摘要：三江平原从1950年以来长期的集约化农业发展导致约3,800,000hm2 自然湿地(We)和1,200,000hm2 自然林地(Fo)转变为农业用地。然而，大面积的土地利用方式改变可能影响土壤磷(P)的含量、形态、植物可用性及吸附特性，进而影响到磷的迁移。本论文以位于三江平原的八五九农场为研究区域，探究三江平原不同土地利用类型表层土壤及土壤剖面磷含量、形态、植物可用性、吸附特性，进而揭示三江平原土地利用变化对土壤磷的影响。三江平原八五九农场表层土壤总磷(TP)含量为324.80～2345.20mg kg-1，平均含量为972.32mg kg-1(n=32)。旱地、水田、林地和湿地土壤TP含量分别是547.07～1573.47mg kg-1、926.80～1062.4mg kg-1、324.8～1047.2mg kg-1、907.2～2345.2 mg kg-1，平均含量分别为958.69mg kg-1、972.51mg kg-1、789.72mg•kg-1、1164.55mg kg-1，湿地表层土壤TP含量明显高于林地、旱地和水田表层土壤TP含量。本研究选择不同土地利用类型代表性土壤剖面(湿地We1点，林地Fo2点，水田P3点和旱地D2点)探究了土壤剖面TP含量变化。结果表明湿地、林地、水田和旱地土壤剖面TP含量分别为644.40～2655.87mg kg-1，381.60～842.40mg kg-1，507.07～941.87mg kg-1和413.53～1169.47mg kg-1。自然湿地TP含量在表层和底层高于中间层，这是由于表层(0～20cm)有机质积累和底层(40～100cm)铁氧化物积累引起的。林地、旱地、水田土壤总磷含量随土壤剖面深度增加而降低。因此，自然湿地转变被农用地40年后，土壤表层(0～20cm)TP含量显著降低(从2302.20mg kg-1降低到891.93mg kg-1)。然而，自然林地转变为农业用地40年后，表层土壤(0～20cm)TP含量升高(从770.60mg kg-1增加到1143.00mg kg-1)，无机磷(Pior)含量也从200.75mg kg-1增加到463.05mg kg-1。总体上，约80%～95%的TP与土壤有机质和铁氧化物结合。因此，这两种固相组成部分是控制三江平原土壤磷的关键因素。三江平原八五九表层土壤中有机磷(Por)含量范围为146.77mg kg-1～2103.05mg kg-1之间，平均含量为516.99mg kg-1(n=32)。湿地、林地、农田表层土壤Por平均含量分别为765.70mg kg-1、514.09mg kg-1和463.12mg kg-1。另一方面，湿地、林地和农田表层无机磷(Pior)平均含量分别为398.84mg kg-1、275.62mg kg-1和499.17mg kg-1。从湿地和林地转变为农业用地并长期耕作后Por含量降低，而Pior含量升高。农田耕作加速了土壤有机质和有机磷的矿化过程，导致表层土壤有机磷含量下降。另一方面，农田施肥及有机磷矿化导致农田表层土壤无机磷含量升高。对于典型湿地和林地土壤剖面，Pior含量与Fe和Mn含量显著相关，并且随剖面深度增加而逐渐升高；但是Por含量与土壤SOM(有机质)含量显著相关，并且随剖面深度增加而逐渐降低。对于农田土壤剖面，尽管Por含量与土壤SOM含量显著相关，但是Pior含量与Fe含量不相关，其原因主要为磷肥输入的影响。而且由于磷肥输入的影响，农田土壤剖面TP含量通常随深度增加而逐渐降低。但是农田土壤剖面Por含量随深度增加表现出先升高而后下降的趋势，而Pior则表现出先下降而后升高的趋势。总体上，农田土壤剖面20～30cm土层Por占TP含量的比例最高；而湿地和林地土壤剖面0～10cm土层 Por占TP含量的比例最高，分别为83.43%和79.39%。因此，湿地和林地转变为耕地影响了土壤剖面Por和Pior的变化。三江平原八五九农场表层土壤中溶解与弱吸附态磷(S/L-P)、铝结合态磷 (Al-P)、铁结合态磷(Fe-P)、闭蓄态(还原溶解态)磷(RS-P)和钙结合态磷(Ca-P)的含量分别是0.02～8.42mg kg-1、10.19～110.23mg kg-1、70.25～447.37mg kg-1、9.41～98.04mg kg-1 和15.53～101.16mg kg-1，其平均含量分别是2.68mg kg-1、58.77mg kg-1、255.17mg kg-1、60.48mg kg-1和45.82mg kg-1(n=32)，占TP比例分别是0.25%、5.88%、25.82%、6.22%和4.48%，其含量大小顺序依次为：Fe-P＞RS-P、Al-P、Ca-P＞S/L-P。对于典型湿地、林地、水田、旱地土壤剖面，磷的化学形态分布与表层土壤相似。而且土壤剖面中，Fe-P含量和Fe含量、Al-P含量和Al含量，Ca-P含量和Ca含量显著相关，表明典型土壤剖面无机磷形态的分布主要决定于剖面铁氧化物、铝氧化物及碳酸盐矿物含量的剖面分布。三江平原八五九农场农田表层土壤植物可利用态磷 (Pi-P)含量范围为10.20～104.32mg kg-1，平均值为50.45mg kg-1(n=23)。Pi-P含量与Pior或活性P(S/L-P，Al-P, Fe-P和Ca-P之和)呈正相关性，但与Por无相关性。与Ca-P，S/L-P和RS-P相比，Fe-P和Al-P含量与Pi-P含量相关性更高，表明Pi-P主要来源于铁，铝氧化物结合态P。耕作区土壤中Pi-P绝对含量和占TP的比例一般高于世界各地土壤中的含量。对于旱地土壤剖面(n=5)，Pi-P含量随剖面深度增加而降低，0～20cm土层中Pi-P含量较高。5个旱地土壤剖面土壤中Pi-P平均含量从表层60.54mg kg-1逐渐减少到底层18.31mg kg-1。根据表层土壤中Pi-P含量判断三江平原农田排水可能导致周围水体富营养化。

ABSTRACT:The Sanjiang Plain is situated in the northeastern part of the Heilongjiang Province in NE China. Long-term intensive agricultural development since the 1950s has converted approximately 3,800,000 hm2 of natural wetlands and 1,200,000 hm2 of natural forestland to cultivated land in the Sanjiang Plain of China. Currently, the most common landscape types are wetland and cultivated land, accounting for 0.9 and 4 million hectares, respectively, as about 27% and 30% of the provincial wetland and cultivated land, respectively. However, investigation on the chemical forms and availability of phosphorus is limited in the soils of the Sanjiang Plain. Little is known on P buffering capacity and loss potential for the soils of Sanjiang Plain. In addition, it is not clear how the large-scale land-use shifts impact the content and chemical forms of phosphorus in the topsoil and soil profile.The objectives of this study were to investigate the contents, chemical forms, and availability of phosphorus in the soils of the Sanjiang Plain, to quantify P sorption and buffering capacity and loss potential for the soils, and to identify the impacts of land-use change on P geochemistry such as chemical forms, availability, sorption properties for the soils.