Year: 2025
Type: article
Abstract: Bundelkhand region of Uttar Pradesh and Madhya Pradesh states faces various problems such as continuous drought, poor management of rainfall, soil loss, single crop focus, poverty and continuous migration of people out of the area.Agriculture in this area is based on rain fed crops such as pulses and oilseeds (Kalia et al., 2021).With its ability to tolerate dry spells and prosper in harsh soil conditions, sorghum is most suited cereal crop for Bundelkhand's semi-arid land where rainfall is uncertain, and water access is scarce.Sorghum is a vital basic food for rural families.It is extremely sensitive to temperature and photoperiod; it is important to study the role of meteorological variables to improve sorghum production and its quality.India is the third-largest producer of sorghum in the world with 6.0 million metric tons (9% of global production) after United States (14% of global production) and Nigeria (10% of global production) (IPAD 2024(IPAD -2025)).Agrometeorological indicators are vital for adapting agricultural planning, which enhances crop management and promotes food security under threat due to climate change.In the past, researchers have employed important statistical methods for crop harvesting challenges due to weather condition and its prediction on pre harvest forecasting (Garde et al., 2015;Maurya et al., 2025).Machine learning models employ algorithms to assess data, spot trends, and develop connects within the dataset, as compared to statistical models that link weather data and agricultural output using preestablished mathematical equations.Machine learning models have been successfully used to predict the yields of a number of crops including cashew, pigeon pea, rice, sorghum, coconut, wheat and soybean (Satpathi et al., 2023;Khan et al., 2023;Setiya et al., 2022).The study focuses on two major sorghum (jowar) producing districts of Uttar Pradesh, Banda (25 0 48' N, 80 0 33' E) and Hamirpur (25 0 95' N,80 0 15'E).Both districts are recognized for being characterized by fertile alluvial and black soils with a high capacity for holding moisture and their area is suitable for growing sorghum because of its favorable soil characteristics and agroclimatic factors (Suman et al., 2024).Time series data on sorghum yield for about 31 years from 1991-92 to 2021-22, especially for the Banda and Hamirpur districts of Uttar Pradesh, was obtained from the Spider Statistics Journal, published by the Directorate of Economics and Statistics, Uttar Pradesh.The data on weather conditions such as maximum temperature (Tmax), minimum temperature (Tmin), rainfall (Rain), wind speed (WS), and relative humidity (RH) was gathered from NASA POWER web portal (https://power.larc.nasa.gov/data-access-viewer/). The weather variables data have been transformed into phenological stage wise averages to process further.In terms of phenology, the average values were computed using the daily weather data.These average values are subsequently employed in the computation of both weighted and unweighted weather indices as per procedure given by Setiya et al., (2022).Using time series data on phenological weather indices and sorghum yield, three multivariate models viz.Artificial Neural Network (ANN), Principal Component Analysis (PCA-ANN) and Least Absolute Shrinkage and Selection Operator (LASSO) were developed for this study.The models were tested and trained on these datasets.Among the complete dataset spanning 31 years, 25 years of data were employed for training of the models, while the remaining 6 years data were utilized for testing of the models.In training stage, the description of the models with their error percentage are given in Table 1.The ANN model consists of an
Source: Journal of Agrometeorology
Authors: RITU SINGH, ANNU, UMESH CHANDRA, HIMANI MAHESHWARI, ARJUN PRASAD VERMA, GAURAV SHUKLA
Language: English
Domain: Life Sciences
SDG: Zero hunger
Open Access status: diamond
OpenAlex ID: W4416973473
FWCI:
Cites:
Cited by:
Related to: