The partitioning of yield gain among component species in intercropping: A global meta-analysis

Context Intercropping could contribute to food security by increasing crop yields, but it remains unclear how the net effect (NE: difference between realized and expected yield) in intercrops is partitioned between the component species and how management and functional species combinations jointly steer this partitioning. Objective This study aims to elucidate how NE in intercropping is partitioned between the component species and how management drives yield gains of each component species and the intercrop as a whole. Methods The dataset was synthesized from 120 articles (from 1950 to 2025). We quantified NE and its partitioning between component species via mixed-effects models, and estimated NE responses to nitrogen input (N), relative density total (RDT), and temporal niche differentiation (TND) using the Random Forest method. Results Intercropping achieved an average NE of 2.12 ± 0.16 Mg ha −1 , with much greater gain for the high-yielding species (NE 1 : 1.99 ± 0.16 Mg ha −1 ) than the low-yielding species (NE 2 : 0.16 ± 0.06 Mg ha −1 ). Species combinations influenced NE and its partitioning. C4 cereal/C3 cereal intercropping had a win-win yield gain with substantial net effects for both component species (C4 cereal: 1.05 ± 0.77 Mg ha −1 ; C3 cereal: 0.77 ± 0.11 Mg ha −1 ). In contrast, C4 cereal/legume intercropping (accounting for 73% of the records) primarily benefited the C4 cereal (NE: 2.43 ± 0.18 Mg ha −1 ) while the net effect of the legume was not significant (0.04 ± 0.06 Mg ha −1 ). In C3 cereal/legume intercropping, a yield gain was found only for the C3 cereal (NE = 0.63 ± 0.14 Mg ha −1 ), while we found no significant yield gain (0.04 ± 0.11 Mg ha −1 ) for the legume. The total NE of intercropping increased with high crop density or high N input in the intercrop compared to the pure stands, and with strong temporal complementarity between species, as in relay intercropping. These three management factors drove yield gains in kg product per ha even more than species choice did. Conclusions C4 cereal/C3 cereal intercropping achieves win-win yield gain, while cereal/legume intercropping benefited mainly the cereal. Tailored management (e.g., high relative density and optimal N input for C4 cereal/legume intercrops) can further increase yield gains from intercropping. Implications The current study revealed the partitioning of yield gain in intercrops and the importance of management in driving this partitioning, which helps to match species combinations and management to obtain high intercropping advantages.