Plant food group diversity and diet-related greenhouse gas emissions in United Kingdom adults: a cross-sectional analysis of the National Diet and Nutrition Survey Year 9

Food systems account for 21 - 37% of global greenhouse gas emissions (GHGE), with animal-source foods driving the majority( 1 ). Plant-based diets are increasingly promoted for their health and environmental benefits, yet the influence of plant food diversity (PFD) - beyond total quantity - on diet-related GHGE remains unclear. This study aimed to examine associations between PFD and both plant food group-related and total-diet GHGE. A cross-sectional analysis was conducted using data from the National Diet and Nutrition Survey Year 9 (2016-2017). Adults (n = 648; mean age 48.8y; 58% female) completed interviewer-checked four-day food diaries. Plant-based ingredients were identified from disaggregated composite foods and classified into eight plant food groups. Diversity scores reflected the total count of unique plant items consumed over four days. Diet-related GHGE were calculated as four-day totals and estimated in two steps: ingredient weights were first derived from disaggregated NDNS codes, then multiplied by cradle-to-retail emission factors (gCO 2 ;-eq/g) compiled from UK life-cycle analysis databases, covering 91% of foods (2) . Multiple linear regression assessed associations between PFD and GHGE across five models with sequential confounder adjustment. The plant food group with the highest diversity was vegetables (mean 7.1, SD 3.2 items), while grains contributed the highest plant food group-related GHGE (mean 3649.4, SD 2235.9 g CO 2 -eq over four days). Across all regression models, greater diversity within each plant food group was positively associated with higher corresponding plant food group-related GHGE (p<0.05). After full confounder adjustment, only vegetable diversity remained positively associated with total-diet GHGE (B 169.1; 95% CI 97.7-240.4; p<0.001). Fruit, legume, grain, fats/oils, drinks, and herbs/spices diversity were not significantly associated with total-diet GHGE after full adjustment (p>0.05), while nuts/seeds diversity showed a borderline inverse association with total-diet GHGE (B -100.2; 95% CI -215.3-3.6; p = 0.058). Greater vegetable diversity was consistently associated with higher total-diet GHGE, whereas no such associations were observed for other plant food groups following confounder adjustment. PFD within each plant food group remained positively associated with its respective plant food group-related GHGE, indicating that more diverse consumption within a category contributes to higher emissions from that group. These findings suggest that the climate impact of plant diversity depends on which plant food groups are diversified and the potential food displacements this may create in the diet. Future research should investigate how seasonality, sourcing, and production practices influence plant group-related GHGE in diverse plant-rich diets.