DEVELOPING PROTOTYPE VOC SENSOR-BASED PRODUCTS FOR DETERMINING SOIL HEALTH ON-FARM

在气候变化的背景下，在2050年养活98亿人将取决于我们保持土壤存活的技能。粮食生产与土壤健康直接相关。为了管理和改善土壤健康，农民需要有关其土壤的化学、物理和生物特性的可靠信息。有一些方法可以用来分析土壤养分和确定土壤的物理性质。农民目前只能使用基于呼吸的方法来衡量微生物对土壤健康的贡献，但这些方法没有提供有关存在的微生物区系的信息，并且受到土壤中其他来源的二氧化碳的影响。下一代测序具有作为土壤健康生物指标的潜力，但成本高，测试需要数小时，获得的数据需要专家进行解释。我们的解决方案是利用土壤生物群释放的挥发性有机化合物(VOC)中包含的丰富信息。这些已被证明是土壤生物群活动的良好指标，但它们的检测和分析目前需要以实验室为基础的仪器和熟练的人员。在前期工作中，我们开发了一种可以检测土壤VOCs的传感器，并证明了它的响应可以与土壤健康相关。在这个项目中，我们将确定这些传感器对各种不同土壤和种植制度的响应。这些将与传统的土壤健康指标和下一代测序数据相关联。机器学习将用于处理获得的数据，以提供一个基于云的数据库，现场传感器可以直接访问该数据库。将调查使用机器人部署传感器和相关的GPS数据，为农民提供其农场土壤健康的全面和精细数据，以便他们能够评估耕作做法对土壤健康的影响，并对这些数据进行调整，以提高土壤健康和生产力。用目前的测试方法(GB 60/样本)测试每平方米的土地数据是不可行的，因为英国农场的平均规模为930,000平方英尺。该项目将由P.E.S.Technologies领导，这是一家初创公司，与Hutchinsons、英国农学专家和Small Robot公司合作，开发了一种用于土壤VOCs的塑料电子传感器。学术合作伙伴将是英国领先的园艺研究组织NIAB-EMR，在挥发性有机化合物分析方面拥有长期经验的自然资源研究所，以及在机器学习方面拥有专业知识的埃塞克斯大学。

Feeding 9.8 billion people in 2050 in a climate change context will depend on our skills to keep soils alive. Food production is directly correlated with soil health. To manage and improve soil health, farmers need reliable information about the chemical, physical and biological properties of their soils. There are methods available to assay soil nutrients and determine the physical properties of soils. Only respiration-based methods are currently available to farmers to measure the microbial contributions to soil health, but these give no information on the microbiota present and are affected by other sources of CO2 in the soil. Next-generation sequencing has potential as a biological indicator of soil health, but the costs are high, the tests take hours to conduct, and the data obtained requires experts in order to interpret it.Our solution is to tap into the wealth of information contained in the volatile organic compounds (VOCs) released by soil biota. These have been demonstrated to be excellent indicators of soil biota activity, but their detection and analysis currently requires laboratory-based instrumentation and skilled personnel. In preliminary work we developed a sensor that can detect soil VOCs and demonstrated that its responses can be correlated with soil health. In this project we will determine the responses of such sensors to a wide range of different soils and cropping systems. These will be correlated with conventional soil health indicators and next-generation sequencing data. Machine learning will be used to process the data obtained to provide a cloud-based database that can be accessed directly by sensors in the field. Use of robots to deploy the sensors with associated GPS data will be investigated to provide farmers with comprehensive and fine-scale data of soil health on their farms so that they can assess the impact of farming practices on soil health and adapt these to increase soil health and productivity. Testing every square meter of land data would be unfeasibly expensive with current testing methods (£60/sample) as the average UK farm size is 930,000 sq. m..The project will be led by P.E.S. Technologies, a start-up company that developed a plastic electronic sensor for soil VOCs, in collaboration with Hutchinsons, UK agronomy specialists, and the Small Robot Company. Academic partners will be NIAB-EMR, the leading UK horticultural research organisation, the Natural Resources Institute with long experience in VOC profiling, and the University of Essex with expertise in machine learning.