed medium spatial dependence,and their maximum correlation distance fluctuated and varied with the study scale and variable attributes.Only the maximum correlation distance of soil clay,silt,sand and bulk density was relatively stable,and its size mainly was 1/3 of the max distance of the study scale.(2)Spatial autocorrelation: It showed a positive correlation among most soil moisture physical characteristics in global space or distance space,and cluster distribution characteristics in space.However,soil porosity(f),saturated hydraulic conductivity(Ks)and the soil aggregates at the soil depth of 20-40 cm mainly showed negative correlation in small scale(area),and the distribution showed the anisotropy characteristics.These variables with negative correlation in space showed pure nugget effect in their variogram analysis with no effective range.From the significance p-value of the test statistics,we concluded that each variable’s spatial autocorrelation would significantly enhance to a very considerable level(P < 0.01)when the sampling frequency increased.It showed that the spatial autocorrelation of variables covered up with varying degrees when the sample size decreased and the sample distance increased.Therefore,to effectively quantify each variable’s spatial relationship,the arrangement of sampling spacing must be considered according to the sampling area.(3)The reasonable sampling number: the actual sampling number of soil clay,silt,sand,bulk density,soil aggregate,soil moisture and soil water retention curve met or exceeded the reasonable number calculated by interval estimation theory.But for soil hydraulic conductivity,the appropriate sampling number under the selected confidence and deviation levels did not meet,and the gap was significant.In addition,the reasonable number of samples for the variables with medium variation in space decreased significantly with the decrease of sampling area except for the soil particle size attributes as it did not change considerably with the scale.The results showed that the layout of the sampling strategy for soil moisture physical characteristics at the field scale could be adjusted appropriately according to the reasonable sampling number of calculated variables.