近红外光谱分析中,异常样本的存在严重影响定标模型的预测性能和适配性。基于 X / Y 联合的ODXY异常样本识别和剔除方法,提出并证明了一种专用于多组分分析的MODXY异常样本识别方法。实验采用80组玉米近红外光谱数据,利用不同异常样本识别方法剔除异常样本后建立玉米含水率、含油率、蛋白质含量和淀粉含量4种组分的偏最小二乘预测模型,采用预测均方差和决定系数作为评价指标比较所建模型的性能,检验MODXY方法在多组分分析中的异常样本识别能力。实验结果表明:在近红外多组分分析中,MODXY方法在大多数情况下具有更好的异常样本识别能力;MODXY方法和ODXY方法均有一定的适用范围,它们更适合于相应组分化学值的相对标准偏差较大的情况。 相似文献
Biodiversity conservation is one of the important ecosystem services that has been negatively impacted by anthropogenic activities. Natural forests (NF) harbor some of the highest species diversity around the world. However, deforestation and degradation have resulted in reduced forest land cover and loss of diversity. Homegarden agroforestry (AF) systems have been proven to be an intermediary for biodiversity conservation. In this study, we evaluate the effectiveness of home garden AF practices to conserve tree species diversity in Bangladesh and compare them with tree species diversity in NF. A total of nine locations were selected for this synthesis from published literature which comprised of five AF sites and four NFs. Shannon?CWeiner Diversity Index (H) was similar for home-garden AF (3.50) and NF (2.99), with no statistical difference between them. Based on non-metric multi-dimensional scaling (NMDS) ordination analysis, the AF and NF plots showed distinct separation. However, Bray?CCurtis dissimilarity index ranged from 0.95 to 0.70 indicating nearly no overlap in species composition to significant overlap between AF and NF. Based on our results, we conclude that AF can serve as an important ecological tool in conserving tree species diversity, particularly on landscapes where NF fragments represent only a small fraction of the total land area. Creating and maintaining AF habitats in such human dominated landscapes should be part of the biodiversity conservation strategy. 相似文献
Climate warming and sea level rise have the potential to change the salt level of soil in tidal wetlands. And it is important to clarify the process and the mechanism of decomposition of soil organic carbon in a tidal wetland under varying salinities. The aim of this study was to evaluate the impacts of soil salinity on the decomposition rate of organic carbon (DROC) and dissolved organic carbon (DOC) in a tidal wetland.
Materials and methods
Two types of soil (surface soil in Suaeda salsa and bare tidal flat) were collected, air-dried, and homogenized. After adding different content of salt (0 g/L, 3 g/L, 6 g/L, 9 g/L, and 12 g/L), the soils were incubated for 28 days at stable room temperature (25?±?2 °C) and added by deionized water to maintain the stability of soil moisture. The gases (CO2 and CH4) emission rates of each salt treatment were measured during 28-day incubation. DROC was determined by the sum of daily CO2-C emission rates and daily CH4-C emission rates in this study.
Results and discussion
Salt addition inhibited the process of gas emissions and DROC. Gases emission rates and DROC of two types of soil showed similar temporal trends, with distinctive drop in the beginning of experiment and no significant decrease followed. Significant difference of DOC among salt treatments was found in the bare tidal flat soil. Variations of partial correlation between DROC and soil salinity and DOC showed similar trends (e.g., in days 9–18, the positive effect of DOC on DROC was greatly promoted (R2?=?0.80, p?<?0.001), and the negative effect of soil salinity was highly improved (R2?=?0.93, p?<?0.001)). Soil properties, in particular DOC, may be primary factors accounting for the discrepancy of gases emission rates and DROC of two types of soil.
Conclusions
Increased soil salinity had a negative effect on DROC during 28-day incubation. The impact of soil salinity and DOC on DROC were varied in different phases of laboratory experiment (soil salinity generally had increasingly negative relationship with DROC, but DOC showed most significantly positive relationship in the middle stage of incubation). Both the formation and consumption of DOC may be valuable for more detail research regarding to decomposition of soil organic carbon.