Assessing Impact of Carbon-smart Farming Practices in Rice with Mobile Crowdsensing

Soil Organic Carbon (SOC) is an important constituent of measures that govern soil health. Sequestration of carbon in the soil is a major growing focus of sustainability initiatives world-wide to reduce carbon footprint in the atmosphere and counter the effects of global warming. India is the world's largest rice producer by area, estimated at about 44.0 million hectares. Sustainable crop and land management practices play a crucial role in enabling more carbon to be absorbed by the soils and potentially generate carbon-credits which could be traded. Mobile Crowdsensing (MCS) can play a pivotal role in this by serving as a channel to acquire the data and generate insights to take in-season decisions and get a compliance-view on carbon-smartness of farm-operations. We present our study to digitally assess through crowdsensing the carbon smartness and stock of different agricultural fields in the Cauvery Delta Zone (CDZ) of Tamil Nadu, India. A set of ten selected farmers were divided into two observation groups CSCP-1 and CSCP-2. CSCP-1 followed organic practices and CSCP-2 followed integrated nutrient management practices with inorganic components. AI in the form of imaging with deep learning was used to validate across both the groups the granularity of operations like tillage that have a significant impact on the level of carbon sequestration. A process model approach was used to simulate the temporal changes in soil organic carbon (SOC) for all the farmers to serve as a continuous feedback on carbon footprint in response to operations reported. SOC levels were obtained for the entire rice season including pre-sowing and post-harvest periods where CSCP-1 farmers sequestrated more carbon than CSCP-2 farmers in general. Results showed an increase of 0.51 tC/ha in carbon stock for CSCP-1 farmers at the end of the season while the corresponding increase was 0.23 tC/ha for CSCP-2 farmers. Within the CSCP-2 group that adopted inorganic practices, farmers following intensive tillage sequestrated even less soil organic carbon at 0.17 tC/ha. MCS coupled with AI and process models therefore can help evolve a real-time carbon-smartness profile for every farm towards better strategic and management decisions.