| Under the influence of regional climatic conditions and large-scale vegetation construction,dried soil layer(DSL)has widely distributed in the Loess Plateau of China and caused a series of ecological and environmental problems,which threats the ecosystem health,the sustainability of artificial vegetation,and achievements of soil water conservation.Studies on the spatiotemporal dynamic characteristics of DSL are beneficial to comprehensively and accurately understanding the development of DSL,and to provide scientific basis for the sustainability of artificial vegetation,and the continuous improvements of the ecological environment in the Loess Plateau of China.This study focused on the DSLs that in a typical north-south transect of the Loess Plateau,and conducted a long-term and high frequent survey of DSL and the related environmental factors.The dynamic characteristics and its dominant controlling factors of DSL were investigated based on classical statistics,geostatistics,principal component analysis,redundancy analysis and partial redundancy analysis at a regional transects scale.On this basis,the indirect and direct prediction model of DSL were established through multiple linear regression and artificial neural network.The main results of this study were summarized as follows:(1)On the scale of the south-north transect of the Loess Plateau,the content of clay,silt and sand in the soil showed a clear zonal distribution pattern with the increase of latitude.In the transect,the variation of soil texture along the horizontal direction was greater than the vertical direction,and the degree of spatial variability in clay and sand content was higher than that of silt content.Soil organic carbon content generally showed a decrease trend from south to north of the transect and generally decreased with increasing depth.Soil organic carbon content maintained a low level with no obvious trend in soil layers below 1.5 m.In the 0~0.4 m soil layer,the saturated hydraulic conductivity showed a trend of decreasing first and then increased with increasing latitude.Compared with saturated hydraulic conductivity,bulk density showed a similar trend with the increase of latitude and the trend of saturated soil water content was the opposite.Trends in field capacity and permanent wilting point along the transect were similar and generally decreased with the increase of latitude,and only slightly higher in the northern Loess Plateau.(2)In different seasons,the average soil water content in the soil profile of the transect were similar and the spatial similarity of soil water content in spring and winter was greater than that in summer.The distribution soil water content in different seasons showed obvious zonality pattern at a regional transect scale and generally showed a trend of decreasing from south to north with the increase of latitude and only rose to some extent at the end of the transect.The differences in profile soil water content in different seasons were mainly existed in the surface layer 0~1.6 m and the soil water condition in autumn was the best.Generally,the soil layer that mainly participating in the water exchange process of the Soil-Plant-Atmosphere-Continuum(SAPC)in the Loess Plateau region was 0~2 m layer.The thickness of the soil water active layer(WALT),the coefficient of variation within the soil water active layer(WAL-CV),and the average soil water content with in the soil water active layer(WAL-SWC)all showed a quadratic distribution with increasing latitude.The WALT in the central Loess Plateau was generally thicker and some sites even exceed 4 m.The spatial distribution of soil water active layer was the result of the joint control of climate factors and local factors,among which,the contribution of climate factors is higher.(3)In different seasons,the spatial distribution of the DSL in the transect had certain similarities.The DSLs were mainly distributed in the central and northern parts of the Loess Plateau,and the DSL of some sampling sites almost penetrated the entire 0~5 m soil profile.On an average level of the study period,the distribution area of DSL accounted for 28.38% of the transect and with an average of formation depth and thickness about 1 and 3 m respectively.In different seasons,summer had the largest distribution area of DSL,and the distribution area was the smallest in autumn which indicate the water condition was the best in autumn.The changes in the DSL indicators showed a certain periodicity over time,but the increase and decrease rules between the indicators do not exactly coincide in time.The change of DSL thickness with time had a certain lag compared with the other two indicators.There was a difference in the degree of temporal variation between the DSL indicators.Overall,the DSL-SDI had the largest degree of variation over time,and the temporal variation in DSLT was the smallest.(4)The distribution sections of DSL in profile and the degree of desiccation in DSL varied with different land uses.The DSL of cropland was the shallowest and mainly distributed in the depth range of 0.12~1.38 m at different time.The profile distribution of DSLs in forest land and grassland were similar,and they all ranged from 0.32 to 3.70 m.The degrees of DSL desiccation under different land use and different times generally followed a decrease trend of cropland,forestland,and grassland.The increase and decrease of the DSL thickness were usually performed at the same time on the upper and lower boundaries.However,the acceleration and deceleration of the upper and lower borders were related to the type of land use.Changes in the boundaries of cropland were usually slower than the lower boundary and the situation in farmland and grassland were the opposite.For the various DSLs indicators in different land use,the variation degree of soil desiccation index with in DSL with time was higher than that of the formation depth of DSL and thickness of DSL.(5)Under extreme conditions,there was a significant difference in the distribution of DSLs and the minimum of the distribution area of DSL in the transect was only 53.8% of the maximum value.Thus,long-term sequence observations were very necessary for the study of DSLs.Permanent DSLs were mainly distributed in the central and northern regions of the Loess Plateau.In consideration of the poor self-recovery ability and serious negative effects,permanent DSLs should be given more attention in the management and repair of DSLs.To reduce its negative impact of permanent DSL on the ecological environment of the Loess Plateau and improve the effectiveness of DSL treatment measures,the measures and efforts to regulate and control DSL should mainly focus on the regions where permanent DSLs mainly distributed.Measures related to local factors,such as changes in land use,optimal selection of vegetation types,reasonable planting densities,and micro-terrain reconstruction,should be strengthened in dry layer restoration measures,and this is the key to repairing permanent DSL.(6)In the simulation of DSL indicators,a high accuracy was obtained by the indirect prediction methods,which predicting the soil desiccation index first and then calculating index of the DSL.Therefore,in the Loess Plateau of China,indirect prediction of DSLs was an effective method for DSL simulation.In addition,regardless of whether use indirect or direct methods to predict DSL indicators,artificial neural networks showed stronger ability to describe complex relationships between independent and dependent variables.Therefore,in the simulation of soil desiccation index and DSL indicators in the Loess Plateau of China,the application of artificial neural network was conducive to improving the simulation accuracy of the model,and its prospects and advantages were more significant than multiple linear regression.Based on the current understanding of the spatial distribution and the influencing factors of DSL,this study fully understand and analyzed the spatial-temporal characteristics of DSLs at a regional transect scale,and clearly pointed out that permanent DSL in the middle of the Loess Plateau should be the key content of DSL regulation and control.Besides,this study also founded that artificial neural network combined with indirect prediction method was the effective way of DSL prediction.These results will help to promote the long-term sustainability of vegetation construction,improve the efficiency of DSL regulation measures,and accurately grasp the serious situation of prevention and control of DSLs in the Loess Plateau.Finally,these results will provide an important scientific basis for the scientific management of soil water and the sustainable advancement of regional ecological environment construction in the Loess Plateau of China. |
PDF Full Text Request