School of Natural Sciences, Bangor University, Deiniol Road, Bangor, LL57 2UW, UK; Sir William Roberts Centre for Sustainable Land Use, Bangor University, Deiniol Road, Bangor, LL57 2DG, UK.
Department of Animal Science, Campus Gaiopolis, University of Thessaly, Larissa, 411 10, Greece; IBERS, Aberystwyth University, Ceredigion, Aberystwyth, SY23 3EB, UK.
J Environ Manage. 2020 Oct 15;272:111054. doi: 10.1016/j.jenvman.2020.111054. Epub 2020 Jul 16.
While milk is a major agricultural commodity, dairy farming also supports a large share of global beef production. In Life Cycle Assessment (LCA) studies of dairy farming systems, dairy-beef production is often ignored or 'allocated off', which may give a distorted view of production efficiencies. This study combines LCA with Data Envelopment Analysis (DEA) to develop an indicator of eco-efficiency for each of 738 UK dairy farms (3624 data points in 15 years) that aggregates multiple burdens and expresses them per unit of milk and dairy-beef produced. Within the DEA framework, the importance (weight) of dairy-beef relative to milk is iteratively increased to quantify the environmental losses from heavily focussing on milk-production, via e.g. higher yields per cow, with consequent lower burdens per unit of milk, yet with lower dairy-beef production levels, where burdens for beef production are externalized. Then, the relationship between DEA eco-efficiency and a series of indicators of dairy farming intensity at animal- and farm-levels was studied with Generalized Additive Models (GAM). For all sets of DEA weights (proportion of deviance explained ranged between 68% and 82%) indicate that milk yield per cow and forage area, and larger dairy herds all have a positive effect on eco-efficiency, while concentrate fed per unit of milk and the forage area both have a negative effect (p < 0.05 for all modelled relationships). These findings suggest that more intensive and consolidated dairy farms can positively impact on eco-efficiency. However, as the DEA weight for dairy-beef relative to milk increases, the relationship between environmental efficiency and farming specialization (expressed as L milk per kg dairy-beef produced) reverses from positive to negative. In conclusion, dairy-beef production is pivotal in determining the wider environmental efficiency of dairy (and ruminant food) systems, and its under-representation in efficiency studies has generated a misleading approach to meeting emission targets.
虽然牛奶是一种主要的农业商品，但奶牛养殖也支撑着全球很大一部分牛肉生产。在奶牛养殖系统的生命周期评估（LCA）研究中，奶牛-牛肉生产通常被忽略或“分配掉”，这可能会对生产效率产生扭曲的看法。本研究将生命周期评估（LCA）与数据包络分析（DEA）相结合，为英国 738 家奶牛场（15 年 3624 个数据点）中的每一家开发了一个生态效率指标，该指标综合了多种负担，并按每单位牛奶和牛奶-牛肉产量表示。在 DEA 框架内，通过例如每头牛更高的产量来集中精力提高牛奶产量，从而降低每单位牛奶的负担，但牛奶-牛肉产量水平较低，从而使牛肉生产的负担外化，奶牛-牛肉相对于牛奶的重要性（权重）被迭代增加，以量化仅关注牛奶生产所带来的环境损失。然后，使用广义加性模型（GAM）研究了 DEA 生态效率与一系列动物和农场层面奶牛养殖强度指标之间的关系。对于所有 DEA 权重集（解释方差的比例在 68%到 82%之间），表明每头牛的牛奶产量和饲料面积以及更大的奶牛群都对生态效率有积极影响，而每单位牛奶的精饲料和饲料面积都有负面影响（所有模型关系的 p 值均<0.05）。这些发现表明，更密集和更集中的奶牛场可以对生态效率产生积极影响。然而，随着奶牛-牛肉相对于牛奶的 DEA 权重增加，环境效率与养殖专业化之间的关系（表示为每生产 1 千克牛奶-牛肉的牛奶量）从正相关变为负相关。总之，奶牛-牛肉生产对于确定奶牛（和反刍动物食品）系统的更广泛环境效率至关重要，而在效率研究中对其代表性不足导致了一种对实现排放目标的误导性方法。
