Agroforestry is gaining interest due to its potential in enhancing climate resilience and sustainability of farming systems. In this meta-analysis, the crop yield in agroforestry system compared to the control (sole crop) from thirty-six experimental field trials in Mediterranean countries was assessed. The response variable Wlog(RR) (i.e., the weighted natural logarithm of the response ratio) was analyzed by the 95% confidence intervals of mean and by fitting eight linear mixed models. Fixed effects, namely the tree cover (low, medium, high), the tree species (ash tree, chestnut, cork oak, holm oak, olive, poplar, walnut), and the crop species (alfalfa, barley, durum wheat, faba bean, forage, oat, pasture, pea, winter wheat) were significant (P = 0.030, P = 0.017, and P = 0.014, respectively), while the system type (alley cropping, silvo-arable, silvo-pastoral) was not. Among management practices (variety, pruning, fertilization, irrigation, crop age classes, imposed warming and drought, harvest time), only the fertilization significantly improved the response variable (P = 0.006), while the interaction of pruning × crop species was marginally significant (P = 0.065). Relatively large study heterogeneity was observed (Q = 72.6, I² = 72%), which is quite common for agronomic meta-analysis. On the contrary, publication bias based on funnel plots and the Trim and Fill method suggested symmetrical distribution of studies. The sensitivity analysis for significant models identified room for improvements. Overall, we observed a negative effect of trees on crop yield that could be ascribed to the competition for light. Nonetheless, facilitation could be expected under extreme climate events, provided that agricultural practices will maximize synergies among tree cover, tree species, crop species, and management. Future works are encouraged to focus on the overall benefit agroforestry can provide at the field and landscape level, along with long-term monitoring to assess the whole lifespan of these systems and other companion planting options and designs in the Mediterranean region.

农林业因其在增强气候适应能力和农业系统可持续性方面的潜力而受到关注。在这项荟萃分析中，评估了地中海国家 36 个田间试验中农林业系统的作物产量与对照（单一作物）的比较。通过均值的 95% 置信区间和拟合八个线性混合模型来分析响应变量 Wlog(RR)（即响应比的加权自然对数）。固定效应，即树木覆盖率（低、中、高）、树种（白蜡树、板栗、栓皮栎、圣栎、橄榄、杨树、核桃）和作物种类（苜蓿、大麦、硬质小麦、蚕豆）豆类、饲料、燕麦、牧草、豌豆、冬小麦）显着（分别为P = 0.030、P = 0.017 和P = 0.014），而制度类型（巷作、林耕、林牧）则不显着。 。在管理措施（品种、修剪、施肥、灌溉、作物年龄等级、变暖和干旱、收获时间）中，只有施肥显着改善了响应变量（P = 0.006），而修剪×作物种类的相互作用则略有显着（P = 0.065）。观察到相对较大的研究异质性（Q = 72.6，I ² = 72%），这在农艺荟萃分析中很常见。相反，基于漏斗图和修剪和填充方法的发表偏倚表明研究的对称分布。对重要模型的敏感性分析确定了改进的空间。总体而言，我们观察到树木对作物产量的负面影响可能归因于对光的竞争。尽管如此，只要农业实践能够最大限度地发挥树木覆盖、树种、作物种类和管理之间的协同作用，极端气候事件下的促进作用就有望实现。鼓励未来的工作重点关注农林业可以在田间和景观层面提供的整体效益，以及长期监测以评估这些系统以及地中海地区其他配套种植选择和设计的整个寿命。