Switchgrass and Miscanthus long-term yield across Michigan and Wisconsin
Data files
Oct 17, 2025 version files 158.72 KB
-
DryadRepository_AllSites_PerennialsYields.csv
156.02 KB
-
README.md
2.69 KB
Abstract
Perennial grasses like switchgrass (Panicum virgatum) and miscanthus (Miscanthus × giganteus) are expected to supply a substantial amount of the United States bioeconomy’s feedstock demand. However, uncertainties around patterns of long-term yields challenge the viability of this potential. To resolve long-term yield patterns, we analyzed over 200 plantings of switchgrass and miscanthus across Michigan and Wisconsin USA measured over 5 to 15 years. During a yield-building phase, peak yields occurred within 4 – 5 years after planting, followed by a 6 – 7 year yield-decline phase in which switchgrass and miscanthus lost 30 – 47% and 14 – 40% of peak yields, respectively. Added nitrogen increased peak-yields by 10 – 20% and attenuated the yield decline by 20 – 50%. A farm-to-gate economic analysis suggests replanting switchgrass and miscanthus 5 and 9 years following their peak yields optimize profit over a 30-year time horizon. This conserved long-term yield pattern has implications for crop management, breeding programs, policy, and global bioenergy with carbon capture and storage modelling.
https://doi.org/10.5061/dryad.6m905qg9x
Description of the data and file structure
The file contains Switchgrass and M. × giganteus long-term yield data were collected from multiple long-term experiments along Michigan and Wisconsin, USA. The 12 different experimental sites were planted through 2007 – 2013 and provided a wide range of conditions experimental conditions, plot sizes, and N fertilization rates. Given the multiple planting years our long-term yield dataset included different experiment lengths; 16 plots with up to 5 years of data, 86 plots with up to 9 years of data, 60 plots with up to 10 – 12 years of data, 32 plot with up to 13 years of data, and 20 plots with up to 15 years of data at the moment of analysis. These multiple planting years also allowed us to compare yields of different ages grown under the same growing seasons to better separate the age and growing season effect.
Files and variables
File: DryadRepository_AllSites_PerennialsYields.csv
Description: End-of-season biomass (yield) harvested yearly from switchgrass and miscanthus plots
Variables
- Experiment: Name of main experiment
- site: Experimental site within each experiment
- Year: Year on which yield was harvested
- crop: name of crop
- Variety: Only for switchgrass. Not applicable for miscanthus (NA). All miscanthus plantings were Miscanthus x giganteus “Illinois clone”.
- Ecotype: Switchgrass ecotype. Not applicable for miscanthus (NA)
- Ntrt: Discrete variable for nitrogen fertilization (N0 for unferitlized, N+ for fertilized plots)
- Nrate: Nitrogen fertilization rate (kg/ha/yr)
- Age: years since planting at the moment of harvest (assumed to be 1 year old after the first growing season)
- replicate: Experimental replicate
- Yield_Mgha: End-of-season biomass (yield) (Mg/ha). NA; Not available data, End-of-season biomass couldn't be harvested, or were not included in the final analysis. Refer to companion code for further specifications.
Code/software
We used R Statistical Software for all statistical analyses based on code developed by Mauricio Tejera-Nieves available at https://github.com/https://github.com/PerennialDr/Long-term-Yield-patterns-of-Perennial-Grasses.
Please direct correspondence to Mauricio Tejera-Nieves (mauri@jii.org)
Access information
Companion manuscript is submitted to the Proceedings of the National Academy of Sciences
Please direct correspondence to Mauricio Tejera-Nieves (mauri@jii.org)
Switchgrass and M. × giganteus long-term Yield data for this analysis was collected from were collected from multiple long-term experiments along Michigan and Wisconsin, USA. The 12 different experimental sites were planted through 2007 – 2013 and provided a wide range of conditions experimental conditions, plot sizes, and N fertilization rates. Given the multiple planting years our long-term yield dataset included different experiment lengths; 16 plots with up to 5 years of data, 86 plots with up to 9 years of data, 60 plots with up to 10 – 12 years of data, 32 plot with up to 13 years of data, and 20 plots with up to 15 years of data at the moment of analysis.