Comparison of N2O emissions from soils at three temperate agricultural sites:: simulations of year-round measurements by four models
Comparison of N2O emissions from soils at three temperate agricultural sites:: simulations of year-round measurements by four models
复制标题
DOI：
10.1023/a:1009780109748
复制
发表时间：
1998-10-01
影响因子：
3.1
通讯作者：
Smith, KA
中科院分区：
文献类型：
--
作者：
Frolking, SE;Mosier, AR;Smith, KA
将四种模式模拟的氧化亚氮（N2O）通量与来自三个国家五个温带农业站点的全年野外测量结果进行了比较。野外地点包括美国科罗拉多州东部一个未施肥、半干旱、N2O通量低的牧场；在苏格兰为青贮而割下的肥沃草地上进行两种肥料处理（尿素和硝酸盐）；在德国的两个施肥、栽培的农田，冬季N2O损失相当高。使用的模型是CENTURY模型的每日微量气体版本，DNDC， ExpertN和NASA-Ames版本的CASA模型。这些模型包括相似的组成部分（土壤物理、分解、植物生长和氮转化），但在某些情况下，这些过程使用了非常不同的算法。所有模型对氮在农业生态系统中的总体循环产生了相似的结果，但模拟的氮微量气体通量差异很大。在大多数情况下，模拟的N2O通量在观测到的年通量的约2因子范围内，但即使模型产生相似的N2O通量，它们也经常产生非常不同的气态氮损失估计数，如一氧化氮（NO）、二氮（N-2）和氨（NH3）。准确模拟土壤湿度似乎是可靠模拟N2O排放的关键要求。所有的模型都模拟了苏格兰站点施肥后低背景通量与高通量的一般模式，但它们不能（或没有设计成）准确捕捉到不同肥料类型对N2O通量的影响。没有一个模型能够可靠地在德国两个站点观测到的短暂冬季解冻期间产生大的一氧化二氮脉冲。除DNDC外，所有模型都模拟了科罗拉多州干旱地区非常低的N2O通量。美国追踪Cas网络（TRAGNET）为该模型和站点的相互比较提供了一种机制。需要对这些模型和其他模型以及其他数据集进行进一步的相互比较；这些措施应包括热带农业生态系统和为可持续性而设计的新的农业管理技术。
Nitrous oxide (N2O) flux simulations by four models were compared with year-round field measurements from five temperate agricultural sites in three countries. The field sites included an unfertilized, semi-arid rangeland with low N2O fluxes in eastern Colorado, USA; two fertilizer treatments (urea and nitrate) on a fertilized grass ley cut for silage in Scotland; and two fertilized, cultivated crop fields in Germany where N2O loss during the winter was quite high. The models used were daily trace gas versions of the CENTURY model, DNDC, ExpertN, and the NASA-Ames version of the CASA model. These models included similar components (soil physics, decomposition, plant growth, and nitrogen transformations), but in some cases used very different algorithms for these processes. All models generated similar results for the general cycling of nitrogen through the agro-ecosystems, but simulated nitrogen trace gas fluxes were quite different. In most cases the simulated N2O fluxes were within a factor of about 2 of the observed annual fluxes, but even when models produced similar N2O fluxes they often produced very different estimates of gaseous N loss as nitric oxide (NO), dinitrogen (N-2), and ammonia (NH3). Accurate simulation of soil moisture appears to be a key requirement for reliable simulation of N2O emissions. All models simulated the general pattern of low background fluxes with high fluxes following fertilization at the Scottish sites, but they could not (or were not designed to) accurately capture the observed effects of different fertilizer types on N2O flux. None of the models were able to reliably generate large pulses of N2O during brief winter thaws that were observed at the two German sites. All models except DNDC simulated very low N2O fluxes for the dry site in Colorado. The US Trace Cas Network (TRAGNET) has provided a mechanism for this model and site intercomparison. Additional intercomparisons are needed with these and other models and additional data sets; these should include both tropical agro-ecosystems and new agricultural management techniques designed for sustainability.