TL;DR
This paper introduces ORCHIDEE-CROP (v0), a new process-based model that integrates crop phenology and harvest into the ORCHIDEE land surface model. This enhanced model significantly improves the simulation of crop development and carbon/energy fluxes across Europe, providing a more realistic representation of agricultural systems in climate models.
What Problem Does This Paper Address?
Understanding how crops respond to changing climate and atmospheric CO2 is crucial for predicting food security and climate feedbacks. Existing land surface models (LSMs) often treat croplands simplistically, either by lumping them with natural vegetation or using static, crop-specific parameterizations that don't fully capture dynamic crop processes. This limits their ability to accurately simulate the complex interactions between agriculture, the carbon cycle, and the climate system. The authors aim to bridge this gap by developing a more sophisticated crop model that can be integrated into a global LSM.
Methods
The core innovation is the integration of a "generic crop phenology and harvest module" into the ORCHIDEEv196 land surface model. This module, based on the STICS agronomic model, simulates key crop development stages, leaf area index (LAI) dynamics, and carbon allocation to grain for yield. It also includes a simplified parameterization for nitrogen fertilization. The model was evaluated against eddy covariance flux tower data and biometric measurements from seven winter wheat and maize sites across Europe.
Datasets
The study utilized observational data from seven eddy-covariance sites across Europe, which provided measurements of:
- Carbon dioxide fluxes (Net Ecosystem Exchange - NEE)
- Latent heat fluxes (LE)
- Sensible heat fluxes (H)
- Leaf Area Index (LAI)
- Aboveground Biomass (AGB)
- Crop Yield
Meteorological data (temperature, precipitation, wind speed, etc.) were also used as model input.
Key Insights and Ideas
- Phenology is Key: Explicitly modeling crop phenology (development stages) is critical for accurately simulating LAI dynamics and, consequently, carbon and energy fluxes. The default ORCHIDEE model, which treated crops like grass, showed poor performance in this regard.
- Generic Approach Works: The use of a "generic" crop module, parameterized from STICS, proved effective for simulating different temperate crops (maize and winter wheat) across diverse European sites, suggesting this approach can be broadly applied.
- Nitrogen Matters, Simply: Even a simplified nitrogen fertilization parameterization improved yield simulations, highlighting the importance of nutrient dynamics in crop models.
- Hydrology Matters Too: The study found that a more detailed 11-layer soil hydrological scheme (11LAY) improved simulations, particularly for crops in drier climates, suggesting that soil moisture dynamics are a significant driver of crop responses.
Key Results
- LAI Improvement: ORCHIDEE-CROP (v0) significantly improved LAI simulations compared to the original ORCHIDEE, reducing the RMSE by nearly half (from 2.38 ± 0.77 to 1.08 ± 0.34 m² m⁻²).
- Good Overall Match: The model achieved a generally good match for aboveground biomass (NRMSE 11.0–54.2%), crop yield, and daily carbon and energy fluxes (NRMSE ~9.0–20.1% for NEE, ~9.4–22.3% for energy fluxes).
- Spatial Gradients Captured: The model successfully captured spatial gradients of carbon and energy fluxes across the European sites, a crucial capability for regional and global simulations.
- Yield Simulation: With fertilization, ORCHIDEE-CROP showed a significant reduction in yield NRMSE (from 47% to 23%) compared to the unfertilized version.
Limitations and Future Work
- Management Simplifications: The authors acknowledge limitations in representing detailed crop management practices (e.g., irrigation, specific fertilization strategies beyond a simple additive response) and the lack of an explicit nitrogen cycle.
- Root Profile: The root profile is assumed to be fixed, and a more mechanistic parameterization is suggested for future improvement.
- Pest, Diseases, Weeds: The current model simplifies pest, disease, and weed impacts, assuming they are controlled.
- Extreme Soil Conditions: Extreme soil conditions like high salinity or acidity are not explicitly handled.
- Future Directions: The authors emphasize the need for explicit nutrition dynamics and more detailed human management modules to improve understanding and prediction of croplands' role in the biosphere-atmosphere continuum. They also suggest coupling with atmospheric circulation models for global applications. The code is still under development but available upon request.
Summary updated