Qualitative crop condition survey reveals spatiotemporal production patterns and allows early yield prediction [Dataset]

Reliable crop monitoring systems provide critical information to detect and track anomalies in the status of crops. These systems are fundamental for the development of integrated methodologies that inform agricultural policy, market analysis, or producer decision-making. They are also used in the development of early warning systems that permit to anticipate drought conditions and trigger action to mitigate short term food shortages or to stabilize the structure and pricing of agricultural markets. Current efforts to develop crop monitoring systems exploit meteorological and crop growth models, and satellite imagery. However, legacy sources of information such as operational crop rating surveys that have long and uninterrupted records receive less attention. We argue that crop rating data, despite its subjective and non-quantitative nature, captures the complexities of assessing the 'status' of a crop better than any model or remote sensing retrieval. This is because crop rating data naturally represents the broad expert knowledge of many individual surveyors spread throughout the country. Crop rating surveys in effect constitute a sophisticated network of "humans as sensors" that provide consistent and accurate information on crop progress. We analyze data from the USDA Crop Progress and Condition (CPC) survey between 1987 and 2019 for four major crops across the US (corn, soybeans, winter wheat, and upland cotton). We show how the original qualitative data can be transformed into a continuous, probabilistic variable better suited to quantitative analysis, and demonstrate it can be used to monitor crop status and provide early predictions of crop yields.