A Spatial Econometric Analysis of Weather Effects on Milk Production

• Published in: Earth (Basel, Switzerland) Vol. 5; no. 3; pp. 477 - 490
• Main Authors: Fan, Xinxin, Ma, Jiechao
• Format: Journal Article
• Language: English
• Published: Basel MDPI AG 01.09.2024
• Subjects: Agricultural production; Bias; Climate change; Cow's milk; Drought; Econometrics; Error analysis; Extreme weather; Feasibility studies; Greenhouse gases; Heat; Heat stress; Heat tolerance; Humidity; Livestock; Mathematical models; Milk; Milk production; Performance prediction; Precipitation; Productivity; Radiation; Regions; Spatial analysis; Spatial data; spatial panel models; Summer; Temperature; Temperature effects; Variables; Weather; weather condition; United States > US
• ISSN: 2673-4834; 2673-4834
• DOI: 10.3390/earth5030026

Greenhouse gas (GHG) emission-induced climate change, particularly occurring since the mid-20th century, has been considerably affecting short-term weather conditions, such as increasing weather variability and the incidence of extreme weather-related events. Milk production is sensitive to such changes. In this study, we use spatial panel econometric models, the spatial error model (SEM) and the spatial Durbin model (SDM), with a panel dataset at the state-level varying over seasons, to estimate the relationship between weather indicators and milk productivity, in an effort to reduce the bias of omitted climatic variables that can be time varying and spatially correlated and cannot be directly captured by conventional panel data models. We find an inverse U-shaped effect of summer heat stress on milk production per cow (MPC), indicating that milk production reacts positively to a low-level increase in summer heat stress, and then MPC declines as heat stress continues increasing beyond a threshold value of 72. Additionally, fall precipitation exhibits an inverse U-shaped effect on MPC, showing that milk yield increases at a decreasing rate until fall precipitation rises to 14 inches, and then over that threshold, milk yield declines at an increasing rate. We also find that, relative to conventional panel data models, spatial panel econometric models could improve prediction performance by leading to smaller in-sample and out-sample root mean squared errors. Our study contributes to the literature by exploring the feasibility of promising spatial panel models and resulting in estimating weather influences on milk productivity with high model predicting performance.