ANNOUNCEMENTS
Within the domain of agriculture, the term "yield" pertains to the quantity of agricultural output attained from a specific land area during a given timeframe. It serves as a metric for productivity, typically expressed as a quantity per unit of land area, such as bushels per acre, kilograms per hectare, or tons per hectare, depending on the crop and geographical location. Farmers endeavor to maximize yield while upholding quality and sustainability in their agricultural systems. They select and cultivate crop varieties based on their specific requirements and local conditions. The growth and development of crops, and consequently their yield, are influenced by environmental factors such as temperature, precipitation, solar radiation, and soil fertility. Favorable conditions, such as adequate moisture, optimal temperature ranges, and appropriate soil nutrients, promote robust plant growth and contribute to higher yields. Conversely, extreme weather events, droughts, excessive rainfall, frost, or heatwaves can have adverse effects on yields.
Iowa is commonly referred to as the "Corn Belt of the USA" due to its status as the leading state in terms of corn production for the country. Kossuth County in Iowa exhibited the highest corn yield, with 58,066,000 bushels in 2020 and 53,238,000 bushels in 2019. On average, the state produces 23,190,297.2 bushels of corn.
Remote sensing datasets play a pivotal role in monitoring crop yield by providing valuable information on crop health, biomass, and spatial variability. Machine learning algorithms can analyze vast amounts of data and uncover intricate patterns and relationships that may not be readily discernible through conventional statistical methods. By leveraging these patterns, machine learning models can yield more accurate predictions of crop yield compared to traditional approaches. Moreover, the adaptability and learning capabilities of machine learning algorithms enable them to continually enhance prediction accuracy by assimilating new data and updating their predictions accordingly.
This study seeks to integrate remote sensing datasets encompassing meteorological, soil, spectral indices, and biophysical parameters with machine learning algorithms, specifically Random Forest and Gradient Boosting, to forecast the yield of corn for the year 2020. Training data has been collected spanning the time period from 2010 to 2019. The outcomes of the machine learning algorithms surpass those of traditional statistical methods for prediction. The study successfully predicted yield for the years 2018, 2019, and 2020, with deviations from reported values ranging between 1.68% and 8.2%.
