TL;DR
French crop yields have seen significant growth and reduced relative variability over the past century. However, this study reveals that yield growth has stagnated since the 1990s for a substantial portion of the cultivated area for several key crops, with maize being a notable exception. This stagnation, particularly in high-yielding regions, raises concerns about future food security and necessitates further investigation into its causes.
What Problem Does This Paper Address?
Ensuring global food security in the face of a growing population and climate change hinges on understanding and predicting crop yield trends. While historical agricultural productivity has seen remarkable increases, there's a growing concern that these gains might be plateauing. Previous research has pointed towards yield stagnation in wheat in France, but evidence for other major crops, like maize, has been contradictory or lacking. This paper addresses this knowledge gap by conducting a comprehensive, sub-national analysis of yield trends and variability for ten major crops across France over more than a century, aiming to definitively identify where and for which crops yield growth has stalled.
Methods
The researchers employed Dynamic Linear Models (DLMs) to analyze yield trends and estimate annual growth rates. DLMs are a robust statistical method that allows for the adjustment of trends over time without making strong assumptions about the functional form of yield trends. For each crop and department, DLMs were used to estimate underlying yield levels and annual growth rates.
To assess yield stagnation, a scoring scheme was developed. Yields were considered stagnating if annual growth rates consistently fell below a chosen threshold (0.25% per year) over a defined period (the last 20 years of data, 1997-2016). This method allowed for the calculation of a "stagnation score" for each crop and department, indicating the likelihood of stagnation.
Inter-annual yield variability was analyzed using the Coefficient of Variation (CV), calculated as the standard deviation of yield residuals divided by the mean yield. Spatial clustering of mean yields was also performed to identify regional differences in productivity over time.
Datasets
The study utilized a comprehensive dataset of over 120,000 yield observations for ten major crops in the 96 mainland French departments from 1900 to 2016. The crops included: barley (spring and winter), durum wheat, soft wheat (spring and winter), maize, oats (spring and winter), potatoes, rapeseed (spring and winter), sugar beet, sunflower, and wine.
Yield data from 1900 to 1988 were digitized from historical French agricultural statistics, while data from 1989 to 2016 were sourced from the Agreste website. Fertilizer data (N, P2O5, K2O) for the period 1946-2013 were also used for correlation analysis.
Key Insights and Ideas
- The "Golden Age" of Yield Growth is Fading: While French crop yields have quadrupled over the 20th century, the rapid growth observed from the 1950s to the 1990s has significantly slowed or stopped for many crops in recent decades.
- Stagnation is Crop and Region Specific: Yield stagnation is not a universal phenomenon. It affects certain crops more than others and is often concentrated in departments with historically higher average yields.
- Maize is the Exception: While many crops show signs of stagnation, maize stands out as an exception, with no evidence of yield growth stalling in recent years.
- Relative Variability is Down, Absolute Variability is Up: Over the long term, the variability of yields relative to the mean has decreased, meaning yields have become more stable as a proportion of their average. However, absolute yield variability (the actual range of yield fluctuations) has increased, likely due to higher maximum yields and better mitigation of severe loss events.
- Northern France Leads, and Stagnation Follows: Higher average yields are consistently observed in Northern France for most crops, and it's in these high-yielding regions that stagnation is more likely to occur. This suggests that reaching physiological yield potentials might be a factor for some crops.
Key Results
- Widespread Stagnation: Since the 1990s, yield growth has stagnated on at least 25% of the cultivated area for winter wheat, barley, oats, durum wheat, sunflower, and wine.
- Maize's Resilience: Maize shows no evidence of yield stagnation across any of its cropping area.
- Wheat Stagnation Confirmed: For total wheat, 70% of the growing area exhibits stagnating yields, aligning with previous findings.
- High-Yielding Areas Affected: Departments with higher average yields are significantly more likely to experience yield stagnation. For example, winter barley in stagnating departments yields 6.28 t/ha compared to 5.58 t/ha in non-stagnating departments.
- Fertilizer Correlation: Yield trends are strongly correlated with nitrogen fertilizer input, but not with phosphate. This suggests that while nitrogen is crucial, other factors might be limiting for certain crops like durum wheat and wine.
Limitations and Future Work
- Causality Remains Elusive: While the study identifies where stagnation is occurring, pinpointing the exact causes is challenging due to data limitations. Potential factors like genetic improvement ceilings, climate change, policy decisions (e.g., Common Agricultural Policy), changes in crop rotation, soil carbon content, and economic considerations (marginal costs of intervention) are discussed but require more granular data for definitive analysis.
- Data Granularity: The lack of crop-specific fertilizer dosage data per department, detailed information on pest and disease outbreaks, and time-resolved data on crop rotation and soil carbon content hinder a more precise understanding of stagnation drivers.
- Irrigation Data: While irrigation is mentioned as a potential factor in yield variability, the study notes that the available data is not granular enough for a definitive sub-national analysis.
- Future Research Directions: The authors emphasize the need for further research into the underlying causes of stagnation, particularly for crops other than wheat and maize. This includes investigating the role of climate change, policy impacts, and detailed management practices. They also suggest that future analyses would benefit from more detailed, crop-specific, and sub-national data on fertilizer use, pest/disease incidence, and irrigation. The study also highlights the importance of considering the split between spring and winter cultivars, as this can significantly influence observed trends.
Summary updated