Attributing and Projecting U.S. Agricultural Economy Growth Due to Changes in Regional Climate and Crop

Partial productivity measures such as crop yields are widely used in most economic studies of climate change impacts on agricultural sustainability. Numerous statistical and numerical studies have highlighted significant climate impacts on crop yields, particularly in regions with increasing temperature, decreasing rainfall, or enhancing variability of the two. The focus of these research evolves over time to incorporate new perspectives and methods. As partial productivity measures cannot account for the overall efficiency and resilience of the agricultural systems at regional to national levels, Total Factor Productivity (TFP) is an excellent indicator of the total performance of the agriculture system. U.S. agriculture is a reliable source of affordable safe food and agriculture products for the world. For the past several decades, U.S. agriculture has been a global leader in TFP and output growth. Quantifying the observed relationships between TFP and climate spatiotemporal variations is critical to understanding whether the current U.S. agricultural productivity will continue into the future under significant climate change as projected. Our previous work has identified the mounting adverse climate impacts that pose a growing threat to the sustainable growth of U.S. agricultural production, assuming the dominant role of seasonal average climate conditions rather than daily variations (Liang et al. 2017). However, crop growth in all stages depends on daily weather conditions and their accumulative consequences, where extreme events such as heat waves, cold surges, floods, and droughts are especially damaging. The intraseasonal climate variations that evidently impose significant effects on final crop yields and the large spatial variations of these effects have not been incorporated into the existing studies assessing the U.S. agricultural TFP growth at the either state or national level. The research on climate impact on TFP and crop yield is a complex and challenging task that requires the integration of multiple disciplines and the use of advanced modeling and analysis tools.

To address these challenges, we propose to develop a new modeling approach to understand and capture the observed relationships among distributed daily weather conditions, county-level annual harvest areas and yields of major crops, state-level and national TFP of the agricultural and all economic sectors, grains and agricultural products price and international trades, and other related factors such as fertilizer application and air pollution. We will use all available data records since 1948 and advanced machine learning algorithms to [1] identify the key periods (starting and ending dates) when the climate statistics (mean, deviation, extreme) affect county-level crop yields significantly; [2] identify the major regions where the averaged climate statistics and crop yields affect state-level TFP significantly; [3] determine the critical metrics of climate, crop, and other features based on the identified key periods and major regions that manifest significant impacts on national TFP growth; and [4] build a robust model that closely links climate variations, crop yields, and other production factors at local-regional scales to the U.S. agricultural TFP growth at the state and national levels. Ultimately, the resulting model will facilitate projections of TFP growth for the coming decades.