Enhancing agricultural sustainability through CNN-LSTM based land suitability assessment for improved production

Agricultural sustainability is an important concern facing global population expansion and environmental challenges. Accurate land suitability evaluation is critical for ensuring efficient and effective land usage. The aim of this research is to assess the following framework that the Food and Agriculture Organization (FAO) developed, if the land in India is suitable for two key crops, namely maize and barley. Soil specimens and physical-chemical properties from 105 genetic soil profile layers are collected. Weather and landscape data were gathered. Deep learning (DL) techniques were used to map the two crops' land suitability classifications after computation. This study suggests a unique method for assessing land potential for increased agricultural output by Convolution Neural Networks-Long Short-Term Memory (CNN-LSTM) networks. The maps generated by the two ways showed significant disparities. Studies on rain-fed corn, for example, showed that DL-based land suitability maps were more accurate than those created using the traditional method. The outcomes demonstrated that when contrasting the CNN-LSTM method with the conventional methodology, the areas with classes C2 were higher and the areas with categories C1 were less projected. Rainfall during the flowering period, steep slopes, poor soil depth, high pH and a high gravel component were the research area's main drawbacks. Land development activities are recommended to boost productivity and build a sustainable agricultural system. This research contributes to balance farm expansion and environmental protection, supporting a sustainable and resilient agricultural system to fulfill the requirements of the world's rising population.