水土流失治理对土壤水文物理特性的影响

发布时间：2019-05-16 17:51

【摘要】：红星农场隶属于黑龙江农垦省北安分局,地势为南高北低,中部和西部均为缓坡且坡长较长。该地区水资源补给主要是降雨,地表水利用率低,存在季节性干旱、渍涝与水土流失并存的问题。水土流失将导致坡耕地黑土层逐年变薄,地力减退,土壤肥力下降,作物产量明显下降等不良结果,因此该地区农业生产受到降雨制约和水土流失严重的问题。针对以上问题,于2013年5月至9月在红星农场径流小区内选取了鼠道(SD)、鼠道+暗管(SA)、鼠道+垄向区田(SQ)、鼠道+暗管+垄向区田(SAQ)四种不同措施处理模式,并设顺坡耕作方法设计对照小区,即常规处理(CK)。研究了不同水土保持技术模式对不同土层深度、不同坡位的土壤水分、土壤容重、土壤持水性能等土壤性质指标的影响。以及多年水土流失治理后土壤团聚体、土壤有机质和土壤三相比的变化。并对不同技术模式条件下的表层土壤水分进行了模拟研究,提出了适合本地区的土壤水分预测模型。研究结果表明:(1)“鼠道+暗管+垄向区田”组合措施的除渍和保水效果均为良好。其他几种模式的除渍效果也优于常规。无论是坡上还是坡下,土壤水分的变异系数按数值大小的排列为SAQSASDSQCK。且这5种措施的土壤含水量变异系数Cv和土层深度的关系为土层越深,其变异系数值越小。即表层土壤含水量的变异系数比深层大。(2)各措施处理土壤容重均随着土层深度的增加而呈逐渐增大,但不同的组合模式之间容重变化有差异。从不同措施处理的土壤容重平均值来看,“鼠道和垄向区田组合”、“鼠道”、“鼠道和暗管组合”,“鼠道、垄向区田和暗管”组合较常规分别减少了2.28%、1.54%、2.29%、3.37%。(3)从土壤水分特征曲线可以看出,“鼠道+暗管+垄向区田”组合的持水性能最好,各措施处理的持水性能依次为鼠道、暗管和垄向区田组合模式(SAQ)、鼠道和暗管组合模式(SA)、鼠道和垄向区田组合模式(SQ)、鼠道(SD)、常规(CK)。且各措施处理随着土层深度增加,其持水性能越差。(4)通过六年的水土流失治理后,不同处理措施下的土壤有机质含量和常规处理比较发现,“鼠道、暗管和垄向区田”组合模式”的治理效果最佳。所研究的4种水土保持措施及组合均对土壤有机质有正面作用。差别程度因措施的不同而不同。“鼠道+暗管+垄向区田”组合模式”的有机质含量高于“鼠道+垄向区田”组合模式”,“鼠道+暗管”组合模式”的有机质高于“鼠道”处理。(5)将计算得到的土壤团聚体稳定性评价指标进行对比研究发现,不同的处理措施对土壤团聚体的稳定性亦产生了一定影响,和常规处理比较,其他4种水土保持模式的土壤团聚体稳定性均较高。其稳定性由大到小依次为“鼠道+暗管+垄向区田”“鼠道+暗管”“鼠道+垄向区田”“鼠道”“常规”。四种措施均可有效减少水土流失。有机质含量高的处理措施,其土壤团聚体稳定性也较好。但二者却并不是显著的相关关系。(6)不同处理的固相容积的体积随土层的加深而逐渐增大,而空气体积随耕层的加深而呈现下降的趋势,“鼠道”、“鼠道+垄向区田”、“鼠道+暗管”、“鼠道+垄向区田+暗管”模式的土壤固相容积均明显小于常规。说明各措施处理均有改善土壤通透性的作用。(7)将GM(1,1)-AR模型和AR模型模拟出的预测值对比,这两种方式得到的预测值相差不大,且都能较准确的模拟和预测土壤含水量。GM(1,1)-AR模型和AR模型这两种方法均可以用来描述水分的具体动态变化,在水分预测上有实际意义,可将其应用和推广。不同坡耕地综合治理技术模式的耕层土壤水分采用AR法模拟的预测值相对误差分别为7.9%(CK)、9.7%(SD)、7.22%(SQ)、5.87%(SA)、3.6%(SAQ);采用GM(1,1)-AR法模拟的预测值相对误差分别为2.77%(CK)、4.9%(SD)、4.7%(SQ)、1.6%(SA)、1.69%(SAQ)。二者平均相对误差相差3.53%。基于小波分解的GM(1,1)-AR模型的拟合和预测效果最优,但是该模型需要对数据拆分分别进行预测,运算过程繁琐,预测结果需要叠加处理,没有AR模型简便,GM(1,1)-AR模型的运算难度明显高于AR模型。在实际应用中,应对具体情况选择适合的预测模型。
[Abstract]:The Red Star Farm is subordinate to the Beian Branch of the Heilongjiang Agricultural Reclamation, with the terrain being low, the middle and the west are gentle slope and the slope is long. The supply of water resources in the area is mainly rainfall, the utilization rate of surface water is low, and the problems of seasonal drought, waterlogging and water and soil loss exist. The soil and water loss will cause the black soil layer of the slope farmland to become thinner year by year, the soil fertility is reduced, the soil fertility is reduced, the crop yield is obviously reduced, and the like, and the agricultural production in the area is seriously affected by the rainfall and the water and soil loss. aiming at the above problems, four different treatment modes of the rat channel (SD), the mouse channel + dark tube (SA), the mouse channel + ridge-direction area field (SQ), the mouse channel + dark tube and the ridge-direction area field (SAQ) are selected in the runoff area of the red-star farm from May to September 2013, And a control cell, i.e., a conventional process (CK), is designed according to a slope cultivation method. The effects of different soil and water conservation technology patterns on soil moisture, soil bulk density and soil water holding performance of different soil and soil layers were studied. And the change of soil aggregate, soil organic matter and soil three after the multi-year water and soil erosion treatment. In this paper, the soil moisture of the surface layer under different technical modes is simulated and the soil moisture prediction model suitable for the region is put forward. The results of the study show that: (1) The removal and water-retaining effects of the "mouse channel + dark tube + ridge-to-region field"-combined measures are good. The stain removal effect of several other modes is also better than that of the conventional ones. The coefficient of variation of soil moisture is as follows: SAQSASDSQCK, whether on the slope or on the slope. The relationship between the coefficient of variation of soil moisture content and the depth of soil in the five measures is the deeper the soil layer, and the smaller the coefficient of variation. That is, the variation coefficient of water content of the surface soil is greater than that of the deep layer. (2) The soil bulk density of each measure is gradually increased with the increase of the depth of the soil layer, but the variation of the bulk