Analysis Of The Spatial-temporal Variability Of Gravel-sand Mulched Field Soil Salinity

Study on temporal and spatial variability of soil salinity, salinization condition and the trend is meaningful for guiding agricultural production. According to gravel-sand mulched field located in northwestern arid region, this article chose bare land(CK) and gravel-sand mulched field(GSM) with different planting years and sampling points were determined through GPS in vacant land and there were 64 sampling points. The sampling points were sampled respectively at the layer of 0~10cm; 10~20cm, 20~30cm and 30~50cm. combing statistics and geostatistics and analyzing spatial distribution pattern, time stability, temporal and spatial variation and reverse succession characteristics of soil salinity of CK and GSM with planting years and achieving following results.(1) Soil mean salinity of GSM is lower than CK, There are significant difference among soil salinity at different layer of GSM and CK. The variation of soil salinity of layerⅡ,Ⅲ,Ⅳ(CK) and MGM belong to lower variability, while others belong to moderate variability. The best semi-variance theory model of layerⅡ(CK) is linear model, layerⅠ,Ⅲ(CK), layerⅢ(NGM), layerⅠ(MGM) and layerⅠ,Ⅱ,Ⅲ(OGM) are exponential model, while others are spherical model. The spatial auto-correlation of soil salinity of layerⅡ(CK) is lower while others are high.(2) As for all sample points, soil salinity and variation coefficient increase with the increase of soil depth, and the spatial variability is weak and moderate. The semi-variance models for soil salinity varied with scales. This meant that the semi-variance models for soil salinity in XXL scale(32m×32m) could be simulated by Gaussian and exponential models; which in XL scale(28m×28m) could be simulated by Gaussian and spherical models; which in L scale(24m×24m) could be simulated by Gaussian, exponential and spherical models; which in M scale(20m×20m) could be simulated by Gaussian models; which in S scale(16m×16m) could be simulated by Gaussian and exponential models. A strong and significant spatial correlation is found in soil salinity in these every scales.(3) Mean soil salinity of each period belonged to moderate lower variability and the correlation of mean soil salinity and coefficient of variation were high, and there were visible positive correlation relationship existed in soil salinity and soil depth. The trend of soil salinity cumulative distribution curve were similar, and there were little difference in simulated value and measured data and space distribution basic trend equated measured value and the trend of semi-variance was significant. The temporal stability of soil salinity decreased and discrete degree of relative deviation between soil salinity and the mean value of each measuring point increased with the increase of soil depth.(4) The results of the temporal stability preliminarily determined the representative sites. The relevance of representative measuring points value and mean soil salinity are high and the difference is small. The mean soil salinity of each period belong to moderate lower variability, and the soil salinization is bare land(CK) > old gravel-sand mulched field(OGM) > new gravel-sand mulched field(NGM) > middle gravel-sand mulched field(MGM). The evolvement of gravel-sand mulched field soil salinization have happened with the increase of planting years. The soil salinity of MGM decreases and reaches to minimum while OGM increases with the increase of planting year and reversal evolvement of soil salinization fades gradually.