In the United States (US), North Dakota is the largest producer of Durum Wheat (Triticum durum), hereinafter referred to as Durum. Durum grain has a high protein content and multiple utilities in food products. We investigated the historical trends in Durum production and yield as influenced by changes in precipitation (precip) and temperature (temp). The study accounted for variations in environmental conditions by running a dynamic crop model in thirteen Durum producing counties.

The climate of North Dakota is representative of the highly productive agricultural lands of the Northern Great Plains, encompassing five US states and two Canadian provinces. The Eastern part of North Dakota has a humid continental climate while the western part is semi-arid. Creating a distinct West-to-East precip gradient across the state. Low mean average temps (cir. +4 °C), and high-temp variability lead to the relatively short growing season (cir. 130 days). Combined with limited rainfall (cir. 350 mm in the E and 560 mm in the W), it makes agriculture highly dependent on temp and precip. Accordingly, climate change has a high potential impact on crop production in the region.

The ALMANAC crop growth model was used to simulate the production of Durum. Model performance was estimated by comparison of simulated yields with historical observations, and was found satisfactory using the Nash--Sutcliffe model efficiency coefficient (E) and Coefficient of determination (r 2) (< 0.50). Uncertainty in projected future climate is addressed using an ensemble of 17 Global Circulation Models (GCMs) run under four scenarios. GCM output data were further downscaled using MarkSim weather, and daily weather was generated for two 30-year periods, characteristic of the 2020's and the 2050's.