基于遥感与作物模型的土壤速效养分时空变异分析

Spatio-Temporal Variability of Soil Available Nutrients Based on Remote Sensing and Crop Model

【目的】研究在均衡施肥模式下,双山农场2012—2016年间土壤速效养分的时空变异特征,分析施肥管理、轮作模式等人为因素和气象、地形等自然因素对速效养分变化的影响,该研究成果可为土壤速效养分管理和作物变量施肥提供参考依据。【方法】以双山基地农场为例,通过时间序列HJ-1 CCD遥感数据和作物模型WOFOST反演得到土壤速效养分。基于以上土壤速效养分数据在不同时间尺度(年际间和生长季内)和不同空间尺度(农场尺度和田块尺度)上的速效养分时空变异分析。利用土壤速效养分分级图来定性分析养分在5年内的空间变异特征;选取相应的统计参量对土壤速效养分的平均含量和变异性进行定量分析;利用线性回归来分析速效养分的变化量与初始含量的关系;以及养分变化曲线图分析速效养分的时间变化特征。【结果】在均衡的农田施肥管理模式下,2012—2016年农场的土壤速效氮(AN)、速效磷(AP)和速效钾(AK)含量平均值变化不大,速效养分高值和低值都在向中间值靠拢,分别集中于280—360 mg·kg-1、38—42 mg·kg-1和160—200 mg·kg-1,该部分面积比例分别增加18.5%、23.1%和23.8%,养分含量整体呈现均一化特征。AN、AP和AK的变异系数分别从2012年的0.314、0.112和0.257变为2016年的0.131、0.034和0.098,速效养分分布的空间差异性在减弱。AN,AP变化量均与初始值呈现极显著的负相关,R 2分别为0.839和0.882,AK的R 2为0.569,其相关性较弱与土壤中的速效钾本身不稳定有关。通过相邻地块的养分变化分析可知,田块尺度上的土壤速效养分也表现出明显的均一化特征,轮作模式是影响田块间养分变化差异的主要因素,对于轮作模式完全相反的#1和#2田块,人工施肥决定了田块内养分变化的总体趋势,对于豆科作物,固氮作用对土壤速效氮的含量变化作用明显。温度升高在一定程度上会促进作物对养分的吸收,但这种影响不足以改变养分变化的总体趋势。土壤的淋洗对养分的空间变异起到决定性作用,特别是强度较大的降水。降水对养分空间变异的影响在地形差异较大的区域,表现得更加明显。【结论】农田的施肥管理和作物轮作模式是土壤速效养分变化的主导因素;其次是气候因素,雨水对土地的淋洗会导致土壤速效养分的流失和下降,该作用在地形差异较大的地区更加明显;速效养分的变化量与养分初始含量关系显著;温度的升高会促进土壤速效养分的降低,其影响力小于降水。以上变化规律均可纳入后期速效养分的预测研究中,对这些影响因子赋予不同的权值,进行预测模型的构建。以期对作物年际间和生长季内的速效养分的变化进行实时动态监测。

【Objective】The spatio-temporal variability of soil available nutrients in Shuangshan farm during 2012-2016 was studied under uniform fertilization mode. The influence of human factors(such as fertilization management and crop rotation model) and natural factors(such as meteorology and topography) on the changes of available nutrients were analyzed, which could be used to provide references for soil available nutrients management and crop variable fertilization.【Method】The time series satellite HJ-1 CCD imagery and the WOFOST crop model were selected to retrieve soil available nutrients. Based on the above soil available nutrients data, spatio-temporal variability of soil available nutrients analysis of available nutrients on different time scales(interannual and within-year) and different spatial scales(farm scale and field scale) was carried out. The hierarchical maps of soil available nutrients were used to qualitatively analyze the spatial variability of nutrients in five years. The specific statistical parameters were used to quantitatively analyze the average content and variability of soil available nutrients. The linear regression was used to analyze the relationship of variation and initial content of available nutrients. And the temporal variation characteristics of available nutrients were analyzed by the change curve of available nutrients. 【Result】The average contents of available nitrogen(AN), available phosphorus(AP) and available potassium(AK) in the farm were not changed obviously from 2012 to 2016. The high and low values of available nutrients were all close to the middle value, which were concentrated at 280-360 mg·kg-1, 38-42 mg·kg-1 and 160-200 mg·kg-1, respectively. And the area of these three parts increased by 18.5%, 23.1% and 23.8%, respectively, which showed uniform characteristics as a whole. The coefficient of variation of AN, AP, and AK changed from 0.314, 0.112, and 0.257 in 2012 to 0.131, 0.034, and 0.098 in 2016, respectively. And the spatial heterogeneity of available nutrients distributions was weakened. The variation of AN and AP were negatively correlated with the initial values in 2012, and the determination coefficient were 0.839 and 0.882, respectively. And the determination coefficient of AK was 0.569, its weak correlation was related to the instability of AK in the soil itself. The nutrients change characteristics of adjacent field indicated that the soil available nutrients on the field scales also showed obvious uniform characteristics. Crop rotation mode was the main factor affecting the difference of nutrients change between fields. For the field #1 and field #2 with the opposite rotation mode, artificial fertilization determined the general trend of nutrients change in the field. For leguminous plants, nitrogen fixation had a significant effect on the content of soil available nitrogen. The increase in temperature would promote the absorption of nutrients by crops to a certain extent, but this effect was not enough to change the general trend of nutrient changes. Soil leaching played a decisive role in the spatial variability of nutrients, especially for stronger precipitation. The effect of precipitation on nutrients spatial variability was more pronounced in areas with large topography variation.【Conclusion】Farmland fertilization management and crop rotations were the dominant factors in the change of soil available nutrients, followed by topographic and climatic factors. The leaching of rainwater would lead to the loss and decline of soil available nutrients, which was more obvious in the area with larger topography. The amount of change in available nutrients was significantly related to the initial nutrients content. The increase of temperature would accelerate the decrease of available nutrients in soil, and its influence was less than that of precipitation. The above rules could be included in the prediction model of available nutrients, and different weights were assigned to these influencing factors to construct the prediction model, to achieve real-time dynamic monitoring of the variation of soil available nutrients in the inter-annual and growing seasons.