A global meta-analysis of yield stability in organic and conservation agriculture

Abstract

One of the primary challenges of our time is to enhance global food production and security. Most assessments in agricultural systems focus on plant yield. Yet, these analyses neglect temporal yield stability, or the variability and reliability of production across years. Here we perform a meta-analysis to assess temporal yield stability of three major cropping systems: organic agriculture and conservation agriculture (no-tillage) vs. conventional agriculture, comparing 193 studies based on 2896 comparisons. Organic agriculture has, per unit yield, a significantly lower temporal stability (-15%) compared to conventional agriculture. Thus, although organic farming promotes biodiversity and is generally more environmentally friendly, future efforts should focus on reducing its yield variability. Our analysis further indicates that the use of green manure and enhanced fertilisation can reduce the yield stability gap between organic and conventional agriculture. The temporal stability (-3%) of no-tillage does not differ significantly from those of conventional tillage indicating that a transition to no-tillage does not affect yield stability.

Fig. 1

Yield and yield stability comparing organic and conventional agriculture. a Mean yield ratio. b Absolute stability ratio. c Relative stability ratio for organic (OA) vs. conventional (CA) agriculture for all crops (Overall) and for crops, for which at least 10 observations were available. Numbers in parentheses denote the number of observations and studies. A ratio of 1 means that there is no difference between organic and conventional managed systems while values <1 indicate higher yield for conventional agriculture. For both stability measures a ratio >1 indicate greater absolute and relative stability for conventional agriculture. Values are mean effect sizes with 95% confidential intervals. Mean yield or stability were deemed significantly different between organic and conventional agriculture if the 95% confidential intervals of the ratios did not overlap one

Fig. 2

Histograms for yield and yield stability ratios. a, b Mean yield ratios. c, d Absolute stability ratios. e, f Relative stability ratios for the dataset comparing organic (OA) and conventional agriculture (CA) (a, c, e) and the dataset comparing no-tillage (NT) and conventional tillage (CT) (b, d, f), respectively. The ratios on the x-axis are on the ln-scale

Fig. 3

Effect of nitrogen input on yield and yield stability comparing organic and conventional agriculture. a Mean yield ratio. b Absolute stability ratio. c Relative stability ratio for organic (OA) vs. conventional (CA) agriculture for different levels of nitrogen input. Numbers in parentheses denote the number of observations and studies. A ratio of 1 means that there is no difference between organic and conventional managed systems while values <1 indicate higher yield for conventional agriculture. For both stability measures a ratio >1 indicate greater absolute and relative stability for conventional agriculture. Values are mean effect sizes with 95% confidential intervals. Mean yield or stability were deemed significantly different between organic and conventional agriculture if the 95% confidential intervals of the ratios did not overlap one

Fig. 4

Yield and yield stability comparing no-tillage and conventional tillage. a Mean yield ratio. b Absolute stability ratio. c Relative stability ratio of no-tillage (NT) vs. conventional tillage (CT) for all crops (Overall) and for crops, for which at least 10 observations were available. Numbers in parentheses denote the number of observations and studies. A ratio of 1 means that there is no difference between no-tillage and conventional tillage while a value <1 indicates higher yield for conventional tillage. For both stability measures ratios >1 indicate greater stability for conventional tillage. Values are mean effect sizes with 95% confidential intervals. Mean yield or stability were deemed significantly different between no-tillage and conventional tillage if the 95% confidential intervals of the ratios did not overlap one

Fig. 5

Effect of crop rotation and residue retention on yield and yield stability comparing no-tillage and conventional tillage. a Mean yield ratio. b Absolute stability ratio. c Relative stability ratio of no-tillage (NT) vs. conventional tillage (CT) for subcategories of observations regarding residue retention (RR) and crop rotation (CR):+RR+CR (residue retention and crop rotation), +RR (only residue retention), +CR (only crop rotation), or –RR–CR (without residue retention or crop rotation). Numbers in parentheses denote the number of observations and studies. A ratio of 1 means that there is no difference between no-tillage and conventional tillage while values <1 indicate higher yield for conventional tillage. For both stability measures values >1 indicate greater stability for conventional tillage. Values are mean effect sizes with 95% confidential intervals. Mean yield or stability were deemed significantly different between no-tillage and conventional tillage if the 95% confidential intervals of the ratios did not overlap one

Fig. 6

Relationship between mean yield ratio and stability ratios. a, b Relationship of the mean yield ratio to the absolute stability ratio. c, d Relationship of the mean yield ratio to the relative stability ratio for the dataset comparing organic (OA) and conventional agriculture (CA) (a, c) and the dataset comparing no-tillage (NT) and conventional tillage (CT) (b, d), respectively. Each dot represents one multiple year observation (MYO) and ratios are on the natural log scale. The regression line was fitted on log-transformed values, i.e. log(y) = a + b × log(x), where y was the respective stability ratio and x was the mean yield ratio. ***denote significance at P < 0.001 for a t-test with H0: b = 0, and n.s. denotes non-significant (P > 0.05)