Simulating alfalfa and pasture yields at regional and national scales in Canada from 1981 to 2019

Abstract

CONTEXTThere has been a great deal of uncertainty in the reported alfalfa and pasture yields in the agricultural database as these commodities are typically used internally on farms and are not sold in the same way as other crops such as wheat, canola or corn. However, yields from alfalfa and pasture crops are important inputs (biological fixation) and outputs (feed) that can impact the estimates of agri-environmental indicators as the land area in Canada under these crops is very large. One of the indicators which uses this data is the residual soil nitrogen (RSN) which is the amount of inorganic N remaining in soils after harvest. High RSN levels can lead to impaired water quality (e.g., eutrophication) and global warming (N2O). Therefore, it is critical to obtain accurate alfalfa and pasture yield estimates. In previous versions of the Canadian Agricultural Nitrogen Budget Model (CANB), only provincial average yield values were used for alfalfa or pasture yields across all of the soils and landscapes. Soils, climates, and crop yields do vary within provinces, and a more accurate estimate of these yields over time and space is required. OBJECTIVEThe objective of this study was to collect recent published yield data and use the Environmental Policy Integrated Climate (EPIC) to estimate yields for the 2780 soil landscapes of Canada (SLC) polygons. METHODSThe EPIC model was used to calibrate and evaluate yields at field scale based on a total of 109, 59 and 47 treatments collected for alfalfa, improved and unimproved pasture yields, respectively, from Canadian publications. Due to satisfactory evaluation, we extended simulation to all SLC polygons based on SLC v3.2 soil and weather databases, representative crops as proxies for planting and harvesting date, and historical rates of fertilization. RESULTS AND CONCLUSIONSThe average simulated alfalfa, improved and unimproved pasture yields were 5.00, 3.85 and 1.22 Mg ha−1, respectively over the period from 1981 to 2019. Generally, the EPIC model showed reliable predictions for alfalfa and pasture yields under variable weather, soil, and management practices. SIGNIFICANCETherefore, it enabled us to provide improved estimates across soil landscapes polygons in Canada and thereby improve our agri-environmental indicator models.