Use, calibration, and validation of Agroecological models for Boreal environments: a review

Authors: FOSTER, DANIEL - Natural Resources Institute Finland (LUKE); HELAMA, SAMULI - Natural Resources Institute Finland (LUKE); HARRISON, MATTHEW - Tasmanian Institute Of Agricultural Research; Rotz, Clarence - Al; CHANG, JINFENG - Zhejiang University; CIAIS, PHILLIPPE - Université Paris-Saclay; PATTEY, ELIZABETH - Agri Food - Canada; VIRKAJARVI, PERTTU - Natural Resources Institute Finland (LUKE); SHURPALI, NARASINHA - Natural Resources Institute Finland (LUKE)

Submitted to: Science of the Total Environment
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 4/12/2022
Publication Date: 5/27/2022
Citation: Foster, D., Helama, S., Harrison, M., Rotz, C.A., Chang, J., Ciais, P., Pattey, E., Virkajarvi, P., Shurpali, N. 2022. Use, calibration, and validation of Agroecological models for Boreal environments: a review. Science of the Total Environment. 1(1):14-30. https://doi.org/10.1002/glr2.12010.
DOI: https://doi.org/10.1002/glr2.12010

Interpretive Summary: Landscapes in Boreal regions may benefit from climate change to play a role in insuring future food security. Although arable cropping systems are growing in popularity in Boreal areas, the agricultural mainstay in such zones will likely continue to be livestock production systems for the foreseeable future. When appropriately calibrated and evaluated, suitable simulation models can produce sufficiently accurate data for agricultural land-use planning and improving farm-level productivity while minimizing impacts on the environment and natural capital. Such models are often used to identify interactions between plant and animal productivity, quantify greenhouse gas (GHG) emissions, soil carbon stocks and sequestration, and fertilizer use and nutrient losses. While models for agricultural systems designed for temperate and tropical agro-ecological zones abound, models developed specifically for boreal zones are very much in their infancy. As such, there is considerable uncertainty around the accuracy and suitability of existing models for boreal areas. Here, we reviewed the performance of extant livestock production systems and GHG models in their ability to simulate GHG emissions and to account for climatic, environmental and management drivers across a range of boreal environments, management systems and enterprise mixes. We identified a substantial dearth of modelling studies in Boreal regions, with the publication of three or less papers per year since the year 2000; in fact, in many years, there were no modelling studies at all. We show that there is significant scope for model refinement and evaluation in boreal regions using field measurements of biomass, animal liveweight gains and GHG emissions, both for individual models as well as in multi-model comparisons. This work would facilitate identification of biophysical, biochemical, and agronomic processes underpinning superior model performance, as well as opportunities for model parsimonisation by simplifying complexity where algorithms contribute little to reliable model performance. We show that the IFSM and BASGRA_N models offer the best performance in terms of grassland production, while DNDC was the most reliable in predicting soil nitrous oxide and ammonia emissions. Our review finds that no model outperforms all others for all metrics, suggesting that the use of multiple models may facilitate simulation of more reliable bandwidths for any given metric. A key conclusion is that both conceptual and mathematical mechanisms used in agroecological models would be worthy of further investigation and comparison, providing improved future models designed for boreal systems.

Technical Abstract: Global warming has caused considerable change in boreal areas with the melting of permafrost exposing landscapes previously unavailable to agriculture. While most assessments report negative impacts of climate change, the combination of milder and shorter winters and elevated atmospheric carbon dioxide may provide substantial and beneficial opportunities for agricultural productivity across vast landscapes in some regions. Intensification of global food supply and climate adaptation must however occur without degrading natural capital, losing biodiversity, increasing greenhouse gas emissions or stimulating other adverse economic, environmental or social trade-offs. Process-based models have proven to be invaluable tools in farm planning and environmental assessment in many agro-ecological regions, but such models remain underutilized in boreal regions. A review was conducted which has built a foundation for future research by identifying some of the more well-known models as candidates for adaptation and application to boreal regions as well as pointing out some of their weaknesses and areas of uncertainty.