“859” farm is located in the northeast of the Sanjiang Plain, where land-use types include natural wetland (We), and natural forest land (Fo), paddy field (P) and dry-farming field (D). Therefore, “859” farm may be a typical area representing the history and current situation of human activity and ecosystem in the Sanjiang Plain and thus is selected as the study area of this investigation. Thirty-two topsoil samples (0 to 10 cm depth) and some soil cores were collected and analyzed for soil properties, total P (TP), organic P (Por), inorganic P (Pior), inorganic P forms, plant-available P (Pi-P), and some major elements.The content of TP in the topsoil averaged 969.18 mg kg-1, in the ranged from 324.80 mg•kg-1 to 2345.2 mg•kg-1. The content of TP in the soils for each land-use type was 547.07 to 1573.47 mg•kg-1 for the dry-farming filed, 926.80 to 1062.4 mg•kg-1 for the paddy field, 324.8 to 1047.2 mg•kg-1 for the natural forest land, and 907.2 to 2345.2mg•kg-1 for the natural wetland, with averages of 958.69, 972.51, 789.72, and 1164.55mg kg-1, respectively. This inferred that the change from the wetland to cultivated land might result in the decrease of TP in topsoil, while the change from the forest land to cultivated land might lead to the increase of TP in the topsoil.The content of TP in the typical soil cores ranged from 644.40 to 2655.87 mg kg-1 for the wetland soil core, 381.60 to 842.40 mg kg-1 for the forest land soil core, 507.07 to 941.87 mg kg-1 for the paddy field soil core, and 413.53 to 1169.47 mg kg-1 for the dry-farming field soil core. The content of TP in the top layers and bottom layers was higher than that in the middle layers, duo to the accumulation of soil organic matter (SOM) in the top layers (0 to 20 cm depth) and the accumulation of iron oxides in the bottom layers (40 to 100 cm depth). However, the content of TP generally decreased with increasing depth for the soil cores of the forest land, paddy field, and dry-faming field. Therefore, the change from natural wetland to cultivated not only impacts the TP content in the topsoil, but also changes the P distribution along the soil profile.The content of Por in the topsoil averaged 516.99 mg kg-1 (146.77 to 2103.05 mg kg-1), while the content of Pior averaged 455.55 mg kg-1 (173.66 to 805.76 mg kg-1). In details, the average contents of Por in the topsoil for the wetland, forest land, and cultivated land were 765.70, 514.09, and 463.12 mg kg-1, respectively; while the average content of Pior were 398.84, 275.62, and 499.17 mg kg-1, respectively. Therefore, the conversion from the wetland and forest land to cultivated field might lead to the decrease of Por content, but the increase of Pior content. For the conversion from the wetland to cultivated filed, the increase of Pior content in the topsoil was less than the decrease of Por content, thus TP content in the topsoil decreased. For the conversion of the forestland to cultivated field, however, the increase of Pior content in the topsoil was greater than the decrease of Por content, thus the content of TP increased. The reasons for the decrease of Por content were mainly degradation of SOM and mineralization of organic P (e.g., in average, SOM content was 13.37%, 9.09%, and 7.47% in the topsoil for the wetland, forest land, and cultivated land, respectively), while the increase in Pior content was mainly due to the input of inorganic P fertilizers, together with mineralization of organic P.For the soil cores of the natural wetland and forest land, Pior content was significantly correlated to Fe and Mn contents and generally increased with depth increase, while Por content was significantly correlated to SOM content and decreased with depth increase. For the soil cores of the cultivated land, whereas Por content was significantly correlated to the SOM content, Pior content