density between different combined modes is different. The "Combination of the mouse track and the ridge-direction area", "mouse track", "the combination of the mouse track and the dark tube" and "mouse track, ridging field and dark tube" were reduced by 2.28%, 1.54%, 2.29% and 3.37%, respectively. (3) From the soil water characteristic curve, it can be seen that the water-holding performance of the "mouse channel + dark tube + ridge-to-region field" combination is the best, and the water-holding performance of each measure treatment is the combination mode (SAQ) of the mouse track, the dark tube and the ridge-direction area, the combination mode (SA) of the mouse track and the dark tube, the combination mode (SQ) of the mouse track and the ridge-direction area, Rat track (SD), routine (CK). And the water-holding performance is worse as the depth of the soil layer is increased. (4) after six years of water and soil loss treatment, the soil organic matter content and the conventional treatment under different treatment measures are found, The control effect of the combined mode of the mouse track, the dark tube and the ridge-to-land field is the best. The four kinds of water and soil conservation measures and the combination of the four soil and water conservation measures in the study have a positive effect on the soil organic matter. The difference of the difference is different from that of the measures. The organic matter content of the "mouse channel + dark tube + ridge-to-land field combined mode" is higher than that of the mouse channel + ridge-to-field "combination mode, and the organic matter in the mouse-channel + dark tube combination mode is higher than that of the mouse track". (5) The stability evaluation indexes of the soil aggregates obtained by the calculation are compared and studied, and the stability of the soil aggregates is also affected by different treatment measures, and the stability of the soil aggregates in the other four soil and water conservation modes is higher than that of the conventional treatment. The stability is as follows: rat track + dark tube + ridge-to-region field "mouse track + dark tube mouse track + ridge-to-region field" mouse track routine ". The four measures can effectively reduce the water and soil loss. The soil aggregate stability is also good in the treatment of high organic matter content. But the two are not a significant correlation. (6) The volume of the solid-phase volume of different treatment gradually increases with the deepening of the soil layer, and the volume of the air decreases with the deepening of the plowing layer, and the solid-phase volume of the soil in the "mouse track", the "mouse-channel + ridge-direction field", the "mouse channel + dark tube" and the "mouse channel + ridging area field + dark tube" mode is obviously less than that of the conventional. It is indicated that the treatment of each measure has the effect of improving the permeability of the soil. And (7) comparing the predicted values of the GM (1,1)-AR model and the AR model, the predicted values obtained in the two ways are not small, and the water content of the soil can be simulated and predicted more accurately. The two methods of GM (1,1)-AR model and AR model can be used to describe the specific dynamic changes of water, which is of practical significance in water prediction and can be applied and extended. The relative error of soil water by AR method was 7.9% (CK), 9.7% (SD), 7.22% (SQ), 5.87% (SA), 3.6% (SAQ), and the relative error of the predicted values simulated by GM (1,1)-AR method was 2.77% (CK), 4.9% (SD), 4.7% (SQ), 1.6% (SA) and 1.69% (SAQ) respectively. The average relative error of the two is 3.53%. The fitting and prediction of the GM (1,1)-AR model based on wavelet decomposition are the best, but the model needs to predict the data splitting separately, the operation process is complicated, the prediction result needs to be superposed, the AR model is simple and the operation difficulty of the GM (1,1)-AR model is obviously higher than the AR model. In the practical application, the appropriate prediction model should be selected for the specific situation.
【学位授予单位】：东北农业大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：S157

【参考文献】