ABSTRACT This study evaluated the ability of fluorescence excitation spectroscopy (FES) to differentiate milk samples from different origins. Three different farming systems were chosen: D-samples originating from low-input biodynamic farms (cows fed on hay or pasture); O-samples from intermediate-input organic farms (cows fed mainly on grass silage); and C-samples from high-input conventional farms (indoor housing, cows fed on maize and grass silage). Milk samples were collected every second month between July 2015 and June 2016 from 12 farms (four farms per system), and a total of 70 samples were obtained. Fat-, protein- and urea-concentrations, somatic-cell count and fatty acid levels (FA) were determined. FES-measurements were performed by exciting the sample with light of different wavelengths and detecting delayed luminescence. Differences between farming systems in each season were checked by ANOVA. Factors of season, system and breed were evaluated in a linear regression model. By linear-discriminant analysis, variables contributing to correct classification were analysed. Milk FAs, especially the concentration of omega-3 (n3) and omega-6 (n6) FAs, were different between farming systems, while conjugated linoleic acid (CLA) and C18:1t11 (tVA)-concentration was mainly influenced by season (pasture). FES-parameters showed slight seasonal variations, but strong farming-system impacts. Differentiation between the three farming systems was possible for 81% of the samples by using FAs as variables. FES-parameters discriminated up to 86% of the samples, and, in combination, 93% of the samples were classified correctly. These results indicated that FES-results contributed to correct discrimination and that FES-results may be linked with qualities different to the FA profile.

中文翻译：

---

使用荧光激发光谱 (FES) 和脂肪酸区分来自低投入生物动力、中等投入有机和高投入传统农业系统的牛奶

摘要 本研究评估了荧光激发光谱 (FES) 区分不同来源牛奶样品的能力。选择了三种不同的养殖系统：来自低投入生物动力农场（以干草或牧场为食的奶牛）的 D 样本；来自中等投入有机农场的 O-样品（主要以青草为食的奶牛）；来自高投入常规农场（室内饲养、以玉米和青贮草为食的奶牛）的 C 样本。2015 年 7 月至 2016 年 6 月期间，每隔一个月从 12 个农场（每个系统 4 个农场）收集牛奶样本，共获得 70 个样本。测定了脂肪、蛋白质和尿素浓度、体细胞计数和脂肪酸水平 (FA)。FES 测量是通过用不同波长的光激发样品并检测延迟发光来进行的。通过方差分析检查每个季节耕作系统之间的差异。在线性回归模型中评估季节、系统和品种的因素。通过线性判别分析，分析了有助于正确分类的变量。牛奶 FAs，尤其是 omega-3 (n3) 和 omega-6 (n6) FAs 的浓度，在不同养殖系统之间存在差异，而共轭亚油酸 (CLA) 和 C18:1t11 (tVA) 浓度主要受季节影响（牧场）。FES 参数显示出轻微的季节性变化，但对农业系统的影响很大。通过使用 FA 作为变量，81% 的样本可以区分三种农业系统。FES 参数区分了多达 86% 的样本，并且结合起来，93% 的样本被正确分类。