was not correlated to Fe content duo to the impact of P fertilizer input. In addition, TP content generally decreased with depth increase along the soil profiles of the cultivated land duo to the impact of P fertilizer input. On the other hand, Por content generally increased first and then decreased along the soil profile of the cultivated land, but Pior content generally decreased first and then increased. Therefore, the conversion of the wetland land and forest land to cultivated land also impacts the distribution of Por and Pior contents along the soil cores.The Pior was further operationally fractionized into Soluble and loosely bound P (S/L-P), Al-bound P (Al-P), Fe-bound P (Fe-P), reductant-soluble P (RS-P), and Ca-bound P (Ca-P). The average contents of them in the topsoil were 2.68, 58.77, 255.17, 60.48, and 45.82 mg kg-1, respectively, accounting for 0.25%, 5.88%, 25.82%, 6.22%, and 4.48% of TP, respectively. The fractions of Pior were in order Fe-P＞RS-P, Al-P, Ca-P＞S/L-P. For the soil cores, the distribution of P chemical forms was similar to that for the topsoil. In addition, the contents of various chemical forms of P were positively correlated with the levels of their corresponding solid phases for the soil cores, e.g., Al-P correlated with Al, Fe-P and RS-P correlated with Fe, and Ca-P correlated with Ca. Plant-available P (Pi-P) in the topsoil of the cultivated land ranged from 10.20 to 104.3 mg kg-1, with an average of 50.45 mg kg-1. Pi-P was significantly correlated to Pior, but not correlated to Por. In details, Pi-P was significantly correlated to Fe-P and Al-P, inferring that Pi-P was mainly extracted from them. The average content of Pi-P and its proportion were usually higher than those for the soils on the world. In addition, P in the runoff estimated by Pi-P content might lead to the eutrophicatio of adjacent surface water. The average content of Pi-P decreased along soil cores (n=5) of the cultivated land from 60.54 mg kg-1 at 0 to 10 cm depth to 18.31 mg kg-1 at 40 to 50 cm depth.P sorption properties for the topsoil of the wetland, forest land, paddy field, and dry-farming field were investigated. Sorption kinetics showed P sorption in the topsoil of the wetland and dry-farming field was characterized with initial fast period and then slow period. The sorption mechamism for the fast sorption period might be specific sorption (chemical sorption or inner-sphere surface complexation sorption) period, while the sorption mechanism for the slow sorption period might be non-specific sorption (physical sorption or outer-sphere surface complexation sorption) period. pH has effect on the P sorption in the topsoil of the wetland and dry-faming land and the sorbed P amount was higher at pH = 7.0 than at pH = 4.0 and pH = 12. Temperature has not significant effect on the P sorption in the topsoil of the wetland and dry-farming field at 5℃, 25℃, and 40℃. The topsoil of the wetland and dry-farming land sorbed more P after SOM was removed by heating. EPC0 at which no net P sorption occurs was 0.0047 to 0.242, 0.059 to 0.077, 0.037 to 0.058, and 0.037 to 0.058 mg L-1 for the topsoil of the dry-farming field, paddy field, forest land, and wetland, respectively. The average EPC0 was in the order: dry-farming field > paddy field > forest land > wetland, indicating that the topsoil of the wetland and forest land has higher P sorption capacity while the topsoil of the cultivated land has higher potential to release P into surface water. The sorption isotherms were well fitted by the Langmuir equation. The maximal sorption capacity (Qmax) of P ranged from 682.50 to 1736.07 mg kg-1, with an average of 1016.10 mg kg-1. The average Qmax was in the order: wetland (1179.14 mg kg-1), forest land (1177.42 mg kg-1) > cultivated land (894.25 mg kg-1).